VOLUME 9 DISEASE CONTROL PRIORITIES • THIRD EDITION Disease Control Priorities: Improving Health and Reducing Poverty VOLUME 9 DISEASE CONTROL PRIORITIES • THIRD EDITION Disease Control Priorities: Improving Health and Reducing Poverty EDITORS Dean T. Jamison Hellen Gelband Susan Horton Prabhat Jha Ramanan Laxminarayan Charles N. Mock Rachel Nugent © 2018 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 21 20 19 18 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org /licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: Jamison, D. T., H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent, editors. 2018. Disease Control Priorities: Improving Health and Reducing Poverty. Disease Control Priorities (third edition), Volume 9. Washington, DC: World Bank. doi:10.1596/978-1-4648-0527-1. License: Creative Commons Attribution CC BY 3.0 IGO Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to reuse a component of the work, it is your responsibility to determine whether permission is needed for that reuse and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org. ISBNs and DOIs: Softcover: Hardcover: ISBN (paper): 978-1-4648-0527-1 ISBN (paper): 978-1-4648-0529-5 ISBN (electronic): 978-1-4648-0528-8 DOI: 10.1596/978-1-4648-0527-1 DOI: 10.1596/978-1-4648-0529-5 Cover photo: © World Bank. Further permission required for reuse. Chapter opener photo: © Frank Spangler/Worldview Images. Used with the permission of Worldview Images. Further permission required for reuse. Library of Congress Cataloging-in-Publication Data has been requested. Contents Foreword by Bill and Melinda Gates ix Introduction by Lawrence H. Summers xi Preface xv Abbreviations xvii PART 1 OBJECTIVES AND CONCLUSION OF DISEASE CONTROL PRIORITIES, THIRD EDITION 1. Universal Health Coverage and Intersectoral Action for Health 3 Dean T. Jamison, Ala Alwan, Charles N. Mock, Rachel Nugent, David A. Watkins, Olusoji Adeyi, Shuchi Anand, Rifat Atun, Stefano Bertozzi, Zulfiqar Bhutta, Agnes Binagwaho, Robert Black, Mark Blecher, Barry R. Bloom, Elizabeth Brouwer, Donald A. P. Bundy, Dan Chisholm, Alarcos Cieza, Mark Cullen, Kristen Danforth, Nilanthi de Silva, Haile T. Debas, Peter Donkor, Tarun Dua, Kenneth A. Fleming, Mark Gallivan, Patricia Garcia, Atul Gawande, Thomas Gaziano, Hellen Gelband, Roger Glass, Amanda Glassman, Glenda Gray, Demissie Habte, King K. Holmes, Susan Horton, Guy Hutton, Prabhat Jha, Felicia Knaul, Olive Kobusingye, Eric Krakauer, Margaret E. Kruk, Peter Lachmann, Ramanan Laxminarayan, Carol Levin, Lai Meng Looi, Nita Madhav, Adel Mahmoud, Jean-Claude Mbanya, Anthony R. Measham, María Elena Medina-Mora, Carol Medlin, Anne Mills, Jody-Anne Mills, Jaime Montoya, Ole Norheim, Zachary Olson, Folashade Omokhodion, Ben Oppenheim, Toby Ord, Vikram Patel, George C. Patton, John Peabody, Dorairaj Prabhakaran, Jinyuan Qi, Teri Reynolds, Sevket Ruacan, Rengaswamy Sankaranarayanan, Jaime Sepúlveda, Richard Skolnik, Kirk R. Smith, Marleen Temmerman, Stephen Tollman, Stéphane Verguet, Damian Walker, Neff Walker, Yangfeng Wu, and Kun Zhao 2. Intersectoral Policy Priorities for Health 23 David A. Watkins, Rachel Nugent, Helen Saxenian, Gavin Yamey, Kristen Danforth, Eduardo González-Pier, Charles N. Mock, Prabhat Jha, Ala Alwan, and Dean T. Jamison 3. Universal Health Coverage and Essential Packages of Care 43 David A. Watkins, Dean T. Jamison, Anne Mills, Rifat Atun, Kristen Danforth, Amanda Glassman, Susan Horton, Prabhat Jha, Margaret E. Kruk, Ole F. Norheim, Jinyuan Qi, Agnes Soucat, Stéphane Verguet, David Wilson, and Ala Alwan v PART 2 PROBLEMS AND PROGRESS 4. Global and Regional Causes of Death: Patterns and Trends, 2000–15 69 Colin Mathers, Gretchen Stevens, Dan Hogan, Wahyu Retno Mahanani, and Jessica Ho 5. Annual Rates of Decline in Child, Maternal, Tuberculosis, and Noncommunicable Disease Mortality across 109 Low- and Middle-Income Countries from 1990 to 2015 105 Osondu Ogbuoji, Jinyuan Qi, Zachary D. Olson, Gavin Yamey, Rachel Nugent, Ole Frithjof Norheim, Stéphane Verguet, and Dean T. Jamison 6. Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 121 Beverley M. Essue, Tracey-Lea Laba, Felicia Knaul, Annie Chu, Hoang Van Minh, Thi Kim Phuong Nguyen, and Stephen Jan PART 3 ECONOMIC EVALUATION RESULTS FROM DISEASE CONTROL PRIORITIES, THIRD EDITION 7. Cost-Effectiveness Analysis in Disease Control Priorities, Third Edition 147 Susan Horton 8. Health Policy Analysis: Applications of Extended Cost-Effectiveness Analysis Methodology in Disease Control Priorities, Third Edition 157 Stéphane Verguet and Dean T. Jamison 9. Benefit-Cost Analysis in Disease Control Priorities, Third Edition 167 Angela Y. Chang, Susan Horton, and Dean T. Jamison PART 4 HEALTH SYSTEM TOPICS FROM DISEASE CONTROL PRIORITIES, THIRD EDITION 10. Quality of Care 185 John Peabody, Riti Shimkhada, Olusoji Adeyi, Huihui Wang, Edward Broughton, and Margaret E. Kruk 11. High-Quality Diagnosis: An Essential Pathology Package 215 Kenneth A. Fleming, Mahendra Naidoo, Michael Wilson, John Flanigan, Susan Horton, Modupe Kuti, Lai Meng Looi, Christopher P. Price, Kun Ru, Abdul Ghafur, Jinaxiang Wang, and Nestor Lago 12. Palliative Care and Pain Control 235 Eric L. Krakauer, Xiaoxiao Kwete, Stéphane Verguet, Hector Arreola-Ornelas, Afsan Bhadelia, Oscar Mendez, Natalia M. Rodriguez, Zipporah Ali, Silvia Allende, James F. Cleary, Stephen Connor, Kristen Danforth, Liliana de Lima, Liz Gwyther, Ednin Hamzah, Dean T. Jamison, Quach Thanh Khanh, Suresh Kumar, Emmanuel Luyirika, Anne Merriman, Egide Mpanumusingo, Diana Nevzorova, Christian Ntizimira, Hibah Osman, Pedro Perez-Cruz, M. R. Rajagopal, Lukas Radbruch, Dingle Spence, Mark Stoltenberg, Neo Tapela, David A. Watkins, and Felicia Knaul 13. Strengthening Health Systems to Provide Emergency Care 247 Teri A. Reynolds, Hendry Sawe, Andrés M. Rubiano, Sang Do Shin, Lee Wallis, and Charles N. Mock vi Contents 14. Community Platforms for Public Health Interventions 267 Melissa Sherry, Abdul Ghaffar, and David Bishai 15. Rehabilitation: Essential along the Continuum of Care 285 Jody-Anne Mills, Elanie Marks, Teri Reynolds, and Alarcos Cieza PART 5 INTERSECTORAL AND INTERNATIONAL TOPICS 16. Development Assistance for Health 299 Eran Bendavid, Trygve Ottersen, Liu Peilong, Rachel Nugent, Nancy Padian, John-Arne Rottingen, and Marco Schäferhoff 17. Pandemics: Risks, Impacts, and Mitigation 315 Nita Madhav, Ben Oppenheim, Mark Gallivan, Prime Mulembakani, Edward Rubin, and Nathan Wolfe 18. The Loss from Pandemic Influenza Risk 347 Victoria Y. Fan, Dean T. Jamison, and Lawrence H. Summers 19. Fiscal Instruments for Health in India 359 Amit Summan, Nicholas Stacey, Karen Hofman, and Ramanan Laxminarayan DCP3 Series Acknowledgments 371 Volume Editors 373 Contributors 375 Advisory Committee to the Editors 381 Reviewers 383 Policy Forum Participants 385 Index 387 Contents vii Foreword Bill and Melinda Gates Gates Foundation, Seattle, Washington, United States During the past 25 years, many countries have achieved stunned to read that 11 million young children were significant improvements in human health and well- dying every year from preventable causes such as pneu- being. Huge problems persist, and terrible inequities monia, diarrhea, malaria, and other infections that are must still be addressed to ease the suffering of the rare or rarely fatal in the developed world. We were world’s poorest and most vulnerable. But that does not shocked by the disparities in health outcomes between diminish several remarkable accomplishments: Since the rich countries and poorer ones. Every page screamed early 1990s, the world has seen substantial reductions out that human life was not being valued as it should be. in extreme poverty; child and maternal mortality; and In addition, our eyes were opened to the fact that the incidence of deadly and debilitating diseases, such most preventable deaths and disability in lower-income as tuberculosis, malaria, and HIV/AIDS. The incidence countries were caused not by hundreds of diseases but of polio has decreased by 99 percent, bringing the world by relatively few, and that the costs of preventing and to the verge of eradicating a major infectious disease for treating them were often low, relative to the benefits. Our only the second time in history. shock turned to excitement. Here were points of leverage Credit for these and other advances in global health where we could work to reduce inequity and help realize belongs to many institutions, governments, and individ- a world where every person has the opportunity to live a uals, including the scholars who organized and contrib- healthy, productive life. uted to the first and second editions of Disease Control DCP2, published in 2006, again advanced the con- Priorities. We hope and expect this third edition also will versation on global health. Where DCP1 focused on the have a large, salutary impact. benefits and costs of interventions against individual The first edition, DCP1, was published by the World diseases, contributors to DCP2 also considered how Bank in 1993. It was the first comprehensive effort to countries might gain greater traction by organizing their systematically assess the effectiveness of interventions efforts around multi-purpose health platforms, ranging against the major diseases of low-income and middle- from village clinics and school-based health programs to income countries. DCP1 also analyzed the relative costs district hospitals with emergency services and surgical of interventions, enabling policy makers and aid donors units. DCP2 showed how investments in health plat- to make smarter decisions about how to allocate scarce forms, especially for community-based primary care, health dollars for the greatest impact. DCP1 helped could magnify impact despite limited budgets. Several bring about dramatic shifts in how countries and the countries, particularly India and Ethiopia, have pursued global community invest in health. this approach with good results. Indirectly, DCP1 also influenced our personal deci- In important and useful ways, this third edition of sion to devote much of our philanthropy to improving Disease Control Priorities further widens the frame for the health of people in poor countries. This came about discussion of health policies and priorities, innovatively because data from DCP1 was a basis for the World addressing the different needs of countries at differ- Bank’s 1993 World Development Report, which focused ent stages in the development of their health systems. on investing in health and catalyzed our thinking about This edition maps out pathways—essential packages of how and where we could make a difference. We were related, cost-effective interventions—that countries can ix consider to speed their progress toward universal health programs, access to treatment for common infec- coverage. DCP3 also draws attention to the catastroph- tions—these pay enormous returns in lives saved and ically impoverishing effects that many medical proce- suffering avoided. Family planning, maternal health dures can have on poor families. This analysis, combined programs, and gender equity benefit communities with data on the lost productivity caused by various and society as a whole. Major infectious diseases diseases, provides insights into how investing in health, can be beaten through collaborative, international particularly in expanded access to health insurance and efforts, as the past 25 years have shown. Overall, prepaid care, can not only save lives but also help allevi- improving the health of the world’s most vulnerable ate poverty and bolster financial security. people remains one of the best investments the global Across the three editions, some conclusions community can continue to make toward realizing a remain constant. Childhood vaccinations, nutrition better, safer world. x Foreword Introduction Lawrence H. Summers Harvard University, Boston, Massachusetts, United States Most economists pride themselves on combining social agencies, civil society, and the academic community. concern with hard analysis. This trait they share with The WDRs are probably the world’s most widely distrib- an important strand of the human rights community uted economic publication. They are prepared by the working on global health. The late Jonathan Mann, to World Bank’s research arm, under the direction of its take a leading example, both argued for an idealistic Chief Economist, a position I had the good fortune to vision of health as a human right for all and created, hold in the period 1991–93. I selected health as the topic from almost nothing, the World Health Organization’s for WDR 1993. (WHO) effective and pragmatic Global Programme Why health? First, health and poverty intertwine against AIDS. Paul Farmer continues to provide global closely, and having a WDR on health provided an oppor- leadership in advocating health as a human right, but he tunity to provide insight into the World Bank’s central rightly emphasizes that advocacy alone remains insuf- goal of reducing poverty. Second, health represents an ficient. In Partners in Health, an organization Farmer area where governments can play a necessary and con- cofounded with Jim Kim (now president of the World structive role. And third, I believed that the potential Bank), Farmer created a vehicle to go beyond advocacy gains from getting health policy right were enormous. and develop the practical dimensions of the aspiration to Thus, the WDR 1993: Investing in Health, was published provide the highest quality of health care in rural Haiti, in June of 1993 (World Bank 1993). Rwanda, and elsewhere. In his essay “Rethinking Health Several features dominated the global health land- and Human Rights,” Farmer points to the importance scape at the time of the WDR 1993. First, and most of research in this agenda: “The purpose of this research visibly, the HIV/AIDS epidemic had emerged from should be to do a better job of bringing the fruits of nowhere to grow into a major problem in Africa and science and public health to the poorest communities” globally. Second, but much less visibly, government (2010, p. 456). Farmer and I may well have a different policies to control undernutrition, excess fertility, and take on the contributions that have been made over time infection had begun to bear fruit. Consolidating and by the World Bank and other international financial expanding the scope of these successes promised enor- institutions. But I think it fair to say that the pragmatic mous gains. As a consequence of success, however, China task of bringing technical knowledge to bear on the and other countries with early progress were already needs of the poor is a shared goal—and a goal that the experiencing substantial relative growth in their older Disease Control Priorities series has sought to advance for populations—and concomitant growth in the incidence over two decades. of cancer, heart disease, and stroke. Intervention against Each year the World Bank’s flagship publication, the these diseases is less decisive and often far more costly World Development Report (WDR), attempts to assemble than intervention against infection. Policy makers thus knowledge and to inspire action that serves the world’s experienced strong pressures to divert resources from poorest communities. These reports develop and take high payoff infection control to responding to noncom- stock of research and other evidence on a specific topic municable diseases. to inform the World Bank’s own policies and to stimulate In response to these features of the health landscape, discourse among member countries, other development the World Bank’s policy staff had initiated a review of xi priorities for disease control. Its purpose was to iden- global risk—but particularly a risk to lower-income tify effective yet affordable responses to the epidemics countries—that warrants inclusion on the macroeco- of HIV/AIDS and noncommunicable disease while nomic policy agenda. expanding successes in control of childhood infection. • Establishing mechanisms for social insurance— Work began on the WDR 1993 while the priorities review insurance that enables income security in old age; was drawing to a close. The detailed analyses of value that provides a financial safety net against permanent for money in that review provided strong intellectual disability, against transitory job loss, and against underpinnings for the WDR 1993. Oxford University inadequate earning power; and that provides finan- Press published the WDR 1993 and the first edition cial protection against medical expenses. DCP3’s of Disease Control Priorities in Developing Countries at extended cost-effectiveness analysis introduces an about the same time (Jamison, Mosley, Measham, and approach to efficient purchase of financial protection Bobadilla 1993; World Bank 1993). against medical expenses. On the occasion of the 20th anniversary of publica- • Allocation of resources within and across those tion of the WDR 1993, The Lancet invited me to chair sectors where efficient levels of investment require a commission to reassess health policies in light of two substantial public finance. These sectors include decades of remarkable change (mostly for the good) in much of physical infrastructure, research, education, health and related institutions around the world. Global environmental protection and population health. Health 2035, the report of the Lancet Commission on Investing in Health (Jamison, Summers, and others DCP3’s methods and conclusions provide critical guid- 2013) took stock of those changes and drew policy ance on resource allocation to and within the health implications for coming decades. Perhaps the most sector. Spending the resources available for health important message from Global Health 2035 is that investments on the wrong interventions is worse than our generation, uniquely in history, has the resources inefficient: it costs lives. As DCP3’s findings make clear, and knowledge to close most of the enormous health huge variation remains in how many lives can be saved gap between rich and poor within a generation. The from a million dollars spent on different interventions. work of the Lancet Commission provided a policy Transferring resources from low- to high-yield health framework for this concluding volume of the third interventions is, therefore, a moral imperative. Nor edition of Disease Control Priorities (DCP3). For evi- should resources available to the health sector be taken dence-oriented decision makers in ministries and in as given. Careful consideration of the social returns to development agencies, and for a broader community, increasing the health sector’s share of national budgets the DCP series has provided (as it did for Global and of national income suggests that, in many coun- Health 2035) a wealth of information relevant to tries, macroeconomic policy makers underinvest in informing policies for improving health and reducing health. health-related poverty. My own career has centered on macroeconomic Let me close by placing DCP3 into a context not just policy and on research to improve macroeconomic of health policy formulation but also of macroeconomic policy. Over the years I have increasingly come to feel policy formulation. Macroeconomic policy encompasses that getting health policy right contributes importantly three major components: to improving the social insurance and public sector investment dimensions of macroeconomic policy. For • Establishing and enforcing an environment for secure this reason, I have closely followed the 20-year evolu- and inclusive economic growth. Creating this envi- tion of the disease control priorities agenda. This new ronment includes finance of domestic and interna- edition continues DCP’s tradition of informing the tional security, enforcement of contracts and property efficient selection of health interventions. And it extends rights, regulation of cross border flows (goods and that agenda to informing choices where health policy services, capital, persons), and establishing the broad can contribute to poverty reduction as well as health structure and regulation of the financial system. improvement. Global warming and the risk of severe pandemics pose particular challenges to long-term economic growth. In chapter 18 of this volume, I report work REFERENCES undertaken with several colleagues that assesses the Fan, V. Y, D. T. Jamison, and L. H. Summers. 2018. “The magnitude of pandemic influenza risk (Fan, Jamison, Loss from Pandemic Influenza Risk.” In Disease Control and Summers 2018). Suffice it to say that low prob- Priorities (third edition): Volume 9, Disease Control ability but potentially devastating pandemics pose a Priorities: Improving Health and Reducing Poverty, xii Introduction edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, Developing Countries. New York: Oxford University Press R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, for the World Bank. DC: World Bank. Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, Farmer, P. 2010. “Rethinking Health and Human Rights: Time S. Berkley, and others. 2013. “Global Health 2035: A World for a Paradigm Shift.” In Partner to the Poor: A Paul Farmer Converging within a Generation.” The Lancet 382 (9908): Reader, chapter 21, 435–70, edited by H. Saussy. Berkley, Los 1898–1955. Angeles, and London: University of California Press. World Bank. 1993. World Development Report: Investing Jamison, D. T., W. H. Mosley, A. R. Measham, and J. L. in Health. New York: Oxford University Press for the Bobadilla, eds. 1993. Disease Control Priorities in World Bank. Introduction xiii Preface Budgets constrain choices. Policy analysis helps deci- financial risk protection objective of health systems. sion makers achieve the greatest value from limited In populations lacking access to health insurance or available resources. In 1993, the World Bank pub- prepaid care, medical expenses that are high relative to lished Disease Control Priorities in Developing Countries income can be impoverishing. Where incomes are low, (DCP1), an attempt to systematically assess the cost- seemingly inexpensive medical procedures can have effectiveness (value for money) of interventions that catastrophic financial effects. DCP3 offers an approach would address the major sources of disease burden (extended cost-effectiveness analysis, or ECEA) to in low- and middle-income countries (LMICs). The explicitly include financial protection as well as the World Bank’s 1993 World Development Report on health distribution across income groups of financial and drew heavily on DCP1’s findings to conclude that spe- health outcomes resulting from policies (for example, cific interventions against noncommunicable diseases public finance) to increase intervention uptake. DCP3 were cost-effective, even in environments where high provides interested policymakers with evidenced-based burdens of infection and undernutrition remained top findings on financial as well as health interventions to priorities. assist with resource allocation. DCP2, published in 2006, updated and extended This volume of DCP3, volume 9, places the findings DCP1 in several aspects, including explicit consider- from the first eight volumes into a framework identi- ation of the implications for health systems of expanded fying an efficient pathway toward essential universal intervention coverage. One way health systems expand health coverage (EUHC) through the identification of coverage is through selected platforms that deliver 21 essential packages that include health interventions, interventions that require similar logistics but address and fiscal and intersectoral policies. The intervention heterogeneous health problems. Platforms often pro- packages are defined by groups with common profes- vide a more natural unit for investment than do sional interests (for example, child health or surgery) individual interventions. Analysis of the costs of pro- and include interventions delivered across a range of viding platforms—and of the health improvements platforms. The volume also provides an up-to-date they can generate in given epidemiological environ- summary of levels and trends in deaths by cause and ments—can help to guide health system investments an early attempt to assess which elements of disease and development. burden most contribute to impoverishment. While DCP3 differs importantly from DCP1 and DCP2 most of DCP3’s 21 packages of interventions are devel- by extending and consolidating the concepts of plat- oped in the first eight volumes, several of the packages forms and by offering explicit consideration of the are presented here, including discussion of pandemic xv preparedness. Along with these new elements, DCP3 series. We convey our acknowledgements elsewhere in this updates the efforts of DCP1 and DCP2 to synthesize volume. Here we express our particular gratitude to the cost-effectiveness analysis of health interventions. Bill & Melinda Gates Foundation for its sustained financial The overall convergence of many countries and support, to the University of Washington’s Department of international development partners around the UN Global Health for hosting DCP3’s Secretariat, and to the Global Goals for 2030 has raised in particular the need World Bank, the original home for the DCP series and for careful analytic work that informs priorities and accomplished publisher of its products. choices. DCP3 stands unique in taking on this chal- lenge, providing analyses of the contributions of 218 Dean T. Jamison health system interventions and 71 intersectoral policies Hellen Gelband grouped into 21 essential packages. Sue Horton DCP3 is a large-scale enterprise involving an interna- Prabhat Jha tional community of authors, editors, peer reviewers, and Ramanan Laxminarayan research and staff assistants who contributed their time Charles N. Mock and expertise to the preparation and completion of this Rachel Nugent xvi Preface Abbreviations ACE Advisory Committee to the Editors AIDS acquired immune deficiency virus ANM auxiliary nurse midwifery ARI acute respiratory illness ART antiretroviral therapy BCA benefit-cost analysis BMI body mass index CCT conditional cash transfer CEA cost-effectiveness analysis CEPI Coalition for Epidemic Preparedness Innovations CHE catastrophic health expenditure CHWs community health workers CLAS local health administration communities COPD chronic obstructive pulmonary disease CPD continuing professional development CRS Creditor Reporting System CVD cardiovascular disease DAC Development Assistance Committee DALY disability-adjusted life year DCP2 Disease Control Priorities in Developing Countries (second edition) DCP3 Disease Control Priorities (third edition) DOTs directly observed treatment, short course ECEA extended cost-effectiveness analysis EP essential package EPHF Essential Public Health Functions EQA external quality assurance EUHC essential universal health coverage FRP financial risk protection FTE full-time equivalent Gavi Gavi, the Vaccine Alliance GBD global burden of disease GBHS global burden of health-related suffering GDP gross domestic product xvii GHE Global Health Estimates GST goods and services tax HICs high-income countries HIV human immunodeficiency virus HIV/AIDS human immunodeficiency virus/acquired immune deficiency syndrome HPP highest-priority package IAMP Interacademy Medical Panel ICD International Classification of Diseases ICER incremental cost-effectiveness ratio ICU intensive care unit IHD ischemic heart disease IHME Institute of Health Metrics and Evaluation IHR International Health Regulations INEGI National Institute of Statistics and Geography (Mexico) IPCC Intergovernmental Panel on Climate Change LC-GAPCPC The Lancet Commission on Global Access to Palliative Care and Pain Control LICs low-income countries LIS laboratory information systems LMICs low- and middle-income countries LPG liquefied petroleum gas MDGs Millennium Development Goals MERS Middle East respiratory syndrome MICs middle-income countries MTB/RIF mycobacterium tuberculosis/rifampicin NAM National Academy of Medicine NCDs noncommunicable diseases NCEF National Clean Energy Fund NGO nongovernmental organization NTCP National Tobacco Control Programme NTDs neglected tropical diseases ODA official development assistance OECD Organisation for Economic Co-operation and Development OOP out of pocket P4P pay for performance PDS public distribution system PEF Pandemic Emergency Financing Facility PEPFAR U.S. President’s Emergency Plan for AIDS Relief POCT point-of-care testing PPA public-private alliance PPP purchasing power parity PT proficiency testing QALYs quality-adjusted life years QIDS Quality Improvement Demonstration Study RBF results-based financing R&D research and development RNTCP Revised National Tuberculosis Control Program SARA Service Availability and Readiness Assessment SARS severe acute respiratory syndrome xviii Abbreviations SDGs Sustainable Development Goals SMU standard mortality unit SRS Sample Registration System SSB sugar-sweetened beverage SSRI selective serotonin reuptake inhibitor STIs sexually transmitted diseases TB tuberculosis UHC universal health coverage UMICs upper-middle-income countries UN United Nations UNAIDS Joint United Nations Programme on HIV/AIDS UNPD United Nations Population Division UPF universal public finance USAID U.S. Agency for International Development VPD vaccine preventable disease VSL value per statistical life VSLr VSL-to-income ratio VSMU value of a standard mortality unit WHO World Health Organization WHO-CHOICE Choosing Interventions That Are Cost-Effective YLGs years of life gained Abbreviations xix Part 1 Objectives and Conclusion of Disease Control Priorities, Third Edition Chapter 1 Universal Health Coverage and Intersectoral Action for Health Dean T. Jamison, Ala Alwan, Charles N. Mock, Rachel Nugent, David A. Watkins, Olusoji Adeyi, Shuchi Anand, Rifat Atun, Stefano Bertozzi, Zulfiqar Bhutta, Agnes Binagwaho, Robert Black, Mark Blecher, Barry R. Bloom, Elizabeth Brouwer, Donald A. P. Bundy, Dan Chisholm, Alarcos Cieza, Mark Cullen, Kristen Danforth, Nilanthi de Silva, Haile T. Debas, Peter Donkor, Tarun Dua, Kenneth A. Fleming, Mark Gallivan, Patricia Garcia, Atul Gawande, Thomas Gaziano, Hellen Gelband, Roger Glass, Amanda Glassman, Glenda Gray, Demissie Habte, King K. Holmes, Susan Horton, Guy Hutton, Prabhat Jha, Felicia Knaul, Olive Kobusingye, Eric Krakauer, Margaret E. Kruk, Peter Lachmann, Ramanan Laxminarayan, Carol Levin, Lai Meng Looi, Nita Madhav, Adel Mahmoud, Jean-Claude Mbanya, Anthony R. Measham, María Elena Medina-Mora, Carol Medlin, Anne Mills, Jody-Anne Mills, Jaime Montoya, Ole Norheim, Zachary Olson, Folashade Omokhodion, Ben Oppenheim, Toby Ord, Vikram Patel, George C. Patton, John Peabody, Dorairaj Prabhakaran, Jinyuan Qi, Teri Reynolds, Sevket Ruacan, Rengaswamy Sankaranarayanan, Jaime Sepúlveda, Richard Skolnik, Kirk R. Smith, Marleen Temmerman, Stephen Tollman, Stéphane Verguet, Damian Walker, Neff Walker, Yangfeng Wu, and Kun Zhao INTRODUCING DISEASE CONTROL Oii 1984), for China (Jamison and others 1984), and in a PRIORITIES, THIRD EDITION New England Journal of Medicine Shattuck Lecture In 1993, the World Bank published Disease Control (Evans, Hall, and Warford 1981). Mexican scholars Priorities in Developing Countries (DCP1), an attempt to (Bobadilla and others 1993; Frenk and others 1989) systematically assess value for money (cost-effectiveness) pointed to the rapid growth of NCDs in Mexico and of interventions that would address the major sources of introduced the concept of a protracted epidemiological disease burden in low- and middle-income countries transition involving a dual burden of NCDs combined (LMICs) (Jamison and others 1993). A major motivation with significant lingering problems of infectious disease. for DCP1 was to identify reasonable responses in highly The dual burden paradigm remains valid to this day. The resource-constrained environments to the growing bur- World Bank’s first (and so far only) World Development den of noncommunicable diseases (NCDs) and of Report (1993) dealing with health drew heavily on find- human immunodeficiency virus/acquired immune defi- ings from DCP1 to conclude that a number of specific ciency syndrome (HIV/AIDS) in LMICs. The World interventions against NCDs (including tobacco control Bank had highlighted the already substantial NCD prob- and multidrug secondary prevention of vascular disease) lem in country studies for Malaysia (Harlan, Harlan, and were attractive even in environments where substantial Corresponding author: Dean T. Jamison, University of California, San Francisco, California, United States; djamison@uw.edu. 3 burdens of infection and insufficient dietary intake major investment case for health (Stenberg and others remained policy priorities (World Bank 1993). 2017) continue to utilize platforms and their costs as Disease Control Priorities, second edition (DCP2), pub- important organizing concepts. lished in 2006, updated and extended DCP1 most notably This chapter conveys the main findings of DCP3, by explicit consideration of implications for health sys- and in particular its conclusions concerning intersec- tems of expanded coverage of high-priority interventions toral policy priorities and essential universal health (Jamison and others 2006). One important link to health coverage (EUHC). Like its two predecessors, DCP3’s systems was through examination of selected platforms broad aim is to assist decision makers in allocating for delivering logistically related interventions that might often tightly constrained budgets so that health address quite heterogeneous sets of problems. Platforms system objectives are maximally achieved. Beyond examined included the district hospital as a whole, the informing policy discourse, the granularity of analy- surgical and emergency room platforms within the dis- sis reported in DCP3’s nine volumes is intended to trict hospital, and school-based platforms for delivering a serve officials within ministries at the implementa- range of services. Platforms often provide a more natural tion level. Beginning with DCP3 volume 1 on Essential unit for investment—and for estimating costs—than do Surgery, DCP3’s first eight volumes (and related individual interventions. Analysis of the costs of provid- overviews of six of them in The Lancet) appeared ing platforms—and of the health improvements they can between 2015 and 2017. This final volume contains generate in a given epidemiological environment— cross-cutting and synthesizing chapters. Box 1.1 lists can thus help guide health system investments and devel- DCP3’s nine volumes and their editors. opment. Both Disease Control Priorities, third edition DCP3 differs importantly from DCP1 and DCP2 in (DCP3), and the World Health Organization’s (WHO) terms of its multivolume format, in terms of extending Box 1.1 DCP3’s Nine Volumes The World Bank has published DCP3 in 2015– Volume 3: Cancer, edited by Hellen Gelband, 2018. In contrast to the single (very large) volume Prabhat Jha, Rengaswamy Sankaranarayanan, and formats of DCP1 and DCP2, DCP3 appeared in Susan Horton, with a foreword by Amartya Sen nine smaller topical volumes, each with its own set Volume 4: Mental, Neurological, and Substance Use of editors. Coordination across volumes is provided Disorders, edited by Vikram Patel, Dan Chisholm, by seven series editors: Dean T. Jamison, Rachel Tarun Dua, Ramanan Laxminarayan, and María Elena Nugent, Hellen Gelband, Susan Horton, Prabhat Medina-Mora, with a foreword by Agnes Binagwaho Jha, Ramanan Laxminarayan, and Charles N. Mock. The topics and editors of the individual volumes Volume 5: Cardiovascular, Respiratory, and Related are as follows: Disorders, edited by Dorairaj Prabhakaran, Shuchi Anand, Thomas Gaziano, Jean-Claude Mbanya, Volume 1: Essential Surgery, edited by Haile T. Debas, Yangfeng Wu, and Rachel Nugent, with a foreword by Charles N. Mock, Atul Gawande, Dean T. Jamison, K. Srinath Reddy Margaret E. Kruk, and Peter Donkor, with a foreword by Paul Farmer Volume 6: Major Infectious Diseases, edited by King K. Holmes, Stefano Bertozzi, Barry R. Bloom, and Volume 2: Reproductive, Maternal, Newborn, and Prabhat Jha, with a foreword by Peter Piot Child Health, edited by Robert E. Black, Ramanan Laxminarayan, Marleen Temmerman, and Neff Walker, Volume 7: Injury Prevention and Environmental with a foreword by Flavia Bustreo Health, edited by Charles N. Mock, Rachel Nugent, box continues next page 4 Disease Control Priorities: Improving Health and Reducing Poverty Box 1.1 (continued) Olive Kobusingye, and Kirk R. Smith, with a Volume 9: Disease Control Priorities: Improving foreword by Ala Alwan Health and Reducing Poverty, edited by Dean T. Jamison, Hellen Gelband, Susan Horton, Prabhat Volume 8: Child and Adolescent Health and Jha, Ramanan Laxminarayan, Charles N. Mock, Development, edited by Donald A. P. Bundy, Nilanthi and Rachel Nugent, with a foreword by Bill and de Silva, Susan Horton, Dean T. Jamison, and Melinda Gates and an introduction by Lawrence H. George C. Patton, with a foreword by Gordon Brown Summers. Figure 1.1 Policies for Health Intersectoral Health sector policies policies (including financial protection policies) Access to and uptake of health interventions Quality of delivery of health interventions To reduce physiological To improve health risk factors outcomes To reduce behavioral To provide financial *** *** and environmental protection from risk factors Stunting Child deaths health costs Overweight Adult premature Anemia deaths Hypertension Short- and long-term disability Dislipidemia Pain and distress High blood glucose Other and consolidating the concept of platforms, and in affect the level of physiological risks and health out- terms of explicit consideration of a broad range of inter- comes directly. The health sector’s role in reducing sectoral and fiscal policies for health. Figure 1.1 illus- behavioral and environmental risk is viewed as modest— trates the division of DCP3’s analyses between rather the health sector’s main role is in reducing intersectoral policies and health sector policies and (some of) the physiological risk factors and reducing the shows examples of the risk factors and conditions that duration and severity of health conditions and their the policies address. Importantly, the DCP3 structure sequelae. Appropriate health sector policies also offer views the role of intersectoral action to be reduction of the potential for reducing health-related financial risks behavioral and environmental risks, which themselves in a population. Universal Health Coverage and Intersectoral Action for Health 5 DCP3 has four major objectives that go beyond pre- sheer magnitude of improvement. As recently as 1910, vious editions. The first is to address explicitly the finan- Chilean life expectancy fell below 32 years. By 2012, life cial risk protection and poverty reduction objective of expectancy exceeded 78 years. Second, time has nar- health systems, as well as other objectives such as provi- rowed cross-country differences. In 1910, world leaders sion of contraception, reduction in stillbirths, and palli- (such as Australia and New Zealand) achieved life ative care or enhancement of the physical and cognitive expectancies almost 30 years greater than Chile, but by development of children. Standard health metrics such 2010 that gap had narrowed to around 4 years. The as the quality-adjusted life year (QALY) and disability- magnitude of Chile’s success has been unusual, but the adjusted life year (DALY) fail to encompass these other broad story it conveys is not. That said, Sub-Saharan objectives of health systems, and DCP3 has endeavored Africa now lags 20 years behind global life expectancy of to be explicit about them and their importance. The 72 years, and countries in other regions (and regions second extension lies in systematic attention to the inter- within large countries) remain similarly disadvantaged. sectoral determinants of health. DCP3’s main purpose is to provide information to help The third major way that DCP3 goes beyond previous close those gaps. editions lies in organizing interventions into 21 essential Income growth in the past century and past decades packages reflecting professional communities. Table 1.1 has contributed to increased life expectancy as has, to a lists DCP3’s 21 packages. DCP3 defines a concept of somewhat greater extent, improvements in education EUHC in the health systems components of the essential levels (Pradhan and others 2017). Most improvements, 21 packages. DCP3 further identifies a subset of EUHC, however, have resulted from an ever-expanding menu of the highest-priority package (HPP), that can potentially drugs, diagnostics, vaccines, and knowledge (Jamison, be afforded by low-income countries (LICs) and that Jha, and others 2013). Nurturing continuation of the offers the most potential achievement (given limited scientific investment therefore remains a policy priority, resources) of health, financial protection, and other as was extensively discussed in DCP2 (Bloom and others objectives. Finally, DCP3 provides estimates for low- and 2006; Mahmoud and others 2006; Meltzer 2006; lower-middle-income countries of incremental and total Weatherall and others 2006). DCP3 has devoted less costs in 2030 for both EUHC and HPP and of the mag- attention to research and development (R&D) than nitude of their impact on mortality. In addition to these did DCP2—in part because of the coverage there. new elements, DCP3 updates the efforts of DCP1 and While R&D is discussed in several places (for example, DCP2 to assemble and interpret the literature on eco- Bundy and others 2017; Trimble and others 2015), a nomic evaluation of health interventions. careful mining of DCP3 for its implications for R&D This chapter introduces the substantive topics remains to be done. addressed by DCP3 and relays our main conclusions. Valuation of mortality decline (or health change Before turning to that, we briefly describe the context in more generally) is excluded from the global system of which DCP3’s analyses have been undertaken. national income and product accounts. Economists have nonetheless expended substantial effort tracing the effect of health improvements on household and national income and in assessing the value of the small CONTEXT reductions in mortality risk that have occurred year by Five considerations set the context for DCP3: (a) the year. Global Health 2035 (GH2035), the report of the 20th-century revolution in human health, (b) the scien- Lancet Commission on Investing in Health (Jamison, tific underpinnings of that revolution, (c) the high Summers, and others 2013), reviewed and extended estimated returns to (carefully chosen) health invest- the literature on the value of health improvements. ments, and (d) the increasing implementation of univer- That literature points to high returns indeed. The sal health coverage (UHC) as a practical goal for domestic Copenhagen Consensus, a project that comparatively finance of health systems. Skolnik (2016) provides fur- assesses returns across all major development sectors, ther discussion of these four issues. A fifth consideration has likewise found high returns: its 2012 assessment concerns evolution in the thinking about the interna- found that 9 of the 15 highest return investments were tional dimension of health finance—development assis- health-related, including all of the top 5 (Kydland and tance for health broadly defined. others 2013). Chile exemplifies the two key elements of the 20th- As national incomes rise, countries typically increase century revolution in human health. One is the the percentage of national income devoted to health. 6 Disease Control Priorities: Improving Health and Reducing Poverty Equally significantly, they increase the proportion of An objective of each DCP3 volume was to define one or health expenditures that are prepaid, usually through more essential packages and the interventions in that public or publicly mandated finance. WHO’s leadership package that might be acquired at an early stage on the in advancing a global UHC agenda has accelerated this pathway to UHC. The essential packages comprise inter- underlying movement of political systems toward UHC. ventions that provide value for money, are implementable, Dr. Tedros Ghebreyesus, WHO’s new Director-General, and address substantial needs. has reaffirmed the WHO commitment to UHC and to Platforms are defined as logistically related delivery the use of evidence and data in support of achieving that channels. DCP3 groups EUHC interventions within goal (Ghebreyesus 2017). GH2035 advocated variants on packages that can be delivered on different types of a pathway toward UHC, “progressive universalism,” that platforms. The temporal character of interventions is emphasized two initial priorities for action: (a) universal critical for health system development. Patients requir- coverage of publicly financed interventions and ing nonurgent but substantial intervention—repair of (b) reductions of user payments at the point of service to cleft lips and palates is an example—can be accumu- very low levels (Jamison, Summers, and others 2013). lated over space and time, enabling efficiencies of high With inevitable constraints on public budgets, these two volume in service delivery. Urgent interventions, priorities point to the need for initial selectivity in the which include a large fraction of essential surgical range of interventions to be publicly financed, the so- interventions, are ideally available 24/7 close to where called benefits package. Many considerations will influ- patients live—with important implications for disper- ence national choices of how benefits packages will evolve sal of relevant platforms and integration of different over time and on the appropriate pathways to universal- services. Nonurgent but continuing interventions to ism. Hence, the importance of maintaining the focus on address chronic conditions (for example, secondary the highest priority health investments as DCP3 is prevention of vascular disease or antiretroviral ther- intended to facilitate. apy for HIV–positive individuals) provide a major and With substantial income growth in most LMICs and quite distinct challenge. One new product of DCP3 has an increasing number of countries committed to public been to explicitly categorize all essential interventions finance of UHC, the role of development assistance is into one of these three temporal categories and to being reexamined (Bendavid and others 2018; Jamison, draw relevant lessons, including concerning cost, for Summers, and others 2013). As the World Bank and health systems. others have long argued, finance ministers will often In total, 71 distinct and important intersectoral poli- reduce domestic allocations to sectors receiving substan- cies for reducing behavioral and environmental risk were tial foreign aid. The challenge to those concerned with identified, and 29 of those were identified as candidates aid effectiveness thus becomes one of identifying and for early implementation. In addition to intersectoral supporting important activities that national finance policies, DCP3 reviews policies that affect the uptake of ministries are likely to underfinance (such as R&D, health sector interventions (such as conditional cash pandemic preparedness, and control of antimicrobial transfers) and the quality with which they are delivered resistance). A recent assessment found that support for (Peabody and others 2018). these international functions already constitutes more than 20 percent of development assistance broadly METHODS defined; the authors make the case that percentage should steadily increase over time (Schäferhoff and oth- DCP3’s authors have thoroughly updated findings from ers 2015). This view of development assistance has clear DCP2 on costs, effectiveness, and cost-effectiveness. implications for the construction of model benefits The literature provides much of specific interest, but packages for domestic finance; other things being equal, formulation of policy, when informed by evidence at all, domestic finance needs to emphasize services having requires expert judgment to fill extensive gaps in the minimal international externalities. literature. The first subsection of this section discusses DCP3’s approach. The second and third subsections discuss methods of economic evaluation and DCP3’s PACKAGES, PLATFORMS, AND POLICIES extension of standard methods to include analysis of the DCP3 defines packages of interventions as conceptually financial protection objectives of health systems. The related interventions—for example, those dealing with final subsection discusses the process of formulation of cardiovascular disease or reproductive health or surgery. DCP3’s packages. Universal Health Coverage and Intersectoral Action for Health 7 Table 1.1 DCP3 ’s Clusters of Essential Packages Packages Age-related cluster 1. Maternal and newborn health; 2. Child health; 3. School-age health and development; 4. Adolescent health and development; 5. Reproductive health and contraception Infectious diseases cluster 6. HIV and STIsa; 7. Tuberculosis; 8. Malaria and adult febrile illnessb; 9. Neglected tropical diseases; 10. Pandemic and emergency preparedness Noncommunicable disease 11. Cardiovascular, respiratory, and related disorders; 12. Cancer; 13. Mental, neurological, and substance and injury cluster use disorders; 14. Musculoskeletal disorders; 15. Congenital and genetic disorders; 16. Injury prevention; 17. Environmental improvementc Health services cluster 18. Surgery; 19. Rehabilitation; 20. Palliative care and pain control; 21. Pathology Note: HIV = human immunodeficiency virus; STIs = sexually transmitted infections. a. Most forms of hepatitis are in part sexually transmitted and hence control of hepatitis is included in this package. b. Dengue is included among adult febrile illnesses. c. Environmental improvements affect the incidence of risk factors both for infectious and for noncommunicable disease. We include them under the noncommunicable disease and injury cluster because the more significant consequences lie there. Use of Evidence (GDP) as a macroeconomic indicator. The health dash- Using research (or other) evidence to guide policy is most board is likewise a natural step beyond use of cost-effec- simply done when randomized controlled trials of the tiveness league tables in constructing health benefit relevant intervention (or mix of interventions) have been packages, an approach consistent with that of Glassman, undertaken on the population of interest in the appropri- Giedion, and Smith (2017). ate ecological setting. Even in high-income countries, such strong evidence is rarely available. In lower-income envi- ronments, the problem of the quality of evidence is Protecting against Financial Risk compounded. As always, evidence must be used to help In populations lacking access to health insurance or pre- decision makers (a) avoid adopting interventions that paid care, medical expenses that are high relative to income don’t work in a given context and (b) avoid rejecting those can be impoverishing (figure 1.2 illustrates mechanisms). that do. Box 1.2 discusses the DCP3 thinking on this issue. Where incomes are low, seemingly inexpensive medical procedures can be catastrophic. WHO’s World Health Report 2010 documented the (very substantial) magnitude Economic Evaluation of medical impoverishment globally and pointed to the The methods and findings of DCP3’s approaches to eco- value of universal health coverage for addressing both nomic evaluation appear in three separate chapters of the health and the financial protection needs of popula- this volume: one on cost-effectiveness, one on benefit- tions (WHO 2010). Most of the literature on medical cost analysis, and one on extended cost-effectiveness impoverishment fails to identify the medical conditions analysis (Horton 2018; Chang, Horton, and Jamison responsible. Essue and others (2018) point to where spe- 2018; Verguet and Jamison 2018). Table 1.2 provides a cific causes of medical impoverishment information are high-level overview. Several of the entries in that table— known, an obviously central point for construction of covering value for money, dashboards, and extended benefits packages. cost-effectiveness analysis—point to the desirability of Although multiple studies document the overall multicriteria decision analysis of the sort explored by magnitude of medical impoverishment, most economic Youngkong (2012) and others. evaluations of health interventions and their finance The bottom row of table 1.2 takes the multioutcome (including those in DCP1 and DCP2) have failed to extended cost-effectiveness analysis (ECEA) approach address the important question of efficiency in the pur- one step further to discussion of the “dashboard” chase of financial protection. In work undertaken for DCP3 uses to help inform and structure setting priories. DCP3, an approach was developed—ECEA—to explicitly This health dashboard concept is a natural extension of include financial protection and equity in economic eval- the dashboard approach that Stigliz, Sen, and Fitoussi uation of health interventions. Smith (2013) has devel- (2010) propose to go beyond gross domestic product oped an approach that addresses the same concern from 8 Disease Control Priorities: Improving Health and Reducing Poverty Box 1.2 Evidence for Policy: From Research Findings to Policy Parameters Analysis in DCP3 proceeds by attempting to make 2. Combined interventions. As in the malaria exam- the best use of the evidence available for informing ple, assume that evidence shows interventions important decisions rather than exclusively using A and B are both effective. What about A + B? what ideally generated evidence has to say (Jamison Is the combination’s effect the sum of the sep- 2015). The distinction is important. An example arate effects? Or are the two substitutes? Hard illustrates. Quite good evidence is available on the evidence on combinations is far more rare than effect of vector control on malaria mortality in evidence on individual interventions. specific environments. Likewise there is strong 3. Changed settings. Assume we have strong evi- evidence concerning treatment efficacy. Very little dence that intervention A works in environment evidence, however, exists on how different mixes of Y, for example, that antimalarial bednets reduce vector control and treatment affect mortality, but all causes of child mortality when mosquitos bite this is the important question for policy. indoors at night, at moderate intensity. Good evi- dence concludes that bednets were effective where Inevitably imperfectly, our task in the Disease Control evaluated, but other, biological considerations sug- Priorities series, beginning with the first edition, has gest that that evidence be rejected in an environ- been to combine the (sometimes) good science ment with very high-biting intensity. Economists about unidimensional intervention in very specific have discussed this point in the context of “external locales with informed judgment to reach reasonable validity.” Ozler (2013) provides a clear overview. conclusions about the effect of intervention mixes 4. Trait-treatment interactions. Finally, patient char- in diverse environments. To put this in a slightly acteristics may differ. Measles immunization in different way: the parameters required for assessing healthy child populations may have been shown policy differ, often substantially, from what has been to have no effect on mortality rates. Generalizing addressed (so far) in the research literature. The that finding to a population with different traits transition from research findings to policy parameters (for example, undernourished or sickly children) requires judgment to complement the research and, might and in this case would generate an unfor- often, a consideration of underlying mechanisms tunate false negative. (for example, use of incentives) that might suggest generalizability (Bates and Glennerster 2017). ∗∗∗ In particular, four types of judgments were often Evidence can be weak. Or, as in the examples needed in the course of DCP3 to make the transition above, evidence can be strong but only partially from research findings to evidence for policy. relevant. Often weak evidence for effectiveness, or Examples illustrate: partially relevant evidence for effectiveness, is like- wise weak evidence concerning lack of effectiveness. 1. Similar interventions. Assume we have evidence Interpreting weak evidence as grounds for reject- that intervention A is effective, and we believe ing an intervention could generate false negatives intervention B is quite similar. (Think of two that cost lives. The attempt in DCP3 has been lipid-lowering agents.) We use judgment to infer to unashamedly combine evidence with informed that intervention B is (or perhaps is not) also judgment in order to judiciously balance false posi- effective. tives and false negatives. a different perspective. ECEA is the approach that DCP3 used to evaluate tobacco taxation and regulatory policies used to address issues of both reduction in financial risk (Verguet and Jamison 2018). An important implication and distribution across income groups of financial as well of the ECEA evaluations of tobacco taxation in China as health outcomes resulting from policies, such as public and in Lebanon was that such taxation, when the full finance, to increase intervention uptake. ECEA has been range of consequences is considered, is progressive in Universal Health Coverage and Intersectoral Action for Health 9 Table 1.2 Economic Evaluation Methods Economic method Costs Consequences 1.1 Cost-effectiveness analysis (CEA) • Social costsa • Changes in specific outcomes (child deaths, Horton (2018) overviews DCP3’s findings on CEA. new HIV infections) Wilkinson and others (2016) and Sanders and others • Changes in aggregated measures (2016) provide recent guidelines for health CEA. (YLL, QALY, DALY) Jamison (2009) provided earlier guidelines that pointed to inclusion of financial protection outcomes and nonfinancial constraints in CEA. 1.2 Value-for-money assessment • Social costsa Important outcomes of health sector intervention Value-for-money assessment of health sector interventions are not measurable in mortality or DALY terms includes CEA but acknowledges the CEA is irrelevant for (and are therefore excluded from CEA) include some health sector outcomes. the following: • Contraception provided • Stillbirths averted • Palliative care • IQ or stature enhanced. 1.3. Extended cost-effectivess analysis (ECEA) • Costs are viewed • Consequences are reported from a Verguet and Jamison (2018) overview of DCP3 ’s findings separately from distributional perspective (for example, on ECEA. perspectives of provider, by gender, income, or membership in a patient, and society. disadvantaged group). See, for example, Asaria, Griffin, Cookson, and others (2015). • Valuation of financial risk protection is included. a 1.4. Benefit-cost analysis (BCA) • Social costs • Changes in income or gross domestic product Chang, Horton, and Jamison (2018) overview of DCP3 ’s • Changes in income plus the monetary value of findings on BCA. change in mortality (or health) 1.5. Economic dashboard • As with ECEA • Poverty reduction consequences or insurance DCP3 ’s judgments about interventions to include in ECEA value are explicitly considered. and in the HPP involved combining multiple strands of • Distribution of costs and consequences across evidence. While intervention cost-effectiveness was income quintiles are explicitly considered. typically most important, in the end judgments involved • Dashboard contains a fuller and more considering a dashboard of information including disease disaggregated list of consequences burden, value for money assessment, ECEA, and BCA. than ECEA, which is itself much more Stiglitz, Sen, and Fitoussi (2010) propose making this comprehensive than CEA. dashboard explicit and the primary guide to decision making in the macroeconomic context. Note: DALY = disability-adjusted life year; DCP3 = Disease Control Priorities third edition; HIV = human immunodeficiency virus; HPP = highest-priority package; IQ = intelligence quotient; QALY = quality-adjusted life year; YLL = years of life lost. a. Social costs refer to the value of real resources used to implement an intervention. For example, if a health ministry needs to pay import taxes on pharmaceuticals, the social cost is the pretax cost not the posttax cost, as the tax simply represents a transfer (from the health to the finance ministry). Taxation itself is often considered by economists to involve a real cost (the so-called deadweight loss from taxation) arising from distortion of prices and hence decisions of actors in the economy. DCP3 follows standard practice in health-related CEA in not considering deadweight losses from taxation. Inclusion of deadweight losses as currently assessed would typically increase the cost per unit of outcome by 50 to 70 percent. terms of health outcomes and unlikely to be regressive in The tobacco ECEAs suggest a more general point terms of financial outcomes (Salti, Brouwer, and Verguet about government policies to provide populations with 2016; Verguet and others 2015). A 13-country ECEA of protection against financial risk. Policy can operate either tobacco taxation found results similar to those from upstream or downstream. Upstream provision of finan- China and Lebanon (Jha and Global Tobacco Economics cial risk protection (FRP) attenuates the need for costly Consortium 2017). medical intervention. Upstream measures include 10 Disease Control Priorities: Improving Health and Reducing Poverty Figure 1.2 Financial Risk Protection Financial Consequences of Pathways architecture inadequate FRP Reduction in savings Ongoing out-of-pocket Impoverishment payments Asset sales Lack of adequate Inadequate health insurance medical care Borrowing (at high rates) Anxiety from bearing financial risk Reduced income or school participation Note: FRP = financial risk protection. prevention, early treatment, and investment in improved appear as an annex to this chapter, and chapters 2 and 3 medical technologies (see Lakdawalla, Malani, and Reif provide a full discussion of methods. Several interven- 2017). Most health systems emphasize downstream mea- tions appear in more than one package as the final lists sures through payment for expensive procedures in the of 71 intersectoral policies, and 218 EUHC interventions hospital. Downstream measures will always be needed. remove this duplication. A consequence is that the cost That said, resource constraints will sharply limit public of EUHC is less than the sum of the costs of the packages finance of downstream financial protection; provision within it. only of downstream measures perverts incentives in the obvious way and in many (but not all) cases upstream measures more efficiently purchase FRP given budget INTERSECTORAL POLICIES FOR HEALTH constraints. Eleven of DCP3’s 21 packages contain a total of 71 inter- sectoral policies. These policies fall into four broad Construction of Packages categories: taxes and subsidies (15 of 71), regulations Editors of DCP3 volumes and authors of specific chap- and related enforcement mechanisms (38 of 71), built ters in volume 9—on rehabilitation (Mills and others environment (11 of 71), and information (7 of 71). 2018), on pathology (Fleming and others 2018), on pal- These policies are designed to reduce the population liative care (Krakauer and others 2018) and on pandemic level of behavioral and environmental risk factors— preparedness (Madhav and others 2018)—constructed tobacco and alcohol use, air pollution, micronutrient the 21 essential packages listed in table 1.1. The series deficiencies in the diet, unsafe sexual behavior, editors and authors of this paper then consolidated those excessive sugar consumption, and others (figure 1.1). policies and formats into a common level of aggregation Watkins, Nugent, and others (2018) provide a thorough and a common structure (for example, screening was overview of DCP3’s findings on intersectoral policy. not considered an intervention by itself but only in con- Here we highlight several of DCP3’s points: junction with the indicated response). This generated a First, at initially low levels of income, the levels of set of harmonized essential packages. The originals many risk factors rise with income, creating headwinds Universal Health Coverage and Intersectoral Action for Health 11 against which health sector policy must proceed. These on the health potential for removing subsidies remains rises are at least potentially countered by sound policy. limited. But the sheer magnitude of some of these subsi- We identify 29 of 71 intersectoral policies to be well dies, as the International Monetary Fund has stressed, worth considering for early adoption. points to the value of careful further analysis. In all Second, for important categories of risk, such as pollu- likelihood, a country’s finance ministry is the most tion and transport risks, there are multiple sources of the important ministry (after health) for improving popula- risk, each of which is addressed through different tion health. And many—not all—of the measures it can modalities. Rather than a clear set of “first priorities,” there take can enhance public sector revenue. are multiple country- or site-specific actions to be taken. Perhaps the single most important point to note is that the ESSENTIAL UNIVERSAL HEALTH COVERAGE success of many high-income countries in reducing these risks to very low levels points to the great potential that The heart of DCP3 consisted of reviewing available evi- these multiple policies can have for dealing, in particular, dence on health sector interventions’ costs, effectiveness, with air pollution and road traffic injuries. ability to be implemented, and capacity to deliver signifi- A third point of importance is that fiscal policies— cant outcomes. DCP3’s nine volumes provide granular finance ministry policies—are likely of key significance. overviews of this evidence, overviews directed to Discussion of these policies has most prominently the implementation community as well as to the policy involved taxes on tobacco, alcohol, and sugar-sweetened community. Chapter 3 of volume 9 provides an integra- beverages. But the possibilities for taxation are broader: tive overview (Watkins, Jamison, and others 2018). sugar production and imports, fossil fuels (or carbon), Figure 1.3 provides a schema of how DCP3 defines and industrial or vehicle emissions. Also of importance EUHC. Beyond EUHC is the full range of available, is reducing expensive subsidies that now exist on fossil efficacious health sector interventions, or UHC. fuels and often on unhealthy food production or While no country publicly finances all interventions, unhealthy child dietary supplements. While health many high-income countries come close and can rea- improvement may be only one of several objectives for sonably be described as having achieved UHC. Short of lowering subsidies, it is an important one. The literature EUHC is what DCP3 labels the HPP. Individual coun- tries’ highest priorities will differ from our model list for multiple reasons. That said, the HPP is intended to pro- vide a useful starting point for national or subnational Figure 1.3 Essential Universal Health Coverage and Highest-Priority assessments. As with EUHC, DCP3 provides estimates Packages for the cost and effects of EUHC. GH2035 (Jamison, Summers, and others 2013) pointed to the possibility of a “grand convergence,” across most countries, in our lifetimes, in levels of under-age-five mortality and major infections. Figure 1.3 illustrates grand convergence in the DCP3 structure. The two following subsections provide our estimates of the costs and mortality-reducing conse- Infection and NCDs quences of EUHC. inadequate and dietary injury intake Costs We generated two estimates of costs for the health system component of each of DCP3’s 21 packages. The first was Grand Highest priority an estimate of how much additional funding it would convergence NCD and injury take—in the 2015 cost and demographic environment— (GC) interventions to implement each package to the extent judged Universal health coverage (UHC) feasible. The packages were designed so that for most Essential universal health coverage (EUHC) cases, “full” implementation, defined as 80 percent effec- Highest priority package (includes GC plus tive coverage, was judged feasible by 2030. The second highest priority NCD and injury interventions) estimate of cost was of total cost for the package, defined as incremental cost plus the amount already (in 2015) Note: NCD = noncommunicable disease. The “grand convergence” agenda for reducing child and infectious disease mortality was advanced by the Lancet Commission on Investing in Health (Jamison, being spent on the intervention. These costs were esti- Summers, and others 2013). mated both for LICs and for lower-middle-income 12 Disease Control Priorities: Improving Health and Reducing Poverty countries. Some interventions were included in several required would require substantial reallocation of public packages, which was a natural outcome of a package sector priorities (Jamison, Summers, and others 2013). formulation process that delineated packages as areas of In principle, projections could be made of changes in concern to specific professional communities, such as both the tradable and nontradable components of cost, surgeons or reproductive health specialists. Eliminating of the responsiveness of costs to demography (and in this duplication resulted in 218 distinct EUHC interven- particular to fertility decline), and on whether improved tions. This implies that the sum of the package costs will transport and other infrastructure might reduce our exceed the cost of providing all packages. The subset of estimates of the cost of expanding coverage to ever-more EUHC that was judged by explicit criteria to be highest difficult-to-reach parts of the population. In a country- priority (the HPP) was costed in the same way as for specific context, this might well be worthwhile. But for EUHC. All these costs are the estimated costs associated purposes of reasonable overall cost estimates we judge with expanding coverage in the 2015 environment, an that adding these layers of assumption would add little environment for which we have substantial, if incom- or nothing to the information content of table 1.3. plete, information without making assumptions about Table 1.4 presents our cost assessments divided along the evolution of costs and epidemiology over time. Costs two other relevant dimensions. Panel a provides esti- should be interpreted as long-term steady state costs, that mates of the costs associated with each platform, and is, costs that include (a) training of staff to replace retire- about half of our calculated costs occur at the health ments and (b) investment to counter depreciation of center level. For EUHC, another 15 to 25 percent each of equipment and facilities. incremental expenditures would go to the first-level Table 1.3 reports the calculated expenditure increases hospital and to the community level. Panel b reports required above baseline and expresses those numbers as intervention cost estimates by degree of urgency. a percentage of gross national income (GNI). (Chapter 3, The health systems implications for increasing interven- volume 9, of DCP3 reports costs by package.) We con- tion coverage differ markedly by urgency. Continuing sider it reasonable to think of the costs in 2030 of EUHC interventions require appropriate community capacity and the HPP in these percentage terms (as well as in for delivery. Examples include antiretroviral therapy or numbers of dollars). Only a small fraction of reasonably antihypertensive therapy. A full half of incremental costs anticipated economic growth in most countries would are needed to finance continuing, very long-term inter- cover the incremental costs of EUHC, although achiev- vention. Urgent interventions—for example, for trauma ing the increased percentage of gross national income or obstructed labor—require that first-level hospitals be Table 1.3 Total and Incremental Annual Costs of Essential UHC and the Highest Priority Package, 2015 (in 2012 US$) Low-income countriesa Lower-middle-income countriesa HPP EUHC HPP EUHC 1. Incremental annual cost (in billions, 2012 US$) $23 $48 $82 $160 b 2. Incremental annual cost per person (in US$) $26 $53 $31 $61 3. Total annual cost (in billions, US$) $38 $68 $160 $280 4. Total annual cost per personc (in US$) $42 $76 $58 $110 5. Incremental annual cost as a share of current 3.1% 6.4% 1.5% 2.9% GNI per personb 6. Total annual cost (as percentage of current 5.1 9.1 2.8 5.2 GNI per person)d Source: Watkins, Jamison, and others 2018. Note: EUHC = essential universal health coverage; GNI = gross national income; HPP = highest-priority package. a. This paper uses the World Bank’s 2014 income classification of countries. As a country’s income changes, its classification can also change; for example, both Bangladesh and Kenya moved from low- to lower-middle income after 2014. b. Incremental annual cost is the estimated cost of going from current to full (80%) coverage of the EUHC and HPP interventions. The total annual cost is the incremental cost plus the cost of the current level of coverage assuming the same cost structure for current as for incremental coverage. Estimated costs are inclusive of estimates for (large) health system strengthening costs and are steady state (or long-term average) costs in that investments to achieve higher levels of coverage and to cover depreciation are included. c. The 2015 population of low-income countries was 0.90 billion. For lower-middle-income countries, it was 2.7 billion. d. The 2015 GNI of low-income countries was $0.75 trillion. For lower-middle-income countries, it was $5.6 trillion. Universal Health Coverage and Intersectoral Action for Health 13 Table 1.4 Incremental Costs of the HPP and EUHC by Platform and by Intervention Urgency, Percent Low-income countries Lower-middle-income countries HPP (percent) EUHC (percent) HPP (percent) EUHC (percent) (a) Incremental costs by platform, percentage of total Population based 0.57 2.3 0.6 2.0 Community 18 16 12 14 Health center 50 52 57 52 First-level hospital 25 25 22 25 Referral and specialty hospitals 6.4 5.2 9.1 6.1 100 100 100 100 (b) Incremental costs by intervention urgency, percentage of total Urgent 35 28 27 24 Continuing 41 48 50 52 Nonurgent 24 24 23 24 100 100 100 100 Source: Watkins, Jamison, and others 2018. Note: EUHC = essential universal health coverage; HPP = highest-priority package. accessible quickly (Reynolds and others 2018). About number of premature deaths by 40 percent, where pre- one-quarter to one-third of incremental costs are required mature is defined as under age 70 years. Subgoals were to to provide this capacity. Nonurgent (but potentially reduce under-age-five and major infectious disease important) interventions (for example, cataract extrac- deaths by two-thirds and NCD and injury deaths by tion) allow patients to be accumulated over space and one-third. time with concomitant potential for efficiency and qual- Our approach in DCP3 followed the approach of ity resulting from high volume. Norheim and others (2015) in broad terms but inserts into it our “counterfactual 2015” analysis. We start with a baseline age distribution of deaths by age and (broad) Mortality Reduction from Essential UHC cause generated from the UNPD’s projected 2030 age DCP3 generated estimates of mortality in 2015, as well as distribution of population and age combined with estimates for a “counterfactual 2015” and of how many cause-specific death rates from 2015 (Mathers and fewer deaths would have occurred following implementa- others 2018). We then estimate the effect of EUHC (and tion of EUHC and the HPP. This analysis thus provides a HPP) on mortality by assuming that the underlying reasoned estimate of the costs and consequences of using— intervention packages are implemented over the 15 years in the 2015 demographic context—today’s medical and from 2015 to 2030. (The packages were designed to public health technology as fully as reasonably possible (as make this assumption reasonable.) The age- and well as associated cost-effectiveness estimates). This subsec- cause-specific mortality rates from counterfactual 2015 tion discusses estimates of mortality reduction. were then applied to the UNPD 2030 age distributions to Norheim and others (2015) developed a struc- give the age distributions of death by cause estimated to ture—40x30—for thinking about mortality reduction result from implementation of EUHC. goals for the Sustainable Development Goal (SDG) These calculations enable comparison of the EUHC period. Their starting point was the United Nations mortality profile to an explicit counterfactual base- Population Division’s (UNPD) projected age distribu- line. Table 1.5 shows these comparisons for EUHC tion of population in 2030 and an age distribution of and for the HPP. What we can see from this compari- deaths generated from that age distribution of popula- son is that full implementation of the HPP could tion and age-specific mortality rates from 2010. The achieve about half of the 40x30 goal. Full implemen- overall 40x30 goal was, then, to reduce the calculated tation of EUHC could achieve about two-thirds of the 14 Disease Control Priorities: Improving Health and Reducing Poverty Table 1.5 Implementation of DCP3’s Essential Packages: Estimated Reduction in Premature Deaths in 2030a (in Millions) Low-income countriesb Lower-middle-income countriesb Projected Expected reduction in Projected Expected reduction in number of 40x30 premature deaths from number of 40x30 premature deaths from Age group or premature reducton premature reducton condition deaths, 2030 targetc HPP EUHC deaths, 2030 targetc HPP EUHC By age group 0–4 2.2 1.5 0.62 0.77 3.3 2.2 1.1 1.3 5–69 5.2 1.5 0.99 1.2 14 4.8 2.2 2.9 0–69 7.4 3.0 1.6 2.0 17 7.0 3.2 4.2 By cause (age 5+)d Group I 1.9 0.76 0.59 0.65 3.2 1.5 0.85 0.94 Tuberculosis 0.34 0.22 0.11 0.13 0.90 0.60 0.29 0.35 HIV/AIDS 0.44 0.29 0.18 0.20 0.48 0.32 0.23 0.26 Malaria 0.087 0.058 0.051 0.051 0.055 0.037 0.026 0.026 Maternal conditions 0.17 0.11 0.075 0.086 0.20 0.13 0.079 0.026 Other diseases 0.90 0.074 0.18 0.18 1.6 0.40 0.22 0.22 Group II 2.5 0.60 0.36 0.53 8.9 2.7 1.3 1.9 Neoplasms 0.65 0.22 0.010 0.039 1.8 0.60 0.10 0.16 Cardiovascular 0.93 0.31 0.24 0.36 4.0 1.3 0.89 1.4 diseases Other diseases 0.93 0.076 0.11 0.13 3.2 0.80 0.28 0.35 Group III 0.77 0.13 0.043 0.060 2.0 0.54 0.070 0.10 Road injuries 0.25 0.085 0.032 0.046 0.57 0.19 0.048 0.069 Other injuries 0.52 0.042 0.010 0.014 1.4 0.36 0.022 0.032 Sources: Watkins, Norheim, and others 2017; Watkins, Qi, and others 2017; Watkins, Jamison, and others 2018. Note: EUHC = essential universal health coverage; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome; HPP = highest-priority package. All estimates are in millions of deaths. The 40x30 reduction target includes a 40 percent reduction in deaths ages 0-69 overall; a two-thirds reduction in under-age-five deaths and adult deaths from tuberculosis, HIV/AIDS, malaria, and maternal conditions; and a one-third reduction in deaths from major noncommunicable diseases. The quantitative targets above reflect these goals; however, targets for the residual categories (“other diseases” and “other injuries”) have been calculated in light of the targets for specific causes of death so that the total number of target deaths for ages 5–69 is sufficient to meet the 40x30 target. a. A death under age 70 is defined as premature. b. This paper uses the World Bank’s income classification of countries. c. A reduction target of 40x30 is defined as a 40 percent reduction in premature deaths by 2030, relative to the number that would have occurred had 2015 death rates persisted to 2030. The United Nations Population Prospects (UN 2017) median population projection for 2030 was used to provide the population totals for calculating deaths by age and sex. d. World Health Organization’s Global Health Estimates provided the 2015 cause distributions of deaths for these calculations (Mathers and others 2018). 40x30 goal. In a sensitivity analysis, Watkins, Norheim, assumption of a 2 percent rate of technical progress in and others (2018) demonstrate that higher levels of one of their scenarios—then the reduction in deaths coverage (on the order of 95 percent) and more opti- from EUHC could be more substantial than shown in mistic assumptions about the quality and efficiency of this table. Such progress is certainly possible, but may intervention delivery could acheive the 40x30 goal be unlikely. Likewise there could be more than antici- in lower-middle-income countries and exceed it by pated reduction in behavioral and environmental risk. about 20 percent in low-income countries. If we were Our model is estimating what is technically and eco- to assume that both tools and implementation capac- nomically feasible given today’s tools. The results are ity improve over the period to 2030—Global Health indeed substantial—and are viable options for deci- 2035 (Jamison, Summers, and others 2013) made an sion makers. But required resources are substantial, Universal Health Coverage and Intersectoral Action for Health 15 and at realistic (that is, 80 percent) coverage levels the DCP3 reached six broad conclusions: goals are incompletely met. The actual decision to commit resources remains, of course, in the hands of 1. DCP3 has found it useful to organize interven- national authorities. tions into 21 essential packages that group the interventions relevant to particular professional communities. Each package can contain both inter- CONCLUSIONS sectoral interventions and health system interven- DCP3 has been a large-scale enterprise involving multi- tions. Specific findings from packages point to the ple authors, editors, and institutions. The first volume attractiveness of widely available surgical capacity, appeared in 2015 and the last of the nine volumes is the value of meeting unmet demand for contracep- being published at the beginning of 2018. The volumes tion, the potential of a multipronged approach to air appear as serious discussion continues about quantify- pollution and the importance of maintaining invest- ing and achieving SDGs, including SDG 3 for health. ment in child health and development far beyond DCP3’s analyses complement those of GH2035 and the first 1000 days. WHO’s recent assessments of the cost of attaining SDG 3 2. Interventions were selected for packages by a (Jamison, Summers, and others 2013; Stenberg and others systematic process using criteria of value for money, 2017). Each of these analyses addresses somewhat different burden addressed, and implementation feasibility. questions (table 1.6), but the broad results they convey are Collectively, the selected interventions are defined mutually supportive. to constitute “essential” universal health coverage Table 1.6 Comparison of Global Health 2035, DCP3, and WHO 2017 Resource Estimates for Costs and Consequences of Large Scale Investment in Health Systems Global Health 2035 DCP3 WHO 2017 1. Countries included 34 low-income and 3 (large) lower- 34 low-income and 49 lower-middle 67 low-, lower-middle, and upper- middle-income countries. Separate income countries. Separate estimates middle-income countries individually estimates for the low- and lower- for the low- and lower-middle-income estimated and then aggregated. middle-income countries groups are countries groups are provided. Reported results are for all included provided. countries combined. 2. Key definitions and Grand convergence (GC) interventions • 21 packages of care (table 1.1) • Investments were modeled for intervention range are defined as ones leading to very are identified in terms that 16 SDGs, including 187 health covered substantial crosscountry convergence include intersectoral and health interventions and a range of in under age 5, maternal, sector interventions (71 distinct health system strengthening tuberculosis, malaria, and HIV/AIDS intersectoral interventions strategies (the latter of which mortality and in the prevalence of and 218 distinct health included investments required to neglected tropical diseases (NTDs). sector interventions). achieve target levels of health • Essential universal health workforce, facilities, and other coverage (EUHC) is defined as health system building blocks). health sector interventions in the • Two scenarios were modeled, 21 packages (covered in national a progress scenario (in which health accounts and potentially coverage is limited by the included in benefits packages). absorptive capacity of current • A highest priority subset of systems to incorporate new EUHC. The highest-priority interventions) and an ambitious package (HPP) includes the scenario (in which most GC interventions but goes countries achieve high levels beyond it, including a limited of intervention coverage and range of interventions against hence SDG targets). noncommunicable diseases (NCDs) and injuries, and cross-cutting areas such as rehabilitation and palliative care. table continues next page 16 Disease Control Priorities: Improving Health and Reducing Poverty Table 1.6 Comparison of Global Health 2035, DCP3, and WHO 2017 Resource Estimates for Costs and Consequences of Large Scale Investment in Health Systems (continued) Global Health 2035 DCP3 WHO 2017 3. Intersectoral action Extensive discussion of intersectoral Intersectoral interventions defined WHO 2017 scenarios include for health actions for health but not included in as those typically managed and some finance of intersectoral modeling grand convergence. financed outside the health sector. interventions, from the health Each of the 21 packages contains the sector perspective, as well as intersectoral interventions deemed their effects on mortality. relevant. The costs and effects of intersectoral action on mortality reduction not explicitly modelled. 4. Intervention Full coverage defined as 85%; rates Full coverage defined as 80%. Full coverage defined as 95% coverage of scale-up defined using historical The HPP differs from EUHC not in for most interventions in the data on “best performers” among coverage rate but in the scope of ambitious scenario, with a range similar groups of countries. interventions included. from 53–99% depending on the intervention. 5. Estimated additional For low-income countries in 2035: Low-income countries, 2030: Low-income countries: $64 billion costs (including US$30 billion annually between 2016 HPP—US$23 billion/year in 2030. requisite investment and 2030. EUHC—US$48 billion/year Lower-middle-income countries: in health system For lower-middle-income countries in Lower middle-income countries, $185 billion in 2030. capacity), in US$ 2035: US$61 billion per year. 2030: HPP—US$82 billion/year (Costs presented in 2014 US$) EUHC—US$160 billion per year. (Costs presented in 2012 US$) 6. Estimated deaths For low-income countries: 4.5 million Low-income countries: 2.0 million Low-income countries: 2.9 million averted a, b, c deaths averted per year between premature deaths averted in 2030. deaths averted in 2030. 2016 and 2030. Lower-middle-income countries: 4.2 Lower-middle-income countries: For lower-middle-income countries: million premature deaths averted 6.1 million deaths averted in 2030. 5.8 million deaths averted per year in 2030. between 2016 and 2030. Sources: Global Health 2035: Jamison, Summers, and others 2013; Boyle and others 2015. DCP3: Watkins, Qi, and others 2017; Watkins, Norheim, and others 2017. Stenberg and others 2017. Note: HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome; SDGs = Sustainable Development Goals. a. DCP3 reports the number of premature deaths averted, that is, deaths under age 70. b. Averted deaths included stillbirths averted in GH2035 and WHO 2017, but not in DCP3. c. For GH2035 and DCP3 the reported deaths averted included only deaths averted among children actually born. Family planning averts unwanted pregnancies and hence potential deaths of children from those pregnancies who were never born. The difference is major. For low-income countries, a GH2035 sensitivity analysis estimated that the more comprehensive figure was 7.5 million deaths averted rather than the 4.5 million shown in the table. The WHO 2017 headline numbers do include deaths averted from pregnancies averted but sensitivity analyses were undertaken. Ambitious scale-up of family planning services accounted for 50 percent of averted child and maternal deaths and over 65 percent of averted stillbirths in the WHO analysis (K. Stenberg 2017, personal communication). or EUHC. A subset of 97 of these interventions, 3. The costs estimated for the HPP and EUHC are substan- selected using more stringent criteria, are suggested tial. The HPP is, however, affordable for LICs prepared as the highest-priority package or HPP, constitut- to commit to rapid improvement in population health, ing an important first step on the path to EUHC. and the EUHC is affordable for lower-middle-income Five platforms—from population-based through countries. Many upper-middle-income countries have the referral hospital—provide the delivery base for yet to achieve EUHC and they, too, might find that 218 health sector interventions. The specific inter- the EUHC interventions are a useful starting point for ventions selected for the HPP and for EUHC and the discussion. definitions of platforms and packages are necessarily 4. The goal of a 40 percent reduction in premature quite generic. Every country’s definitions and selec- deaths by 2030 (Norheim and others 2015), 40x30, tions will differ from these and from each other’s. represents a goal for mortality reduction closely Nonetheless, we view DCP3’s selections as a poten- mirroring the quantitative content of SDG 3. Our tially useful model—as a starting point for what are calculations suggest that implementing EUHC or appropriately country-specific assessments. the HPP by 2030 will make substantial progress Universal Health Coverage and Intersectoral Action for Health 17 toward 40x30. Higher levels of coverage than we have • Low-income countries (LICs) = US$1,045 or less assumed here would be required to reach 40x30, but • Middle-income countries (MICs) are subdivided: this might be a realistic target for some early-adopter (a) lower-middle-income = US$1,046 to US$4,125. UHC countries. (b) upper-middle-income (UMICs) = US$4,126 to US$12,745. • High-income countries (HICs) = US$12,746 or more. 5. DCP3 has shown that it is possible to identify the main sources of health-related financial risk and impover- ishment to estimate the value of risk reduction and to use ECEA to help achieve efficiency in purchase of REFERENCES risk reduction. Although DCP3 has made a beginning Asaria, M., S. Griffin, R. Cookson, S. Whyte, and P. Tappenden. in applying these methods, much remains to be done. 2015. “Distributional Cost-Effectiveness Analysis of Health 6. In addition to the aggregate conclusions of DCP3 just Care Programmes: A Methodological Case Study of the summarized, each volume provides rich detail on policy UK Bowel Cancer Screening Programme.” Health Economics options and priorities. This granularity in the volumes 24: 742–54. makes them of use to the implementation level of gov- Bates, M. A, and R. Glennerster. 2017. “The Generalizability ernment ministries as well as the policy level. Puzzle.” Stanford Social Innovation Review (summer). https://ssir.org/articles/entry/the_generalizability_puzzle. Bendavid E., T. Ottersen, P. Liu, R. Nugent, N. Padian, and others. 2018. “Development Assistance for Health.” In Disease ACKNOWLEDGMENTS Control Priorities (third edition): Volume 9, Improving We wish to acknowledge three institutions that have Health and Reducing Poverty, edited by D. T. Jamison, played key roles in DCP3. One is the World Bank, original H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, home for the DCP series and accomplished publisher of and R. Nugent. Washington, DC: World Bank. its products. In the World Bank, Carlos Rossel and Mary Bloom, B. R., C. M. Michaud, J. R. La Montagne, and L. Simonsen. 2006. “Priorities for Global Research and Fisk oversaw the editing and publication of the series and Development of Interventions.” In Disease Control served as critical champions for DCP3. The second is the Priorities in Developing Countries, second edition, Interacademy Medical Panel (IAMP) and its U.S. affiliate, edited by D. T. Jamison, J. G. Breman, A. R. Measham, the National Academy of Medicine (NAM). IAMP/NAM G. Alleyne, M. Claeson, D. B. Evans, P. Jha, A. Mills, and have organized a peer review process to cover chapters in P. Musgrove. Washington, DC: World Bank and Oxford the nine volumes; they established an Advisory Committee University Press. to the Editors (ACE), chaired by Anne Mills, that has been Bobadilla, J. L., J. Frenk, R. Lozano, T. Frejka, C. Stern, and of enormous value. The Department of Global Health of others. 1993. “The Epidemiologic Transition and Health the University of Washington has provided a congenial Priorities.” In Disease Control Priorities in Developing home for DCP for the past five years. We wish in particu- Countries, first edition, edited by D. T. Jamison, W. H. lar to acknowledge the intellectual and practical support Mosley, A. R. Measham, and J. L. Bobadilla. New York: Oxford University Press. of the department’s two chairs during that period: King Boyle, C. F., C. Levin, A. Hatefi, S. Madriz, and N. Santos. Homes and Judith Wasserheit. We also wish to acknowl- 2015. “Achieving a ‘Grand Convergence’ in Global Health: edge Brianne Adderley, Shamelle Richards, and Nazila Modeling the Technical Inputs, Costs, and Impacts from Dabestani for their management, administrative, and 2016 to 2030.” PLOS ONE 10 (10). research support to the production of DCP3. Bundy, D. A. P., N. de Silva, S. Horton, D. T. Jamison, and G. C. Patton, eds. 2017. Child and Adolescent Health and Development. Volume 8, Disease Control Priorities (third ANNEX edition), edited by D. T. Jamison, R. Nugent, H. Gelband, S. Horton, P. Jha, R. Laxminaryan, and C. N. Mock. The following annex to this chapter is available at http:// Washington, DC: World Bank. www.dcp-3.org/DCP. Chang, A., S. Horton, and D. T. Jamison. 2018. “Benefit-Cost Analysis in Disease Control Priorities, Third Edition.” In • Annex 1A: Essential Packages as They Appear in Disease Control Priorities (third edition): Volume 9, DCP3 Volumes 1 through 9 Improving Health and Reducing Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, NOTE DC: World Bank. World Bank Income Classifications as of July 2014 are as Essue, B. M., T.-L. Laba, F. M. Knaul, A. Chu, H. V. Minh, follows, based on estimates of gross national income (GNI) and others. 2018. “Economic Burden of Chronic Ill per capita for 2013: Health and Injuries for Households in Low- and 18 Disease Control Priorities: Improving Health and Reducing Poverty Middle-Income Countries.” In Disease Control Priorities Jamison, D. T., W. H. Mosley, A. R. Measham, J. L. Bobadilla, (third edition): Volume 9, Improving Health and Reducing eds. 1993. Disease Control Priorities in Developing Countries, Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, first edition. New York: Oxford University Press. P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent. Jamison, D. T., P. Jha, V. Malhotra, and St. Verguet. 2013. Washington, DC: World Bank. “Human Health: The Twentieth-Century Transformation Evans, J. R., K. L. Hall, and J. Warford. 1981. “Shattuck of Human Health—Its Magnitude and Value.” In How Lecture; Health Care in the Developing World: Problems Much Have Global Problems Cost the World?: A Scorecard of Scarcity and Choice.” New England Journal of Medicine from 1900 to 2050, edited by B. Lomborg. New York: 305: 1117–27. Cambridge University Press. Fleming, K., M. Naidoo, M. Wilson, J. Flanigan, S. Horton, Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, and others. 2018. “High Quality Diagnosis: An Essential S. Berkley, and others. 2013. “Global Health 2035: A World Pathology Package.” In Disease Control Priorities (third Converging within a Generation.” The Lancet 382 (9908): edition): Volume 9, Improving Health and Reducing Poverty, 1898–955. edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, Jha, P., and Global Tobacco Economics Consortium. 2017. R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, The Health, Poverty and Financial Consequences of a Large DC: World Bank. Tobacco Price Increase among 0.5 Billion Male Smokers in Frenk, J., J. L. Bobadilla, J. Sepúlveda, and M. Lopez-Cervantes. 13 Low- and Middle-Income Countries. Toronto: Centre for 1989. “Health Transition in Middle-Income Countries: New Global Health Research. Challenges for Health Care.” Health Policy and Planning Krakauer, E., Z. Ali, H. Arreola, A. Bhadelia, S. Connor, and 4 (1): 29–39. others. 2018. “Palliative Care.” In Disease Control Priorities Glassman, A., U. Giedion, and P. C. Smith. 2017. What’s (third edition): Volume 9, Improving Health and Reducing In, What’s Out: Designing Benefits for Universal Health Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, Coverage. Washington, DC: Center for Global Development. P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent. Ghebreyesus, T. A. 2017. “All Roads Lead to Universal Health Washington, DC: World Bank. Coverage.” Commentary. The Lancet, July 17. Kydland, F. E., R. Mundell, T. Schelling, V. Smith, and N. Stokey. Harlan, W. R., L. C. Harlan, and W. L. Oii. 1984. “Changing Disease 2013. “Expert Panel Ranking” In Global Problems, Smart Patterns in Developing Countries: The Case of Malaysia.” Solutions: Costs and Benefits, edited by B. Lomborg, 701–16. In Health Information Systems, edited by P. Leaverton and Cambridge, UK: Cambridge University Press. L. Massi. New York: Praeger Scientific. Lakdawalla, D., A. Malani, and J. Reif. 2017. “The Insurance Horton, S. E. 2018. “Cost-Effectiveness in Disease Control Priorities, Value of Medical Innovation.” Journal of Public Economics Third Edition.” In Disease Control Priorities (third edition): 145: 94–102. Volume 9, Improving Health and Reducing Poverty, edited by Madhav, N., B. Oppenheim, M. Gallivan, P. Mulembakani, D. T. Jamison, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, E. Rubin, and others. 2018. “Pandemics: Risks, Mitigation, and C. N. Mock, and R. Nugent. Washington, DC: World Bank. Costs.” In Disease Control Priorities (third edition): Volume Jamison, D. T. 2009. “Cost-Effectiveness Analysis: Concepts and 9, Improving Health and Reducing Poverty, edited by Applications.” In Oxford Textbook of Public Health (fifth D. T. Jamison, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, edition). Volume 2, The Methods of Public Health, edited by C. N. Mock, and R. Nugent. Washington, DC: World Bank. R. Detels, J. McEwen, R. Beaglehole, and H. Tanaka, 767–82. Mahmoud, A., P. M. Danzon, J. H. Barton, and R. D. Mugerwa. Oxford, UK: Oxford University Press. 2006. “Product Development Priorities.” In Disease ———. 2015. “Disease Control Priorities: Improving Health Control Priorities in Developing Countries, second edition, and Reducing Poverty.” The Lancet. Comment. doi: 10.1016/ edited by D. T. Jamison, J. G. Breman, A. R. Measham, S0140-6736(15)60097-6. G. Alleyne, M. Claeson, D. B. Evans, P. Jha, A. Mills, and Jamison, D. T., A. Alwan, C. N. Mock, R. Nugent, D. A. Watkins, P. Musgrove. Washington, DC: Oxford University Press and others. 2018. “Universal Health Coverage and and World Bank. Intersectoral Action for Health.” In Disease Control Priorities Mathers, C., G. Stevens, D. Hogan, A. Mahanani, and J. Ho. (third edition): Volume 9, Improving Health and Reducing 2018. “Global and Regional Causes of Death: Patterns Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, and Trends, 2000–15.” In Disease Control Priorities (third P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent. edition): Volume 9, Improving Health and Reducing Poverty, Washington, DC: World Bank. edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, Jamison, D. T., J. G. Breman, A. R. Measham, G. Alleyne, R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, M. Claeson, D. B. Evans, P. Jha, A. Mills, and P. Musgrove, DC: World Bank. eds. 2006. Disease Control Priorities in Developing Countries, Meltzer, D. 2006. “Economic Approaches to Valuing Global second edition. Washington, DC: Oxford University Press Health Research.” In Disease Control Priorities in Developing and World Bank. Countries, second edition, edited by D. T. Jamison, Jamison, D. T., J. R. Evans, T. King, I. Porter, N. Prescott, and J. G. Breman, A. R. Measham, G. Alleyne, M. Claeson, others. 1984. “China: The Health Sector.” Country Study. D. B. Evans, P. Jha, A. Mills, and P. Musgrove. Washington, World Bank, Washington, DC. DC: World Bank and Oxford University Press. Universal Health Coverage and Intersectoral Action for Health 19 Mills, J. A., E. Marks, T. Reynolds, and A. Cieza. 2018. Needs in 67 Low-Income and Middle-Income Countries.” “Rehabilitation: Essential Along the Continuum of Care.” In The Lancet Global Health. doi:http://dx.doi.org/10.1016 Disease Control Priorities (third edition): Volume 9, Improving /S2214-109X(17)30263-2. Health and Reducing Poverty, edited by D. T. Jamison, Stigliz, J., A. Sen, and J. P. Fitoussi. 2010. Mis-Measuring Our H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, Lives: Why GDP Doesn’t Add Up. New York and London: and R. Nugent. Washington, DC: World Bank. The New Press. Norheim, O. F., P. Jha, K. Admasu, T. Godal, R. H. Hum, and Trimble, E. L., P. Rajaraman, A. Chao, T. Gross, and others. 2015. others. 2015. “Avoiding 40% of the Premature Deaths in “Need for National Commitments to Cancer Research to Each Country, 2010–30: Review of National Mortality Guide Public Health Investment and Practice.” In Disease Trends to Help Quantify the UN Sustainable Development Control Priorities (third edition): Volume 3, Cancer, edited Goal for Health.” The Lancet 385 (9964): 239–52. by H. Gelband, P. Jha, R. Sankaranarayanan, and S. Horton. Ozler, B. 2013. “Learn to Live without External Validity.” Washington, D.C.: World Bank. Development Impact (blog). World Bank, Washington, UN (United Nations). 2017. World Population Prospects: The 2017 DC. https://blogs.worldbank.org/impactevaluations/learn Revision. New York: Population Division, United Nations -live-without-external-validity. Department of Economic and Social Affairs. Peabody, J., R. Shimkhada, O. Adeyi, H. Wang, E. Broughton, Verguet, S., C. L. Gauvreau, S. Mishra, M. MacLennan, and and others. 2018. “Quality of Care.” In Disease Control others. 2015. “The Consequences of Tobacco Tax on Priorities (third edition): Volume 9, Improving Health and Household Health and Finances in Rich and Poor Smokers Reducing Poverty, edited by D. T. Jamison, H. Gelband, in China: An Extended Cost-Effectiveness Analysis.” The S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and Lancet Global Health 3 (4): e206–e216. R. Nugent. Washington, DC: World Bank. Verguet, S., and D. T. Jamison. 2018. “Health Policy Analysis: Pradhan, E., E. M. Suzuki, S. Martínez, M. Schäferhoff, and oth- Applications of Extended Cost-Effectiveness Analysis ers. 2017. “The Effects of Education Quantity and Quality Methodology in Disease Control Priorities: (third edition).” In on Child and Adult Mortality: Their Magnitude and Their Disease Control Priorities (third edition): Volume 9, Improving Value.” In Disease Control Priorities (third edition): Volume Health and Reducing Poverty, edited by D. T. Jamison, 8, Child and Adolescent Health and Development,” edited by H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, D. A. P. Bundy, N. de Silva, S. E. Horton, D. T. Jamison, and and R. Nugent. Washington, DC: World Bank. G. C. Patton. Washington, D.C.: World Bank. Verguet, S., R. Laxminarayan, and D. T. Jamison. 2015. Reynolds, T., H. Sawe, A. M. Rubiano, S. D. Shin, and others. “Universal Public Finance of Tuberculosis Treatment in 2018. “Strengthening Health Systems to Provide Emergency India: An Extended Cost-Effectiveness Analysis.” Health Care.” In Disease Control Priorities (third edition): Volume 9, Economics 24 (3): 318–32. Improving Health and Reducing Poverty, edited by D. T. Jamison, Watkins, D. A., D. T. Jamison, A. Mills, R. Atun, K. Danforth, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and others. 2018. “Universal Health Coverage and and R. Nugent. Washington, DC: World Bank. Essential Packages of Care.” In Disease Control Priorities Salti, N., E. D. Brouwer, and S. Verguet. 2016. “The Health, (third edition): Volume 9, Improving Health and Reducing Financial, and Distributional Consequences of Increases in Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, the Tobacco Excise Tax among Smokers in Lebanon.” Social P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent. Science and Medicine 170 (December): 161–69. Washington, DC: World Bank. Sanders, G. D., P. J. Neumann, A. Basu, D. W. Brock, and others. Watkins, D. A., O. F. Norheim, P. Jha, and D. T. Jamison. 2016. “Recommendations for Conduct, Methodological 2017. “Mortality Impact of Acheiving Essential Universal Practices, and Reporting of Cost-Effectiveness Analyses: Health Coverage in Low- and Lower Middle-Income Second Panel on Cost-Effectiveness in Health and Medicine.” Countries.” DCP3 Working Paper no. 21. World Bank, Journal of the American Medical Association 316 (10): Washington, DC. 1093–103. Watkins, D. A., R. A. Nugent, H. Saxenian, G. Yamey, and others. Schäferhoff, M., S. Fewer, J. Kraus, E. Richter, L. H. Summers, 2018. “Intersectoral Policy Priorties for Health.” In Disease and others. 2015. “How Much Donor Financing for Health Control Priorities (third edition): Volume 9, Improving Is Channelled to Global Versus Country-Specific Aid Health and Reducing Poverty, edited by D. T. Jamison, Functions?” The Lancet 386 (10011): 2436–41. H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, Skolnik, R. 2016. Global Health 101 (third edition). Burlington, and R. Nugent. Washington, DC: World Bank. MA: Jones & Bartlett Learning. Watkins, D. A., J. Qi, S. E. Horton, E. Brouwer, and Smith, P. C. 2013. “Incorporating Financial Protection others. 2017. “Costs and Affordability of Essential into Decision Rules for Publicly Financed Healthcare Universal Health Coverage in Low- and Middle-Income Treatments.” Health Economics 22 (2): 180–93. Countries.” DCP3 Working Paper no. 20. World Bank, Stenberg, K., O. Hanssen, T. Tan-Torres Edejer, M. Bertram, Washington, DC. and others. 2017. “Financing Transformative Health Weatherall, D., B. Greenwood, H. L. Chee, and P. Wasi. 2006. Systems Towards Achievement of the Health Sustainable “Science and Technology for Disease Control: Past, Present, Development Goals: A Model for Projected Resource and Future.” In Disease Control Priorities in Developing 20 Disease Control Priorities: Improving Health and Reducing Poverty Countries, second edition, edited by D. T. Jamison, Reference Case for Economic Evaluation: An Aid to J. G. Breman, A. R. Measham, G. Alleyne, M. Claeson, Thought.” Value Health 19 (8): 921–28. D. B. Evans, P. Jha, A. Mills, and P. Musgrove. Washington, World Bank. 1993. World Development Report 1993: Investing DC: World Bank and Oxford University Press. in Health. New York: Oxford University Press. https://open WHO (World Health Organization). 2010. The World Health knowledge.worldbank.org/handle/10986/5976. Report 2010: Health Systems Financing: The Path to Youngkong, S. 2012. “Multi-Criteria Decision Analysis for Universal Coverage. Geneva: WHO. Including Health Interventions in the Universal Health Wilkinson, T., M. J. Sculpher, K. Claxton, P. Revill, and Coverage Benefit Package in Thailand.” PhD dissertation, others. 2016. “The International Decision Support Initiative Radboud University, Nijmegen, the Netherlands. Universal Health Coverage and Intersectoral Action for Health 21 Chapter 2 Intersectoral Policy Priorities for Health David A. Watkins, Rachel Nugent, Helen Saxenian, Gavin Yamey, Kristen Danforth, Eduardo González-Pier, Charles N. Mock, Prabhat Jha, Ala Alwan, and Dean T. Jamison INTRODUCTION mechanisms comprising top-level policy makers to Many aspects of population health can be addressed enable health-related decisions to be made across gov- solely by services delivered through the health sector. ernment sectors (Buss and others 2016). The goal is to These services include health promotion and prevention create benefits across sectors by taking actions to sup- efforts as well as treatment and rehabilitation for specific port population health and beyond that, to ensure that diseases or injuries. At the same time, policies initiated even “nonhealth” policy decisions and implementation by or in collaboration with other sectors, such as agricul- have beneficial, or at least neutral, effects on determi- ture, energy, and transportation, can also reduce the nants of health. Intersectoral involvement increases the incidence of disease and injury, often to great effect. arsenal of available tools to improve health, helps ensure These policies can make use of several types of instru- that government policies are not at cross-purposes to ments, including fiscal measures (taxes, subsidies, and each other, and can generate sizable revenue (as in the transfer payments); laws and regulations; changes in the case of tobacco and alcohol taxes). built environment (roads, parks, and buildings); and Many countries do not practice a Health in All information, education, and communication campaigns Policies approach, and doing so is especially challenging (see chapter 1 of this volume, Jamison and others 2018). when there are extreme resource constraints, low capac- In addition, a range of non–health sector social services ity, and weak governance and communication structures can mitigate the consequences of ill health and provide (Khayatzadeh-Mahani and others 2016), as in many financial protection. These intersectoral policies that low- and middle-income countries (LMICs). As an promote or protect health, when implemented as part of alternative in these settings, a ministry of health could a coherent plan, can constitute a whole-of-government engage other sectors opportunistically and strategically approach to health (UN 2012). on specific issues that are likely to produce quick suc- Ideally, a whole-of-government approach to health cesses and have substantial health effects (WHO 2011a). would involve the systematic integration of health con- Thus, a concrete menu of policy options that are highly siderations into the policy processes of all ministries. effective, feasible, and relevant in low-resource environ- This collaborative approach is often termed Health in ments is needed. This need is particularly relevant in All Policies (Khayatzadeh-Mahani and others 2016). light of the ambitious targets specified in the United Some governments have achieved such collaboration Nations Sustainable Development Goals (SDGs) for by employing ministerial commissions or other 2030 (UN 2015). Rachel Nugent, Helen Saxenian, and Gavin Yamey are co-second authors for this chapter. Corresponding author: David A. Watkins, University of Washington, Seattle, Washington, United States; davidaw@uw.edu. 23 The Disease Control Priorities series has consistently in this chapter, DCP3 seeks to reinforce the importance of stressed the importance of intersectoral action for health these policy instruments and provide a template for and the feasibility of intersectoral action in LMICs. action for ministers of health when engaging other sectors Disease Control Priorities in Developing Countries, second and heads of state. edition (DCP2) (Jamison and others 2006), included chapters that emphasized intersectoral policies for spe- cific diseases, injuries, and risk factors, and it also HEALTH CONDITIONS AND RISK FACTORS included a chapter devoted to fiscal policy (Nugent and AMENABLE TO INTERSECTORAL ACTION Knaul 2006). Disease Control Priorities, third edition (DCP3), has reinforced many of these messages—usually Most of this chapter discusses policies that influence the with newer and stronger evidence—and has also distribution of selected risk factors for diseases and explored some emerging topics and new paradigms, injuries across the population (Jamison and others 2018). particularly for control of noncommunicable disease Risk factors fall into three broad categories: risk factors. Volume 7 of DCP3 is especially noteworthy in this respect: it provides a list of 111 policy recommen- 1. Individual personal characteristics. Important char- dations for prevention of injuries and reduction of envi- acteristics include an individual’s genetics (including ronmental and occupational hazards, 109 of which are epigenetic factors arising very early), age, height, almost entirely outside the purview of health ministers body mass index, blood lipid profile, blood pressure, to implement (Mock and others 2017). and many others. Although age and genetics cannot Despite the political barriers to developing an inter- be modified, they may provide information to guide sectoral agenda for health, this chapter contends that not medical treatment and behavior. only is intersectoral action a good idea for health—it is a 2. Diseases. Some diseases increase the risk of other dis- must. Much of the reduction in health loss globally over eases or increase their severity. Important examples the past few decades can be attributed to reductions in include diabetes, hepatitis, severe mood disorders, and risk factors such as tobacco consumption and unsafe malaria. In some cases, the burden from diseases as risk water that have been implemented almost exclusively by factors well exceeds their intrinsic burden. Diabetes is actors outside the health sector (Hutton and Chase 2017; one of the most prominent examples in this regard Jha and others 2015). An environment that increases (Alegre-Díaz and others 2016). health risks at early stages of industrial and urban 3. Behavior and environment. Important examples of growth often, although not always, gives way to a cleaner behavioral risk factors include diets that contribute natural environment at higher levels of per capita to adiposity and vascular risk; diets that contribute to income. Yet these risks can be associated with dramatic undernutrition; lack of exercise; unsafe sex; and abuse health losses along the way (Mock and others 2017). of addictive substances such as tobacco, alcohol, Furthermore, the health risks produced by advanced and narcotics. Important environmental risk factors industrialization—such as unhealthy diet and physical include air and water pollution and unsafe occupa- inactivity—require policy interventions across multiple tional and transport conditions. sectors if they are not to worsen substantially with eco- nomic development. This chapter’s main focus is on instruments of pol- This chapter is based on a close look at the intersectoral icy intended to change the third category of risk factors: policies recommended across the DCP3 volumes, and it behavior and environment. Changes in behavior and proposes 29 concrete early steps that countries with highly environment can influence disease incidence or severity constrained resources can take to address the major risks either directly or by modifying other risk factors. that can be modified. The chapter also touches on broader Interventions that address both individual personal social policies that address the consequences of ill health characteristics and diseases as risk factors are covered in and stresses that the need for such policies will increas- chapter 3 of this volume (Watkins and others 2018). ingly place demands on public finance. This chapter can be viewed as a complement to chapter 3 of this volume (Watkins and others 2018) concerning health sector inter- Conceptual Model for Interactions among ventions in the context of universal health coverage. It also Health Risks provides illustrative examples of successful health risk Behavioral and environmental risk factors can be disag- reduction through intersectoral policy and discusses vari- gregated into multiple specific risks, illustrating sources ous aspects of policy implementation. By synthesizing and pathways of risk exposure. The more disaggre- non–health sector policies separately and in greater depth gated set of risk factors outlined in figure 2.1 has two 24 Disease Control Priorities: Improving Health and Reducing Poverty Figure 2.1 Conceptual Model of Interactions among Key Risk Factors and Diseases That Can Be Modified Environmental • Lead or other chemical contamination • Climate change • Occupational hazards • Unsafe roads and vehicles • Outdoor air pollution • Household air pollution • Unsafe water and poor sanitation Cancers Injuries Infections Behavioral and dietary Cardiovascular • Excessive nutrient intake and respiratory disorders • Inadequate nutrient intake • Suboptimal breastfeeding • Risky sexual behavior • Physical inactivity • Tobacco use • Harmful alcohol use Mental, neurological, and substance use disorders • Harmful use of injection drugs and addictive substances striking features. First, multiple risk factors can overlap Magnitude of Health Loss from Specific Risk Factors and interact to influence the incidence of specific dis- There are theoretical and practical challenges to quanti- eases or injuries; for example, smoking, dietary risks, and fying the effect of specific risk factors on fatal and non- physical inactivity can all contribute to the development fatal outcomes. Comparative risk assessment is the most of ischemic heart disease (Ajay, Watkins, and Prabhakaran commonly used approach for this purpose, and its 2017). Second, single risk factors can be responsible for limitations have been reviewed elsewhere (Hoorn and a substantial fraction of cases of multiple diseases or others 2004). Whereas expanded direct measurement of injuries; for example, air pollution from outdoor sources deaths by cause has led to greater precision in mortality can lead to chronic obstructive pulmonary disease and estimates in recent years, especially in LMICs (Jha 2014), asthma, among other conditions (Smith and Pillarisetti methods and data sources that can be used to quantify 2017). One implication of these interactions is that risk factor–attributable mortality are much less devel- aggressive targeting of a few major risk factors, such as oped and subject to greater uncertainty. Nonetheless, for tobacco smoke and air pollution, can greatly improve priority setting, information on mortality patterns by population health. broad cause group and the relative proportion of cases Intersectoral Policy Priorities for Health 25 that can be attributed to modifiable risk factors, the lat- exceed WHO-recommended limits on ambient particu- ter of which is taken from comparative risk assessment late matter. Further, 91 percent and 56 percent of house- studies, is useful. The data shown in table 2.1 suggest holds in these two income groups, respectively, still used that perhaps one-fourth or more of the 57 million solid fuels for cooking in homes in 2013. Water, sanita- deaths globally in 2015 can be attributed to one or more tion, and hygiene indicators appear to be more favorable: behavioral or environmental risk factors. 34 percent and 11 percent, respectively, lack access to In addition, several environmental and behavioral improved water sources; and 71 percent and 48 percent, risk factors have been studied for their effects on life respectively, lack access to improved sanitation. These expectancy. Air pollution studies have estimated life expec- proportions have declined significantly over the past tancy losses of 3.3 years in India (Sudarshan and others decade (Hutton and Chase 2017). 2015) and 5.5 years in northern China (Chen and others As for the behavioral cluster of risk factors, insuffi- 2013). (It is important to note that the methodological cient physical inactivity appears to be the most prevalent challenges to estimating the relative risks from air pollu- risk, particularly among adolescents, with estimates rang- tion appear to be considerable in settings where there is ing from 78 to 85 percent across World Bank income widespread exposure [Lipfert and Wyzga 1995]). The losses groups in 2010. The prevalence of risky sexual behavior from unsafe water and sanitation appear to be somewhat among reproductive-age individuals in low-income and smaller—ranging from one month in more-developed lower-middle-income countries was an estimated areas of Mexico to one year or more in the least-developed 74 percent and 30 percent, respectively, over 2007–13. areas (Stevens, Dias, and Ezzati 2008). In the behavioral The prevalence of tobacco smoking—likely the most risk factor cluster, tobacco studies have estimated that hazardous behavior of all—was about 17–18 percent smokers in India, Japan, the United Kingdom, and the among adults in low- and lower-middle-income coun- United States have about 10 years’ lower life expectancy tries in 2012 (WHO 2016b). than their nonsmoking peers (Jha and Peto 2014). A U.S. study estimated that physical inactivity, defined as sitting for more than three hours a day, decreases life expectancy Distal Determinants of Health by three years (Katzmarzyk and Lee 2012). Inadequate individual or household income constrains Yet another way of appreciating the importance of access to clean water, adequate sanitation, safe shelter, various risk factors is simply to compare estimates of the medical services, and other goods and services poten- proportion of the population exposed to specific risks. tially important for health. Inadequate education results The World Health Organization (WHO) Global Health in less likelihood that individuals will acquire informa- Observatory database contains estimates of the preva- tion relevant to their health-related behaviors or use that lence of a number of important risk factors (WHO information well. For these reasons, income, education, 2016b). In the environmental cluster, 95–99 percent of and other social (or socioeconomic) determinants of cities across low- and lower-middle-income countries health have received much attention for many years. Table 2.1 Magnitude of Effect of Top Environmental and Behavioral Risk Factors on Major Causes of Death, 2015 Number of deaths Share of deaths attributable globally in 2015 to one or more behavioral or Risk category (millions) environmental risks (%) Top risk factors Communicable, maternal, 12 30 Unsafe water, sanitation, and handwashing; perinatal, and nutritional maternal and child nutritional risks; unsafe conditionsa sex; air pollution; tobacco smoke Noncommunicable diseases 40 24 Dietary risks; tobacco smoke; air pollution; alcohol and drug use; low physical activity; occupational hazards Injuriesb 5 20 Alcohol and drug use Sources: GBD Risk Factors Collaborators (Forouzanfar and others 2016). Note: Mortality data are taken from World Health Organization (WHO) Global Health Estimates database (Mathers and others 2018, chapter 4 of this volume). Risk factor proportions are taken from the Global Burden of Disease (GBD) 2015 Study (Forouzanfar and others 2016) because similar data were not available from Mathers and others (2018). The table includes risk factors that were estimated to be responsible for 1 percent or more of total deaths globally. a. For alternative estimates of the attributable burden of maternal and child nutritional risks, see the 2013 Lancet series on “Maternal and Child Nutrition” (Lancet 2013). b. Unsafe roads are not included as a risk factor in the GBD 2015 project (Forouzanfar and others 2016); however, the WHO estimates that about 1.3 million road injury deaths occurred in 2015, comprising about 2 percent of all deaths in 2015 (Mathers and others 2018). 26 Disease Control Priorities: Improving Health and Reducing Poverty Two recent studies extend cross-country time-series A review of the full range of potential social deter- studies dealing with income and education (Jamison, minants or the health outcomes they affect is beyond Murphy, and Sandbu 2016; Pradhan and others 2017). the scope of this chapter. However, these findings are Three broad conclusions emerge from this literature: highlighted to note two implications for intersectoral action on health. First, the level of female education 1. Countries’ income levels are highly statistically sig- appears to be a quantitatively important social determi- nificant but quantitatively small factors in terms nant of mortality reduction, so discussions of intersec- of influencing reductions in both adult and child toral policies for health need to stress the importance of mortality. female education. Second, discrimination and violation 2. Level and quality of education are both statistically of human rights lead to worse health outcomes and significant and quantitatively important. Pradhan need to be considered in conversations with ministers and others (2017) concluded that about 14 percent of of justice and law enforcement. the decline in under-five mortality between 1970 and 2010 resulted from improvements in education levels. INTERSECTORAL POLICY PACKAGES Likewise, about 30 percent of the decline in adult mortality resulted from improvement in education. Essential Intersectoral Policies 3. Female education is far more important than male Chapter 1 of this volume (Jamison and others 2018) education for reducing both adult and child mortality. describes the 21 packages of disease interventions pre- sented throughout the nine DCP3 volumes that con- Aside from income and education, social norms and tain 327 interventions in total. Of these, 218 are health attitudes can greatly affect health. For example, discrim- sector specific and are covered in chapter 3 of this ination and stigma have been shown to increase the volume (Watkins and others 2018). The remaining risks of acquiring sexually transmitted infections, suf- 119 intersectoral interventions are discussed in this fering from mental disorders, and incurring injuries chapter. from interpersonal violence (Drew and others 2011; Annex 2A presents the contents of the intersectoral Piot and others 2015). In some countries, legalized dis- component of DCP3’s essential packages of interventions. crimination persists against vulnerable groups such as These policy interventions varied across packages in terms men who have sex with men and transgender people. of their level of specificity, and in a number of cases (such Even in countries without harsh legal arrangements, as tobacco taxation) they were duplicated across packages. pervasive discrimination—for example, against The authors of this chapter critically reviewed this list of indigenous groups—can greatly limit access to needed policies and consolidated and harmonized them. This health and other social services (Davy and others 2016). process led to a list of 71 harmonized intersectoral inter- Emerging evidence suggests that providing legal and ventions that were grouped by risk factor and type of human rights protections to vulnerable and stigmatized policy instrument (annex 2B). groups can reduce health risks or improve health Annex 2C provides a few important additional char- outcomes. Conversely, the lack of such protections can acteristics of the interventions contained in the harmo- increase health risks and worsen outcomes. For example, nized list. These include criminalization of sex work and same-sex relations is associated with increased risk of human immunodefi- • The risk factor(s) or cause(s) of death or disability ciency virus (HIV) among commercial sex workers and addressed men who have sex with men, through mechanisms such • The ministry primarily responsible for implementa- as increased risk of sexual violence and decreased provi- tion of the policy sion and uptake of HIV prevention services (Beyrer and • Whether there are health sector interventions that are others 2012; Shannon and others 2015). At the same equally or more effective (that is, to serve as so-called time, decriminalization can “avert incident infections substitutes—in which cases a health sector approach through combined effects on violence, police harass- may be more feasible than an intersectoral approach ment, safer work environments, and HIV transmission in limited resource settings) pathways” (Piot and others 2015). In general, criminal- • Where relevant, notable costs and benefits of the ization of same-sex relations and certain health intervention to other sectors conditions—such as drug addiction and abortion— • SDG target(s) addressed. often leads to worse health outcomes and cannot be supported on health grounds (Godlee and Hurley 2016; The vast majority of interventions in annexes 2A and Sedgh and others 2016). 2B were featured in volume 7 of DCP3. Major areas of Intersectoral Policy Priorities for Health 27 focus in this volume were air pollution, road injuries, economically more tractable than a comprehensive and a number of individually small but collectively approach. Further, epidemiological and economic important environmental toxins such as lead, mercury, conditions will dictate that some intersectoral interven- arsenic, and asbestos. This volume also included a tions can await a more urgent need for their number of interventions focused on occupational implementation. Nonetheless, initiating a subset of health, primarily by reducing occupational injury. intersectoral interventions as soon as possible to achieve Volumes 3, 4, and 5 of DCP3 contained a number of significant progress during the 2015–30 SDG period is interventions focused on noncommunicable disease important. The focus could be on those policies that are risk, particular from addictive substances and excessive likely to provide the best value for money and to be nutrient intake. The most common types of policy feasible in a wide range of settings. instruments recommended were legal and regulatory Table 2.2 outlines the authors’ distillation of the con- instruments (38 of 71), followed by fiscal instruments tents of annex 2B into an early intersectoral package. (15 of 71). This package draws on policy interventions that the authors have reviewed and determined to have the strongest evidence and the highest likely magnitude of An Early Intersectoral Package health effect. (The specific interventions are shown in The 71 interventions listed in annex 2B constitute a boldface in annex 2B.) In some cases, the policies have demanding menu for policy makers, especially in low- quickly and directly resulted in a measurable decline in resource settings. Even in well-resourced settings, mortality, with notable examples being in the area of an incremental approach to implementation of the household air pollution (box 2.1) and suicide preven- essential intersectoral package may be politically or tion (box 2.2). Table 2.2 Components of an Early Intersectoral Package of Policy Instruments Key health risk Policy Instrument Air pollution 1. Indoor air pollution: subsidize other clean household energy sources, including liquid propane Fiscal gas (LPG), for the poor and other key populations. 2. Indoor air pollution: halt the use of unprocessed coal and kerosene as a household fuel. Regulatory 3. Indoor air pollution: promote the use of low-emission household devices. Information and education 4. Emissions: tax emissions and/or auction off transferable emission permits. Fiscal 5. Emissions: regulate transport, industrial, and power generation emissions. Regulatory 6. Fossil fuel subsidies: dismantle subsidies for and increase taxation of fossil fuels (except LPG). Fiscal 7. Public transportation: build and strengthen affordable public transportation systems in Built environment urban areas. Addictive 8. Substance use: impose large excise taxes on tobacco, alcohol, and other addictive substances. Fiscal substance use 9. Substance use: impose strict regulation of advertising, promotion, packaging, and availability of Regulatory tobacco, alcohol, and other addictive substances, with enforcement. 10. Smoking in public places: ban smoking in public places. Regulatory Inadequate 11. School feeding: finance school feeding for all schools and students in selected geographical Fiscal nutrient intake areas. 12. Food quality: ensure that subsidized foods and school feeding programs have adequate Regulatory nutritional quality. 13. Iron and folic acid: fortify food. Regulatory 14. Iodine: fortify salt. Regulatory table continues next page 28 Disease Control Priorities: Improving Health and Reducing Poverty Table 2.2 Components of an Early Intersectoral Package of Policy Instruments (continued) Key health risk Policy Instrument Excessive nutrient 15. Trans fats: ban and replace with polyunsaturated fats. Regulatory intake 16. Salt: impose regulations to reduce salt in manufactured food products. Regulatory 17. Sugar sweetened beverages: tax to discourage use. Fiscal 18. Salt and sugar: provide consumer education against excess use, including product labeling. Information and education Road traffic 19. Vehicle safety: enact legislation and enforcement of personal transport safety measures, including Regulatory injuries seatbelts in vehicles and helmets for motorcycle users. 20. Traffic safety: set and enforce speed limits on roads. Regulatory 21. Traffic safety: include traffic calming mechanisms into road construction. Built environment Other risks 22. Pesticides: enact strict control and move to selective bans on highly hazardous pesticides. Regulatory 23. Water and sanitation: enact national standards for safe drinking water, sanitation, and hygenic Regulatory behavior within and outside households and institutions. 24. Hazardous waste: enact legislation and enforcement of standards for hazardous waste disposal. Regulatory 25. Lead exposure: take actions to reduce human exposure to lead, including bans on leaded fuels and Regulatory on lead in paint, cookware, water pipes, cosmetics, drugs, and food supplements. 26. Agricultural antibiotic use: reduce and eventually phase out subtherapeutic antibiotic use in Regulatory agriculture. 27. Emergency response: create and exercise multisectoral responses and supply stockpiles to Regulatory respond to pandemics and other emergencies. 28. Safe sex: remove duties and taxes on condoms, then introduce subsidies in brothels and for key Fiscal at-risk populations. 29. Exercise: take initial steps to develop infrastructure enabling safe walking and cycling. Built environment Box 2.1 Bans on Household Coal Use Coal has been used for household cooking and heat- risk factor for cancer and cardiac and respiratory ing for around 1,000 years, especially in places such diseases in adults and children. as China and the United Kingdom where coal is easy to mine. The famous 1952 “London smog” (smoke Bans on coal use, and successful enforcement of these and fog) episode, which killed 12,000 people, was bans have been followed by a reduction in premature mostly the result of indoor burning of coal for heat- deaths from these conditions. For example, during ing (Bell, Davis, and Fletcher 2004). the six years after the Irish government banned the sale of coal in 1990, the age-standardized cardiovas- Household coal use has diminished in high-income cular death rate fell by 10.3 percent and the age- countries. Today, it is mostly confined to LMICs, standardized respiratory death rate by 15.5 percent especially China and other countries in the Western (Clancy and others 2002). These reductions suggest Pacific region, where it constitutes around 20 percent that Dublin experienced about 243 fewer cardiovas- of all household fuel use (Duan and others 2010). cular deaths and 116 fewer respiratory deaths per Indoor burning of coal and other solid fuels is a year after the coal ban. Intersectoral Policy Priorities for Health 29 Box 2.2 Preventing Suicide in Sri Lanka by Regulating Pesticides From 1950 to 1995, suicide rates in Sri Lanka remaining WHO Class I (“extremely” or “highly” increased eightfold to a peak of 47 per 100,000 in toxic) organophosphate pesticides, and (c) the 1998 1995, the highest rate in the world (Gunnell and ban on endosulfan, a Class II (“moderately hazard- others 2007). Around two-thirds of the suicide ous”) pesticide that farmers had been using in place deaths during this period were due to self-poisoning of Class I pesticides (figure B2.1.1, panel a). with pesticides (Abeyasinghe 2002). Consensus is lacking on the chief contributors to the changing An ecological analysis of time trends in suicide and rates of suicide in Sri Lanka, but these are likely to suicide risk factors in Sri Lanka from 1975 to 2005 include periods of civil war and economic reces- found that these bans coincided with marked sion, changes in the rates of mental illness and its declines in the suicide rates of both men and women treatment, and the easy availability of hazardous (figure B2.1.1, panel a). Time trends in the data on agrochemicals (Abeyasinghe 2002; Gunnell and suicide method showed that the large reduction in others 2007). suicide was mostly due to a reduction in self- poisoning (figure B2.1.1, panel b). Further support In the 1980s and 1990s, a series of legislative activi- for this interpretation came from in-hospital ties led to the stepwise banning of the most toxic of mortality data, which showed a halving in death the pesticides being used for self-poisoning. This rates from pesticide self-poisoning—from 12.0 per legislation included (a) the 1984 ban on methyl 100,000 population in 1998 to 6.5 per 100,000 parathion and parathion, (b) the 1995 ban on the population in 2005. Figure B2.1.1 Suicide Rates in Relation to Selected Events in Sri Lanka, by Gender and Method, 1975–2005 a. Suicide rates, by gender b. Suicide rates, by method All class I pesticides banned (1995) All class I pesticides baned (1995) 90 Endosulfan banned (1998) 40 Parathion and methyl Suicide rate per 100,000 population parathion banned (1984) Suicide rate per 100,000 population 80 35 Endosulfan Parathion banned (1984) 70 banned (1998) 30 60 25 50 Outbreak of civil war 20 40 1981 census: revised Ceasefire 15 30 population estimates in civil war 10 20 10 2001 census: revised 5 population estimates 0 0 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05 75 77 79 81 83 85 87 89 91 93 95 97 99 01 03 05 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 19 19 19 19 19 19 19 19 19 19 19 19 19 20 20 20 Year Year Male age-standardized rate Self-poisoning and other methods Burning Female age-standardized rate Hanging Shooting Drowing Sharp cutting instruments Source: Gunnell and others 2007. 30 Disease Control Priorities: Improving Health and Reducing Poverty A few general themes emerge from table 2.2: sectors in mitigating the consequences of illness and injury for the fraction of cases that are not effectively • Nearly all of the policies address risks that produce large preventable by addressing the major risk factors. negative externalities such as polluted air (including Projection studies from high-income and selected from tobacco), unsafe driving, and environmental middle-income countries raise concerns that, even in toxins, to name a few. The presence of such externali- countries with high-performing health systems, spend- ties justifies the use of aggressive fiscal and regulatory ing on long-term care for individuals with chronic phys- measures to correct the economic inefficiencies that ical or mental disability is significant and likely to result from the failure of households or firms to take continue increasing (de la Maisonneuve and Oliveira negative externalities into account in their decision Martins 2013). A recent study from the Netherlands making. found that health expenditure increases dramatically • Many of the policies attempt to regulate or alter with age and nearness to death, with about 10 percent of markets for unhealthy and often addictive substances aggregate expenditure devoted to individuals in their last such as tobacco, alcohol, and processed foods. These year of life (Bakx, O’Donnell, and van Doorslaer 2016). might be seen as important first steps toward a Studies from other settings such as the United States more comprehensive approach to reduce disease risks validate these findings (Bekelman and others 2016). Yet that would eventually include greater incentives for another concerning result of the Dutch study is that healthy eating and physical activity. Greater incen- about one-third of total health expenditure in recent tives for healthy eating and physical activity are likely years was on long-term care, and the distribution of this to be much more disruptive and potentially expensive share of expenditure was skewed toward a relatively to fully incorporate into a whole-of-government pol- small number of individuals with severe disability (Bakx, icy but could lead to greater and more sustained gains O’Donnell, and van Doorslaer 2016). These expendi- in healthy life years as incomes grow. tures were also persistent over time, highlighting the • These policies require cross-cutting engagement with chronic, often lifelong, nature of ill health. a few key ministries, including finance, justice, envi- Several sources of long-term disabilities have been ronment, agriculture, and trade. Ministers of health observed to accompany economic growth and popula- could seek to develop productive relationships across tion aging, including vision and hearing loss, dementias, these key sectors early in the process. disability from cerebrovascular disease, and injuries related to advanced age. These conditions are no longer limited to high-income countries; most LMICs are now POLICIES TO ADDRESS THE CONSEQUENCES experiencing substantial health burden related to popu- lation aging (WHO 2011b). In many cases, these trends OF ILLNESS OR INJURY are superimposed on continued high levels of disability Globally, estimates of overall life expectancy have at younger ages—for example, disabilities resulting from exceeded estimates of healthy life expectancy by several severe injuries (which can result from interpersonal vio- years on average over the past few decades, suggesting lence, falls, or transport injury), severe psychiatric disor- that nonfatal health losses are a significant—and in ders, and intellectual disability (Kassebaum and others many countries, growing—concern for global health 2016). The growing population, elderly and nonelderly, (WHO 2016b). One group has estimated that at the needing long-term care in LMICs will inevitably require same time that global mortality has declined in absolute a greater response from government in the form of terms, absolute levels of disability have increased over broad-based social support measures. time, particularly in regions that have experienced sig- Support for those individuals with long-term disabil- nificant social and economic development (Kassebaum ity will need to include health sector–based interventions and others 2016). Thus, the general conclusion is that such as home health services, institutional care (for although rapid declines in child and adult mortality example, in skilled nursing facilities), and palliative care, have facilitated population growth and aging, these but it will need more than the health sector can provide changes have not been matched by improvements in to care adequately for the whole person. Intersectoral overall rates of disability. In part, this phenomenon can policies can be developed to provide these individuals be attributed to unchanged or increased levels of non- with assistance in obtaining affordable food, housing, communicable disease and injury risk factors that could and transportation, all of which are instrumental to pre- potentially be addressed using intersectoral measures, as venting further health loss. These policies usually fall described previously. At the same time, an equally under the category of transfer payments and may be important question is the role of health and nonhealth delivered directly as grants (nonwage income) or through Intersectoral Policy Priorities for Health 31 more targeted efforts such as subsidized housing or by country and would depend on the proportion of nutrition programs. the population in extreme poverty and the sorts of These transfer payments provide an important oppor- benefits (such as income, food, and transportation) tunity for ministries of health to work with ministries of included in the social support package. In low-income social development and others to care for the whole indi- countries, such a comprehensive program would vidual. In some settings, intersectoral collaboration has probably be unaffordable at current levels of govern- led to large-scale anti-poverty, social welfare, and ment spending. cash-transfer programs that integrate key social support The following three general points can be emphasized measures and enable effective uptake of health interven- for all countries, even those that are not currently able to tions (Watkins and others 2018). There are examples of implement fiscal policies that address long-term care: successful social support programs that effectively inte- grate health interventions, including support for older 1. The need for long-term care is increasing in nearly all adults. One of these is Mexico´s Prospera program, which countries because of population aging and high rates has been in operation since the late 1990s and covers the of nonfatal health loss. majority of the population living in poverty (Knaul and 2. Long-term care accounts for a significant fraction others 2017). of government expenditure in high-income set- As a result, DCP3 recommends that, as resources tings, and LMICs need to start preparing for this permit, countries consider income and in-kind social transition. support for individuals living with long-term disabil- 3. To address the needs of disabled persons ade- ity or severe, life-limiting illness (Krakauer and oth- quately, non–health sectors will need to be engaged ers 2018). Unfortunately, there is a limited evidence and willing to assume a large part of the fiscal base on which to design and implement social sup- responsibility. port measures in LMICs. Further, the feasibility and sustainability of broad-based social support pro- This last point suggests that countries could begin grams in low-income and lower- middle-income to develop a more inclusive notion of national health countries, in particular, are unknown. For example, accounts. Mexico’s experience in developing inclusive Krakauer and others (2018) produce preliminary national health accounts can be instructive for other estimates of social support costs for individuals in LMICs (box 2.3). In light of the critical gaps in current need of palliative care. These costs could vary widely evidence and the rapid shifts in disease burden in Box 2.3 Inclusive National Health Accounts: The Case of Mexico National health accounts (NHAs) show that Mexico satellite accounts to estimate the value at market spent 5.7 percent of its gross domestic product (GDP) prices of informal health activities generated by on health in 2015. This share is low compared with economic agents. These satellite accounts are an average of 9.3 percent among Organisation for sizable: the value of unpaid work related to Economic Co-operation and Development countries health care performed by households alone can and an average of 8.2 percent for the Latin American add an extra 18.6 percent to the traditional GDP region. However, the real figure is probably much estimates for the health sector. An even more larger because a significant part of health-related inclusive figure of the costs of ill health would economic activities, in particular those related to add income transfers of voluntary and legally long-term illness and injuries, goes unreported or mandated sick leave and disability insurance. unaccounted for by official NHA figures. Figures from the main social security institutions would add another 9.2 percent, bringing total The National Institute of Statistics and Geography health spending estimates closer to 7.3 percent (INEGI) acknowledged this concern by producing of GDP. box continues next page 32 Disease Control Priorities: Improving Health and Reducing Poverty Box 2.3 (continued) Conservative estimates from the satellite accounts of reducing work hours. Because long-term care for the combined value of (a) unpaid household the elderly or the chronically ill is not reimbursed members’ activities aimed at preventing ill health and by social or public health insurance schemes, fam- caring for and maintaining health both within and ilies must step in and find ways to provide care, outside the household and (b) the volunteer work for sometimes for long periods of time. The institu- nonprofit organizations averages 1 percent of GDP tional response from the health system has been over the past 10 years (INEGI 2017). According to slow regarding long-term care. Elderly or chroni- INEGI, the value of 69 percent of total hours and cally ill patients receive hospital care for acute 82 percent of unpaid work comes from household events, but the supply of publicly funded long- members undertaking mostly specialty care of term care or nursing homes to care for them over chronic ailments. Moreover, 70 percent of unremu- longer periods is very limited, and services pro- nerated caregivers are women (INEGI 2017). vided by existing private nursing homes need to A more inclusive approach toward NHA also helps be paid for out of pocket. estimate the economic consequences of ill health that are increasingly being borne outside of institu- Although social security institutions and other social tional settings. In 2015, approximately half of the assistance programs run day centers, which can burden of disease in Mexico was related to years lived include meals, families are by far the main provider of with disability, out of which mental and substance long-term care for the elderly (OECD 2007). Mexico’s abuse and musculoskeletal disorders accounted for omission in reporting expenditure on long-term care 40 percent (Kassebaum and others 2016), and an only reflects this institutional void. Part of the value estimated 16 percent of the adult population had of the informal long-term care provided by families is diabetes (OECD 2016). This burden has not only included in the satellite health accounts, but a signifi- increased pressure in an already overwhelmed and cant amount of nursing home services paid for out of underfunded public health care system but also cre- pocket by families possibly still goes unregistered. ated significant pressure on social security institu- As health needs become more complex and require tions. Not surprisingly, about half of total health care that goes beyond the traditional clinical and spending is from private sources, most of it paid out acute care settings, a broader perspective is needed to of pocket. Moreover, figures on the value of cash tease apart the economic and organizational impli- benefits for temporary disability (resulting from ill- cations. Mexico’s satellite accounts illustrate one step ness or accident, whether work or nonwork related, in this direction, highlighting the need to broaden and maternity leave) paid through the main social the range of types of care and providers considered security schemes—the Mexican Social Security when estimating the production value of the health Institute and the Institute of Social Security and sector’s share of GDP is necessary. Informal care Services for State Workers—amount to at least undertaken by families and by nursing homes and 9.2 percent of total health spending. Adding pen- other types of long-term care facilities needs to be sions for permanent disability would include this accounted for, even if this means considering a mix value. None of these figures are currently being of medical and other services (such as psychological accounted for as health-related spending neither in and nutrition services). Yet the indirect costs of ill- the NHA nor in the satellite accounts. ness are also important, as confirmed by the large Naturally, families also face increased pressure as value of income transfers for temporary disability. they seek ways to care for these patients, whether These should also be considered for a more inclusive by reorganizing household members’ roles and NHA. More comprehensive estimates of the produc- timetables, investing to adapt their homes to bet- tion value of the health sector would increase aware- ter suit their needs, hiring nonfamily caregivers, ness and inform policy formulation to better prepare or sometimes even quitting their own jobs or for the long-term care transition. Intersectoral Policy Priorities for Health 33 LMICs, the issue of long-term care could be regarded • Careful consideration of the social, cultural, economic, as one of the most important priorities for policy and political context research over the coming years. • Emphasis on generating political will and commit- ment from all relevant sectors at the national and subnational levels IMPLEMENTATION OF AN INTERSECTORAL • Design and reinforcement of accountability mecha- nisms, which also integrate into the monitoring and AGENDA FOR HEALTH evaluation process. Translation of the Intersectoral Package into Action The DCP3 intersectoral package, including the early- In addition, it stresses that historically major policy priority actions outlined in table 2.2, is intended to pro- change has tended to occur at times of political or eco- vide a list of policy actions outside the health sector that nomic transition or crisis and that ministries of health could substantially improve population health through a should take advantage of these times to put their priori- whole-of-government approach. Of course, the applica- ties on the agenda (WHO 2011b). tion of this intersectoral package will vary according A number of countries have overcome barriers to to epidemiological and demographic considerations. implementation by mainstreaming intersectoral For instance, low- and lower-middle-income countries approaches to health. A common theme in these suc- might place a higher priority on controlling indoor cesses is that the government, including the health sector, sources of air pollution, improving maternal and child recognized the legitimacy of intersectoral action for nutrition through food fortification, and scaling up health, as the following examples show: water and sanitation measures. Upper-middle-income and high-income countries would probably devote more • Iran has established several national mechanisms for efforts toward reductions in dietary risks. Most LMICs bringing sectors together to improve health, includ- could consider implementing stronger road safety and ing the National Coordination Council for Healthy tobacco control measures. All countries could work col- Cities and Healthy Villages (Sheikh and others 2012). lectively to address climate change, antimicrobial resis- The council oversees community-based health tance, and other global threats. improvement initiatives based on strategies such as The WHO (2011b) has produced a practical guide expanding access to financial credit, social services, and to intersectoral engagement that includes a 10-step sanitation. process for building and sustaining cross-sectoral • Vietnam has established a national intersectoral coordina- collaboration. The guide—“Intersectoral Action on tion mechanism, the National Traffic Safety Committee, Health: A Path for Policy-Makers to Implement with representatives from 15 ministries and agencies, to Effective and Sustainable Action on Health”— advise the prime minister on improving road safety. The highlights three cross-cutting themes relevant to committee played a key role in the passage of Vietnam’s implementation: national mandatory helmet law (box 2.4). Box 2.4 Reducing Road Traffic Deaths in Vietnam through Helmet Laws Nearly half of all road deaths worldwide are among Head injuries from motorcycle crashes are a com- groups of individuals who are the least protected— mon cause of morbidity and mortality. A Cochrane pedestrians, cyclists, and motorcyclists (WHO systematic review of 61 observational studies con- 2015). The risk to these different groups shows large cluded that motorcycle helmets reduce the risk of regional variations. For example, in Sub-Saharan head injury by around 69 percent and death by Africa pedestrians and cyclists are at highest risk, around 42 percent (Liu and others 2008). Several whereas in Southeast Asia motorcyclists are at countries in Southeast Asia have seen significant greatest risk. reductions in the rate of head injuries and deaths box continues next page 34 Disease Control Priorities: Improving Health and Reducing Poverty Box 2.4 (continued) Figure B2.4.1 Share of Motorcycle Drivers and Passengers Wearing Helmets in Vietnam, 2007 and 2008 100 Percentage wearing helmets 80 60 40 20 0 Drivers Passengers Drivers Passengers Drivers Passengers Yen Bai Da Nang Binh Duong Before helmet law (November 2007) After helmet law (June 2008) Source: Passmore, Nguyen, and others 2010. Note: Figure shows extent of motorcycle helmet wearing in three provinces of Vietnam before and after the introduction of mandatory helmet-wearing legislation. among motorcyclists after the introduction of laws and 18 percent, respectively (Passmore, Tu, and that made motorcycle helmet use mandatory (Hyder others 2010). and others 2007). For example, after Vietnam’s man- datory motorcycle helmet law went into effect in An extended cost-effectiveness analysis of the 2007 December 2007, an observational time-series study helmet policy suggests that it prevented about 2,200 using data from a random selection of the road net- deaths and 29,000 head injuries in the year following work in three provinces (Yen Bai, Da Nang, and Binh its introduction (Olson and others 2016). The analy- Duong) found significant increases in helmet wear- sis found that the wealthy owned the greatest number ing among both motorcycle riders and their passen- of motorcycles, so they accrued a larger share of the gers (Passmore, Nguyen, and others 2010), as shown absolute health and financial benefits from the law. in figure B2.4.1. Surveillance data from 20 rural and However, the policy probably prevented a larger urban hospitals found that the risk of road traffic number of cases of poverty among the poor and head injuries and deaths decreased by 16 percent middle class as well. • Thailand has vigorously promoted nationwide inter- A Key Role for Ministries of Finance sectoral action on health, including the use of health As shown in table 2.2 and annexes 2A and 2B, many of impact assessments. Such assessments are important the essential intersectoral policies in DCP3 are fiscal in tools for the health sector to engage other sectors by nature. Even the nonfiscal instruments proposed have identifying the possible positive and negative health implications for non–health sector budgets and thus consequences of other sectoral policies (Kang, Park, involve ministries of finance to a degree. By tracking the and Kim 2011). They have been conducted for a wide anticipated effects of interventions on government and range of policies or plans, including biomass power private revenues and expenditures outside the health plant projects, patents on medicines, coal mining, sector, annex 2C provides ministries of health with some and industrial estate development (Phoolcharoen, sense of where opportunity and opposition may arise on Sukkumnoed, and Kessomboon 2003). fiscal grounds. Intersectoral Policy Priorities for Health 35 Estimating the costs and consequences of intersectoral added sugar in a product group would also require intervention can be challenging for a variety of reasons, information on the nutrient content in those foods. and evaluation of all-of-society costs and benefits of Closely tied to what to tax is the issue of substitution health-related policies is outside the scope of DCP3. effects—that is, how demand for another product might Health economic evaluations usually implement change when the price of the newly taxed product cost-effectiveness and cost-utility analyses from a health changes (Fletcher, Frisvold, and Tefft 2013). For example, sector perspective on costs. In some cases, cost- if sugar-sweetened beverages are taxed, the decrease effectiveness analysis has been used to evaluate intersec- in sugar intake from reduced consumption of sugar- toral interventions. However, this perspective is quite sweetened beverages might be offset by increased con- limited because many of the important economic costs sumption of fruit juice or confectionary products. At the and benefits of these interventions lie outside the health same time, not all substitution effects are negative: sector. Fortunately, interest in benefit-cost analysis has recently implemented soda taxes in Mexico were associ- grown within health economics of late, and this approach ated with increased consumption of bottled water is ideal for evaluating intersectoral policies (see chapter 9 (Colchero and others 2016, Colchero and others 2017). of this volume, Chang, Horton, and Jamison 2018). In some cases, substitution effects might mutually rein- In volume 7 of DCP3, Watkins and others (2017) force public health goals ultimately. For example, tobacco summarize benefit-cost studies, including program costs, taxes appear to decrease binge drinking, presumably of interventions focusing on injury prevention and envi- because tobacco and alcohol use disorders co-occur in ronmental hazards, which are among the health topics many individuals (Young-Wolff and others 2014). with a significant benefit-cost literature. Although the Hence, when designing taxes, policy makers need to con- costs reviewed in volume 7 are neither totally representa- sider substitution effects and balance these against tive nor exhaustive, they can provide a rough sense of the implementation feasibility. For example, a broader magnitude of intersectoral costs. These range from neg- nutrient tax on sugar or on added sugar in processed ative costs in the case of taxes to less than US$1 per capita foods would decrease the substitution effects relative to per year for regulation and legislation to more than a tax on sugar-sweetened beverages alone, but it may not US$10 per capita per year for certain education interven- be easily implemented in many settings given the high tions or built-environment modifications (Watkins and tax administration requirements. others 2017). Several other tax design considerations are worth noting briefly: Taxation-Based Strategies This chapter strongly recommends taxation-based strat- • The type of tax is important to determine, and expe- egies for addressing harmful substance use and selected rience suggests that excise taxes can be more effective environmental hazards because of their clear effect on than sales taxes (IARC 2011). Tobacco taxes provide behavioral change and the positive revenue implications an important example in this regard. Tax rates can be for governments. Tobacco, alcohol, carbon emissions, simplified and based on the quantity of cigarettes, not and unhealthy food products may all be considered as their price (the latter of which is easier for the tobacco candidates for taxation. Although tobacco and alcohol industry to manipulate). A related goal is to preempt were originally taxed solely to generate revenue— downward substitution, when smokers switch to perhaps as early as the 1300s (Crooks 1989)—the long cheaper cigarette brands in response to a tax-rate hike history of these taxes can provide insights into how to on the brands they had previously smoked. Specific implement a variety of taxes to improve health. The excises, as opposed to ad valorem (value-based) fundamental question to answer first is what to tax. For excises or other taxes, are more effective at doing so. example, is it more effective to tax sugar as a nutrient The second strategy is to merge the multiple tobacco per se, to tax specific products such as sugar-sweetened tax tiers that are used in most LMICs. This way, tax beverages, or to opt for a hybrid approach (for example, hikes raise prices by the same large amount on all a tax based on the amount of added sugar in a particular brands at once, pushing smokers to quit completely class of products, such as sugar-sweetened beverages)? rather than switch (Marquez and Moreno-Dodson The pros and cons of any specific tax target need to be 2017). evaluated in terms of consumption habits, possible sub- • The amount of tax needs to be large enough to change stitution effects (as discussed below), and the adminis- behavior. For example, the WHO recommends that trative costs and feasibility of tax implementation given the cigarette excise tax make up at least 70 percent a country’s tax administration. Taxing the amount of of the final consumer price and that it be designed 36 Disease Control Priorities: Improving Health and Reducing Poverty to keep up with inflation and overall affordability connections between the DCP3 intersectoral package (WHO 2011c). and the SDG targets—especially the nonhealth-related • Tax evasion and avoidance are common problems that SDGs, which are of particular interest to other sectors. can be mitigated by having effective tax administration These connections and other international agreements measures and harmonized tax rates within a country that have intersectoral implications (for example, the and with neighboring countries (WHO 2011c.). WHO Framework Convention on Tobacco Control and • Tax effectiveness may improve as part of a compre- the United Nations Convention on the Rights of the hensive approach that includes public education, Child) can be leveraged both to engage other sectors on regulations, and other types of policies that support health issues and to put into place good accountability behavior change (WHO 2016a). and reporting mechanisms for specific policies. This • Public and industry opposition to taxes needs to approach suggests a strong relationship with ministries be anticipated and countered. A traditional tactic of foreign affairs that are accountable for the implemen- of industry groups is to argue that taxes will hurt tation of these agreements (WHO 2011b). employment and have a regressive effect on the poor. The SDGs contain strong language on poverty allevi- Yet low-income groups are generally more respon- ation (for example, SDG 1) and equity (for example, sive to these taxes and are likely receive more of the SDGs 5 and 10).1 One new scientific contribution long-term health and economic benefits from the tax of DCP3 has been the development of extended (Chaloupka and others 2012). cost-effectiveness analysis (ECEA), which considers not only the health outcomes but also the financial risk pro- Subsidy-Related Strategies tection and distributional (equity-enhancing) effects of Recognizing the role that subsidies can play in increasing policies (as further discussed in chapter 8 of this volume, or reducing health risks is also important. In many Verguet and Jamison 2018). Although ECEA most natu- countries, fossil fuels are heavily subsidized, representing rally serves as a tool to prioritize various health services a major economic barrier to clean energy (Coady and for public finance (covered in chapter 3 of this volume, others 2015). In some countries, broad food subsidies Watkins and others 2018), several ECEAs have also been (such as on bread, milk, or other products) are conducted on intersectoral policies, including tobacco entrenched, but these measures are ineffective in pro- taxation (Verguet and others 2015), regulation of salt in moting a healthy diet and may actually incentivize over- processed foods (Watkins and others 2016), and manda- consumption in environments, such as in the Arab tory helmet laws (Olson and others 2016). These ECEAs Republic of Egypt, that are experiencing forms of mal- show that intersectoral policies can—by reducing disease nutrition currently (IFPRI 2013). Similarly, agricultural risk and hence reducing an individual’s need for health subsidies in some countries greatly influence food con- care—prevent medical impoverishment, and in some sumption, both in the producing country and in its cases they can be pro-poor (meaning the poor benefit trading partners, sometimes to the detriment of health disproportionately to their population share from the (Fields 2004; Russo and Smith 2013). combined health and financial benefits of such interven- In light of anticipated revenue streams and country tions). One area of future work would be to integrate the experiences, a potential expansion path can be conceived ECEA approach into health impact assessment or for the rollout of fiscal policies directed toward a given benefit-cost analysis to illustrate the disaggregated non- substance. A first step would be to remove subsidies— health benefits of intersectoral policies, particularly especially important in the case of fossil fuels and unhealthy when those benefits speak to SDG targets or goals. foods—or, at the very least, to prevent subsidies from being added. The next step would be to add taxes on the substance. The final step would be to add subsidies for ANNEXES healthier substitutes. The first two steps would generate revenue and create fiscal space for subsidies, including The following annexes to this chapter are available at those that preferentially affect vulnerable populations. http://www.dcp-3.org/DCP. • Annex 2A: Intersectoral Policies of DCP3’s 21 Essential Intersectoral Action in the SDG Era Packages One method for increasing political will and account- • Annex 2B: Essential Intersectoral Policies Covered in ability is to design policies explicitly linked to interna- This Chapter tional agreements to which governments are already • Annex 2C: Characteristics of Essential Intersectoral signatories. Annex 2C demonstrates wide-reaching Policies Covered in This Chapter Intersectoral Policy Priorities for Health 37 NOTES Chaloupka, F. J., A. Yurekli, and G. T. Fong. 2012. “Tobacco Taxes as a Tobacco Control Strategy.” Tobacco Control 21 (2): World Bank Income Classifications as of July 2014 are as 172–80. follows, based on estimates of gross national income (GNI) Chang, A., S. Horton, and D. Jamison. 2018. “Benefit-Cost per capita for 2013: Analysis in Disease Control Priorities, Third Edition.” In Disease Control Priorities (third edition): Volume 9, Disease • Low-income countries (LICs) = US$1,045 or less Control Priorities: Improving Health and Reducing Poverty, • Middle-income countries (MICs) are subdivided: edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, (a) lower-middle-income = US$1,046 to US$4,125. R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, (b) upper-middle-income (UMICs) = US$4,126 to US$12,745. DC: World Bank. • High-income countries (HICs) = US$12,746 or more. Chen, Y., A. Ebenstein, M. Greenstone, and H. Li. 2013. “Evidence on the Impact of Sustained Exposure to Air 1. SDG 1: “End poverty in all its forms everywhere”; SDG 5: Pollution on Life Expectancy from China’s Huai River “Achieve gender equality and empower all women and Policy.” Proceedings of the National Academy of Sciences of girls”; and SDG 10: “Reduce inequality within and among the United States of America 110 (32): 12936–41. countries.” Clancy, L., P. Goodman, H. Sinclair, and D. W. Dockery. 2002. “Effect of Air-Pollution Control on Death Rates in Dublin, Ireland: An Intervention Study.” The Lancet 360 (9341): 1210–14. REFERENCES Coady, D., I. Parry, L. Sears, and B. Shang. 2015. “How Large Abeyasinghe, R. “Psychiatric Aspects of Pesticide Poisoning.” Are Energy Subsidies?” IMF Working Paper WP/15/105, 2002. In Pesticides: Health Impacts and Alternatives. International Monetary Fund, Washington, DC.. Proceedings of a Workshop held in Colombo 24 January Colchero, M. A., B. Popkin, J. Rivera, and S. W. Ng. 2016. 2002, edited by L. A. M. Smit, 11–15. Working Paper 45. “Beverage Purchases from Stores in Mexico under the Sri Lanka: International Water Management Institute. Excise Tax on Sugar Sweetened Beverages: Observational Ajay, V. S., D. A. Watkins, and D. Prabhakaran. 2017. Study.” BMJ 352: h6704/ “Relationships among Major Risk Factors and the Burden ———. 2017. “In Mexico, Evidence of Sustained Consumer of Cardiovascular Diseases, Diabetes, and Chronic Lung Response Two Years after Implementing a Sugar-Sweetened Disease.” In Disease Control Priorities (third edition): Volume Beverage Tax.” Health Affairs. doi:10.1377. 5, Cardiovascular, Respiratory, and Related Disorders, edited Crooks, E. 1989. Alcohol Consumption and Taxation. London: by D. Prabhakaran, S. Anand, T. Gaziano, J.-C. Mbanya, The Institute of Fiscal Studies. Y. Wu, and R. Nugent. Washington, DC: World Bank. Davy, C., S. Harfield, A. McArthur, Z. Munn, and A. Brown. Alegre-Díaz, J., W. Herrington, M. Lopez-Cervantes, L. Gnatiuc, 2016. “Access to Primary Health Care Services for R. Ramirez, and others. 2016. “Diabetes and Cause-Specific Indigenous Peoples: A Framework Synthesis.” International Mortality in Mexico City.” New England Journal of Medicine Journal for Equity in Health 15 (1): 163. 375 (20): 1961–71. De la Maisonneuve, C., and J. Oliveira Martins. 2013. “Public Bakx, P., O. O’Donnell, and E. van Doorslaer. 2016. “Spending Spending on Health and Long-Term Care: A New Set of on Health Care in the Netherlands: Not Going So Dutch.” Projections.” OECD Economic Policy Papers 6, Organisation Fiscal Studies 37 (3–4): 593–625. for Economic Co-operation and Development, Paris. Bekelman, J. E., S. D. Halpern, C. R. Blankart, J. P. Bynum, Drew, N., M. Funk, S. Tang, J. Lamichhane, E. Chavez, and oth- J. Cohen, and others. 2016. “Comparison of Site of Death, ers. 2011. “Human Rights Violations of People with Mental Health Care Utilization, and Hospital Expenditures for and Psychosocial Disabilities: An Unresolved Global Crisis.” Patients Dying with Cancer in 7 Developed Countries.” The Lancet 378 (9803): 1664–75. Journal of the American Medical Association 315 (3): 272–83. Duan, X., J. Zhang, H. Adair-Rohani, N. Bruce, H. Solomon, Bell, M. L., D. L. Davis, and T. Fletcher. 2004. “A Retrospective and K. R. Smith. 2010. “WHO Guidelines for Indoor Assessment of Mortality from the London Smog Episode of Air Quality: Household Fuel Combustion—Review 8: 1952: The Role of Influenza and Pollution. Environmental Household Coal Combustion: Unique Features of Exposure Health Perspectives 112 (1): 6–8. to Intrinsic Toxicants and Health Effects.” World Health Beyrer, C., P. S. Sullivan, J. Sanchez, D. Dowdy, D. Altman, and Organization, Geneva. http://www.who.int/indoorair others. 2012. “A Call to Action for Comprehensive HIV /guidelines/hhfc/Review_8.pdf. Services for Men Who Have Sex with Men.” The Lancet Fields, S. 2004. “The Fat of the Land: Do Agricultural Subsidies 380 (9839): 424–38. Foster Poor Health?” Environmental Health Perspectives Buss, P. M., L. E. Fonseca, L. A. Galvao, K. Fortune, and 112 (14): A820–23. C. Cook. 2016. “Health in All Policies in the Partnership for Fletcher, J., D. Frisvold, and N. Tefft. 2013. “Substitution Patterns Sustainable Development.” Revista Panamericana de Salud Can Limit the Effects of Sugar-Sweetened Beverage Taxes on Pública 40 (3): 186–91. Obesity.” Preventing Chronic Disease 10 (February 7): 120–95. 38 Disease Control Priorities: Improving Health and Reducing Poverty Forouzanfar, M., and others. 2016. “Global, Regional, and Jha, P. 2014. “Reliable Direct Measurement of Causes of Death National Comparative Risk Assessment of 79 Behavioural, in Low- and Middle-Income Countries.” BMC Medicine Environmental and Occupational, and Metabollic Risks 12: 19. or Clusters of Risks, 1990–2015: A Systematic Analysis Jha, P., M. MacLennan, A. Yurekli, C. Ramasundarahettige, for the Global Burden of Disease Study 2015.” The Lancet K. Palipudi, and others. 2015. “Global Hazards of Tobacco 388 (10053): 1659–724. and the Benefits of Smoking Cessation and Tobacco Tax.” In Godlee, F., and R. Hurley. 2016. “The War on Drugs Has Failed: Disease Control Priorities (third edition): Volume 3, Cancer, Doctors Should Lead Calls for Drug Policy Reform.” Britsh edited by H. Gelband, P. Jha, R. Sankaranarayanan, and Medical Journal 355: i6067. S. Horton, 175–94. Washington, DC: World Bank. Gunnell, D., R. Fernando, M. Hewagama, W. D. Priyangika, Jha, P., and R. Peto. 2014. “Global Effects of Smoking, of F. Konradsen, and M. Eddleston. 2007. “The Impact of Quitting, and of Taxing Tobacco.” New England Journal of Pesticide Regulations on Suicide in Sri Lanka.” International Medicine 370 (1): 60. doi:10.1056/NEJMra308383. Journal of Epidemiology 36 (6): 1235–42 Jha, P., C. Ramasundarahettige, V. Landsman, B. Rostron, Hoorn, S. V., M. Ezzati, A. Rodgers, A. D. Lopez, and M. Thun, and others. 2013. “21st-Century Hazards of C. J. L. Murray. 2004. “Estimating Attributable Burden of Smoking and Benefits of Cessation in the United States.” Disease from Exposure and Hazard Data.” In Comparative New England Journal of Medicine 368 (4): 341–50. Quantification of Health Risks: Global and Regional Burden of Kang, E., H. J. Park, and J. E. Kim. 2011. “Health Impact Diseases Attributable to Selected Major Risk Factors, Volume Assessment as a Strategy for Intersectoral Collaboration.” 2, edited by M. Ezzati, A. D. Lopez, A. Rodgers, and C. J. L. Journal of Preventive Medicine and Public Health 44 (5): 20109. Murray, 2129–40. Geneva: World Health Organization. Kassebaum, N. J., M. Arora, R. M. Barber, Z. A. Bhutta, Hutton, G., and C. Chase. 2017. “Water Supply, Sanitation, J. Brown, and others. 2016. “Global, Regional, and National and Hygiene.” In Disease Control Priorities (third edition): Disability-Adjusted Life Years (DALYs) for 315 Diseases Volume 7, Injury Prevention and Environmental Health, and Injuries and Healthy Life Expectancy (HALE) for edited by C. N. Mock, O. Kobusingye, R. Nugent, and 195 Countries and Territories, 1990–2015: A Systematic K. Smith. Washington, DC: World Bank. Analysis for the Global Burden of Diseases, Injuries, and Hyder, A.A., H. Waters, T. Phillips, and J. Rehwinkel. 2007. Risk Factors (GBD) 2015 Study.” The Lancet 388 (10053): “Exploring the Economics of Motorcycle Helmet Laws— 1603–58. Implications for Low and Middle-Income Countries.” Asia- Katzmarzyk, P. T., and I. M. Lee. 2012. “Sedentary Behaviour Pacific Journal of Public Health 19 (2): 16–22. and Life Expectancy in the USA: A Cause-Deleted Life Table IARC (International Agency for Research on Cancer). 2011. Analysis.” BMJ Open 2 (4): e000828. Effectiveness of Tax and Price Policies for Tobacco Control. Khayatzadeh-Mahani, A., Z. Sedoghi, M. H. Mehrolhassani, Volume 14, IARC Handbooks. Lyon, France: IARC. and V. Yazdi-Feyzabadi. 2016. “How Health in All Policies IFPRI (International Food Policy Research Institute). 2013. Are Developed and Implemented in a Developing Country? “Food Subsidies in Egypt: A Help or Hindrance?” IFPRI A Case Study of a HiAP Initiative in Iran.” Health Promotion blog, June 24, http://www.ifpri.org/blog/food-subsidies International 31 (4): 769–81. -egypt-help-or-hindrance. Knaul, F. M., P. E. Farmer, E. L. Krakauer, L. de Lima, A. INEGI (National Institute of Statistics and Geography, Government Bhadelia, and others. 2017. “Alleviating the Access Abyss in of Mexico). 2017. “Cuenta Satélite del Sector Salud de México, Palliative Care and Pain Relief: An Imperative of Universal 2015” [Health Sector Satellite of Mexico, 2015]. Press release, Health Coverage. Report of the Lancet Commission on March 23. Global Access to Palliative Care and Pain Control.” The Jamison, D. T., A. Alwan, C. N. Mock, R. Nugent, D. A. Watkins, Lancet. doi:10.1016/S0140-6736(17)32513-8. and others. 2018. “Universal Health Coverage and Intersectoral Krakauer, E., X. Kwete, H. Arreola-Ornelas, A. Bhadelia, O. Action for Health: Findings from Disease Control Priorities, Mendez, and others. 2018. “Palliative Care in Response to Third Edition.” In Disease Control Priorities (third edition): the Global Burden of Health-Related Suffering.” In Disease Volume 9, Disease Control Priorities: Improving Health and Control Priorities (third edition): Volume 9, Improving Reducing Poverty, edited by D. T. Jamison, H. Gelband, Health and Reducing Poverty, edited by D. T. Jamison, H. S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, R. Nugent. Washington, DC: World Bank. and R. Nugent. Washington, DC: World Bank. Jamison, D. T., J. G. Breman, A. R. Measham, G. Alleyne, Lancet. 2013. “Maternal and Child Nutrition.” Executive M. Claeson, D. B. Evans, P. Jha, A. Mills, and P. Musgrove, Summary of The Lancet Maternal and Child Nutrition eds. 2006. Disease Control Priorities in Developing Countries, Series, The Lancet, London. http://www.thelancet.com/pb second edition. Washington, DC: Oxford University Press /assets/raw/Lancet/stories/series/nutrition-eng.pdf. and World Bank. Lipfert, F. W., and R. E. Wyzga. 1995. “Air Pollution and Jamison, D. T., S. M. Murphy, and M. E. Sandbu. 2016. “Why Mortality: Issues and Uncertainties.” Journal of the Air & Has Under-5 Mortality Decreased at Such Different Rates in Waste Management Association 45 (12): 949–966, DOI: 10.1 Different Countries?” Journal of Health Economics 48: 16–25. 080/10473289.1995.10467427. Intersectoral Policy Priorities for Health 39 Liu, B.C., R. Ivers, R. Norton, S. Boufous, S. Blows, and S.K. Lo. Volume 8, Child and Adolescent Health and Development, 2008. “Helmets for Preventing Injury in Motorcycle Riders.” edited by D. A. P. Bundy, N. de Silva, S. Horton, D. T. Cochrane Database of Systematic Reviews 23 (1): CD004333. Jamison, and G. Patton. Washington, DC: World Bank. doi:10.1002/14651858.CD004333.pub3. Russo, M., and D. Smith. 2013. “Apples to Twinkies 2013: Marquez, P., and B. Moreno-Dodson. 2017. “Tobacco Control Comparing Taxpayer Subsidies for Fresh Produce and Program.” Brief, World Bank, Washington, DC, July 19. Junk Food.” CALPIRG (California Public Interest Research http://www.worldbank.org/en/topic/health/brief/tobacco. Group), Sacramento, July. Mathers, C., G. Stevens, D. Hogan, A. Mahanani, and J. Ho. Sedgh, G., J. Bearak, S. Singh, A. Bankole, A. Popinchalk, and 2018. “Global and Regional Causes of Death: Patterns others. 2016. “Abortion Incidence between 1990 and 2014: and Trends, 2000–15.” In Disease Control Priorities (third Global, Regional, and Subregional Levels and Trends.” edition): Volume 9, Improving Health and Reducing Poverty, The Lancet 388 (10041): 258–67. edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, Shannon, K., S. A. Strathdee, S. M. Goldenberg, P. Duff, R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, P. Mwangi, and others. 2015. “Global Epidemiology of DC: World Bank. HIV among Female Sex Workers: Influence of Structural Mock, C. N., O. Kobusingye, R. Nugent, and K. R. Smith, Determinants.” The Lancet 385 (9962): 55–71. eds. 2017. “Injury Prevention and Environmental Health: Sheikh, M. R., M. M. Afzal, S. Z. Ali, A. Hussain, and R. Shehzadi. Key Messages from This Volume.” In Disease Control 2012. “Multisectoral Development for Improved Health Priorities (third edition): Volume 7, Injury Prevention and Outcomes: Evaluation of Community-Based Initiatives in Environmental Health, edited by C. N. Mock, O. Kobusingye, the Islamic Republic of Iran.” Eastern Mediterranean Health R. Nugent, and K. Smith. Washington, DC: World Bank. Journal 16 (12): 1231–36. Nugent, R., and F. Knaul. 2006. “Fiscal Policies for Health Smith, K. R., and A. Pillarisetti. 2017. “Household Air Pollution Promotion and Disease Prevention.” In Disease Control from Solid Cookfuels and Health.” In Disease Control Priorities in Developing Countries, second edition, edited Priorities (third edition): Volume 7, Injury Prevention and by D. T. Jamison, J. G. Breman, A. R. Measham, G. Alleyne, Environmental Health, edited by C. N. Mock, O. Kobusingye, M. Claeson, D. B. Evans, P. Jha, A. Mills, and P. Musgrove, R. Nugent, and K. Smith. Washington, DC: World Bank. eds., 211–24. Washington, DC: World Bank and Oxford Stevens, G. A., R. H. Dias, and M. Ezzati. 2008. “The Effects University Press. of 3 Environmental Risks on Mortality Disparities across OECD (Organisation for Economic Co-operation and Mexican Communities.” Proceedings of the National Development). 2007. “Conceptual Framework and Academy of Sciences of the United States of America 105 (44): Methods for Analysis of Data Sources for Long-Term Care 16860–65. Expenditure.” Final report, OECD, Paris. Sudarshan, A., A. Sugathan, J. Nilekani, M. Greenstone, N. Ryan, ———. 2016. “OECD Reviews of Health Systems: and R. Pande. 2015. “Lower Pollution, Longer Lives: Life Mexico 2016.” OECD Publishing, Paris. http://dx.doi Expectancy Gains if India Reduced Particulate Matter .org/10.1787/9789264230491-en. Pollution.” Economic and Political Weekly 50 (8): 40–46. Olson, Z., J. A. Staples, C. Mock, N. P. Nguyen, A. M. Bachani, UN (United Nations). 2012. “Political Declaration of the High- and others. 2016. “Helmet Regulation in Vietnam: Impact Level Meeting of the General Assembly on the Prevention on Health, Equity and Medical Impoverishment.” Injury and Control of Non-Communicable Diseases.” A/Res/66/2, Prevention 22 (4): 233–38. UN, New York. Passmore, J. W., L. H. Nguyen, N. P. Nguyen, and J.-M. Olivé. ———. 2015. “Sustainable Development Goals.” Sustainable 2010. “The Formulation and Implementation of a National Development Knowledge Platform, UN, New York. http:// Helmet Law: A Case Study from Viet Nam.” Bulletin of the sustainabledevelopment.un.org/?menu=1300. World Health Organization 88 (10): 783–87. Verguet, S., C. L. Gauvreau, S. Mishra, M. MacLennan, Passmore, J., N. T. H. Tu, M. A. Luong, N. D. Chinh, and S. M. Murphy, and others. 2015. “The Consequences of N. P. Nam. 2010. “Impact of Mandatory Motorcycle Helmet Tobacco Tax on Household Health and Finances in Rich and Wearing Legislation on Head Injuries in Viet Nam: Results Poor Smokers in China: An Extended Cost-Effectiveness of a Preliminary Analysis. Traffic Injury Prevention 11: 202–6. Analysis.” The Lancet Global Health 3 (4): e206–16. Phoolcharoen, W., D. Sukkumnoed, and P. Kessomboon. 2003. Verguet, S., and D. T. Jamison. 2018. “Health Policy Assessment: “Development of Health Impact Assessment in Thailand: Applications of Extended Cost-Effectiveness Analysis Recent Experiences and Challenges.” Bulletin of the World Methodology in Disease Control Priorities, Third Edition.” Health Organization 81 (6): 465–67. In Disease Control Priorities (third edition): Volume 9, Piot, P., S. S. Abdool Karim, R. Hecht, H. Legido-Quigley, Disease Control Priorities: Improving Health and Reducing K. Buse, and others. 2015. “Defeating AIDS—Advancing Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, Global Health.” The Lancet 386 (9989): 171–218. P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent. Pradhan, E., E. Suzuki, S. Martinez, M. Schaferhoff, and Washington, DC: World Bank. D. Jamison. 2017. “The Effects of Education Quantity and Watkins, D. A., N. Dabestani, R. Nugent, and C. Levin. Quality on Child and Adult Mortality: Their Magnitude 2017. “Interventions to Prevent Injuries and Reduce and Their Value.” In Disease Control Priorities (third edition): Environmental and Occupational Hazards: A Review of 40 Disease Control Priorities: Improving Health and Reducing Poverty Economic Evaluations from Low- and Middle-Income ———. 2011b. “Intersectoral Action on Health: A Path for Countries.” In Disease Control Priorities (third edition): Policy-Makers to Implement Effective and Sustainable Volume 7, Injury Prevention and Environmental Health, Action on Health.” Guidance booklet, WHO Centre for edited by C. N. Mock, O. Kobusingye, R. Nugent, and Health Development, Kobe, Japan. K. Smith. Washington, DC: World Bank. ———. 2011c. WHO Technical Manual on Tobacco Tax Watkins, D. A., D. Jamison, A. Mills, R. Atun, K. Danforth, and Administration. Geneva: World Health Organization. others. 2018. “Universal Health Coverage and Essential ———. 2015. Global Status Report on Road Safety. Geneva: Packages of Care.” In Disease Control Priorities (third WHO. http://www.who.int/violence_injury_prevention edition): Volume 9, Disease Control Priorities: Improving /road_safety_status/2015/en/. Health and Reducing Poverty, edited by D. T. Jamison, ———. 2016a. “Fiscal Policies for Diet and Prevention of H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Noncommunicable Diseases: Technical Meeting Report, Mock, and R. Nugent. Washington, DC: World Bank. 5–6 May 2015 Geneva, Switzerland.” Geneva: World Health Watkins, D. A., Z. D. Olson, S. Verguet, R. A. Nugent, Organization. and D. T. Jamison 2016. “Cardiovascular Disease and ———. 2016b. Global Health Observatory (GHO) database, Impoverishment Averted Due to a Salt Reduction Policy WHO, Geneva. http://www.who.int/gho/en/. in South Africa: An Extended Cost-Effectiveness Analysis.” Young-Wolff, K., K. Kasza, A. Hyland, and S. McKee. 2014. Health Policy and Planning 31 (1): 75–82. “Increased Cigarette Tax Is Associated with Reductions WHO (World Health Organization). 2011a. “Global Health in Alcohol Consumption in a Longitudinal U.S. Sample.” and Aging.” WHO and U.S. National Institute of Aging, Alcoholism: Clinical and Experimental Research 38 (1): Geneva and Washington, DC. 241–48. Intersectoral Policy Priorities for Health 41 Chapter 3 Universal Health Coverage and Essential Packages of Care David A. Watkins, Dean T. Jamison, Anne Mills, Rifat Atun, Kristen Danforth, Amanda Glassman, Susan Horton, Prabhat Jha, Margaret E. Kruk, Ole F. Norheim, Jinyuan Qi, Agnes Soucat, Stéphane Verguet, David Wilson, and Ala Alwan INTRODUCTION face catastrophic or impoverishing health expenditure Health systems have several key objectives; the most when seeking acute or chronic disease care (Xu and fundamental is to improve the health of the population. others 2007). These financial risks can result in further In addition, they are concerned with the distribution of health loss and reduced economic prosperity for house- health in the population—for example, with health holds and populations (Kruk and others 2009; McIntyre equity—and they strive to be responsive to the needs of and others 2006). the population and to deliver services efficiently (WHO The current universal health coverage (UHC) move- 2007). Notably, they also seek to provide protection ment emerged in response to a growing awareness of against the financial risks that individuals face when the worldwide problems of low access to health ser- accessing health services. Ideally, this financial risk vices, low quality of care, and high levels of financial protection (FRP) is accomplished through mechanisms risk (Ji and Chen 2016). UHC is now a core tenet such as risk pooling and group payment that ensure of United Nations (UN) Sustainable Development prepayment of most, if not all, health care costs (Jamison Goal (SDG) 3.1 UHC was preceded by the aspirational and others 2013). notion of a minimum standard of health for all, An effective health system is one that meets these enshrined in the Universal Declaration of Human objectives by providing equitable access to affordable, Rights (adopted by the UN General Assembly in 1948) high-quality health care—including treatment and cura- and the declaration of Alma-Ata in 1978, and many tive services as well as health promotion, prevention, and HICs have provided universal coverage for decades. rehabilitation services—to the entire population. The World Health Assembly endorsed the modern Unfortunately, most countries lack health systems that concept of UHC as an aspiration for all countries in meet this standard. Shortfalls in access, quality, effi- 2005. Subsequent World Health Reports by the World ciency, and equity have been documented extensively, Health Organization (WHO) expanded on various both in low- and middle-income countries (LMICs) and technical aspects of UHC, and in 2015, UHC was in some high-income countries (HICs) (WHO 2010). adopted as a subgoal (target 3.8) of SDG 3 (UN 2016; In addition, in many countries, households routinely WHO 2013b). Corresponding author: David A. Watkins, University of Washington, Seattle, Washington, United States; davidaw@uw.edu. 43 Mechanisms and approaches, summarized elsewhere summarize the evidence for their effectiveness or (WHO 2010; WHO 2013b), have been proposed or cost-effectiveness. In this regard, DCP3 provides guid- attempted as specific means of achieving UHC, but the ance on priority health interventions for UHC in objectives of UHC are the same in all settings, regard- LMICs in the form of a model health benefits package less of approach: improving access to health services that is based on DCP3’s 21 essential packages (see (particularly for disadvantaged populations), improv- chapter 1 of this volume, Jamison and others 2018). ing the health of individuals covered, and providing This chapter proposes a concrete set of priorities for FRP (Giedion, Alfonso, and Díaz 2013). There are three UHC that is grounded in economic reality and is fundamental dimensions to UHC—proportion of pop- intended to be appropriate to the health needs and con- ulation covered, proportion of expenditures prepaid, straints of LMICs, particularly low-income countries and proportion of health services included in UHC— and lower-middle-income countries. It develops a model that any given health care reform strategy seeks to benefits package referred to as essential UHC (EUHC) achieve in some prioritized order (Busse, Schreyögg, and identifies a subset of interventions termed the and Gericke 2007). Recent reports, including the Lancet highest-priority package (HPP). The chapter presents a Commission on Investing in Health and the WHO case that all countries, including low-income countries, Making Fair Choices consultation, have endorsed a could strive to fully implement the HPP interventions by “progressive universalist” approach to public finance the end of the SDG period (2030), and many middle- of UHC (Jamison and others 2013; WHO 2014).2 income countries could strive to achieve full implemen- Progressive universalism makes the case, on the basis of tation of EUHC. The chapter also presents estimates of efficiency and equity, for an expansion pathway through the EUHC and HPP costs and mortality consequences. It the three UHC dimensions that prioritizes full popula- concludes with a discussion of measures that improve tion coverage and prepayment, albeit for a narrower the uptake and quality of health services and with some scope of services than could be achieved at lower cover- remarks on the implications of EUHC and the HPP for age levels or through cost-sharing arrangements. (It has health systems. been argued that full population coverage and full pre- The chapter does not, however, prescribe one correct payment are necessary conditions to ensure that UHC approach to UHC, nor does it attempt to review the wide leaves no one behind [WHO 2014].) array of delivery mechanisms, policy instruments, and If progressive universalism is the preferred approach financial arrangements that support the transition to to UHC, then a critical question for health planners is UHC; these have been covered in detail elsewhere which health interventions should be included. HICs are (WHO 2010; World Bank 2016). Rather, this chapter able to provide a wide array of health services, but stresses that the UHC priority-setting process is contex- LMICs have the resources to deliver a smaller set of tual, depending on political economy as well as local services, necessitating a more explicit and systematic costs, budgets, and demographic and epidemiological approach to priority setting (Glassman and others factors—all of which influence the value for money of 2016). In this spirit, the Making Fair Choices report rec- specific interventions. ommended that UHC focus on interventions that are Because the development and refinement of a bene- the most cost-effective, improve the health of the worst fits package is an incremental and iterative process, off, and provide FRP (WHO 2014). The extended many ministries of health probably will not use DCP3’s cost-effectiveness analysis (ECEA) approach developed recommendations as a template for their packages but for this third edition of Disease Control Priorities (DCP3) rather as an aid in reviewing existing services, identify- assesses policies in these dimensions and can help iden- ing outliers, and considering services that are not cur- tify efficient, fair pathways to UHC. Chapter 8 of this rently provided. The DCP3 model benefits package can volume provides an overview of ECEA methods and thus serve as a starting point for deliberation on a new results of ECEAs undertaken in conjunction with health benefits package or refinement of an existing DCP3 (Verguet and Jamison 2018). package. However, as construed here, it would not be a The set of prioritized health services publicly perfect package for a particular country. To translate the financed through a UHC scheme has been termed a DCP3 findings into an actionable UHC agenda at the health benefits package (Glassman and others 2016). The national or subnational level will require context- limited experience of LMICs with benefits packages specific technical analyses and public consultation, suggests that such packages can be part of a coherent ideally as part of a clearly articulated political agenda and efficient approach to health system strengthening, and an institutionalized priority-setting process that but many countries lack the technical capacity to can govern public and donor resource allocation in the review a broad range of candidate interventions and health sector. 44 Disease Control Priorities: Improving Health and Reducing Poverty FROM ESSENTIAL PACKAGES TO ESSENTIAL additional steps to arrive at a final list of EUHC UNIVERSAL HEALTH COVERAGE interventions: Development of an Essential UHC Package • First, instances of duplicate or redundant inter- Identification of interventions for the HPP and EUHC ventions were removed. Although duplicate inter- began by compiling all of the interventions described in ventions were removed in the construction of the DCP3’s essential packages. As described in chapter 1 of this EUHC list, each essential package retained all of its volume (Jamison and others 2018), the essential packages interventions. of volumes 1 through 9 of DCP3 contain 327 interven- • Second, the authors worked with the editors respon- tions that have been deemed to accomplish the following: sible for each of these packages to revise intervention descriptions, when needed, to add specificity or clar- • Provide good value for money in multiple settings. ity for a nonspecialist audience. On the advice of the • Address a significant disease burden. editors of DCP3 volumes 4 (Patel and others 2015) • Be feasible to implement in a range of LMICs. and 6 (Holmes and others 2017), only a subset of best-practice interventions from these two volumes (Note that 119 of the interventions in these essential was included in the EUHC package. This chapter also packages are intersectoral in nature, as discussed in aggregated a number of specific health services into chapter 2 of this volume, Watkins and others [2018]. single interventions that would always be delivered Some interventions in DCP3 are not easily classified as together in practice, such as screening of at-risk indi- health sector or intersectoral; these were generally viduals for a given disease plus treatment of individu- included in the present chapter as health sector interven- als who have screened positive for that disease. tions by default. Examples of such interventions include • The authors deemed some interventions not to maternal and infant nutrition [that is, food as medicine] be specific health services but rather measures to and vector control.) increase intervention uptake or quality. These inter- The interventions recommended in these essential ventions were removed from the EUHC list and are packages reflect the synthesis of a wide range of epi- discussed as a group later in this chapter. demiological and economic evidence instilled with the • Finally, the authors mapped all interventions to a expert judgment required to extrapolate these findings standard typology of health system platforms that to settings and policy questions for which data are very reflects the consensus of editors and members of the limited. Most of the economic evidence takes a health DCP3 Advisory Committee (box 3.1). The grouping sector perspective on costs and draws on estimates of of interventions into platforms is intended to illus- incremental value for money in settings where the trate how they could be integrated with each other number and scale of current health services are limited. and within existing health systems. Still, as summarized in chapter 7 of this volume (Horton 2018), the quality and applicability of eco- Annex 3C presents the final contents of the EUHC nomic evidence in these studies vary widely, requiring package, by platform. The EUHC package includes additional deliberation and judgment as described later 218 unique interventions, including 13 interventions at in this chapter. the population level, 59 at the community level, 68 at Notably, this chapter includes essential packages for health centers, 58 at first-level hospitals, and 20 at two additional groups of conditions: congenital and referral and specialized hospitals. Annex 3D, which genetic disorders (annex 3A) and musculoskeletal disor- accompanies annex 3C, examines issues related to spe- ders (annex 3B). These conditions had been treated cific EUHC interventions. These issues include prices extensively in Disease Control Priorities in Developing and their impact on cost-effectiveness in cases where Countries, second edition (DCP2) (Jamison and others prices are rapidly changing, health system require- 2006) and were touched upon in various volumes of ments such as integration of urgent intervention across DCP3, but they were deemed not to require dedicated delivery platforms, and considerations of feasibility in chapters. The essential packages for these two groups of certain settings. conditions reflect the key messages of the relevant sec- tions of DCP2, with updated information on burden of disease and economic evidence in LMICs, particularly Identifying a Highest-Priority UHC Package over the past decade. The EUHC list of 218 unique interventions still consti- After compiling the contents of DCP3’s 21 essential tutes an ambitious agenda for many countries, and packages, the authors of this chapter took several achieving full coverage of EUHC by 2030, the end of the Universal Health Coverage and Essential Packages of Care 45 Box 3.1 Defining Delivery Platforms for Essential UHC in DCP3 : A Standardized Typology DCP3 volumes 1–9 present interventions in 21 pack- clients, meeting people where they live. It includes a ages tailored to various “platforms,” defined as logis- wide variety of delivery mechanisms. Specific sub- tically related delivery channels. Thus, a platform is platforms include the following: the level of a health system at which interventions can be appropriately, effectively, and efficiently delivered. • Health outreach and campaigns (such as vac- These platforms, and the interventions that are deliv- cination campaigns, mass deworming, and ered through them, were determined by the editors face-to-face health information, education, and of the individual volumes. To compile a single list of communication) unique interventions in Essential Universal Health • Schools (including school health days) Coverage and group them by platform, the authors • Community health workers, who may be based of this chapter harmonized the definitions of the primarily in the community but also connected platforms and, in some cases, reallocated interven- to first-level care providers, with ties to the rest tions to platforms different from those that appeared of the system. elsewhere in the DCP3 volumes. Health centers: The health center level captures two This platform model is a pragmatic typology types of facility. The first is a higher-capacity health rather than a comprehensive description of the myriad facility staffed by a physician or clinical officer and health facilities currently serving clients in low- and often a midwife to provide basic medical care, minor middle-income countries. Contextual factors, includ- surgery, family planning and pregnancy services, ing local culture, disease burden, resources, and geog- and safe childbirth for uncomplicated deliveries. raphy, will influence both the types of services provided (In annexes 3C and 3F, this sort of health center is at each level and the way in which patients interact denoted with an asterisk.) The second is a lower- with a health care system. With changes in technology capacity facility (for example, health clinics, phar- and delivery know-how, it is likely and desirable that macies, dental offices, and so on) staffed primarily by existing modalities of health care delivery will evolve a nurse or mid-level health care provider, providing and adapt over time. A platform’s definition will also services in less-resourced and often more remote evolve as a country’s health system becomes more settings. advanced and offers a wider array of health services, particularly at lower levels of the system. First-level hospitals: A first-level hospital is a facil- ity with the capacity to perform surgery and pro- The five platforms of a health system as defined in vide inpatient care. This platform also includes this chapter are as follows: outpatient specialist care and routine pathology Population-based health interventions: This platform services that cannot be feasibly delivered at lower captures all nonpersonal or population-based levels, such as newborn screening. DCP3 contends health services, such as mass media and social that a primary goal for all countries to achieve marketing of educational messages, as typically during the Sustainable Development Goals era delivered by public health agencies. (Note that could be to ensure most patients have access to nonhealth-system platforms related to fiscal and fully resourced, high-quality, first-level hospitals—a intersectoral policies—for example, taxes, subsidies, goal that, although aspirational, could be feasible regulatory policies, and changes in the built envi- by 2030. ronment—are discussed in chapter 2 of this volume Referral and specialized (second- and third-level) [Watkins and others 2018].) hospitals: This platform includes general and special- Community services: The community platform ist hospitals that provide secondary and tertiary encompasses efforts to bring health care services to services. 46 Disease Control Priorities: Improving Health and Reducing Poverty SDG period, would be challenging for most low-income noted separately. These issues are noted where pertinent countries. Further, as has been highlighted throughout in annexes 3D and 3F. DCP3, there is great heterogeneity in the strength of evi- A few additional remarks should be made on DCP3’s dence and the magnitude of the health impact of these shift from the criterion of cost-effectiveness to the essential interventions. broader criterion of value for money. In general, DCP3 Some helpful guidance comes from the WHO Making has drawn upon cost-effectiveness or cost-utility analy- Fair Choices consultation, which outlined the principle of ses to assess interventions that primarily affect health priority classes—namely, that health services could be outcomes, including disability and premature mortality. grouped into three classes (high, medium, or low priority) In these cases, referring to the cost-effectiveness of an based on their relative merits in the dimensions of cost- intervention, measured by cost per adult or child effectiveness, priority given to the worse off, and FRP death averted or cost per disability-adjusted life year (Chan 2016; WHO 2014). In this spirit, this chapter (DALY) averted, is appropriate. At the same time, sev- develops an illustrative HPP that parallels the high- eral important types of health sector interventions pre- priority class described in Making Fair Choices. It looks at dominantly produce outcomes that are not easily the HPP through the lens of low-income countries, tak- measured in deaths, DALYs, or quality-adjusted life- ing into consideration their aggregate epidemiological years (QALYs); these include met need for family plan- and demographic patterns as well as typical resource ning, reductions in stillbirth rates, palliative care and constraints. relief of suffering, and remediation of intellectual losses associated with illness or poor nutritional status. In Identifying the Highest-Priority UHC these cases, metrics such as cost per death or DALY Interventions: Three Key Dimensions averted do not apply. As a result, the more general term To identify the subset of EUHC interventions that could value for money is used here to refer to the relative be included in the HPP, the authors appraised each attractiveness of interventions in terms of relevant EUHC intervention in three dimensions: value for outcomes. Outside of a benefit-cost analysis framework, money, priority given to the worse off, and FRP afforded. the commensurability of different value for money indi- Annex 3E provides details on the methods and data cators (for example, cost per death averted versus cost used in this appraisal process, and annex 3F displays the per case of met need for contraception) is a matter of authors’ assessments of each EUHC intervention in judgment and may require further empirical study these dimensions. (see chapter 9 of this volume, Chang, Horton, and Value for money. To assess value for money, the Jamison [2018]). authors considered cost-effectiveness estimates where Another limitation of the use of cost-effectiveness cost-effectiveness was a relevant metric of value for and value-for-money criteria is the potential disconnect money. In these cases, the geometric mean of incremen- between modeled estimates and real-world impact. If the tal cost-effectiveness ratios was calculated from the eco- quality of care in practice lags what is captured in effec- nomic evaluation literature in LMICs (see chapter 7 of tiveness studies, cost-effectiveness ratios will be higher this volume, Horton 2018). In the cases of EUHC inter- than reported in the literature. Variations in observed ventions not covered in chapter 7, other databases of clinical practice suggest that differential benefits from cost-effectiveness studies were searched for relevant health care are likely within and between populations. estimates. The authors also noted the major drivers of Unfortunately, the quality of health services in LMICs is cost-effectiveness in cases where interventions would an understudied topic and is generally not considered in not be uniformly cost-effective in LMICs. These drivers economic evaluations (Akachi and Kruk 2017; Kruk and include epidemiological context (such as high- versus others 2017). In the assessments presented in annex 3F, moderate-transmission areas for malaria), price varia- the authors have attempted to account for potential real- tions in key technologies (such as vaccines for which world reductions in value-for-money caused by low certain countries may be eligible for subsidies), and the quality of care, particularly for complex and longitudinal quality and generalizability of the cost-effectiveness services in low-income countries. (Measures that can data. These factors were then synthesized into a sum- ensure the quality of EUHC interventions are discussed mary assessment of cost-effectiveness that placed later in this chapter.) interventions into one of five categories. Where cost- Despite all the important limitations discussed above, effectiveness was not a relevant metric of value for the DCP3 perspective is that estimates of cost-effectiveness money, the appropriate outcome and the efficiency and value-for-money are critical inputs to the priority- of the intervention in achieving the outcome were setting process. Universal Health Coverage and Essential Packages of Care 47 Priority given to the worse off. To assess whether an Criteria for Inclusion in the Illustrative intervention gave priority to the worse off, the authors Highest-Priority UHC Package identified the principal health condition addressed by A working concept of the HPP can be defined as the sum each intervention. An indicator for the “worse off” was of all interventions that meet the following criteria, bal- developed that attempted to identify individuals who, by anced against each other: virtue of having a particular disease or injury, would have a much lower level of lifetime health. This indicator was • Very good value for money in low-income countries. In termed “health-adjusted average age of death” (annex 3E). cost-effectiveness terms, this is on the order of less In brief, this measure estimated the additional fatal and than US$5,000–US$7,500 per death averted, depend- nonfatal health loss experienced by an individual affected ing on average age of death (with a higher willingness by a specific cause of death or disability or both, as com- to pay for child and adolescent deaths averted), or pared to the average levels of health in the population. In less than US$200–$300 per DALY averted (or QALY essence, the measure identified causes that would be very gained). This range of cost-effectiveness values draws severe or result in extremely premature mortality or from the growing literature on health care opportu- both. Because the focus of the illustrative HPP is low- nity costs, which suggests that a figure approximating income settings, aggregate epidemiological estimates half of gross domestic product (GDP) per capita per for low-income countries as a group were used as the DALY averted is a realistic level of willingness-to-pay reference population for constructing this indicator. for health care interventions in LMICs (Ochalek, Estimates of health-adjusted average age of death by Lomas, and Claxton 2015). (DCP3 does not explic- cause were assigned to ordinal groups using cutoffs itly endorse this particular threshold—or the health described in annex 3E and then mapped to specific inter- care opportunity cost approach in general—as a ventions that addressed each cause. normative one but rather uses it in this chapter as an The criterion of priority to the worse off is one vari- example of a typical threshold that might be imple- ant on the more general notion of “pro-poor” UHC. mented in a highly resource-constrained country.) There is broad agreement that UHC schemes in LMICs For interventions where cost-effectiveness is not a should strive first and foremost to serve the needs of relevant metric of value for money, an assessment was marginalized and low-income groups (Bump and oth- made by the authors as to whether the intervention ers 2016). To accomplish this, some UHC reforms have would be likely to efficiently lead to health outcomes focused on expanding all health services to the poorest important in low-income countries that are not areas, while others have identified interventions against captured in DALYs (for example, averted stillbirths, a set of “diseases of poverty” (such as tuberculosis or averted unwanted pregnancies, and provision of pal- neglected tropical diseases) as priorities for public liative care). As a matter of both value for money and finance. Whereas this chapter’s approach shares more ethical obligation, full coverage of basic palliative care in common with the latter than the former, it takes a services was included in the HPP by default. lifecourse perspective on ill health and gives greater • Priority given to the worst off. This criterion is met by weight, for example, to selected noncommunicable dis- an intervention being directed against a cause of dis- eases (such as schizophrenia, congenital disorders, or ease or injury that has a low health-adjusted average childhood cancers) and injuries than might be given age of death. within a “diseases of poverty” framework that is ori- • Likely to provide a high degree of FRP. This criterion is ented to communicable diseases. met by an intervention receiving a high score on the Financial risk protection. A qualitative approach composite indicator for FRP. was taken to assess FRP. The authors used a compos- • Part of the “grand convergence” agenda proposed ite indicator for FRP derived from expert judgments by the Lancet Commission on Investing in Health. in three dimensions: (a) likelihood of medical impov- These interventions—in the domains of reproduc- erishment in the absence of public finance of the tive, maternal or neonatal, and child health; human intervention, based on unit cost data; (b) urgency of immunodeficiency virus and acquired immune defi- need for the intervention with unpredictable, severe, ciency syndrome (HIV/AIDS); tuberculosis; and acute events generally conferring higher financial malaria—underwent careful scrutiny for this report. risk; and (c) average age of death and level of disabil- They largely overlap with the essential packages ity, with more FRP provided by interventions that of DCP3 volumes 2 and 6: Reproductive, Maternal, improve the health of wage earners or address dis- Newborn, and Child Health (Black and others 2016) eases that cause high levels of disability, all else being and Major Infectious Diseases (Holmes and others equal (WHO 2014). 2017), respectively, although they are more selective. 48 Disease Control Priorities: Improving Health and Reducing Poverty Three additional remarks can be made on the criteria criteria to specific real-world policy questions would above. First, the exact thresholds for including an inter- involve (a) gathering more local information on demo- vention in a country’s HPP are context specific and graphics, disease burden, and costs which would influ- should be weighed against social preferences. For ence local estimates of value for money and of who are instance, how to compare cases of poverty averted to the “worst off,” and (b) conducting local or regional deaths averted is not obvious; UHC priority setting exer- studies that could quantify tradeoffs across each of these cises will reasonably differ as to how they weigh health criteria, such as the comparability of a child death and nonhealth outcomes. A scheme that seeks to priori- averted and a case of poverty averted. Empirical tize the needs of the poor but is relatively resource- advances in these areas could facilitate their incorpora- constrained may include more interventions that score tion into multi-criteria decision analysis as described by high on priority given to the worse off and fall below a Youngkong (2012) and others. strict willingness-to-pay threshold—reflecting high Interventions that fulfill the criteria above are shown health care opportunity costs. Thus, policy makers may in boldface in annex 3C and also noted alongside the be somewhat less likely to include interventions that appraisals in annex 3F. In all, 97 of 218 interventions provide significant FRP but not much health for money. could be classified as high priority according to the four At the same time, different levels of willingness to pay criteria above. Although the proposed HPP includes a may be defined for different health outcomes (Cairns preponderance of maternal and child health interven- 2016); for example, a country that is committed to tions and interventions against HIV/AIDS and tubercu- tackling HIV/AIDS (especially with aid from foreign losis in adults, a significant number of interventions also donors) may decide to include HIV-related interven- primarily address noncommunicable diseases (NCDs) tions despite their being somewhat less cost-effective and injuries. In terms of the scope of health conditions than interventions for other conditions. DCP3 does not addressed, these interventions go far beyond the take a position on the ethics of a choice like this but high-priority interventions typically included in the simply advocates for transparency and public account- global NCD discourse (WHO 2011). ability in the priority-setting process (that is, for explicit statements about trade-offs) as well as for consideration of health care opportunity costs (inefficiencies) and the COSTS OF ESSENTIAL UHC AND THE HPP possibility of failure in achieving stated levels of coverage Estimating the potential costs and health effects of pack- because of budget constraints. ages of health interventions is technically challenging in Second, the last criterion listed above is predicated on the face of limitations of current data, uncertainty about the analytic work conducted for the Lancet Commission future demographic and epidemiological patterns, and on Investing in Health. Before the commission issued its lack of established methods and tools that span disease 2013 report, “Global Health 2035: A World Converging groups. This chapter presents estimates of costs and con- within a Generation” (Jamison and others 2013), not all sequences of EUHC and the HPP, treating low-income of the interventions included in its “grand convergence” and lower-middle-income countries in the aggregate. package had the same rigorous evidence of value for These estimates are not intended to be normative or money. However, the commission’s original analysis precise, but rather illustrative of the magnitude and bal- deemed them to be effective and important to imple- ance of costs and health benefits that a given country ment as a package, and their costs and benefits were might expect. estimated for the commission as such. Hence, the com- The authors took a comparative statics approach to mission’s finding that the grand convergence package estimating cost and health gains from EUHC and the was affordable and cost-beneficial influenced this HPP, estimating the change in costs and mortality chapter’s judgment of the individual interventions’ value patterns that would be expected following an instanta- for money when implemented as part of a package, espe- neous increase in the coverage of services in the EUHC cially regarding interventions for which other economic and HPP lists and holding constant all other factors (for evidence was not available. example, demographics, epidemiology, and local prices) Finally, it is acknowledged that the design and imple- that might influence costs. The perspective taken on mentation of the criteria in this chapter required a con- costs was that of the ministry of health, which was siderable amount of judgment and de-emphasized assumed to be the payer for EUHC and the HPP. quantitative precision and comparability of criteria. To For this analysis, “universal” coverage was defined as some extent this is an artifact of the DCP3 process, 80 percent coverage; other groups have chosen targets which is intended to be illustrative rather than prescrip- ranging from 80 percent to 100 percent depending on tive for a wide range of local contexts. Applying these the costing perspective, intervention, and health Universal Health Coverage and Essential Packages of Care 49 condition (Black and others 2016; WHO 2013a). The defined—base case, worst case, and best case. For a set rationale for our 80 percent target is that the authors of key parameters in the costing model, a base case, determined it would be unrealistic and infeasible in worst case, and best case value was identified. The over- nearly all cases to achieve greater than 80 percent inter- all best and worst case estimates of UHC costs were vention coverage during the SDG period. obtained by simultaneously varying the values of all Watkins, Qi, and others (2017) present in detail the the key parameters to their most optimistic and pessi- methods, data, and assumptions behind this chapter’s mistic values, respectively. The point estimates and costing exercise. Costs were decomposed into the follow- uncertainty ranges presented subsequently reflect these ing three categories: direct costs of service delivery at the three scenarios. point of care—for example, personnel, drugs, and equip- Table 3.1 presents potential annual EUHC costs by ment; costs of facility-level ancillary services required to package, including per capita and total population deliver these services—for example, rents, building estimates of current spending, incremental costs, and maintenance, and laboratory and radiology services total costs (that is, the sum of current spending and (sometimes referred to as overhead or indirect costs); and incremental costs, where total costs reflect 80 percent program costs that support health services but occur above coverage). The largest single cost component of EUHC and separate from facility-level costs and are not easily is health system costs, comprising about 40 percent of allocable to specific services—for example, administration, total costs at full coverage. The second largest cost com- logistics, and surveillance activities. We refer to the first ponent is the service delivery costs related to the cardio- category of cost as “service delivery costs” and the second vascular, respiratory, and related disorders package. In and third categories together as “health system costs.” both country groups, the service delivery costs related to For each intervention, representative datasets that HIV/AIDS and STIs, malaria, and adult febrile illness contained relevant unit cost estimates were identified, were also very high. In lower-middle-income countries, and then costs were adjusted to “average” costs in low- the service delivery costs related to mental, neurological, and lower-middle-income countries using assumptions and substance use disorders were relatively high. It is also about the proportion of health care based on traded noteworthy that the share of incremental costs attrib- goods and, for the nontraded proportion, gradients in uted to NCDs is higher than the share of total costs health care worker salaries across various countries and attributed to NCDs. This finding reflects low levels of between low-income and lower-middle-income coun- current spending on NCDs and suggests that, in order to tries on average. Care was taken to extract unit cost achieve EUHC, all countries will need to pay particular estimates that reflected long-run average costs. Most unit attention to the incremental investments required to cost studies included ample detail on service delivery scale up NCD services. costs but did not factor in health system costs, so these Table 3.2 presents the potential total and incremental were added as markups on service delivery costs using annual costs of EUHC and the HPP in low- and lower- supplementary datasets and assumptions (Boyle and middle-income countries, including uncertainty ranges others 2015, Seshadria and others 2015). derived from the best- and worst-case scenario analyses The next step was to identify the population in need of described previously. The total cost per person of sustain- the intervention. Previously published estimates of inci- ing the HPP and EUHC at full coverage would be US$42 dence or prevalence of various causes of disease or injury and US$76, respectively, in low-income countries and were compiled and mapped against the EUHC interven- US$58 and US$110, respectively, in lower-middle-income tions (Vos and others 2016; WHO 2016).3 In some cases, countries. Getting to full implementation of the HPP and additional adjustments were made to estimates of popula- EUHC would require, annually, an additional 3.1 percent tion in need; for example, the proportion of the population and 6.4 percent, respectively, of current income in low- requiring screening for diabetes (based on risk level) was income countries and 1.5 percent and 2.9 percent, respec- first estimated and then divided by three to reflect the rec- tively, in lower-middle-income countries. ommendation for screening once every three years on To put these cost estimates in context, combined average. The final step was to estimate current coverage of annual per capita health expenditure by government and each intervention using coverage indicators from the donors in low- and lower-middle-income countries is WHO Global Health Observatory database or reasonable currently US$25 and US$31, respectively, with out-of- proxies for coverage (WHO 2016). pocket spending by the population being about as large As described by Watkins, Qi, and Horton (2017), the again (WHO 2016). Assuming that the objective of UHC authors attempted to quantify major sources of uncer- is to successfully crowd out out-of-pocket spending at tainty in the cost estimates. Three scenarios were the point of care through prepayment mechanisms and 50 Disease Control Priorities: Improving Health and Reducing Poverty Table 3.1 Costs of Essential UHC in Low-Income and Lower-Middle-Income Countries, by DCP3 Intervention Package Current annual Incremental spending, annual cost, Total annual Total annual cost, Current annual population (US$ Incremental annual population (US$ cost, per population (US$ Share of total spending, per capita billions) cost, per capitaa billions)a capitab billions)c costs (%)d Panel a. Low-income countries Age related 1. Maternal and newborn health $1.3 $1.2 $1.8 $1.6 $3.1 $2.8 6.1 (MNH) 2. Child health (CHH) $2.3 $2.1 $1.2 $1.0 $3.4 $3.1 6.7 3. School-age health and $0.094 $0.085 $0.20 $0.18 $0.30 $0.27 0.58 development (SAH) 4. Adolescent health and $0.31 $0.28 $0.44 $0.40 $0.75 $0.68 1.5 development (AHD) 5. Reproductive health and $0.82 $0.74 $0.38 $0.34 $1.2 $1.1 2.3 contraception (RHC) Infectious diseases 6. HIV and STIs (HIV) $3.6 $3.2 $4.0 $3.6 $7.6 $6.8 15 7. Tuberculosis (TB) $0.34 $0.31 $0.15 $0.13 $0.49 $0.44 0.95 8. Malaria and adult febrile $2.4 $2.1 $2.6 $2.4 $5.0 $4.5 9.7 illness (MAL) Universal Health Coverage and Essential Packages of Care 9. Neglected tropical diseases $0.33 $0.30 $0.31 $0.28 $0.63 $0.57 1.2 (NTD) 10. Pandemic and emergency $0.016 $0.014 $0.71 $0.63 $0.75 $0.68 1.5 preparedness (PAN) Noncommunicable disease and injury 11. Cardiovascular, respiratory, $0.67 $0.60 $13 $11 $13 $12 26 and related disorders (CVD) 12. Cancer (CAN) $0.21 $0.19 $2.5 $2.2 $2.7 $2.4 5.2 13. Mental, neurological, and $0.49 $0.44 $1.8 $1.6 $2.3 $2.1 4.5 substance use disorders (MNS) 14. Musculoskeletal disorders $0.75 $0.67 $1.2 $1.1 $1.5 $1.4 3.0 (MSK) table continues next page 51 52 Disease Control Priorities: Improving Health and Reducing Poverty Table 3.1 Costs of Essential UHC in Low-Income and Lower-Middle-Income Countries, by DCP3 Intervention Package (continued) Current annual Incremental spending, annual cost, Total annual Total annual cost, Current annual population (US$ Incremental annual population (US$ cost, per population (US$ Share of total spending, per capita billions) cost, per capitaa billions)a capitab billions)c costs (%)d 15. Congenital and genetic $0.59 $0.53 $1.2 $1.1 $1.8 $1.7 3.6 disorders (CGD) 16. Injury prevention (IPR) $0.0044 $0.0039 $0.039 $0.035 $0.044 $0.039 0.085 17. Environmental $0.050 $0.045 $0.049 $0.044 $0.10 $0.089 0.19 improvement (ENV) Health services 18. Surgery (SUR) $1.6 $1.5 $1.3 $1.1 $2.9 $2.6 5.6 19. Rehabilitation (RHB) $0.10 $0.089 $1.5 $1.3 $1.6 $1.4 3.1 20. Palliative care and pain $0.11 $0.10 $1.6 $1.5 $1.7 $1.6 3.4 control (PCP) 21. Pathology (PTH) $0.71 $0.64 $1.8 $1.7 $2.6 $2.3 5.1 Totals Total service delivery costs $16 $14 $36 $32 $51 $46 (sum of costs by package) De-duplicated service delivery costs $12 $11 $31 $28 $43 $39 60 Total health system costs $7.9 $7.1 $20 $18 $29 $26 40 Total cost (sum of service delivery $20 $18 $51 $46 $72 $65 100 and health systems)c table continues next page Table 3.1 Costs of Essential UHC in Low-Income and Lower-Middle-Income Countries, by DCP3 Intervention Package (continued) Current annual Incremental Current annual spending, Incremental annual cost, Total annual Package spending, per population (US$ annual cost, per population (US$ Total annual cost, population share of total capita billions) capitaa billions)a cost, per capitab (US$ billions)b costs Panel b. Lower-middle-income countries Age related 1. Maternal and newborn health (MNH) $1.6 $4.4 $2.1 $5.5 $3.7 $9.9 5.3 2. Child health (CHH) $3.0 $8.1 $0.99 $2.6 $4.0 $11 5.8 3. School-age health and development (SAH) $0.083 $0.22 $0.21 $0.57 $0.29 $0.79 0.42 4. Adolescent health and development (AHD) $0.37 $0.99 $0.53 $1.4 $0.90 $2.4 1.3 5. Reproductive health and contraception (RHC) $1.6 $4.4 $0.45 $1.2 $2.1 $5.6 3.0 Infectious diseases 6. HIV and STIs (HIV) $2.6 $7.0 $4.1 $11 $6.7 $18 9.6 7. Tuberculosis (TB) $0.34 $0.91 $0.19 $0.50 $0.53 $1.4 0.76 8. Malaria and adult febrile illness (MAL) $4.1 $11 $2.3 $6.2 $6.4 $17 9.1 9. Neglected tropical diseases (NTD) $0.37 $1.0 $0.39 $1.0 $0.74 $2.0 1.1 10. Pandemic and emergency 0.094 0.25 $0.66 $1.8 $0.75 $2.0 1.1 preparedness (PAN) Universal Health Coverage and Essential Packages of Care Noncommunicable disease and injury 11. Cardiovascular, respiratory, and $9.4 $25 $15 $40 $24 $65 35 related disorders (CVD) 12. Cancer (CAN) $0.64 $1.7 $1.8 $4.7 $2.4 $6.4 3.5 13. Mental, neurological, and substance $1.8 $4.8 $3.7 $9.8 $5.47 $15 7.8 use disorders (MNS) 14. Musculoskeletal disorders (MSK) $1.1 $3.0 $2.1 $5.6 $2.8 $7.5 4.0 15. Congenital and genetic disorders (CGD) $0.74 $2.0 $1.3 $3.5 $2.0 $5.4 2.9 16. Injury prevention (IPR) $0.021 $0.055 $0.11 $0.30 $0.13 $0.36 0.19 17. Environmental improvement (ENV) $0.11 $0.30 $0.10 $0.26 $0.16 $0.42 0.23 table continues next page 53 54 Disease Control Priorities: Improving Health and Reducing Poverty Table 3.1 Costs of Essential UHC in Low-Income and Lower-Middle-Income Countries, by DCP3 Intervention Package (continued) Current annual Incremental Current annual spending, Incremental annual cost, Total annual Package spending, per population (US$ annual cost, per population (US$ Total annual cost, population share of total capita billions) capitaa billions)a cost, per capitab (US$ billions)b costs Health services 18. Surgery (SUR) $1.6 $4.2 $0.97 $2.6 $2.6 $6.8 3.7 19. Rehabilitation (RHB) $0.41 $1.1 $2.9 $7.6 $3.3 $8.7 4.7 20. Palliative care and pain control (PCP) $0.071 $0.19 $0.50 $1.3 $0.57 $1.5 0.81 21. Pathology (PTH) $1.0 $2.6 $2.1 $5.6 $3.6 $9.7 5.2 Totals Total service delivery costs (sum of costs $30 $81 $40 $110 $70 $190 by package) De-duplicated service delivery costs $16 $44 $35 $93 $60 $160 60 Total health system costs $11 $29 $23 $62 $40 $110 40 Total cost (sum of service delivery and $27 $73 $58 $160 $101 $270 100 health systems)c Source: Watkins, Qi, and others 2017. Note: All dollar amounts are in U.S. dollars. DCP3 = Disease Control Priorities, third edition; HIV = human immunodeficiency virus; STIs = sexually transmitted infections; UHC = universal health coverage. a. Incremental cost of scaling is from current coverage to 80 percent coverage. b. Cost is at 80 percent coverage. c. Total costs are the sum of “de-duplicated service delivery costs” and “total health system costs.” The de-duplicated service delivery costs are lower than the total service delivery costs because a number of interventions are included in more than one DCP3 essential package. d. Two types of shares are presented in this column. First, the shares of costs presented for each of the 21 essential packages use, as the denominator, the de-duplicated service delivery costs, so the sum of these shares exceeds 100 percent because of duplication; however the share of any given package can be interpreted as the remaining fraction of the total EUHC service delivery cost if the interventions in all other packages were removed. Second, the shares of costs presented in the totals section reflect the relative proportion of EUHC costs related to service delivery and to health system strengthening, with the sum of these two being the total cost of EUHC. Table 3.2 Total and Incremental Annual Costs of Essential UHC and the Highest-Priority Package (HPP) in 2015 Lower-middle-income Low-income countries countries HPP EUHC HPP EUHC a 1. Incremental annual cost (US$ billions) 23 48 82 160 (9.2 to 51) (20 to 100) (32 to 180) (66 to 350) 2. Incremental annual cost per person (US$) 26 53 31 61 (10 to 57) (22 to 110) (12 to 67) (25 to 130) 3. Total annual cost (US$ billions)a 38 68 160 280 (19 to 71) (34 to 130) (81 to 280) (150 to 500) 4. Total annual cost per person (US$) 42 76 58 110 (21 to 79) (37 to 140) (30 to 100) (54 to 190) 5. Incremental annual cost as a share of current GNI (%)b 3.1 6.4 1.5 2.9 (1.2 to 6.9) (2.6 to 13) (0.57 to 3.2) (1.2 to 6.2) 6. Total annual cost as a share of current GNI (%)b 5.1 9.1 2.8 5.2 (2.5 to 9.5) (4.5 to 17) (1.4 to 4.8) (2.6 to 9.1) Source: Watkins, Qi, and others 2017. Note: EUHC = Essential Universal Health Coverage; GNI = gross national income; UHC = Universal Health Coverage. Incremental annual cost is the estimated cost of going from current to full implementation (80 percent population coverage) of the EUHC and HPP interventions. The total annual cost is the incremental cost plus current spending assuming the same cost structure for current and incremental investments. Estimated costs are inclusive of estimates for (large) health system strengthening cost and are steady-state (or long-run average) costs in that investments to achieve higher levels of coverage and to cover depreciation are included. a. The 2015 population of low-income countries was 0.90 billion. For lower-middle-income countries, it was 2.7 billion. Population sizes were estimated using data from UN DESA 2017 according to the country classifications listed at the end of this chapter. b. The 2015 GNI of low-income countries was $0.75 trillion and for lower-middle income countries it was $5.4 trillion. Aggregate GNI figures were estimated using data from the World Bank.4 pooled contributions, these cost estimates suggest adults and substantial investments in NCDs and injury that current government and donor spending will need care at health centers and first-level hospitals. approximately to double or triple to finance the HPP or Finally, DCP3’s cost estimates are in line with those EUHC packages. These implied shortfalls are compara- estimated by others. Earlier work based on the WHO ble to a recent costing exercise in Ethiopia (Ethiopia, Commission on Macroeconomics and Health and Ministry of Health 2015) that estimated that a 30–80 the High Level Taskforce for Innovative International percent increase in available resources would be required Financing of Health Systems suggested that the mini- to finance universal coverage of a very basic package of mum total annual public expenditure on UHC in LMICs essential health services in Ethiopia. would need to be about US$86 per capita or 5 percent of The incremental cost of reaching full coverage is current GDP per capita, whichever is larger (McIntyre, significant; probably feasible in lower-middle-income Meheus, and Rottingen 2017). A more recent costing countries but unlikely to be feasible in low-income exercise by WHO has suggested that the incremental countries without additional external support. For annual public expenditure on UHC in LMICs would comparison, the annual incremental cost of the need to be US$58 (ranging US$22–US$167) per capita Lancet Commission on Investing in Health’s grand con- (in 2014 U.S. dollars) across LMICs in order to achieve vergence package was about 1 percent of current per full implementation by 2030 (Stenberg and others 2017). capita income overall as compared to 2–3 percent of (The WHO study only reported incremental costs, not current per capita income in this chapter’s HPP total costs. Watkins, Qi, and others [2017] compare the (Jamison and others 2013). The higher cost of DCP3’s contents of the WHO’s package and DCP3’s EUHC and HPP results from the inclusion of a wider scope of HPP.) Taken together, these figures also suggest that, interventions, including both the reproductive, mater- if resources for UHC do not increase in low-income nal, neonatal, and child health interventions in the countries, even the HPP—however attractive on health Lancet Commission on Investing in Health package and and efficiency grounds—would need to be significantly additional interventions for major infectious diseases in reduced in scope. Universal Health Coverage and Essential Packages of Care 55 HEALTH CONSEQUENCES OF ESSENTIAL lower-middle-income countries, with relatively more prog- UHC AND THE HPP ress in low-income countries. However, at 80 percent cover- age and usual levels of delivery quality, the HPP and EUHC Watkins, Norheim, and others (2017) present in detail would achieve roughly half and two-thirds, respectively, of the data sources, methods, and assumptions that are the mortality reduction target. used to estimate the mortality impact of EUHC and the There are two sets of factors that influence the short- HPP. In brief, the overall framework for the impact fall in mortality reduction. First, 80 percent is a partic- assessment was the supplementary SDG 3 target pro- ularly modest target for some conditions, such as posed by Norheim and others (2015) of a 40 percent childhood illnesses and HIV/AIDS and tuberculosis reduction in deaths under age 70 years by 2030. This among adults. Scaling up the child health and infec- chapter projects total deaths in 2030—by age group, tious diseases packages to 95% or higher coverage, with gender, and cause—using UN Population Division esti- more optimistic assumptions about the quality of deliv- mates of population size (UN DESA 2017) and ery, would facilitate countries reaching the mortality cause-specific mortality rates (by age group and gender) target at least for these conditions. Second, lower- using the WHO’s most recent Global Health Estimates middle-income countries face greater challenges in database (Mathers and others 2018) reaching the target because of the predominance of Estimates of mortality reduction from specific noncommunicable diseases and injuries. The HPP and HPP and EUHC interventions implemented a hybrid EUHC interventions for these conditions, particularly approach. For under-five years, maternal, HIV/AIDS, for neoplasms, are relatively less effective even at and tuberculosis deaths, the analysis drew on the impact high levels of coverage. In addition, these countries modeling undertaken for the Commission on Investing face demographic and epidemiologic headwinds, with in Health (Boyle and others 2015). For NCDs and inju- greater increases in total deaths and in the share of pro- ries, as well as for selected causes of death from infec- jected deaths in 2030 due to noncommunicable diseases tious disease in adults, the authors identified a subset of and injuries. The findings of this analysis suggest that, interventions for which there was strong evidence for a particularly in lower-middle-income countries, meet- large relative effect on cause-specific mortality. These ing the target will be feasible only if health sector inter- relative reductions in mortality were then applied to ventions against NCDs and injuries are complemented cause-specific mortality rates, focusing on deaths in the by strong intersectoral policies such as tobacco taxation groups ages 5–69 years. The impact estimates were then and control, reduction of air pollution, and road safety adjusted to reflect the proportion of deaths that would that can reduce the risk of incidence of fatal and nonfa- be affected by an increase in intervention coverage. tal NCDs and injuries. These sorts of interventions are Effect sizes were also adjusted downward to account for addressed in greater detail in chapter 2 of this volume suboptimal quality of delivery, including imperfect (Watkins and others 2018). adherence. The adjusted effect sizes were then applied to projected 2030 estimates of deaths, by cause, in low- IMPLEMENTING ESSENTIAL UHC income and lower-middle-income countries. Table 3.3 presents these estimates of the potential The primary focus of this chapter and of DCP3 as a mortality consequences of the HPP and EUHC in 2030. whole has been to develop detailed essential packages of They can be regarded as conservative estimates: other care. At the same time, the interventions contained in EUHC and HPP interventions can reduce mortality as EUHC and the HPP would translate to gains in popula- well as disability (the latter of which is not the focus of tion health only through expanded uptake and improved this analysis). A subset of NCD interventions also efficiency and quality of health care (figure 1.1 in reduces mortality over the age of 70 years, although chapter 1 of this volume, Jamison and others 2018). these deaths are not counted toward the target. Finally, Further, EUHC and the HPP require health systems that many EUHC and HPP interventions have well-known have adequate human and material resources to deliver a nonhealth benefits, such as increased productivity, edu- wide range of services. This section of the chapter dis- cational attainment, economic benefits to women result- cusses some important considerations for implementing ing from reduced fertility rates, and so on, that make the EUHC and the HPP. These include reducing barriers to suite of societal benefits of UHC even larger. the uptake of priority health services, improving the The impact estimates in table 3.3 suggest that HPP and quality of services provided, strengthening the building EUHC implementation will facilitate substantial prog- blocks of health systems, and supporting the institution- ress toward the SDG 3 target in both low-income and alization of priority setting. 56 Disease Control Priorities: Improving Health and Reducing Poverty Table 3.3 Premature Deaths Averted in 2030, by Age Group and Cause, through Full Implementation of EUHC and the HPP, Low-Income and Lower-Middle- Income Countries Low-income countriesb Lower-middle-income countriesb Projected Expected reduction in premature Projected Expected reduction in premature number of deaths from number of deaths from Age group or premature 40x30 reduction premature 40x30 reduction condition deaths, 2030a targetc HPP EUHC deaths, 2030a targetc HPP EUHC By age group 0–4 2.2 1.5 0.62 0.77 3.3 2.2 1.1 1.3 5–69 5.2 1.5 0.99 1.2 14 4.8 2.2 2.9 0–69 7.4 3.0 1.6 2.0 17 7.0 3.2 4.2 d By cause (age 5+) I. Group I 1.9 0.76 0.59 0.65 3.2 1.5 0.85 0.94 Tuberculosis 0.34 0.22 0.11 0.13 0.90 0.60 0.29 0.35 HIV/AIDS 0.44 0.29 0.18 0.20 0.48 0.32 0.23 0.26 Malaria 0.087 0.058 0.051 0.051 0.055 0.037 0.026 0.026 Maternal conditions 0.17 0.11 0.075 0.086 0.20 0.13 0.079 0.092 Other diseases 0.90 0.074 0.18 0.18 1.6 0.40 0.22 0.22 II. Group II 2.5 0.60 0.36 0.53 8.9 2.7 1.3 1.9 Neoplasms 0.65 0.22 0.010 0.039 1.8 0.60 0.10 0.16 Universal Health Coverage and Essential Packages of Care Cardiovascular 0.93 0.31 0.24 0.36 4.0 1.3 0.89 1.4 diseases Other diseases 0.93 0.076 0.11 0.13 3.2 0.80 0.28 0.35 III. Group III 0.77 0.13 0.043 0.060 2.0 0.54 0.070 0.10 Road injuries 0.25 0.085 0.032 0.046 0.57 0.19 0.048 0.069 Other injuries 0.52 0.042 0.010 0.014 1.4 0.36 0.022 0.032 Source: Watkins, Norheim, and others 2017. Note: All estimates are in millions of deaths. The 40x30 reduction target includes a 40 percent reduction in deaths 0–69 overall; a two-thirds reduction in under-five deaths and adult deaths from tuberculosis, HIV/AIDS, malaria, and maternal conditions; and a one-third reduction in deaths from major noncommunicable diseases. The quantitative targets above reflect these goals; however, targets for the residual categories (“other diseases” and “other injuries”) have been calculated in light of the targets for specific causes of death so that the total number of target deaths 5–69 is sufficient to meet the 40 x 30 target. a. A death under age 70 years is defined as premature. b. See unnumbered endnote for World Bank classification of countries by income group. UN and WHO data were aggregated according to these groupings. c. A reduction target of 40 x 30 is defined as a 40 percent reduction in premature deaths by 2030, relative to the number that would have occurred had 2015 death rates persisted to 2030. The UN Population Prospects (UN DESA 2017) median population projection for 2030 was used to provide the population totals for calculating deaths by age and sex. d. WHO’s Global Health Estimates (Mathers and others 2018) provided the 2015 cause distributions of deaths for these calculations. 57 Reducing Barriers to Intervention Uptake longitudinal interventions (such as chronic management Ng and others (2014) have proposed the concept of of HIV/AIDS) and acute care interventions (such as “effective coverage” as a quantitative indicator of the fracture reduction and fixation) need to be decentralized effect of UHC. The concept goes beyond the usual as much as possible because of the frequency or urgency notion of coverage, which is often measured as the prob- of contact with the health system. Such services, which ability that specific health services are available at a given make up nearly 75 percent of the recommended EUHC facility. Effective coverage, in contrast, incorporates mea- interventions, require highly decentralized facilities at sures of intervention uptake by those in need as well as high density in communities, including in hard-to-reach measures of the quality of the care provided, and thus it populations, to reach universal coverage. The interven- considers the actual health gain that an intervention is tions on the community, health center, and first-level likely to produce in the population. Although the use of hospital platforms can build a foundation for efficient quantitative indicators for UHC continues to stimulate primary health care (annex 3C). At the same time, rou- international debate, the principle that the health impact tine, one-off services (such as immunization programs of UHC is bounded by effective coverage—constraints or cataract surgery) can often be efficiently delivered on access to and quality of care—is intuitive. Hence, a through stand-alone, targeted programs appropriate to UHC scheme and associated package can truly claim to the epidemiology of the country or region (Atun and be “universal” only once full effective coverage has been others 2010). Finally, complex, high-risk services (such achieved. as chemotherapy treatment of childhood leukemia) Removing or reducing key barriers to intervention generally need to be centralized, with strong referral uptake is crucial to achieving full effective coverage. systems, to ensure sufficient quality. Barriers to intervention uptake fall into four broad Sociocultural and legal barriers, which may be inter- types: economic, geographic, sociocultural, or legal. twined in cause and effect, vary according to both the Economic barriers feature prominently in the UHC characteristics of the intervention and the country con- discourse, and they can be partially remediated through text. Disease stigma may influence individuals’ willingness public finance. Still, public finance usually addresses only to seek care or—consciously or unconsciously—providers’ the direct cost of care. Direct nonmedical costs such as attitudes toward these individuals. Low knowledge or transportation and food expenses that are borne by indi- health literacy can also impede intervention uptake, and viduals are not easily remedied by prepayment, nor are this has been a major focus of information, education, the economic consequences of taking time off work or and communication interventions. Finally, there may be school to receive care. Despite currently limited evidence, legal barriers to care, or mandates to provide certain kinds these sorts of barriers may be more amenable to intersec- of care, that have little to do with stigma or culture. For toral action (for example, paid sick leave and subsidized example, restrictions on prescribing by nurses or mid- public transportation for visits to health facilities) than level practitioners may reduce the opportunities to changes in the delivery or financing of health care. for individuals with chronic illness to receive needed In addition, social development policies and other medications. approaches complementary to public finance may be Table 3.4 provides examples from DCP3 of measures needed to improve access to marginalized groups, partic- that have been used to expand access to care, either by ularly in countries with high levels of political, economic, reducing access barriers or by inducing demand for and social inequality. Ideally, health insurance should be health care. integrated with broader social protection measures that are implemented outside the health sector. At a mini- mum, the spirit of the progressive universalist approach Improving the Quality of Essential UHC to UHC implies that user fees should be reduced as much In addition to affordability and availability, the quality of as possible or eliminated entirely, and in some cases, services is also critical to the success of UHC schemes. If additional steps—such as cash transfers or other finan- users do not perceive services as valuable, public support cial incentives for the poor—could be considered. will falter, undermining the politics of implementing Geographic barriers arise when the distribution of UHC (Savedoff and others 2012). Low quality of care health facilities does not match the distribution of the can thus reduce the positive health impact of otherwise population’s health needs. The EUHC package’s plat- effective and cost-effective interventions. From an eco- form structure allows health planners to identify what nomic standpoint, low quality suggests that more money sorts of health facilities are most needed and what sort of needs to be spent on a health service than the estimates capacity is required at those facilities. In general, of cost-effectiveness would imply. As discussed in 58 Disease Control Priorities: Improving Health and Reducing Poverty Table 3.4 Selected Examples of Measures to Address Barriers to Health Care Access, LMICs Barrier type Examples Economic Bus fares to support attendance at STI clinics Conditional cash transfers for antenatal care Geographic Decentralization of chronic disease care, for example, for HIV and diabetes Extension of antenatal care using community health workers Mobile units to provide screening and care for HIV and tuberculosis Sociocultural Information and education about cervical cancer and the benefits of screening Ensuring that health care providers of the same sex are available when requested Educational campaigns to reduce stigma concerning mental health Legal Easing legal restrictions on access to family planning measures Legal measures to ensure confidential reporting of and care following episodes of intimate partner violence Sources: Black and others 2016; Gelband and others 2015; Patel and others 2015; Prabhakaran and others 2017; Holmes and others 2017. Note: LMICs = low- and middle-income countries; STI = sexually transmitted infection. chapter 10 of this volume (Peabody and others 2018), • National essential medicines and diagnostics lists and health planners can improve outcomes and reduce inef- formularies ficiency in spending on the UHC intervention package • Use of community health workers and technologies by integrating into routine health care four types of (such as mHealth) to promote medication adherence measures that ensure high quality: • Creation of high-volume, specialized centers to deal with complex but not urgent problems • Measuring activities and providing feedback • Adequate control of pain, including pain related to • Identifying relevant standards for these measures acute injuries or severe life-limiting illnesses. using scientific evidence, guidelines, and best practices • Ensuring that providers are adequately trained to Implications of EUHC for the Building deliver the intervention with adequate management Blocks of Health Systems and oversight Once consensus has been reached on a health benefits • Motivating and aligning providers through incen- package such as the HPP or EUHC, with political and tives, which may be either financial (such as results- public buy-in, the next step would be to implement this based financing) or nonfinancial (such as reputation agenda within the context of the current health system. enhancement among peers). Using the WHO health systems framework (WHO 2007) as a point of reference, the most critical implications of In some cases, investments in improving quality can the EUHC package for health systems can be identified, translate to improvements in health over a shorter time particularly leadership and governance challenges, UHC frame than introducing a new health technology or financing issues, health workforce constraints, gaps in policy. Costs related to quality improvement are covered medical product and technology availability, and limited in the EUHC and HPP cost estimates as part of health information and research functions. system costs (see table 3.1). The following are some examples from DCP3 of measures that have been used Leadership and Governance to improve the quality of care for specific health A recent case series of early-adopter UHC countries conditions: highlighted the importance of leadership and gover- nance as well as the strategic use of social and economic • Clinical checklists for complex tasks such as surgical crises as opportunities for moving forward with UHC procedures reforms (Reich and others 2016). National UHC plans • Hospital infection control policies and procedures and strategies would rely on strong regulatory mea- • Clinical guidelines for specific syndromes or diseases, sures and bureaucracy. As mentioned, well-considered including guidance on reducing unnecessary antibi- management of private interests and agendas (such as otic use donors, industries, and advocacy groups) can help ensure Universal Health Coverage and Essential Packages of Care 59 that an economically efficient and equitable form of complex and oriented toward management of NCDs, UHC moves forward. At the same time, mechanisms for specialized systems and providers will also be required in feedback and response can ensure that governments are many cases (Samb and others 2010). The EUHC and the accountable to constituents (Kieslich and others 2016). HPP interventions include a limited number of special- In addition, management competence at a subna- ized and referral services that reflect these future needs, tional level is incredibly important in ensuring that but the human and material resources required to deliver health services are delivered effectively. In particular, these services at any reasonable level of coverage can take large clinics and first-level and referral hospitals require years to develop. Hence, low-income countries could robust administrative capacity and health information consider adding capacity for specialized services that management systems. A variety of studies have demon- provide good value for money, such as specialized sur- strated that the quality of management is critical to the gery and cancer centers (Gelband and others 2015; Mock delivery of high-quality health services (Mills 2014). and others 2015), as a first step during the SDG period toward more advanced, comprehensive health systems. UHC Financing Issues around financing UHC have been reviewed by Medical Product and Technology Availability others and are not treated in detail here (WHO 2010; Implementing EUHC will also require greater availabil- World Bank 2016). Nevertheless, it is important to ity of existing medical products and technologies. recognize that all early-adopter countries, regardless of Problems and proposed solutions to gaps in access to income level, have faced challenges in raising sufficient essential medicines have been reviewed by others and public revenues for UHC (Reich and others 2016). This are not dealt with here (Howitt and others 2012; Wirtz chapter provides some general conclusions on the likely and others 2017). However, DCP3’s model benefits magnitude of UHC costs (table 3.2), which in most packages could provide a useful input to the revision of countries suggests a need for increases in both total national formularies and essential medicines lists. health expenditure and the government’s share of total Procurement bodies and local agencies that regulate and health expenditure. Conversely, the HPP would need to manage supply chains could then be strengthened along be reduced substantially or disinvestment in interven- the lines of these essential medicines so that they reach tions would be needed if resource levels could not be the last mile and make UHC truly universal. Additionally, increased. This costing exercise also suggests that many DCP3 has stressed the importance of using generic low-income countries would need to continue relying medications throughout (Patel and others 2015; on development assistance for health as a supplement to Prabhakaran and others 2017). Generic medications public finance for priority conditions, such as HIV/ nearly always have equivalent clinical effectiveness and AIDS. Notably, countries from around the world have can be a major factor ensuring the affordability and successfully employed a wide range of public, private, sustainability of UHC. and hybrid financing models to achieve UHC (Reich and others 2016). Financing models are usually path depen- Information and Research dent, but the key objective in any case is to divert out-of- As critical as information and research are to health sys- pocket payments into pooled and prepayment tems, they are often the most neglected of all health sys- mechanisms and to establish fairness in risk pooling. In tem functions in limited-resource settings. In particular, addition, measures such as price negotiation with indus- strong disease surveillance programs can inform the try and local health technology assessment are crucial to priorities for UHC and track progress. Box 3.2 summa- managing cost escalation and maximizing efficiency of rizes some of the major information needs in limited- public expenditure (Nicholson and others 2015). resource settings, emphasizing disease surveillance. Although research is often perceived as a global pub- Health Workforce lic good rather than a specific national priority for Short- to medium-run constraints on the health work- limited-resource settings, a local research agenda could force are probably among the most important bottle- prioritize the validation of interventions and policies necks in implementation of UHC reforms (Reich and that have been tried in other settings but that likely vary others 2016; Stenberg and others 2017). DCP3 has high- significantly in effectiveness and cost-effectiveness lighted numerous examples of task sharing that allow for because of differences in culture, language, disease epi- broader coverage of essential health services, such as the demiology, and health system arrangements. In the long use of midlevel providers and general physicians for basic term, many countries could begin to develop completely first-level hospital surgical procedures (Mock and others novel interventions guided by local experience. 2015). At the same time, as health systems become more Developing local capacity to conduct health technology 60 Disease Control Priorities: Improving Health and Reducing Poverty Box 3.2 Health System Information and Research Needs in Limited-Resource Settings Routine, reliable, low-cost, long-term surveillance are surveillance functions. However, effective mod- vital to maintaining public health and providing effec- els have been implemented successfully in some tive medical care. Health surveillance systems are also countries, often at low cost. In India, for exam- critical to tracking trends in health conditions of the ple, the Registrar General has created the Million population, detecting new epidemics and outbreaks Death Study in which a verbal autopsy instrument (such as Ebola and Zika virus infection), evaluating the is added to its Sample Registration System to success of control programs, and improving account- obtain cause-of-death data, by age, from about ability for health expenditures. Surveillance supports 1.4 million nationally representative homes from five objectives, although, unfortunately, systems cover- every state. The overall system costs less than US$1 ing all five functions are rare in most LMICs: per person annually. The Million Death Study has transformed disease control in India by enhancing • Monitoring of population health status (the most the amount and quality of health data available for important aspect of which is premature mortal- public health officials (Jha 2014). ity) to guide policy choices • Efficiency in use of resources A variety of new approaches could be taken to • Disease surveillance to aid control programs expand surveillance to support the core goals of • Epidemic alert to enable rapid response and UHC and increase the demand for such surveil- containment lance. These include increasing global assistance • Identification of new risk factors or intermediate allocations from development agencies, expand- determinants of disease ing monitoring for NCDs in particular, and pro- moting international health audit days. More Currently, no low-income country has adequate information on these opportunities can be found coverage of these key and often quite different in annex 3G. assessment and health policy analysis, while still aspira- At the same time, experience from all parts of the tional for a number of LMICs, will ensure that the UHC world has shown that setting priorities can also evolve in agenda is realized in the most effective, efficient, and an inefficient and potentially inequitable manner (Kieslich equitable manner possible. and others 2016). Political calculus, inertia, efforts of prominent disease advocates, and donor priorities, among other influences, can at times create inefficiencies and The Role of Priority-Setting Institutions increase inequalities if not well managed. In contrast, pub- This chapter has argued that UHC in some form can be lic sector priorities need to account for the preferences and realized in nearly every country and that an array of expectations of the local population, which may deviate highly cost-effective, currently available interventions from what clinicians or technocrats would predict or can be efficiently employed in limited resource settings extrapolate from other settings (Larson and others 2015). to help countries reach most, if not all, of the SDG 3 Robust, transparent, and publicly accountable priority- goals and targets. By using economic tools and evidence, setting institutions are essential in all countries, but most countries can develop health benefits packages that LMICs do not yet have these sorts of institutions. Notable address their major health concerns on the basis of alloc- country examples from across the development spectrum ative efficiency, equity, and feasibility. Benefits packages can provide a template for building local capacity for designed in this way provide good value for money. By health policy analysis and health technology assessment in dramatically improving population health, they could LMICs (Li and others 2016). Academic organizations and also, over time, foster economic development and sup- partnerships such as the International Decision Support port other social goals, including poverty reduction. Initiative also play an important role in building local Universal Health Coverage and Essential Packages of Care 61 capacity to conduct health technology assessment and ACKNOWLEDGMENTS policy analysis in lower resource settings.5 As resources increase within a country, the possibili- The authors wish to thank the following individuals for ties for what a UHC scheme could include will grow as their contributions to the background materials for this well. Glassman and others (2016) have described the chapter: Matthew Schneider and Carol Levin, who con- process of defining a health benefits package as cyclical, tributed to the cost analyses for HIV/AIDS and surgery, with iterative improvements and revisions over time as respectively; Kjell Arne Johansson and Matthew Coates, well as expansions in the services offered. At the same who produced the indicators for conditions that affect time, Making Fair Choices argued that, when an existing the worse off; and the volume and series editors of DCP3 package of interventions is not yet universally available, and the Advisory Committee to the Editors, who provided it is fairer to focus on achieving full coverage of that input on the conceptualization of this chapter and—in package before adding interventions to the package many cases—critically reviewed draft tables and annexes. (WHO 2014). In practice, this principle can be difficult to follow, and in some cases, novel interventions are ANNEXES arguably worth considering on efficiency grounds if they result in significant economies of scope. Yet within the The following annexes to this chapter are available at context of DCP3, the ethical principle suggests that, in http://www.dcp-3.org/DCP. general, all countries could first strive to achieve full coverage of the HPP (that is, of the most cost-effective • Annex 3A: An Essential Package of Interventions to interventions in a given setting), begin to add the EUHC Address Congenital and Genetic Disorders interventions incrementally, and then expand to a • Annex 3B: An Essential Package of Interventions to broader range of interventions similar to those available Address Musculoskeletal Disorders in upper-middle-income or high-income settings. • Annex 3C: Essential Universal Health Coverage: For most low-income countries, implementing and Interventions and Platforms scaling up a package like the HPP would likely be the • Annex 3D: Notes on the Essential UHC Interventions focus during the SDG period. (Low-income countries in Annex 3C that wish to offer a broader set of interventions than • Annex 3E: Methods for Appraisal of Essential UHC what is outlined in the HPP could continue to deliver Interventions this set of interventions; however, lower-priority inter- • Annex 3F: Findings from the Appraisal of Essential ventions would need to be identified from among this UHC Interventions set and financed through copayment or cost recovery • Annex 3G: The Role of Surveillance in Achieving UHC mechanisms until public budgets were sufficient to cover the entire set [WHO 2014].) For lower-middle-income countries, the initial focus might be reaching full cover- NOTES age of the HPP (if full coverage has not already been World Bank Income Classifications as of July 2014 are as achieved), then moving toward full EUHC. The focus for follows, based on estimates of gross national income (GNI) most upper-middle-income and high-income countries per capita for 2013: might be ensuring full EUHC, which in some cases may require disinvesting from interventions and technologies • Low-income countries (LICs) = US$1,045 or less that provide less value for money. • Middle-income countries (MICs) are subdivided: These sorts of actions undoubtedly require strong (a) lower-middle-income = US$1,046 to US$4,125. political commitment and mechanisms for managing (b) upper-middle-income (UMICs) = US$4,126 to US$12,745. special interests (Reich and others 2016). Nevertheless, • High-income countries (HICs) = US$12,746 or more. this chapter argues that EUHC is a relevant and useful notion for all countries regardless of income, because it 1. SDG 3, titled “Good Health and Well-Being,” provides the following: “Ensure healthy lives and promote well-being represents the aspects of health care that are likely to for all at all ages” (UN 2016). provide the best value for money and thus be the most 2. The “Making Fair Choices consultation” refers to the efficient use of the next health care dollar. For LMICs WHO Consultative Group on Equity and Universal Health in particular, EUHC could provide an economically Coverage, the author of Making Fair Choices on the Path to grounded and realistic pathway to UHC and facilitate Universal Health Coverage (WHO 2014). progress toward a “grand convergence” in global health 3. Estimates from Vos and others (2016) were used because during the SDG period (Jamison and others 2013). similar data were not available from WHO. 62 Disease Control Priorities: Improving Health and Reducing Poverty 4. Current GNI data by country aggregated using the 2014 Giedion, U., E. A. Alfonso, and Y. Díaz. 2013. “The Impact country classification, see http://data.worldbank.org of Universal Coverage Schemes in the Developing World: /indicator/NY.GNP.ATLS.CD?page=1. A Review of the Existing Evidence.” Universal Health 5. For more information, see the International Decision Support Coverage (UNICO) Studies Series No. 25, World Bank, Initiative website, http://www.idsihealth.org/who-we-are Washington, DC. /about-us. Glassman, A., U. Giedion, Y. Sakuma, and P. C. Smith. 2016. “Defining a Health Benefits Package: What Are the Necessary Processes?” Health Systems and Reform 2 (1): 39–50. Holmes, K. K., S. Bertozzi, B. Bloom, and P. Jha, eds. 2017. Major REFERENCES Infectious Diseases. Volume 6, Disease Control Priorities Akachi, Y., and M. Kruk. 2017. “Quality of Care: Measuring (third edition), edited by D. T. Jamison, H. Gelband, a Neglected Driver of Improved Health.” Bulletin of the S. Horton, P. Jha, and R. Laxminarayan. Washington, DC: World Health Organization 95 (6): 465–72. doi:http://dx World Bank. .doi.org/10.2471/BLT.16.180190. Horton, S. 2018. “Cost-Effectiveness Analysis in Disease Control Atun, R., T. de Jongh, F. Secci, K. Ohiri, and O. Adeyi. 2010. Priorities, Third Edition.” In Disease Control Priorities (third “A Systematic Review of the Evidence on Integration of edition): Volume 9, Disease Control Priorities: Improving Targeted Health Interventions into Health Systems.” Health Health and Reducing Poverty, edited by D. T. Jamison, Policy and Planning 25 (1): 1–14. H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, Black, R. E., R. Laxminarayan, M. Temmerman, and and R. Nugent. Washington, DC: World Bank. N. Walker, eds. 2016. Reproductive, Maternal, Newborn, Howitt, P., A. Darzi, G. Z. Yang, H. Ashrafian, R. Atun, and and Child Health. Volume 2, Disease Control Priorities others. 2012. “Technologies for Global Health.” The Lancet (third edition), edited by D. T. Jamison, R. Nugent, 380 (9840): 507–35. H. Gelband, S. Horton, P. Jha, and R. Laxminarayan. Jamison, D. T., A. Alwan, C. N. Mock, R. Nugent, D. A. Washington, DC: World Bank. Watkins, and others. 2018. “Universal Health Coverage Boyle, C. F., C. Levin, A. Hatefi, S. Madriz, and N. Santos and Intersectoral Action for Health: Findings from Disease 2015. “Achieving a ‘Grand Convergence’ in Global Health: Control Priorities, Third Edition.” In Disease Control Modeling the Technical Inputs, Costs, and Impacts from Priorities (third edition): Volume 9, Disease Control Priorities: 2016 to 2030.” PLoS One 10 (10): e0140092. Improving Health and Reducing Poverty, edited by D. T. Bump, J., C. Cashin, K. Chalkidou, D. Evans, and others. 2016. Jamison, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, “Implementing Pro-Poor Universal Health Care Coverage.” C. N. Mock, and R. Nugent. Washington, DC: World Bank. The Lancet Global Health 4 (1): e14–e16. doi:http://dx.doi Jamison, D. T., J. G. Breman, A. R. Measham, G. Alleyne, .org/10.1016/S2214-109X(15)00274-0. M. Claeson, D. B. Evans, P. Jha, A. Mills, and P. Musgrove, Busse, R., J. Schreyögg, and C. Gericke. 2007. “Analyzing eds. 2006. Disease Control Priorities in Developing Changes in Health Financing Arrangements in High-Income Countries, second edition. Washington, DC: World Bank Countries: A Comprehensive Framework Approach.” HNP and Oxford University Press. Discussion Paper, World Bank, Washington, DC. Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, S. Berkley, Cairns, J. 2016. “Using Cost-Effectiveness Evidence to Inform and others 2013. “Global Health 2035: A World Converging Decisions as to Which Health Services to Provide.” Health within a Generation.” The Lancet 382 (9908): 1898–55. Systems and Reform 2 (1): 32–38. Jha, P. 2014. “Reliable Direct Measurement of Causes of Death Chan, M. 2016. “Making Fair Choices on the Path to Universal in Low- and Middle-Income Countries.” BMC Medicine Health Coverage.” Health Systems & Reform 2 (1): 5–7. 12 (19). https://doi.org/10.1186/1741-7015-12-19. Chang, A. Y., S. Horton, and D. T. Jamison. 2018. “Benefit- Ji, J. S., and L. Chen. 2016. “UHC Presents Universal Challenges.” Cost Analysis in Disease Control Priorities, Third Edition.” Health Systems and Reform 2 (1): 1–14. In Disease Control Priorities (third edition): Volume 9, Kieslich, K., J. Bump, O. F. Norheim, S. Tantivess, and P. Littlejohns. Disease Control Priorities: Improving Health and Reducing 2016. “Accounting for Technical, Ethical, and Political Factors Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, in Priority Setting.” Health Systems and Reform 2 (1): 51–60. P. Jha, R. Laxminarayanm, C. N. Mock, and R. Nugent. Kruk, M., A. Chukwuma, G. Mbaruku, and H. H. Leslie. Washington, DC: World Bank. 2017. “Variation in Quality of Primary-Care Services in Ethiopia, Ministry of Health. 2015. “HSTP: Health Sector Kenya, Malawi, Namibia, Rwanda, Senegal, Uganda, and the Transformation Plan, 2015/16–2019/20.” Strategy and plan- United Republic of Tanzania.” Bulletin of the World Health ning document for Second Growth and Transformation Organization 95 (6): 408–18. doi:http://dx.doi.org/10.2471 Plan (GTP II), Addis Ababa. /BLT.16.175869. Gelband, H., P. Jha, R. Sankaranarayanan, C. L. Gavreau, and Kruk M. E., E. Goldmann, S. Galea. 2009. “Borrowing and S. Horton. 2015. “Summary.” In Disease Control Priorities Selling to Pay for Health Care in Low- and Middle-Income (third edition): Volume 3, Cancer, edited by H. Gelband, Countries.” Health Affairs 28 (4): 1056–66. P. Jha, R. Sankaranarayanan, and S. Horton. Washington, Larson, E., D. Vail, G. M. Mbaruku, A. Kimweri, L. P. Freedman, DC: World Bank. and M. E. Kruk. 2015. “Moving toward Patient-Centered Universal Health Coverage and Essential Packages of Care 63 Care in Africa: A Discrete Choice Experiment of Preferences Peabody, J., R. Shimkhada, O. Adeyi, H. Wang, E. Broughton, and for Delivery Care among 3,003 Tanzanian Women.” PLoS M. Kruk. 2018. “Quality of Care.” In Disease Control Priorities One 10 (8): e0135621. (third edition): Volume 9, Disease Control Priorities: Improving Li, R., K. Hernandez-Villafuerte, A. Towse, I. Vlad, and Health and Reducing Poverty, edited by D. T. Jamison, K. Chalkidou. 2016. “Mapping Priority Setting in Health in H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, 17 Countries across Asia, Latin America, and Sub-Saharan and R. Nugent. Washington, DC: World Bank. Africa.” Health Systems and Reform 2 (1): 71–83. Prabhakaran, D., S. Anand, T. Gaziano, J.-C. Mbanya, Y. Wu, Mathers, C., G. Stevens, D. Hogan, A. Mahanani, and J. Ho. and R. Nugent, eds. 2017. Cardiovascular, Respiratory, and 2018. “Global and Regional Causes of Death: Patterns and Related Conditions. Volume 5, Disease Control Priorities Trends, 2000–15.” In Disease Control Priorities (third edi- (third edition), edited by D. T. Jamison, R. Nugent, tion): Volume 9, Disease Control Priorities: Improving Health H. Gelband, S. Horton, P. Jha, R. Laxminarayan, and and Reducing Poverty, edited by D. T. Jamison, H. Gelband, C. N. Mock. Washington, DC: World Bank. S. Horton, P. Jha, R. Laxminarayanm, C. N. Mock, and Reich, M. R., J. Harris, N. Ikegami, A. Maeda, C. Cashin, and R. Nugent. Washington, DC: World Bank. others. 2016. “Moving Towards Universal Health Coverage: McIntyre, D., F. Meheus, and J. A. Rottingen. 2017. “What Level Lessons from 11 Country Studies.” The Lancet 387 (10020): of Domestic Government Health Expenditure Should We 811–16. Aspire to for Universal Health Coverage?” Health Economics Samb, B., N. Desai, S. Nishtar, S. Mendis, H. Bekedam, and Policy and Law 12 (2): 12–37. others. 2010. “Prevention and Management of Chronic McIntyre, D., M. Thiede, G. Dahlgren, and M. Whitehead. 2006. Disease: A Litmus Test for Health-Systems Strengthening “What Are the Economic Consequences for Households of in Low-Income and Middle-Income Countries.” The Lancet Illness and of Paying for Health Care in Low- and Middle- 376 (9754): 1785–97. Income Country Contexts?” Social Science and Medicine Savedoff, W., D. de Ferranti, A. Smith, and V. Fan. 2012. “Political 62 (4): 858–65. and Economic Aspects of the Transition to Universal Health Mills, A. 2014. “Health Care Systems in Low- and Middle- Coverage.” The Lancet 3880 (9845): 924–32. Income Countries.” New England Journal of Medicine Seshadria, S. R., P. Jha, P. Sati, C. Gauvreau, U. Ram, and 370: 552–57. doi:10.1056/NEJMra1110897. R. Laxminarayan. 2015. “Karnataka’s Roadmap to Mock, C. N., P. Donkor, A. Gawande, D. T. Jamison, M. E. Kruk, Improved Health: Cost Effective Solutions to Address and H. T. Debas. 2015. “Essential Surgery: Key Messages Priority Diseases, Reduce Poverty and Increase Economic of This Volume.” In Disease Control Priorities (third Growth.” Report for the Government of Karnataka, Azim edition): Volume 1, Essential Surgery, edited by H. T. Debas, Premji University, Bangalore. P. Donkor, A. Gawande, D. T. Jamison, M. E. Kruk, and Stenberg, K., O. Hanssen, T. Edejer, M. Bertram, C. Brindley, C. N. Mock. Washington, DC: World Bank. and others. 2017. “Financing Transformative Health Ng, M., N. Fullman, J. L. Dieleman, A. D. Flaxman, C. J. Murray, and Systems Towards Achievement of the Health Sustainable S. S. Lim. 2014. “Effective Coverage: A Metric for Monitoring Developmenbt Goals: A Model for Projected Resource Universal Health Coverage.” PLoS Med 11 (9): e1001730. Needs in 67 Low-Income and Middle-Income Countries.” Nicholson, D., R. Yates, W. Warburton, and G. Fontana. The Lancet Global Health 5 (9): e875–e887. 2015. “Delivering Universal Health Coverage: A Guide UN (United Nations). 2016. “Sustainable Development Goals: 17 for Policymakers.” Report of the WISH Universal Health Goals to Transform Our World. Goal 3: Ensure Healthy Lives Coverage Forum 2015, World Innovation Summit for and Promote Well-Being for All at All Ages.” Website, UN, Health (WISH), Doha, Qatar. New York. http://www.un.org/sustainabledevelopment/health/. Norheim, O. F., P. Jha, K. Admasu, T. Godal, R. J. Hum, and UN DESA (United Nations Department of Economic and others. 2015. “Avoiding 40% of the Premature Deaths in Social Affairs). 2017. “World Population Prospects: 2017 Each Country, 2010–30: Review of National Mortality Revision, Key Findings and Advance Tables.” Report ESA/P/ Trends to Help Quantify the UN Sustainable Development WP/248, UN DESA Population Division, New York. Goal for Health.” The Lancet 385 (9964): 239–52. Verguet, S., and D. T. Jamison. 2018. “Health Policy Analysis: Ochalek, J., J. Lomas, and K. Claxton. 2015. “Cost per DALY Applications of Extended Cost-Effectiveness Analysis Averted Thresholds for Low- and Middle-Income Countries: Methodology in Disease Control Priorities, Third Edition.” Evidence from Cross Country Data.” Centre for Health In Disease Control Priorities (third edition): Volume 9, Economics (CHE) Research Paper 122, University of York, Disease Control Priorities: Improving Health and Reducing York, U.K. https://www.york.ac.uk/media/che/documents Poverty, edited by D. T. Jamison, H. Gelband, S. Horton, /papers/researchpapers/CHERP122_cost_DALY_LMIC P. Jha, R. Laxminarayan, C. N. Mock, and R. Nugent. _threshold.pdf. Washington, DC: World Bank. Patel, V., D. Chisholm, T. Dua, R. Laxminarayan, and Vos, T., C. Allen, M. Arora, R. M. Barber, Z. A. Bhutta, and M. E. Medina-Mora, eds. 2015. Mental, Neurological, others. 2016. “Global, Regional, and National Incidence, and Substance Use Disorders. Volume 4, Disease Control Prevalence, and Years Lived with Disability for 310 Acute Priorities (third edition), edited by D. T. Jamison, R. Nugent, and Chronic Diseases and Injuries, 1990–2015: A Systematic H. Gelband, S. Horton, P. Jha, R. Laxminarayan, and Analysis for the Global Burden of Disease Study 2015.” The C. N. Mock. Washington, DC: World Bank. Lancet 388 (10053): 1545–1602. 64 Disease Control Priorities: Improving Health and Reducing Poverty Watkins, D. A., O. F. Norheim, P. Jha, and D. T. Jamison. 2017. ———. 2013a. Global Action Plan for the Prevention and “Mortality Impact of Achieving Essential Universal Health Control of Noncommunicable Diseases 2013–2020. Geneva: Coverage in Low- and Lower-Middle-Income Countries.” WHO. Working Paper 21 for Disease Control Priorities (third ———. 2013b. The World Health Report 2013: Research for edition), Department of Global Health, University of Universal Health Coverage. Geneva: WHO. Washington, Seattle. ———. 2014. Making Fair Choices on the Path to Universal Watkins, D. A., R. A. Nugent, H. Saxenian, G. Yarney, Health Coverage: Final Report of the WHO Consultative Group K. Danforth, E. González-Pier, C. N. Mock, P. Jha, A. Alwan, on Equity and Universal Health Coverage. Geneva: WHO. and D. T. Jamison. 2018. “Intersectoral Policy Priorities ———. 2016. Global Health Observator database. WHO, for Health.” In Disease Control Priorities (third edition): Geneva. http://www.who.int/gho/en/. Volume 9, Disease Control Priorities: Improving Health and Wirtz, V. J., H. V. Hogerzeil, A. L. Gray, M. Bigdeli, Reducing Poverty, edited by D. T. Jamison, H. Gelband, C. P. de Joncheere, and others. 2017. “Essential Medicines S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and for Universal Health Coverage.” The Lancet 389 (10067): R. Nugent. Washington, DC: World Bank. 403–76. Watkins, D. A., J. Qi, and S. E. Horton. 2017. “Costs and World Bank. 2016. “Universal Health Coverage Study Series Affordability of Essential Universal Health Coverage in (UNICO).” Studies from 23 countries, World Bank, Low- and Middle-Income Countries.” Working Paper 20 Washington, DC. http://www.worldbank.org/en/topic/health for Disease Control Priorities (third edition), Department of /publication/universal-health-coverage-study-series. Global Health, University of Washington, Seattle. Xu, K., D. B. Evans, G. Carrin, A. M. Aguilar-Rivera, WHO (World Health Organization). 2007. Everybody’s P. Musgrove, and T. Evans. 2007. “Protecting Households Business: Strengthening Health Systems to Improve Health from Catastrophic Health Spending.” Health Affairs 26 (4): Outcomes. WHO’s Framework for Action. Geneva: WHO. 972–83. ———. 2010. The World Health Report. Health Systems Youngkong, S. 2012. “Multi-Criteria Decision Analysis for Financing: The Path to Universal Coverage. Geneva: WHO. Including Health Interventions in the Universal Health ———. 2011. Global Status Report on Noncommunicable Coverage Benefit Package in Thailand.” PhD dissertation, Diseases 2010. Geneva: WHO. Radboud University, Nijmegen, Netherlands. Universal Health Coverage and Essential Packages of Care 65 Part 2 Problems and Progress Chapter 4 Global and Regional Causes of Death: Patterns and Trends, 2000–15 Colin Mathers, Gretchen Stevens, Dan Hogan, Wahyu Retno Mahanani, and Jessica Ho INTRODUCTION The GHE 2015 present results for 183 WHO member states with a population of 90,000 or greater in 2015. One of the six core functions of the World Health The GHE 2015 cause-of-death estimates by country, Organization (WHO) is monitoring the health situation, region, and world for 2000–15 confirm and expand trends, and determinants in the world. Global, regional, previous WHO analyses of global health trends. In par- and country statistics on population and health indica- ticular, the WHO published an assessment of progress tors are important for assessing progress toward goals toward achievement of the UN Millennium Development for development and health and for guiding the alloca- Goals (MDGs) at the end of 2015 (WHO 2015b), fol- tion of resources. Timely data are needed to monitor lowed by the World Health Statistics 2016: Monitoring progress on increasing life expectancy and reducing age- Health for the SDGs (WHO 2016d), which focused on and cause-specific mortality rates. In particular, timely progress and challenges for achieving the SDGs for 2030. data are needed to monitor progress toward reaching the The SDGs expand the focus of health targets from the health-related targets within the Sustainable unfinished MDG agenda for child and maternal mortality Development Goals (SDGs), which will require regular and priority infectious diseases to a broader agenda includ- reporting on child mortality; maternal mortality; and ing NCDs, injuries, health emergencies, and health risk mortality owing to noncommunicable diseases (NCDs), factors as well as a strong focus on universal health cover- suicide, air pollution, road traffic injuries, homicide, age (UN Statistics Division 2017; WHO 2016d). The GHE natural disasters, and conflict. 2015 estimates of trends and levels of mortality by cause This chapter summarizes global and regional patterns will contribute to WHO and UN monitoring and report- of causes of death for 2015 and trends for 2000–15 using ing of progress toward the SDG health goals and targets. the 2015 Global Health Estimates (GHE 2015) released by the WHO at the beginning of 2017 (WHO 2017a). The GHE 2015 statistics provide a comprehensive, com- METHODS parable set of cause-of-death estimates from 2000 onward, consistent with and incorporating estimates Categories of Analysis from the United Nations (UN) and interagency and the The GHE 2015 provide estimates of the total number of WHO data for population, births, all-cause deaths, and deaths in 2000–15 for 177 detailed categories of disease specific causes of death. and injury as well as for all causes. The categories of Corresponding author: Colin Mathers, Department of Information, Evidence, and Research, World Health Organization, Geneva; mathersc@who.int. 69 cause are specified in the International Statistical • Global Burden of Disease 2015 (GBD 2015) esti- Classification of Diseases and Related Health Problems mates for other causes in countries lacking usable (known as the International Classification of Diseases, vital registration data or other nationally represen- or ICD) tenth revision codes (WHO 1990), as shown in tative sources of information on causes of death annex 4A. Deaths are estimated for the neonatal period (IHME 2016). (1 to 27 days), the postneonatal period (1 to 11 months), 1 to 4 years, and 5-year age groups starting at age 5 to 85 Figure 4.1 provides an overview of the data and pro- years and above. cesses used to produce the GHE 2015. Annex 4A provides This chapter uses World Bank classifications of a more detailed summary, which covers the processes national income (gross national income per capita) as of involved in the use of death registration data submitted July 2014 to classify countries into four income to the WHO Mortality Database (WHO 2016c). categories: low, lower middle, upper middle, and high. Death Registration Data Used Directly All-Cause Mortality Death registration data, with medical certification of the The WHO life tables were revised and updated for cause of death and the cause of death coded using the 183 member states for 1990–2015 (WHO 2016b), ICD, are the preferred source of information for monitor- drawing on the World Population Prospects: 2015 Revision ing mortality by cause, age, and sex. However, there are (UN 2015), recent and unpublished analyses of all-cause major gaps in the coverage of death registration data and mortality and mortality from human immunodefi- persistent issues in the quality of such data. In 2015, nearly ciency virus/acquired immune deficiency syndrome half of all deaths worldwide were registered in a national (HIV/AIDS) for countries with high HIV/AIDS preva- death registration system with information on cause of lence (Avenir Consulting 2016; UNAIDS 2016), vital death (figure 4.2), an improvement from about one-third registration data (WHO 2016c), and United Nations in 2005. However, only 38 percent of all global deaths are Inter-agency Group for Child Mortality Estimation esti- currently reported to the WHO Mortality Database mates of levels and trends for under-age-5 mortality (WHO 2016c). Of these reported deaths, 43 percent are (UN-IGME 2015). Methods and data sources are for high-income countries (HICs), 44 percent are for documented in more detail in annex 4A. The WHO upper-middle-income countries, 13 percent are for lower- life tables are available in the WHO Global Health middle-income countries, and less than 1 percent are for Observatory (2016). low-income countries, (LICs). Only about 28 percent of Total deaths by age and sex were estimated for each all global deaths are reported to the WHO by ICD code, country by applying death rates in the WHO life tables and only 23 percent are reported to the WHO with mean- to the estimated de facto resident population pre- ingful information on their underlying cause. pared by the UN Population Division in its 2015 revision Two main dimensions of quality impede the use of (UN 2015). death registration data for public health monitoring: (a) low level of completeness and (b) missing, incom- plete, or invalid information on the underlying cause of Causes of Death death. “Completeness” is defined as the percentage of all The GHE 2015 are consistent with UN agency, inter- deaths in the de facto resident population that are regis- agency, and WHO estimates for population, births, all- tered and compiled nationally. The quality of informa- cause deaths, and specific causes of death, including the tion on underlying cause of death is summarized by the following: proportion of deaths coded to so-called garbage codes, which do not provide information on valid underlying • The most recent vital registration data for all coun- disease or injury causes of death. tries where the quality of data is assessed as usable Since 2010, the WHO has been summarizing the • UN estimates of levels and trends for all-cause usability of death registration data for estimating causes mortality for older children and adults and UN of death in a population with a usability score calculated interagency estimates of neonatal, infant, and child as follows: mortality (Percentage usable) = Completeness (%) • WHO programs and interagency groups’ updated × (1 − Proportion garbage). (4.1) estimates for specific causes of death, including maternal, HIV/AIDS, tuberculosis, malaria, cancers, Death registration data reported to the WHO were road traffic injuries, and homicide used to estimate causes of death for 69 countries 70 Disease Control Priorities: Improving Health and Reducing Poverty Figure 4.1 Overview of the Processes Involved in Preparing the Global Health Estimates Dataset for Causes of Death in 183 WHO Member States, 2000–15 HIV/AIDS Tuberculosis Review and revise VR Adjust outliers country causes with and country- Malaria Inputs WHO/UN inputs as cause-specific Child causes needed coding issues Foodborne Maternal causes Cancers Road injury GHE 2015 GHE 2015 Homicide initial estimates final estimates 2000–2015 GBD 2015 other cause fractions for non-VR countries WHO all-cause envelopes Shapes and colours Input data Mortality shocks (conflict, disasters, Expand MCEE cause list GHE estimates for Process Ebola) to GHE cause list and child deaths ages <5 impute sex breakdown 2000–2015 Results WHO-MCEE estimates of child WHO/UN Interagency COD 2000–2015 Expand SRS-MDS cause GBD 2015 study GHE estimates list to GHE cause list for India ages 5+ Verbal autopsy data and impute time series Indian SRS-MDS 2000–2015 GHE–Global health 2000–2015 2001–03 and 2010–13 estimates Note: COD = cause of death; GBD 2015 = Global Burden of Disease 2015; GHE 2015 = 2015 Global Health Estimates; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome; MCEE = Maternal and Child Epidemiology Estimation Collaboration; SRS-MDS = Sample Registration System–Million Death Study; UN = United Nations; VR = vital registration; WHO = the World Health Organization. meeting the following inclusion criteria: (a) at least five Figure 4.2 Number of Global Deaths in 2015, by Expected Registration years of data were available during 2005–15, and (b) at or Reporting Status least 65 percent of deaths were usable for 2000 to the latest available year (WHO 2016c). The following short 60 Number of deaths (millions) list of garbage codes was used to compute the usable 50 percentage: 40 • Symptoms, signs, and ill-defined conditions (ICD 10 30 codes R00–R99) 20 • Injuries undetermined whether intentional or unin- 10 tentional (ICD 10 Y10–Y34, Y87.2) • Ill-defined cancers (C76, C80, and C97) 0 Total Registered Reported to Reported Reported • Ill-defined cardiovascular diseases (I46, I47.2, I49.0, with cause WHO to WHO by to WHO by I50, I51.4, I51.5, I51.6, I51.9, and I70.9). of deatha ICD code meaningful ICD code Deaths coded to these and various other garbage codes were redistributed to valid underlying causes of death. Note: Reports to the World Health Organization (WHO) are projected on the basis of 2010 data to allow for reporting lag. ICD = International Classification of Diseases. Estimates for India were based on WHO analyses of a. Local death registration, in the absence of a state or national system to compile data, is excluded, data from the Sample Registration System (SRS) for two as is registration with cause of death based on verbal autopsy. Global and Regional Causes of Death: Patterns and Trends, 2000–15 71 periods: 2001–03 (Registrar General of India 2009) and • Conflict and natural disasters: the WHO and the 2010–13 (Registrar General of India and CGHR 2015). Centre for Research on the Epidemiology of Disasters. Estimates for China drew on death registration data for For methods, see WHO (2016b). 2013 (China CDC 2016) together with IHME analyses of trends in causes of death (GBD 2016). Additional adjustments and revisions were applied to GBD 2015 estimates for schistosomiasis, rabies, leprosy, Causes of Death for Children under Age 5 liver cancer, alcohol use disorders, drug use disorders, For countries lacking usable death registration data, and liver cirrhosis, as described in annex 4A. neonatal deaths and deaths at age 1–59 months were estimated for 15 major causes identifiable from verbal autopsy studies using methods described by Liu and Other Causes of Death for Countries Lacking Death others (2015). These categories were expanded to the full Registration Data GHE list of causes using nested cause fraction results Estimates of mortality and causes of death were released predicted from the GBD 2015 study. in 2016 (GBD 2015 Mortality and Causes of Death For China, estimates of causes of death for children Collaborators 2016) by the Institute of Health Metrics under age 5 were based on a separate analysis of data and Evaluation (IHME) as part of the GBD 2015 study from the Maternal and Child Health Surveillance System (IHME 2016). The WHO has drawn on the GBD 2015 (WHO 2016b). For India, a separate multiple-cause analyses for selected causes for member states lacking model was used to prepare state-level estimates based on comprehensive death registration data. about 40 subnational community-based verbal autopsy For major causes of death except HIV/AIDS and studies (WHO and MCEE 2016). measles, the IHME used ensemble modeling to create a weighted average of many individual covariate-based Cause-Specific Estimates from the WHO and models (ranging from hundreds to thousands in some UN Agencies cases) for each specific cause. The overall out-of-sample The GHE 2015 incorporate the latest updated WHO and predictive validity of the ensemble is usually not much UN interagency assessments of levels and trends for the different from that of the top-ranked model, but ranges following specific causes of death: of uncertainty are generally much wider and more plau- sible than for single models. To ensure that the results of all the single-cause models summed to the all-cause • Tuberculosis: Global Tuberculosis Report 2016 (WHO mortality estimate for each age-sex-country-year group, 2016a) the IHME applied a final step to rescale the cause- • HIV/AIDS: UNAIDS (2016); WHO (2016b) specific estimates. This step effectively squeezed or • Malaria: World Malaria Report 2016 (WHO 2016e) expanded causes with wider uncertainty ranges more • Vaccine-preventable child causes: Patel and others than those with narrower uncertainty ranges. The GBD (2016); WHO (2017b) 2015 results (IHME 2016) were resqueezed to the WHO • Other major child causes: the WHO and the Maternal all-cause envelopes to produce a set of so-called prior and Child Epidemiology Estimation collaboration estimates for the GHE categories of cause by age, sex, (WHO and MCEE 2016) country, and year. • Foodborne diseases: the WHO Foodborne Disease Burden Epidemiology Reference Group (Torgerson and others 2015) • Ebola virus infection: WHO estimates of direct deaths Final Adjustments owing to infections and indirect deaths owing to IHME results for priority causes such as HIV/AIDS, measles outbreaks and reduced coverage of treatment tuberculosis, malaria, cancers, maternal mortality, and for HIV/AIDS and malaria (see annex 4A) child mortality differ to varying degrees from those of the • Maternal mortality: UN Maternal Mortality Estimation WHO and UN agency partners. In part, these variations Inter-Agency Group (MMEIG 2015) reflect not only differences in modeling strategies but also • Cancers: International Agency for Research on Cancer the inclusion by IHME of data from verbal autopsy stud- (Ferlay and others 2013) ies, mapped to ICD categories using IHME-developed • Road injuries: Global Status Report on Road Safety computer algorithms. We carried out an adjustment 2015 (WHO 2015a) process to ensure that the estimated number of deaths • Homicide: Global Status Report on Violence Prevention tallied across causes to the estimated total number of 2014 (WHO 2014a) deaths by age, sex, country, and year for all countries. 72 Disease Control Priorities: Improving Health and Reducing Poverty Levels of Evidence and Uncertainty and death rates for most diseases and disorders in this General guidance on the quality and uncertainty of these group of countries declined substantially between 2000 cause-of-death estimates for 2000–15 is provided with and 2015. regard to the quality of data inputs and methods used. NCDs caused 70 percent of deaths globally in 2015, Most of the inputs to the GHE 2015 have explicit uncer- with regional figures ranging from 43 percent in LICs to tainty ranges. The two main exceptions are the UN 87 percent in HICs. In terms of the absolute number of Population Division’s World Population Prospects 2015 deaths, however, 74 percent of global NCD-related life tables (UN 2015) and the Globocan cancer mortality deaths occurred in low- and middle-income countries estimates (IARC 2013). The Globocan 2012 database (LMICs). provides information on sources of data and quality of Injuries claimed nearly 5 million lives in 2015 inputs for seven categories of incidence data and six cat- (8.8 percent of total deaths). More than a quarter egories of mortality data as well as six estimation meth- (27 percent) of these deaths were due to road traffic inju- ods for mortality (IARC 2013). The GBD 2015 estimates ries. LICs had the highest mortality rate for road traffic of deaths by cause, age, sex, country, and year also include injuries, with 25.0 deaths per 100,000 population, com- estimates of 95 percent uncertainty ranges that take into pared with a global rate of 18.3. More than 90 percent of account some, but not all, sources of uncertainty. road traffic deaths occur in LMICs, which account for Based on the uncertainty ranges estimated for the 82 percent of the world’s population but only 54 percent inputs, explicit uncertainty ranges for the GHE 2015 are of the world’s registered vehicles. Several factors are at available on the WHO website (see box 4.1). work, including poorly designed or implemented regula- tions, inadequate road and vehicle quality, and a higher proportion of vulnerable road users (pedestrians, cyclists, RESULTS and motorcyclists). Broad Patterns of Causes of Death in 2015 In 2015, a total of 56.4 million deaths occurred in the Leading Causes of Death in 2015 world; of these, 7.0 million occurred in LICs and Figure 4.4 shows the 10 leading causes of death for 20.4 million occurred in lower-middle-income countries. the world and for country income groups in 2015. Just under half (46 percent) of all deaths in LICs were The 10 leading causes of death globally were 6 NCDs, caused by Group I conditions, which include communi- 3 infectious diseases, and road injuries, which collec- cable diseases, maternal causes, conditions arising dur- tively accounted for more than half of all deaths. ing the perinatal period, and nutritional deficiencies Ischemic heart disease (IHD) and stroke killed (figure 4.3). For HICs that have passed through the epi- 15 million people in 2015; these two diseases have demiological transition, Group I conditions accounted been the biggest killers globally in the past 15 years. for less than 7 percent of deaths. For LICs, Group I Whereas 7 of the 10 leading causes in low-income conditions accounted for 65 percent of deaths in 2000, countries were Group I conditions, all but 1 of the Box 4.1 Datasets Available for the WHO Global Health Estimates 2015 The WHO Global Health Estimates provide a num- (http://www.who.int/healthinfo/global_burden ber of datasets: _disease/estimates/en/index2.html) • Files with uncertainty (http://terrance.who.int • Regional and country spreadsheets of deaths by /mediacentre/data/ghe/) cause, age, and sex, 2000–15 (http://www.who • Life expectancy and life tables by country, .int/healthinfo/global_burden_disease/estimates region, and world (http://www.who.int/gho /en/index1.html) /mortality_burden_disease/life_tables/en/) • Regional and country spreadsheets of disability- • Global Health Estimates technical paper series -adjusted life years, years of life lost, and years (http://www.who.int/healthinfo/global_burden lost to disability by cause, age, and sex, 2000–15 _disease/data_sources_methods/en/). Global and Regional Causes of Death: Patterns and Trends, 2000–15 73 Figure 4.3 Overall Mortality Rates, by Cause and Country Income Group, 2000 and 2015 1,400 1,200 1,000 Death rate per 100,000 population 800 600 400 200 0 2000 2015 2000 2015 2000 2015 2000 2015 2000 2015 Low-income Lower-middle- Upper-middle- High-income World countries income countries income countries countries Intentional injuries Other noncommunicable diseases Maternal, perinatal, and nutritional Other unintentional injuries Chronic respiratory diseases conditions Road injury Cardiovascular diseases and diabetes Other infectious diseases Cancers HIV/AIDS, tuberculosis, malaria Note: HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. 10 leading causes of death in HICs were NCDs. Road 1 million in 2000. Total deaths attributable to diabe- injuries were among the 10 leading causes of death for tes are more than double this number, because countries at all income levels except HICs. The colors diabetes raises the risk of cardiovascular and other of the bars in figure 4.4 indicate causes for which over- diseases. Estimated deaths from dementia more than all death rates are increasing (red) or decreasing doubled between 2000 and 2015, making dementia (green). Increases in overall (crude) death rates (total the seventh-leading cause of death globally in 2015. In the deaths divided by total population) may reflect the case of dementia and diabetes, aging and rising death effect of population aging as well as changes in age- rates contribute to the rise in overall number of specific risks of death. Population aging is often a dom- deaths. Rising reported death rates for these two causes inant factor for diseases with death rates that rise with may also reflect an increase in diagnosis or recording as age, such as most cancers, cardiovascular diseases, and an underlying cause of death rather than an increase in dementia, even when age-specific death rates are falling. the age-specific risk of mortality. One important exception is the substantial decline in Lower respiratory infections remained the deadliest the death rates of IHD and stroke in HICs. communicable disease, causing 3.2 million deaths Chronic lung disease claimed 3.2 million lives in worldwide in 2015. The diarrhea death rate almost 2015, while lung cancer (along with tracheal and halved between 2000 and 2015, but the disease still bronchus cancers) caused 1.7 million deaths. Diabetes caused 1.4 million deaths in 2015. Similarly, the tubercu- killed 1.6 million people in 2015, up from less than losis death rate fell during the same period, but the 74 Disease Control Priorities: Improving Health and Reducing Poverty Figure 4.4 The 10 Leading Causes of Death, for the World and by Country Income Group, 2015 a. Global b. Low-income countries Ischemic heart disease Lower respiratory infections Stroke Stroke Lower respiratory infections Ischemic heart disease COPD Diarrheal diseases Lung cancers HIV/AIDS Diabetes mellitus Tuberculosis Dementias Preterm birth complications Diarrheal diseases Malaria Tuberculosis Road injuries Road injuries Birth asphyxia and birth trauma 0 50 100 150 200 0 50 100 150 200 CDR per 100,000 population CDR per 100,000 population c. Lower-middle-income countries d. Upper-middle-income countries Ischemic heart disease Ischemic heart disease Stroke Stroke Lower respiratory infections COPD COPD Lung cancers Tuberculosis Lower respiratory infections Diarrheal diseases Diabetes mellitus Preterm birth complications Dementias Diabetes mellitus Road injuries Cirrhosis of the liver Liver cancer Road injuries Stomach cancer 0 50 100 150 200 0 50 100 150 200 CDR per 100,000 population CDR per 100,000 population e. High-income countries Ischemic heart disease Stroke Dementias Size of bar indicates crude death rate (CDR) Lung cancers per 100,000 population COPD Color indicates % change Lower respiratory infections in CDR within the region Colon and rectum cancers between 2000 and 2015: Diabetes mellitus Kidney diseases –70.0% 170.0% Breast cancers 0 50 100 150 200 CDR per 100,000 population Note: The colors of the bars indicate causes for which overall death rates are increasing (red) or decreasing (green). CDR = crude death rate; COPD = chronic obstructive pulmonary disease; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. disease was still among the top 10 causes of death in Cause-Specific Trends from 2000 to 2015 2015, with a death toll of 1.4 million. HIV/AIDS Tables 4.1 to 4.10 provide summary tabulations of dropped out of the top 10 causes of death globally, deaths by cause, age, and sex for the world and for coun- falling from 1.5 million deaths in 2000 to just under try income groups for 2000 and 2015. More detailed 1.1 million in 2015. However, it remains the fifth-leading results at the country and regional levels are also avail- cause of death in LICs. able on the WHO website (see box 4.1). Global and Regional Causes of Death: Patterns and Trends, 2000–15 75 Table 4.1 Deaths from Selected Causes in the World, by Age and Sex, 2015 thousands Both Both sexes Sex sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 7,344 3,705 3,639 671 1,244 1,802 1,987 1,248 393 All causes 56,441 30,177 26,264 5,992 1,303 2,687 5,780 14,628 26,051 I. Communicable, maternal, 11,959 6,317 5,642 4,843 638 792 1,525 1,620 2,540 perinatal, and nutritional conditions A. Infectious and parasitic diseases 5,706 3,195 2,512 1,452 463 513 1,202 1,115 961 a 1. Tuberculosis 1,373 927 446 69 28 87 328 513 348 3. HIV/AIDS 1,060 617 443 87 46 134 602 178 13 4. Diarrheal diseases 1,389 684 705 526 103 90 95 207 368 b 5. Vaccine-preventable diseases 273 139 134 207 34 10 8 9 5 6. Meningitis and encephalitis 405 209 196 116 91 69 44 47 37 7. Acute hepatitisc 145 77 68 8 14 24 32 41 28 8. Malaria 439 228 211 312 31 27 29 22 18 9. Other infectious and parasitic diseases 623 314 309 129 116 71 65 97 145 B. Respiratory infections 3,913 2,122 1,791 36 14 36 122 989 2,716 C. Maternal conditions 303 — 303 — — 155 148 — — D. Neonatal conditions 2,311 1,292 1,019 2,311 — — — — — 1. Preterm birth complications 1,058 586 472 1,058 — — — — — 2. Birth asphyxia and birth trauma 691 386 305 691 — — — — — 3. Neonatal sepsis and infections 405 240 166 405 — — — — — 4. Other neonatal conditions 157 81 76 157 — — — — — E. Nutritional deficiencies 439 215 224 160 35 23 26 60 135 II. Noncommunicable diseases 39,544 20,541 19,003 783 312 778 3,049 12,001 22,622 A. Malignant neoplasms 8,763 4,982 3,781 37 49 152 947 3,498 4,080 3. Stomach cancer 754 490 263 — — 5 64 300 384 4. Colon and rectum cancers 774 418 356 1 1 7 61 267 438 5. Liver cancer 788 554 235 — 2 11 112 347 315 7. Lung cancer 1,695 1,174 521 1 — 4 95 724 870 9. Breast cancer 571 1 570 — — 8 131 247 185 Other cancers 4,182 2,345 1,836 36 46 117 484 1,611 1,889 C. Diabetes mellitus 1,586 729 856 2 4 18 101 582 879 E. Mental and behavioral disorders 317 233 84 — — 49 115 109 44 4. Alcohol use disordersd 129 108 21 — — 9 46 59 15 5. Drug use disorderse 168 117 51 — — 39 63 43 23 table continues next page 76 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.1 Deaths from Selected Causes in the World, by Age and Sex, 2015 (continued) Both Both sexes Sex sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 2,011 812 1,199 17 29 60 66 209 1,629 1. Alzheimer’s disease and other 1,542 557 985 — — — 2 109 1,432 dementias H. Cardiovascular diseases 17,689 8,850 8,839 39 37 194 1,072 5,136 11,210 3. Ischemic heart disease 8,756 4,603 4,153 2 2 62 528 2,586 5,576 4. Stroke 6,241 2,990 3,250 12 15 57 314 1,845 3,997 I. Respiratory diseases 3,913 2,122 1,791 36 14 36 122 989 2,716 J. Digestive diseases 2,347 1,355 991 27 47 119 386 853 914 2. Cirrhosis of the liver 1,162 762 400 8 20 59 258 517 301 K. Genitourinary diseases 1,382 701 681 18 17 55 134 407 751 1. Kidney diseases 1,129 580 549 12 14 44 113 349 598 N. Congenital anomalies 647 340 307 509 58 34 20 16 10 Other noncommunicable diseasesf 888 415 473 96 56 62 87 201 388 III. Injuries 4,939 3,319 1,619 366 352 1,118 1,206 1,007 889 A. Unintentional injuries 3,527 2,322 1,204 344 304 646 749 731 752 1. Road traffic injury 1,342 1,014 328 73 70 353 400 307 140 2. Other unintentional injuries 2,184 1,308 877 271 234 293 349 425 612 B. Intentional injuries 1,412 997 415 22 48 472 457 276 137 1. Suicide 788 504 284 — 13 221 241 200 114 2. Homicide and collective violence 624 493 131 22 35 251 216 76 23 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 77 Table 4.2 Deaths from Selected Causes in Low-Income Countries, by Age and Sex, 2015 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 896 446 449 128 223 249 193 83 20 All causes 6,997 3,712 3,285 1,902 460 652 945 1,362 1,676 I. Communicable, maternal, 3,248 1,706 1,542 1,588 265 283 427 334 349 perinatal, and nutritional conditions A. Infectious and parasitic diseases 1,730 948 782 569 195 178 329 256 203 a 1. Tuberculosis 326 222 104 25 10 16 66 119 90 3. HIV/AIDS 334 179 155 35 22 46 177 49 4 4. Diarrheal diseases 413 214 199 177 42 37 32 47 77 b 5. Vaccine-preventable diseases 103 52 52 78 14 5 3 2 1 6. Meningitis and encephalitis 142 76 66 47 36 27 14 11 8 c 7. Acute hepatitis 18 10 8 1 2 4 4 4 3 8. Malaria 225 115 110 155 19 17 16 10 9 9. Other infectious and parasitic diseases 171 82 88 51 51 26 17 14 12 B. Respiratory infections 291 155 136 8 5 10 18 85 165 C. Maternal conditions 117 — 117 — — 61 56 — — D. Neonatal conditions 632 359 273 632 — — — — — 1. Preterm birth complications 242 136 106 242 — — — — — 2. Birth asphyxia and birth trauma 223 124 99 223 — — — — — 3. Neonatal sepsis and infections 130 80 50 130 — — — — — 4. Other neonatal conditions 36 19 17 36 — — — — — E. Nutritional deficiencies 160 89 71 80 19 12 9 14 25 II. Noncommunicable diseases 3,014 1,517 1,497 192 90 174 377 932 1,249 A. Malignant neoplasms 549 268 281 8 9 28 124 235 144 3. Stomach cancer 31 18 13 — — 1 6 15 9 4. Colon and rectum cancers 27 14 13 — — 1 6 12 8 5. Liver cancer 35 22 13 — — 2 9 17 7 7. Lung cancer 46 28 17 — — — 5 22 17 9. Breast cancer 43 — 43 — — 2 18 18 6 Other cancers 367 185 183 8 8 23 81 151 97 C. Diabetes mellitus 138 70 68 — 1 4 13 47 73 E. Mental and behavioral disorders 18 14 5 — — 4 7 5 3 4. Alcohol use disordersd 9 8 1 — — 2 4 3 1 e 5. Drug use disorders 8 5 3 — — 3 3 1 2 table continues next page 78 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.2 Deaths from Selected Causes in Low-Income Countries, by Age and Sex, 2015 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 116 54 62 6 6 18 10 12 65 1. Alzheimer’s disease and other 66 25 42 — — — — 6 60 dementias H. Cardiovascular diseases 1,229 581 648 9 12 39 113 404 653 3. Ischemic heart diseases 474 248 226 — 1 9 43 162 258 4. Stroke 521 236 285 3 4 14 42 174 284 I. Respiratory diseases 291 155 136 8 5 10 18 85 165 J. Digestive diseases 285 170 114 8 13 26 59 96 82 2. Cirrhosis of the liver 146 92 53 2 4 12 36 58 34 K. Genitourinary diseases 107 54 53 5 5 13 15 29 41 1. Kidney diseases 77 39 38 3 4 10 10 21 29 N. Congenital anomalies 156 83 73 119 21 10 4 2 1 f Other noncommunicable diseases 124 68 57 30 19 20 15 17 23 III. Injuries 735 489 246 121 105 195 141 95 77 A. Unintentional injuries 583 381 202 115 96 132 96 76 68 1. Road traffic injury 224 158 66 28 21 67 50 36 22 2. Other unintentional injuries 360 223 136 87 76 65 46 40 45 B. Intentional injuries 151 108 43 6 9 63 45 19 10 1. Suicide 66 44 22 — 3 25 18 12 8 2. Homicide and collective violence 85 64 21 6 6 38 26 7 2 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 79 Table 4.3 Deaths from Selected Causes, in Lower-Middle-Income Countries, by Age and Sex, 2015 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 2,669 1,361 1,307 290 538 717 682 355 87 All causes 20,422 11,064 9,358 3,308 665 1,317 2,646 5,606 6,880 I. Communicable, maternal, 6,323 3,339 2,984 2,745 328 420 781 902 1,146 perinatal, and nutritional conditions A. Infectious and parasitic diseases 3,143 1,767 1,376 785 238 277 624 663 556 a 1. Tuberculosis 905 604 301 40 16 63 227 343 218 3. HIV/AIDS 425 252 173 44 17 55 243 62 4 4. Diarrheal diseases 858 413 445 303 56 50 58 147 243 b 5. Vaccine-preventable diseases 160 82 78 120 20 5 4 6 3 6–7. Meningitis and encephalitis 217 109 107 59 49 37 24 27 20 8. Acute hepatitisc 107 55 52 6 11 18 25 28 19 9a. Malaria 199 106 93 146 11 10 12 12 9 Other infectious and parasitic 273 146 126 67 57 39 32 39 40 diseases B. Respiratory infections 1,437 779 658 24 7 17 63 465 862 C. Maternal conditions 165 — 165 — — 84 81 — — D. Neonatal conditions 1,381 766 615 1,381 — — — — — 1. Preterm birth complications 669 367 302 669 — — — — — 2. Birth asphyxia and birth trauma 385 216 169 385 — — — — — 3. Neonatal sepsis and infections 237 137 99 237 — — — — — 4. Other neonatal conditions 91 46 45 91 — — — — — E. Nutritional deficiencies 188 85 104 63 13 9 13 36 54 II. Noncommunicable diseases 12,065 6,383 5,681 389 160 385 1,366 4,330 5,435 A. Malignant neoplasms 1,768 916 852 14 21 62 337 831 503 3. Stomach cancer 116 75 41 — — 2 21 58 36 4. Colon and rectum cancers 125 69 56 — — 4 22 56 44 5. Liver cancer 140 94 46 — 1 4 26 69 40 7. Lung cancer 199 147 52 — — 2 24 110 63 9. Breast cancer 181 180 — — 4 60 84 32 Other cancers 1,006 530 476 14 19 45 185 455 288 C. Diabetes mellitus 643 292 351 1 2 8 48 263 321 E. Mental and behavioral disorders 76 59 17 — — 19 28 21 8 d 4. Alcohol use disorders 26 22 4 — — 3 10 9 3 5. Drug use disorderse 47 36 11 — — 15 16 11 4 table continues next page 80 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.3 Deaths from Selected Causes, in Lower-Middle-Income Countries, by Age and Sex, 2015 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 374 169 206 7 16 26 26 52 248 1. Alzheimer’s disease and other 261 107 154 — — — 1 28 232 dementias H. Cardiovascular diseases 5,640 2,992 2,649 17 18 103 537 2,068 2,897 3. Ischemic heart disease 3,117 1,749 1,368 2 1 35 292 1,162 1,625 4. Stroke 1,813 894 919 6 8 28 142 660 968 I. Respiratory diseases 1,437 779 658 24 7 17 63 465 862 J. Digestive diseases 1,008 589 418 15 31 80 214 369 299 2. Cirrhosis of the liver 545 362 183 6 15 42 147 220 115 K. Genitourinary diseases 538 305 232 9 10 32 78 196 213 1. Kidney diseases 455 259 197 7 8 25 67 172 177 N. Congenital anomalies 310 161 149 257 27 14 6 4 2 f Other noncommunicable diseases 270 121 149 45 29 26 28 60 83 III. Injuries 2,034 1,341 693 174 176 512 500 374 298 A. Unintentional injuries 1,479 962 517 163 152 295 314 289 268 1. Road traffic injury 517 404 113 29 31 147 159 109 43 2. Other unintentional injuries 962 558 404 134 121 148 155 180 225 B. Intentional injuries 554 379 176 11 25 218 186 85 30 1. Suicide 298 183 115 — 7 123 99 50 19 2. Homicide and collective violence 257 196 61 11 18 95 87 35 12 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 81 Table 4.4 Deaths from Selected Causes in Upper-Middle-Income Countries, by Age and Sex, 2015 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 2,473 1,252 1,221 179 337 590 747 486 134 All causes 17,124 9,343 7,781 693 156 555 1,531 4,963 9,227 I. Communicable, maternal, perinatal, and nutritional conditions 1,606 863 743 465 43 75 248 266 510 A. Infectious and parasitic diseases 617 356 262 95 29 51 204 145 94 a 1. Tuberculosis 110 77 33 4 2 6 25 43 30 3. HIV/AIDS 248 146 103 7 7 29 151 51 4 4. Diarrheal diseases 84 44 40 45 4 3 5 9 18 5. Vaccine-preventable diseasesb 10 5 5 8 1 — — — — 6–7. Meningitis and encephalitis 36 19 17 8 6 4 5 7 5 8. Acute hepatitisc 15 9 6 — — 1 3 6 4 9a. Malaria 15 8 8 11 1 1 1 1 1 Other infectious and parasitic diseases 98 48 51 11 7 6 13 29 32 B. Respiratory infections 1,430 781 649 4 2 7 30 314 1,073 C. Maternal conditions 20 — 20 — — 10 10 — — D. Neonatal conditions 261 146 115 261 — — — — — 1. Preterm birth complications 125 70 54 125 — — — — — 2. Birth asphyxia and birth trauma 77 42 35 77 — — — — — 3. Neonatal sepsis and infections 36 21 15 36 — — — — — 4. Other neonatal conditions 24 13 11 24 — — — — — E. Nutritional deficiencies 66 33 33 17 2 2 3 8 34 II. Noncommunicable diseases 14,066 7,476 6,590 164 51 160 887 4,348 8,455 A. Malignant neoplasms 3,474 2,153 1,322 12 16 48 351 1,416 1,631 3. Stomach cancer 417 281 136 — — 2 28 169 218 4. Colon and rectum cancers 256 143 113 — — 2 20 90 144 5. Liver cancer 464 339 125 — 1 5 71 206 181 7. Lung cancer 817 580 236 — — 2 48 338 428 9. Breast cancer 138 — 138 — — 1 33 64 39 Other cancers 1,383 809 574 12 15 37 150 549 621 C. Diabetes mellitus 532 234 298 — 1 4 31 206 290 E. Mental and behavioral disorders 90 62 28 — — 10 27 33 20 d 4. Alcohol use disorders 35 31 4 — — 2 12 16 5 5. Drug use disorderse 43 26 17 — — 7 12 12 12 table continues next page 82 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.4 Deaths from Selected Causes in Upper-Middle-Income Countries, by Age and Sex, 2015 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 593 243 350 4 5 11 18 77 477 1. Alzheimer’s disease and other dementias 497 190 306 — — — 1 51 445 H. Cardiovascular diseases 6,507 3,245 3,262 12 7 39 279 1,837 4,332 3. Ischemic heart disease 2,809 1,426 1,383 — — 14 125 773 1,897 4. Stroke 2,756 1,380 1,377 3 2 12 103 823 1,813 I. Respiratory diseases 1,430 781 649 4 2 7 30 314 1,073 J. Digestive diseases 617 363 255 4 3 11 79 250 271 2. Cirrhosis of the liver 309 201 108 1 1 5 51 153 99 K. Genitourinary diseases 447 212 235 3 2 10 35 138 259 1. Kidney diseases 375 180 195 2 2 8 30 121 212 N. Congenital anomalies 138 73 65 109 9 7 6 4 3 f Other noncommunicable diseases 237 111 126 15 7 13 31 72 99 III. Injuries 1,452 1,005 448 63 62 320 396 349 261 A. Unintentional injuries 988 678 311 59 50 172 252 250 205 1. Road traffic injury 483 367 117 14 16 113 156 130 54 2. Other unintentional injuries 505 311 194 45 34 59 95 121 151 B. Intentional injuries 464 327 137 4 12 148 144 99 56 1. Suicide 228 129 98 — 2 41 60 74 50 2. Homicide and collective violence 236 198 38 4 10 106 84 25 6 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 83 Table 4.5 Deaths from Selected Causes in High-Income Countries, by Age and Sex, 2015 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 1,307 645 662 74 146 247 365 324 152 All causes 11,899 6,058 5,841 90 22 164 658 2,698 8,269 I. Communicable, maternal, 781 409 373 45 2 13 69 118 534 perinatal, and nutritional conditions A. Infectious and parasitic diseases 216 124 92 4 1 7 46 51 107 a 1. Tuberculosis 31 23 8 — — 2 10 9 11 3. HIV/AIDS 53 41 12 — — 4 31 17 1 4. Diarrheal diseases 34 13 21 1 — — 1 4 29 b 5. Vaccine-preventable diseases 1 — — 1 — — — — — 6–7. Meningitis and encephalitis 10 5 5 1 — — 1 3 3 c 8. Acute hepatitis 5 3 2 — — — 1 2 2 9a. Malaria — — — — — — — — — Other infectious and parasitic 81 38 44 1 — 1 3 16 61 diseases B. Respiratory infections 755 407 348 — — 2 11 125 617 C. Maternal conditions 2 — 2 — — 1 1 — — D. Neonatal conditions 37 21 16 37 — — — — — 1. Preterm birth complications 22 13 10 22 — — — — — 2. Birth asphyxia and birth trauma 6 3 3 6 — — — — — 3. Neonatal sepsis and infections 3 2 1 3 — — — — — 4. Other neonatal conditions 6 3 3 6 — — — — — E. Nutritional deficiencies 25 10 15 — — — — 2 21 II. Noncommunicable diseases 10,400 5,165 5,234 37 11 60 420 2,391 7,482 A. Malignant neoplasms 2,972 1,646 1,326 3 4 14 134 1,015 1,803 3. Stomach cancer 189 116 73 — — — 9 59 121 4. Colon and rectum cancers 365 192 174 — — 1 13 109 243 5. Liver cancer 149 98 51 — — — 6 56 87 7. Lung cancer 633 418 215 — — — 18 253 362 9. Breast cancer 209 1 208 — — — 20 81 108 Other cancers 1,426 821 605 2 4 12 68 457 883 C. Diabetes mellitus 273 133 140 — — 1 10 66 195 E. Mental and behavioral disorders 133 99 34 — — 16 52 51 14 d 4. Alcohol use disorders 58 47 12 — — 2 19 31 6 e 5. Drug use disorders 70 50 20 — — 14 32 18 6 table continues next page 84 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.5 Deaths from Selected Causes in High-Income Countries, by Age and Sex, 2015 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 927 347 581 1 2 5 13 68 839 1. Alzheimer’s disease and other 718 235 483 — — — — 24 694 dementias H. Cardiovascular diseases 4,313 2,032 2,281 1 1 13 142 827 3,328 3. Ischemic heart disease 2,356 1,180 1,176 — — 3 67 489 1,796 4. Stroke 1,150 481 669 — — 2 27 187 933 I. Respiratory diseases 755 407 348 — — 2 11 125 617 J. Digestive diseases 437 233 204 — — 2 34 138 262 2. Cirrhosis of the liver 162 107 55 — — 1 23 85 53 K. Genitourinary diseases 290 130 160 — — 1 7 44 238 1. Kidney diseases 222 103 119 — — 1 6 35 180 N. Congenital anomalies 43 23 20 25 2 3 4 6 4 f Other noncommunicable diseases 257 115 141 6 1 3 13 50 182 III. Injuries 718 484 234 8 8 90 169 189 253 A. Unintentional injuries 476 301 175 7 6 47 87 116 212 1. Road traffic injury 118 85 33 2 2 27 35 33 21 2. Other unintentional injuries 357 216 142 6 4 20 53 84 191 B. Intentional injuries 242 183 59 1 2 44 82 73 41 1. Suicide 196 148 48 — 1 31 63 63 38 2. Homicide and collective violence 46 35 11 1 1 13 18 10 3 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 85 Table 4.6 Deaths from Selected Causes in the World, by Age and Sex, 2000 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 6,122 3,082 3,040 606 1,241 1,587 1,609 813 266 All causes 52,135 27,617 24,517 10,063 1,644 2,993 5,937 12,016 19,481 I. Communicable, maternal, 16,160 8,384 7,776 8,715 901 1,053 1,879 1,617 1,995 perinatal, and nutritional conditions A. Infectious and parasitic diseases 8,608 4,615 3,993 3,551 697 717 1,515 1,223 906 1. Tuberculosisa 1,667 1,108 559 100 50 133 425 590 369 3. HIV/AIDS 1,463 754 709 222 29 227 788 181 16 4. Diarrheal diseases 2,177 1,061 1,116 1,206 166 115 111 234 345 5. Vaccine-preventable diseasesb 1,040 527 513 802 172 33 13 13 7 6–7. Meningitis and encephalitis 560 289 271 281 96 65 40 44 33 c 8. Acute hepatitis 131 71 60 19 16 23 24 30 19 9a. Malaria 859 440 419 749 23 24 26 23 15 Other infectious and parasitic 711 366 345 171 144 97 88 108 102 diseases B. Respiratory infections 3,672 1,976 1,696 61 17 44 157 1,043 2,350 C. Maternal conditions 425 — 425 — — 220 205 — — D. Neonatal conditions 3,232 1,817 1,415 3,232 — — — — — 1. Preterm birth complications 1,340 731 609 1,340 — — — — — 2. Birth asphyxia and birth trauma 1,120 637 483 1,120 — — — — — 3. Neonatal sepsis and infections 540 325 215 540 — — — — — 4. Other neonatal conditions 232 124 108 232 — — — — — E. Nutritional deficiencies 475 234 241 207 43 27 28 53 117 II. Noncommunicable diseases 31,391 16,128 15,263 914 321 778 2,875 9,623 16,880 A. Malignant neoplasms 6,950 3,840 3,110 37 60 149 916 2,789 2,998 3. Stomach cancer 739 460 280 — — 6 80 303 350 4. Colon and rectum cancers 578 292 285 — 1 6 51 202 318 5. Liver cancer 662 450 212 — 4 14 128 282 233 7. Lung cancer 1,255 886 370 1 — 5 101 556 592 9. Breast cancer 445 2 443 — — 6 113 185 140 Other cancers 3,272 1,751 1,521 36 55 112 442 1,262 1,365 C. Diabetes mellitus 958 431 527 3 4 17 80 365 489 E. Mental and behavioral disorders 267 206 60 — — 45 116 81 26 4. Alcohol use disordersd 143 119 24 — — 11 64 56 12 e 5. Drug use disorders 105 79 26 — — 31 45 19 9 table continues next page 86 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.6 Deaths from Selected Causes in the World, by Age and Sex, 2000 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 1,008 437 571 20 30 64 67 130 698 1. Alzheimer’s disease and other 654 243 411 — — — 1 64 589 dementias H. Cardiovascular diseases 14,425 7,009 7,416 60 43 210 989 4,164 8,958 3. Ischemic heart disease 6,883 3,531 3,352 4 3 67 468 1,989 4,353 4. Stroke 5,407 2,479 2,927 21 17 63 304 1,590 3,412 I. Respiratory diseases 3,672 1,976 1,696 61 17 44 157 1,043 2,350 J. Digestive diseases 1,880 1,110 769 39 47 109 355 655 674 2. Cirrhosis of the liver 905 603 302 12 18 54 230 379 212 K. Genitourinary diseases 898 467 431 23 19 53 110 259 434 1. Kidney diseases 709 368 341 16 15 42 90 212 333 N. Congenital anomalies 687 355 331 575 50 29 16 9 7 f Other noncommunicable diseases 647 296 351 97 49 58 69 129 246 III. Injuries 4,583 3,105 1,478 434 422 1,163 1,183 775 606 A. Unintentional injuries 3,228 2,150 1,078 409 375 675 726 544 500 1. Road traffic injury 1,118 829 289 75 90 320 339 200 95 2. Other unintentional injuries 2,110 1,321 789 334 284 356 387 344 405 B. Intentional injuries 1,355 955 400 25 47 488 457 231 106 1. Suicide 748 479 269 — 15 240 245 162 87 2. Homicide and collective violence 607 476 131 25 32 248 212 70 20 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 87 Table 4.7 Deaths from Selected Causes in Low-Income Countries, by Age and Sex, 2000 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 636 317 319 101 166 176 125 55 12 All causes 7,735 4,030 3,705 3,145 535 706 1,084 1,162 1,102 I. Communicable, maternal, 4,998 2,554 2,444 2,856 343 383 689 401 326 perinatal, and nutritional conditions A. Infectious and parasitic diseases 3,070 1,599 1,470 1,446 269 258 571 321 203 1. Tuberculosisa 380 252 128 34 15 24 86 129 92 3. HIV/AIDS 744 349 396 129 20 101 393 92 9 4. Diarrheal diseases 683 352 330 409 59 44 36 57 78 5. Vaccine-preventable diseasesb 359 182 177 281 60 11 4 3 1 6–7. Meningitis and encephalitis 200 108 93 120 33 23 11 9 5 8. Acute hepatitisc 17 9 8 1 2 5 4 3 2 9a. Malaria 474 241 233 426 12 12 12 7 6 Other infectious and parasitic 212 107 106 47 69 39 25 21 11 diseases B. Respiratory infections 212 111 101 11 6 9 17 72 97 C. Maternal conditions 158 — 158 — — 82 77 — — D. Neonatal conditions 797 451 346 797 — — — — — 1. Preterm birth complications 306 170 135 306 — — — — — 2. Birth asphyxia and birth trauma 297 165 132 297 — — — — — 3. Neonatal sepsis and infections 151 92 59 151 — — — — — 4. Other neonatal conditions 43 23 20 43 — — — — — E. Nutritional deficiencies 202 109 93 86 25 15 14 25 36 II. Noncommunicable diseases 2,087 1,036 1,051 175 79 137 276 691 729 A. Malignant neoplasms 389 177 212 6 7 19 83 183 91 3. Stomach cancer 26 15 11 — — 1 5 13 7 4. Colon and rectum cancers 18 9 9 — — 1 4 9 4 5. Liver cancer 26 16 10 — — 1 6 13 5 7. Lung cancer 28 17 11 — — — 3 16 9 9. Breast cancer 29 — 29 — — 1 12 12 3 Other cancers 262 121 141 6 6 15 54 119 62 C. Diabetes mellitus 70 37 33 — 1 3 8 28 29 E. Mental and behavioral disorders 11 9 2 — — 3 5 3 1 4. Alcohol use disordersd 7 6 1 — — 2 3 2 1 e 5. Drug use disorders 4 3 1 — — 1 1 — — table continues next page 88 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.7 Deaths from Selected Causes in Low-Income Countries, by Age and Sex, 2000 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 79 40 40 6 5 14 7 8 38 1. Alzheimer’s disease and other 39 15 24 — — — — 4 35 dementias H. Cardiovascular diseases 803 369 433 9 11 32 83 286 380 3. Ischemic heart diseases 275 142 133 — 1 8 29 102 136 4. Stroke 356 155 201 3 4 11 32 132 174 I. Respiratory diseases 212 111 101 11 6 9 17 72 97 J. Digestive diseases 235 139 96 10 14 25 50 79 57 2. Cirrhosis of the liver 108 66 42 2 4 10 29 42 21 K. Genitourinary diseases 71 38 34 6 5 10 11 19 22 1. Kidney diseases 50 26 24 4 4 8 7 13 14 N. Congenital anomalies 126 66 61 100 15 7 3 1 1 Other noncommunicable diseasesf 90 51 39 27 16 14 10 12 12 III. Injuries 650 440 210 113 113 186 119 70 48 A. Unintentional injuries 472 311 161 106 96 106 70 52 42 1. Road traffic injury 148 105 43 17 19 47 32 20 11 2. Other unintentional injuries 324 206 118 89 77 59 37 31 30 B. Intentional injuries 178 129 50 7 17 80 50 18 6 1. Suicide 52 34 18 — 3 22 14 9 4 2. Homicide and collective 127 95 32 7 14 58 36 9 2 violence Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 89 Table 4.8 Deaths from Selected Causes in Lower-Middle-Income Countries, by Age and Sex, 2000 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 2,103 1,072 1,032 261 488 582 491 224 57 All causes 19,067 10,121 8,946 5,414 829 1,352 2,283 4,339 4,850 I. Communicable, maternal, 8,403 4,329 4,074 4,803 492 521 804 887 897 perinatal, and nutritional conditions A. Infectious and parasitic diseases 4,474 2,395 2,079 1,863 381 345 628 709 548 a 1. Tuberculosis 1,042 684 358 54 28 91 265 381 224 3. HIV/AIDS 383 204 179 59 7 60 203 50 4 4. Diarrheal diseases 1,335 627 708 689 101 67 68 166 243 b 5. Vaccine-preventable diseases 650 329 321 501 106 21 8 9 5 6–7. Meningitis and encephalitis 294 147 147 134 54 36 22 27 21 8. Acute hepatitisc 82 42 40 16 13 16 13 14 9 9a. Malaria 359 187 173 302 10 12 13 15 9 Other infectious and parasitic 329 176 153 107 62 43 37 46 33 diseases B. Respiratory infections 1,272 711 561 41 7 21 80 489 634 C. Maternal conditions 236 — 236 — — 123 112 — — D. Neonatal conditions 1,887 1,052 835 1,887 — — — — — 1. Preterm birth complications 774 412 361 774 — — — — — 2. Birth asphyxia and birth trauma 639 367 272 638 — — — — — 3. Neonatal sepsis and infections 332 199 133 332 — — — — — 4. Other neonatal conditions 142 74 69 142 — — — — — E. Nutritional deficiencies 182 83 99 95 15 9 9 18 36 II. Noncommunicable diseases 8,945 4,683 4,262 419 150 368 1,087 3,184 3,738 A. Malignant neoplasms 1,259 617 642 11 20 51 266 578 333 3. Stomach cancer 102 64 39 — — 2 20 50 30 4. Colon and rectum cancers 77 40 37 — — 2 14 34 26 5. Liver cancer 105 67 39 — 1 4 23 48 29 7. Lung cancer 131 97 34 — — 1 18 72 39 9. Breast cancer 123 — 123 — — 3 44 55 20 Other cancers 721 349 371 11 19 37 147 319 189 C. Diabetes mellitus 326 151 175 1 2 8 35 141 139 E. Mental and behavioral disorders 54 42 11 — — 14 21 14 5 4. Alcohol use disordersd 23 20 3 — — 3 10 8 2 5. Drug use disorderse 27 21 6 — — 10 10 5 2 table continues next page 90 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.8 Deaths from Selected Causes in Lower-Middle-Income Countries, by Age and Sex, 2000 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 254 122 132 6 16 33 32 42 124 1. Alzheimer’s diseases and other 136 56 79 — — — 1 23 112 dementias H. Cardiovascular diseases 4,147 2,123 2,023 24 19 108 400 1,485 2,111 3. Ischemic heart diseases 2,185 1,190 994 3 1 37 210 786 1,148 4. Stroke 1,441 683 758 10 8 30 111 525 757 I. Respiratory diseases 1,272 711 561 41 7 21 80 489 634 J. Digestive diseases 755 461 294 24 27 66 167 266 205 2. Cirrhosis of the liver 385 261 124 9 13 36 106 145 76 K. Genitourinary diseases 361 208 154 11 11 29 57 121 132 1. Kidney diseases 279 157 122 8 8 23 46 96 98 N. Congenital anomalies 300 151 149 257 22 11 6 2 1 f Other noncommunicable diseases 218 98 120 44 24 26 25 45 54 III. Injuries 1,719 1,109 610 192 187 463 393 268 215 A. Unintentional injuries 1,268 808 460 182 171 271 247 203 193 1. Road traffic injury 342 263 79 29 34 101 94 59 26 2. Other unintentional injuries 926 544 381 153 138 170 153 145 168 B. Intentional injuries 451 301 149 10 16 193 146 65 21 1. Suicide 270 165 106 — 7 125 86 40 13 2. Homicide and collective violence 180 137 43 10 9 68 60 25 8 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 91 Table 4.9 Deaths from Selected Causes in Upper-Middle-Income Countries, by Age and Sex, 2000 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 2,184 1,106 1,078 173 428 575 634 290 84 All causes 14,130 7,751 6,380 1,369 244 695 1,668 3,828 6,326 I. Communicable, maternal, 2,053 1,129 924 983 63 134 309 229 334 perinatal, and nutritional conditions A. Infectious and parasitic diseases 892 512 380 234 45 104 264 151 94 1. Tuberculosisa 193 130 62 12 7 15 55 64 40 3. HIV/AIDS 302 175 126 33 2 63 168 32 3 4. Diarrheal diseases 150 78 72 106 6 4 6 10 17 5. Vaccine-preventable diseasesb 30 16 14 20 6 1 1 1 1 6–7. Meningitis and encephalitis 53 27 26 25 8 5 6 5 4 c 8. Acute hepatitis 24 15 9 1 1 2 6 9 5 9a. Malaria 25 12 13 21 1 1 1 1 1 Other infectious and parasitic 117 59 58 16 13 13 22 29 24 diseases B. Respiratory infections 1,548 800 748 9 3 11 45 362 1,117 C. Maternal conditions 29 — 29 — — 14 15 — — D. Neonatal conditions 493 283 211 493 — — — — — 1. Preterm birth complications 227 129 98 227 — — — — — 2. Birth asphyxia and birth trauma 175 100 75 175 — — — — — 3. Neonatal sepsis and infections 51 30 21 51 — — — — — 4. Other neonatal conditions 40 23 16 40 — — — — — E. Nutritional deficiencies 72 35 36 26 3 3 4 8 28 II. Noncommunicable diseases 10,706 5,666 5,039 270 77 201 954 3,366 5,838 A. Malignant neoplasms 2,718 1,612 1,106 17 28 62 386 1,086 1,137 3. Stomach cancer 399 254 146 — — 3 40 164 192 4. Colon and rectum cancers 164 82 81 — — 2 18 57 86 5. Liver cancer 410 286 123 — 3 8 91 169 139 7. Lung cancer 543 380 163 — — 3 52 232 256 9. Breast cancer 99 — 99 — — 1 31 42 25 Other cancers 1,103 609 494 16 25 45 155 422 439 C. Diabetes mellitus 325 139 186 1 1 5 26 135 158 E. Mental and behavioral disorders 68 50 18 — — 10 27 21 10 4. Alcohol use disordersd 31 27 4 — — 2 13 12 3 e 5. Drug use disorders 24 17 8 — — 7 9 5 4 table continues next page 92 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.9 Deaths from Selected Causes in Upper-Middle-Income Countries, by Age and Sex, 2000 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 304 133 171 6 6 12 15 41 223 1. Alzheimer’s disease and other 233 94 139 — — — — 26 207 dementias H. Cardiovascular diseases 4,637 2,326 2,311 25 11 53 304 1,409 2,835 3. Ischemic heart diseases 1,717 898 819 1 1 17 124 523 1,052 4. Stroke 2,131 1,062 1,069 8 4 18 119 668 1,315 I. Respiratory diseases 1,548 800 748 9 3 11 45 362 1,117 J. Digestive diseases 494 299 194 5 5 15 91 191 186 2. Cirrhosis of the liver 259 173 86 1 1 7 62 118 69 K. Genitourinary diseases 260 129 131 6 3 12 34 83 123 1. Kidney diseases 224 111 113 4 3 10 29 74 104 N. Congenital anomalies 209 112 97 182 11 7 4 2 2 f Other noncommunicable diseases 143 67 77 18 8 13 22 36 47 III. Injuries 1,372 955 417 116 104 360 405 234 154 A. Unintentional injuries 933 651 282 109 93 204 256 159 111 1. Road traffic injury 452 334 118 26 32 121 158 82 33 2. Other unintentional injuries 481 317 164 83 61 83 99 77 78 B. Intentional injuries 439 304 135 7 11 155 149 74 42 1. Suicide 222 123 99 — 4 57 70 54 37 2. Homicide and collective violence 217 181 36 7 7 98 79 20 6 Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 93 Table 4.10 Deaths from Selected Causes in High-Income Countries, by Age and Sex, 2000 thousands Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. Population (millions) 1,200 588 612 70 158 255 360 244 113 All causes 11,202 5,715 5,487 135 35 241 902 2,686 7,203 I. Communicable, maternal, 706 371 335 73 3 15 78 100 438 perinatal, and nutritional conditions A. Infectious and parasitic diseases 173 110 63 8 2 9 52 41 60 a 1. Tuberculosis 52 42 11 1 — 4 20 15 13 3. HIV/AIDS 35 26 8 — — 3 24 7 — 4. Diarrheal diseases 10 4 6 2 — — — 1 6 5. Vaccine-preventable diseasesb 1 1 1 1 — — — — — 6–7. Meningitis and encephalitis 13 7 6 2 1 2 2 3 3 c 8. Acute hepatitis 9 5 4 — — — 2 3 3 9a. Malaria 1 — — 1 — — — — — Other infectious and parasitic 52 24 28 1 1 1 4 11 35 diseases B. Respiratory infections 641 355 286 1 1 3 15 120 502 C. Maternal conditions 3 — 3 — — 1 1 — — D. Neonatal conditions 55 32 23 55 — — — — — 1. Preterm birth complications 33 19 14 33 — — — — — 2. Birth asphyxia and birth trauma 10 6 4 10 — — — — — 3. Neonatal sepsis and infections 5 3 2 5 — — — — 4. Other neonatal conditions 7 4 3 7 — — — — E. Nutritional deficiencies 20 7 13 — — 1 2 17 II. Noncommunicable diseases 9,654 4,743 4,911 50 15 72 559 2,383 6,575 A. Malignant neoplasms 2,585 1,434 1,150 3 5 17 181 942 1,437 3. Stomach cancer 212 127 84 — — 1 15 75 121 4. Colon and rectum cancers 319 161 158 — — 1 15 101 202 5. Liver cancer 121 81 40 — — 1 9 52 60 7. Lung cancer 554 392 162 — — — 29 236 288 9. Breast cancer 194 2 192 — — — 27 75 91 Other cancers 1,185 671 514 3 5 14 86 402 675 C. Diabetes mellitus 237 104 133 — — 2 11 61 163 E. Mental and behavioral disorders 134 105 29 — — 17 63 43 10 d 4. Alcohol use disorders 81 65 16 — — 4 38 34 6 5. Drug use disorderse 50 39 11 — — 13 24 9 3 table continues next page 94 Disease Control Priorities: Improving Health and Reducing Poverty Table 4.10 Deaths from Selected Causes in High-Income Countries, by Age and Sex, 2000 (continued) Both sexes Sex Both sexes Male Female Age group Total Total Total 0–4 yrs. 5–14 yrs. 15–29 yrs. 30–49 yrs. 50–69 yrs. 70+ yrs. F. Neurological conditions 371 143 228 1 3 4 12 38 312 1. Alzheimer’s disease and other 246 78 168 — — — — 11 235 dementias H. Cardiovascular diseases 4,838 2,190 2,648 2 2 17 203 983 3,632 3. Ischemic heart diseases 2,706 1,301 1,404 — — 5 106 578 2,017 4. Stroke 1,479 580 899 — 1 4 42 266 1,166 I. Respiratory diseases 641 355 286 1 1 3 15 120 502 J. Digestive disease 396 211 185 — — 3 47 119 227 2. Cirrhosis of the liver 153 103 50 — — 1 33 73 46 K. Genitourinary diseases 205 93 113 — — 2 9 36 157 1. Kidney diseases 156 73 83 — — 1 8 29 117 N. Congenital anomalies 51 27 24 35 3 3 4 4 3 Other noncommunicable diseasesf 196 81 115 8 2 4 13 37 132 III. Injuries 842 601 241 13 17 154 265 203 190 A. Unintentional injuries 555 380 175 12 14 94 153 129 154 1. Road traffic injury 177 127 49 3 5 51 55 38 25 2. Other unintentional injuries 379 253 125 9 9 44 98 91 129 B. Intentional injuries 287 221 66 1 3 60 113 74 37 1. Suicide 204 157 47 — 1 36 76 59 32 2. Homicide and collective 83 64 19 1 2 24 37 15 4 violence Note: — = fewer than 500 deaths are attributable to a specific cause; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. a. Tuberculosis deaths in HIV-negative people. Tuberculosis deaths in HIV-positive people are included in the HIV/AIDS category. b. Pertussis, diphtheria, measles, and tetanus are included here. c. Liver cancer and cirrhosis deaths resulting from past hepatitis infection are not included here. d. Only direct deaths because of alcohol intoxication are included. e. Only direct deaths because of drug overdose or adverse reaction for licit and illicit drugs are included. f. Benign neoplasms; endocrine, blood, and immune disorders; sense organ diseases; skin diseases; musculoskeletal diseases; oral conditions; and sudden infant death syndrome are included. Global and Regional Causes of Death: Patterns and Trends, 2000–15 95 Figure 4.5 displays the trends in global death rates for aging and epidemiological change act in opposite direc- specific causes from 2000 to 2015, covering NCDs, tions, resulting in a relatively small increase in the num- Group I conditions, and injuries. Trends include those ber of deaths overall from Group I causes and NCDs and for dementia, already noted; for HIV/AIDS, where the a decline in deaths from injuries from 2000 to 2015. scale-up of antiretroviral treatment coverage has had a Table 4.11 summarizes average annual rates of change significant effect; and for falls, where population aging is for cause-specific death rates over the period 2000 to driving much of the increase in deaths. 2015 for the world and for countries grouped by income. The relative contributions of population growth, For children under age 15 years, death rates from leading aging, and epidemiological change (changes in age- infectious causes have declined for all groups of coun- specific death rates) to overall growth in the number of tries by more than 4 percent per year, while death rates deaths from 2000 to 2015 are summarized in figure 4.6 from preterm birth complications have declined in all for HICs and for LMICs. Population growth and epi- groups, but at a lower rate of about 2 to 4 percent. demiological improvement have been the dominant fac- For younger adults ages 15–49 years, death rates from tors in mortality for LMICs over the past 15 years, acting major causes are declining across all income groups, with in opposite directions and resulting in an overall increase the exception of road injuries, where rates are almost flat of 34 percent for total NCD-related deaths and 13 percent or rising in LMICs and declining significantly in HICs. for injury-related deaths. The 28 percent decline in For older adults ages 50–69 years, NCD mortality Group I–related deaths is driven by epidemiological rates are declining slowly in most regions at 1–2 percent improvement. Population aging is an important factor per year, with the exception of mortality from IHD, for only NCD mortality, but it is likely to become more which is increasing in low-income and upper-middle- important over the next 15 years. For HICs, population income countries, and mortality from IHD, stroke, Figure 4.5 Trends in Global Mortality Rates for Selected Causes, 2000–15 a. Noncommunicable diseases b. Group I conditions c. Injuries 60 30 120 50 25 100 Death rate per 100,000 population Death rate per 100,000 population Death rate per 100,000 population 40 20 80 30 15 60 40 20 10 20 10 5 0 0 0 2000 2005 2010 2015 2000 2005 2010 2015 2000 2005 2010 2015 IHD Cirrhosis Pneumonia Tuberculosis Road injury Falls Stroke Stomach cancer Neonatal HIV Suicide Drowning COPD Kidney diseases conditions Malaria Homicide Conflict Digestive Dementia Diarrheal Maternal diseases Lung cancer Colorectal cancer Diabetes Breast cancer Note: COPD = chronic obstructive pulmonary disease; HIV = human immunodeficiency virus; IHD = ischemic heart disease. 96 Disease Control Priorities: Improving Health and Reducing Poverty Figure 4.6 Decomposition of Changes in Annual Number of Deaths, by Country Income Group and Major Cause, 2000–15 a. Low- and middle-income countries b. High-income countries 100 100 80 80 60 60 Change (percent) Change (percent) 40 40 20 20 0 0 –20 –20 –40 –40 –60 –60 Group I Noncommunicable Injuries Group I Noncommunicable Injuries diseases diseases Total change Population growth Population aging Epidemiological change Table 4.11 Average Annual Rate of Change in Cause-Specific Death Rates, by Selected Causes within Age Groups, for the World and by Country Income Group, 2000–15 percent Lower-middle Upper-middle Age and cause World Low income income income High income Ages 0–14 years Diarrheal diseases −6.0 −6.9 −6.0 −5.8 −5.6 Malaria −6.3 −8.0 −5.4 −4.4 −6.5 Lower respiratory infections −4.8 −5.0 −4.7 −6.1 −5.9 Preterm birth complications −2.2 −3.1 −1.7 −4.1 −2.9 Ages 15–49 years HIV/AIDS −3.2 −7.6 −0.9 −2.3 1.8 Tuberculosis −3.1 −4.4 −3.1 −5.7 −4.5 Maternal conditions −3.3 −4.5 −4.1 −3.1 −1.4 Road injury −0.2 0.0 1.2 −0.9 −3.5 Self-harm −1.4 −1.2 −1.4 −2.1 −1.1 Interpersonal violence −1.3 −0.8 −1.3 −0.9 −3.7 Ages 50–69 years Malignant neoplasms −1.0 −0.7 −0.5 −1.3 −0.8 Ischemic heart disease −1.1 0.9 −0.6 0.2 −2.7 Stroke −1.7 −0.2 −1.4 −1.5 −3.5 Chronic obstructive pulmonary disease −2.2 0.2 −1.4 −3.9 −1.1 Note: HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. Global and Regional Causes of Death: Patterns and Trends, 2000–15 97 and COPD, which are declining at around 3 to 4 percent DISCUSSION per year in HICs. Globally, life expectancy has been improving at a rate of more than three years per decade since 1950, with the Gains in Life Expectancy exception of the 1990s (UN 2015). During that period, Figure 4.7 decomposes the gains in life expectancy from progress on life expectancy stalled in Africa because of the 2000 to 2015 to identify the contribution of major rising HIV/AIDS epidemic and in Europe because causes using the methods of Beltran-Sanchez, Preston, of higher mortality in many former Soviet republics and Canudas-Romo (2008). For LICs, 88 percent of the following the collapse of the Soviet Union. Gains in life nine-year increase in life expectancy is due to declines in expectancy accelerated in most regions from 2000 onward, Group I cause-specific death rates, particularly for HIV/ and overall life expectancy rose 5.0 years overall between AIDS, tuberculosis, malaria, diarrheal diseases, lower 2000 and 2015, with an even larger increase of 9.4 years in respiratory infections (mainly pneumonia), and neona- Sub-Saharan Africa (WHO Global Health Observatory tal causes (mainly complications of prematurity, birth 2016). Almost 90 percent of the increase in life expectancy trauma, and neonatal infections). At the other end of the in Sub-Saharan Africa is the result of lower death rates for epidemiological spectrum, in HICs, 96 percent of the Group I causes, the main focus of the MDG health targets 3.7-year gain in life expectancy is associated with a and of global health policies over the MDG period. In reduction in mortality from NCDs (62 percent) and contrast, the increase of 3.7 years in life expectancy in injuries (33 percent). HICs (corresponding to an average increase of 2.5 years per decade or 6 hours per day) was dominated by decreases in NCD death rates, particularly for cardiovas- cular disease. Rates of premature deaths (ages 50–69 years) Figure 4.7 Gains in Life Expectancy at Birth Because of Improved from IHD and stroke decreased 36 percent and 47 percent, Outcomes for Major Causes of Death, for the World and by Country respectively, from 2000 to 2015. Income Group, 2000–15 The global average increase in life expectancy at birth since 2000 exceeds the overall average increase in life 11 expectancy achieved by the best-performing countries over the past century (Oeppen and Vaupel 2002). The 9 world as a whole is catching up with those countries, and improvements in outcomes for all major causes of death Gains in life expectancy (years) have contributed to these huge gains. The gap between 7 life expectancy for HICs and LICs has narrowed, from 26 years in 2000 to 19 years in 2015, a decrease of 7 years. 5 Prospects for Accelerated Improvement to Achieve 3 the 2030 Sustainable Development Agenda The post-2015 SDGs include 13 cause-specific or 1 age-specific mortality targets (WHO 2017c), with 0 many focusing on reducing or ending preventable –1 deaths. Achievement of the major SDG targets for child, maternal, and infectious diseases and for NCDs tri me nt - tri me ou dle ld nt - s ou dle or rie un co un co s e c id rie W es would result in a projected increase in global average es e c mid co -in co h-in om er-m w om er- g Lo inc pp Hi life expectancy of about 4 years by 2030. The gap in inc ow U L average life expectancy between HICs and LICs would Other injury Neonatal, maternal, and nutritional narrow from about 19 years in 2015 to about 14 years Homicide conditions in 2030 (WHO 2014b). Road traffic injuries Other infectious diseases Norheim and others (2015) have proposed an over- Other noncommunicable diseases Respiratory infections arching target for health of reducing the number of Chronic respiratory disease HIV/AIDS, tuberculosis, malaria deaths before age 70 years—both globally and in every Cancer Diarrheal diseases country—by 40 percent by 2030. Countries at different Cardiovascular diseases and diabetes stages of development could, depending on their epi- Note: HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. demiological priorities, achieve this kind of gain in 98 Disease Control Priorities: Improving Health and Reducing Poverty premature mortality by reducing mortality from HIV/ Figure 4.8 Premature Deaths (under Age 70 Years) That Would Have AIDS, malaria, and tuberculosis or reducing causes of Been Averted by Achievement of SDG Mortality Targets, for the World child deaths or NCD-related deaths under age 70 years. and by Country Income Group, 2015 Concerted action to reduce NCD-related deaths before Number of deaths under age 70 years per 1,000 population age 70 years would also reduce NCD death rates for peo- 7 ple ages 70 years and over. 6 Applying the SDG targets to the estimated number of deaths in 2015 by cause, age, and sex can approximate the 5 effect of attaining the SDG health-related targets for number of deaths under age 70 years. In 2015, there were 4 an estimated 30.3 million deaths under age 70 years; if the SDG mortality targets had been achieved in 2015, the 3 number would have been reduced to 19.6 million deaths.1 This represents a 35 percent reduction (almost 11 million 2 premature deaths averted)—close to the proposed 40 percent target. Of these averted deaths, 5 million from 1 infectious diseases, malnutrition, and child and maternal mortality (the MDG causes) would have been avoided, 0 with a further 5 million from NCDs and 900,000 from Low-income Low- and Upper-middle- High-income World injuries also avoided. Figure 4.8 shows the rates of pre- countries middle-income income countries countries countries mature deaths (under age 70 years) per 1,000 population in 2015 for the world and for country income groups, Injury prevented MDG causes prevented together with the estimated number of deaths that would NCD prevented Remaining deaths have been averted by achievement of the SDG mortality Note: NCD = noncommunicable disease; MDG = Millennium Development Goal. targets in 2015. The achievement of SDG mortality tar- gets would have dramatically narrowed cross-income variations in the rate of premature deaths. data or assessing the level of underreporting of deaths in surveys or censuses. In many low-income countries and lower-middle- Uncertainty of Estimates and Limitations income countries, estimates for many causes of death Comparable information about the number of deaths are predicted from available data on causes of death, and mortality rates by cause, age, sex, country, and year using covariates such as gross domestic product and provides important information for discussing priorities educational attainment. Even in HICs with relatively and for monitoring and evaluating progress toward complete health statistics information systems, data global health goals. However, serious problems exist with quality is problematic for many causes. Approximate the quality and availability of information on levels and estimates of cause-specific uncertainty ranges are avail- causes of death, particularly in LICs, where the mortality able in datasets that can be downloaded from the WHO burden is highest. For this reason, there is considerable website (see box 4.1). uncertainty in most cause-of-death estimates. Although death registration data are generally the Demographic methods of assessing the completeness best form of available information on causes of death, of death registration all involve strong assumptions or such data have considerable limitations, even in well- information about migration and are prone to error functioning systems with medical certification of cause resulting from age misstatement in registration or census of death. The so-called garbage codes represent a sub- data and to differential completeness of successive cen- stantial proportion of deaths in some countries, and suses. These errors can result in considerable uncertainty methods for reassigning these deaths to valid causes are in estimates for countries with partially complete regis- highly uncertain and generally not based on empirical tration systems, even before one considers the quality of data. The assignment of underlying cause of death is cause-of-death assignment. both limited by the information provided on the death All-cause mortality estimates in countries without certificate and quite sensitive to the order in which well-functioning death registration systems rely heavily diagnoses are written. For most causes of death, variabil- on census and survey data (particularly sibling survival ity (owing to differences in physician practice when data) and model life tables. Yet no consensus has been certifying a death) in the assignment of valid underlying reached on the methods for analyzing sibling survival causes of death has not been addressed to date. Global and Regional Causes of Death: Patterns and Trends, 2000–15 99 Additionally, some diseases and injuries have specific The type and complexity of models used for GHE vary problems that create difficulty in judging the underlying widely by research and institutional group and by health cause of death (for example, diabetes and heart disease, estimate. More complex models are necessary to generate Alzheimer’s disease and heart disease, and drug or alco- more comprehensive uncertainty intervals. These models hol overdose). Finally, HIV/AIDS and other stigmatized require greater expertise and time and greater computa- causes of death, such as suicide, are routinely certified tional resources to run. In cases of available, high-quality incorrectly; incorrect certification rates vary substan- data, estimates from different institutions are generally in tially across settings. agreement. Discrepancies are more likely to arise for coun- For many countries without a functioning death reg- tries where data are poor and for conditions where data are istration system, particularly in Africa, there is strong sparse and potentially biased. This situation is best reliance on verbal autopsy studies. Most studies are not addressed through improving the primary data. nationally representative samples and, even when con- The WHO and the UN devote considerable attention ducted well, have substantial limitations with respect to to estimates for several high-priority areas, including sensitivity and specificity of identifying specific causes of neonatal, child, and maternal mortality; HIV/AIDS; death. Considerable variation also exists in verbal tuberculosis; malaria; major causes of child death; road autopsy instruments and in analysis and cause assign- injuries; homicides; and cancers. In all of these cases, ment methods. Validation studies are challenging and input data for the particular area are scrutinized by difficult to generalize to other settings. specialists in that area, including academic collaborators; The WHO GHE bring together single-cause analy- household survey technical staff involved with data ses from several WHO departments, interagency col- collection; and country experts, including through the laborations, and other sources and estimates drawn WHO country consultation mechanism. from the GBD 2015 study. These estimates are updated The GHE 2015 draw on these WHO and UN agency using different timetables and varying methods and or interagency statistics and place them in a consistent assumptions in some cases. Ensuring consistency comprehensive context for all causes, drawing on death across cause analyses that are created by various registration data and GBD 2015 analyses for causes and sources is more difficult than for large comprehensive for countries lacking both death registration data and estimates, such as GBD 2015, that are prepared by a investment by the UN system in detailed estimates. Over single academic group. In addition, preparing separate time, some convergence has occurred between GBD and estimates of total mortality and cause-specific mortal- WHO estimates for some causes, although major differ- ity can lead to incompatible cause-specific and total ences remain in areas such as adult malaria mortality. mortality estimates. However, the WHO continues to produce its own GHE, partly because of differences in the estimates of all-cause mortality (envelope) and of mortality for Differences from Other Global Cause-of-Death some major causes. In addition, the WHO has been Assessments unwilling to rely on third-party statistics for which it is Academic institutions are increasingly publishing esti- not responsible or accountable to member states and for mates in parallel to those of the WHO, using different which it does not have, in many cases, access to the data methods that may result in substantially different and methods used. results. The Lancet has become a regular channel for The GHE 2015 use the latest UN Population Division publication of global, regional, and country statistics on life tables to provide envelopes, with some adjustments key health indicators and the burden of disease. Rudan for countries with high HIV/AIDS prevalence and for and Chan (2015) recently characterized this practice as countries with relatively complete death registration a competitive situation that is challenging the position data. The UN life tables are less systematic than the GBD of the WHO. project (which uses its own model life table system), in Over recent years, investigation into differences in the part because of greater investment both in closely exam- estimates for the same indicator has led to improve- ining and assessing available country data and context ments in the data inputs and estimation methods used and in ensuring consistency of estimated deaths with by UN agencies and by the GBD 2015 study. The exis- population, fertility, and migration estimates. For coun- tence of divergent estimates for the same indicator also tries with high HIV/AIDS prevalence, the UN Population has led to increased awareness of major data gaps, espe- Division works with the Joint United Nations Programme cially in LMICs. Lack of reliable data suggests greater use on HIV/AIDS (UNAIDS) to maximize consistency of of data from other—often higher-income—countries HIV/AIDS estimates and all-cause mortality trends and and covariates to predict country statistics. age patterns. In its most recent updates, the GBD 2015 100 Disease Control Priorities: Improving Health and Reducing Poverty study also uses UNAIDS models and inputs but has Table 4.12 Comparison of Estimates of Total Global Deaths, modified key assumptions regarding survival owing to 1990, 2000, and 2015 antiretroviral treatment. It also models HIV/AIDS mor- millions tality as part of its overall model life table analysis in a way that may not adequately account for the complexity Year 1990 2000 2015 of time and age patterns for the HIV/AIDS epidemic. World The GBD model life tables differ most significantly Global Health Estimates (WHO) 48.9 52.1 56.4 from the UN estimates in three ways: Global Burden of Disease estimates (IHME) 47.9 52.1 55.8 • Much lower estimates of older child mortality Africa • Different estimates of all-cause mortality in countries Global Health Estimates (WHO) 7.9 9.8 9.2 with high HIV/AIDS prevalence Global Burden of Disease estimates (IHME) 6.8 8.5 8.0 • Slower time trends, with lower mortality rates in the Note: GHE 2015; WHO = World Health Organization; IHME = Institute for Health Metrics and 1990s in some parts of the world. Evaluation. In the latest update, some of these differences are reduced, but the GBD 2015 estimate of 8.0 million deaths The WHO and other UN agencies will continue to for the WHO African region is still much lower than the prepare and report on global health indicators to fulfill UN estimate of GHE 2015 of 9.2 million deaths their mandate from member states and to be account- (table 4.12). The GBD 2015 estimates for African deaths able to those states through a transparent process, repro- are consistently lower by close to 1.1 million across 1990– ducible methods, and country involvement. For many 2015. In contrast, GHE 2015 and GBD 2015 estimates of years, this involvement has occurred mainly in the con- deaths in children under age 5 years have converged glob- text of WHO or UN expert groups; this work is now also ally and in most regions. Past GBD estimates have oscil- taking place in independent academic research institu- lated above and below the UN interagency estimates. tions, notably through the IHME’s work on the global There are also significant differences (at the global, burden of disease. The resulting debates on data inter- regional, and country levels) for some major causes of pretation, methods, and results can be healthy and death. These differences include HIV/AIDS mortality, productive and can lead to improvements in global for which the GBD 2015 has converged somewhat by health statistics, as long as the focus on methodological using the UNAIDS Spectrum model but has changed sophistication does not come at the expense of working some input parameters. The parameters also include together to improve the essential investments in data malaria mortality, which has seen some convergence for collection, analysis, and resulting use in LMICs. child mortality. However, significant differences remain for adult mortality, with the high GBD 2015 estimates for rates of adult malaria deaths not deemed plausible by CONCLUSIONS many experts in malaria. Some convergence has occurred in other areas, such as maternal mortality, tuberculosis, The results presented here document major changes and causes of child death. Pathogen-specific estimates during the MDG era. On the whole, progress toward the for diarrhea and pneumonia mortality have also con- MDGs has been remarkable, including, for instance, verged, largely as a result of revisions to GBD methods. poverty reduction, improved education, and increased There are some more specific causes where the WHO access to improved drinking water. Progress on the three and the GBD assessments differ (for example, road traf- health goals and targets has also been considerable. fic injuries and homicides), in part because of different Globally, the HIV/AIDS, tuberculosis, and malaria epi- data inclusion and adjustment criteria. For example, demics have been “turned around,” and child mortality both GBD 2015 and the WHO use death registration and maternal mortality have decreased greatly (53 percent data and police or justice system data for homi- and 44 percent, respectively, since 1990), despite falling cides. Despite the intense effort put into assessing and short of the MDG targets. Large reductions in mortality adjusting data from incomplete death registration, GBD have occurred in Sub-Saharan Africa since the early 2015 has not yet put the same effort into assessing and 2000s, coinciding with increased coverage of HIV/AIDS adjusting data from police or justice systems, resulting in treatment, methods of malaria control, and scale-up of low estimates in some countries (for example, estimated vaccination coverage. Despite this progress, major chal- homicide rates are lower for Burkina Faso and Nigeria lenges remain in achieving further progress on child and than for Japan). maternal mortality and on infectious diseases such as Global and Regional Causes of Death: Patterns and Trends, 2000–15 101 HIV/AIDS, tuberculosis, malaria, neglected tropical dis- Phillipe Boucher, Louisa Degenhardt, Jacques Ferlay, eases, and hepatitis. Patrick Gerland, Prabhat Jha, Joy Lawn, Li Liu, Mary The rate of increase in life expectancy in LICs over the Mahy, Bruno Masquelier, Shefali Oza, Francois Pelletier, past 15 years has exceeded the rate of growth observed Juergen Rehm, John Stover, and Danzhen You. The for life expectancy in the countries with the highest life WHO funded this work. expectancies. Longer life expectancies and population aging have resulted in an increased focus on NCDs and their risk factors in LMICs and in HICs. Three-quarters ANNEXES of NCD-related deaths occurred in LMICs in 2015. The two annexes to this chapter are available at http:// Over the past four decades, death rates from cardio- www.dcp-3.org/DCP. vascular disease and smoking-associated cancers have declined substantially in most HICs, and rates for pre- • Annex 4A. Global and Regional Causes of Death mature deaths from cardiovascular disease at ages 30 to 2000–15: Data and Methods 69 declined 28 percent in HICs over the period 2000– • Annex 4B. Global and Regional Burden of Disease 15, more than three times the decrease seen in LMICs. 2000–15: Methods and Summary Results Public health action to address risk factors such as tobacco smoking and air pollution, along with the scale-up of health system coverage for individual-level NOTES risk factor interventions, are important priorities in the SDG era, particularly for LMICs. Weak health systems World Bank Income Classifications as of July 2014 are as are a major obstacle in many countries, resulting in follows, based on estimates of gross national income (GNI) major deficiencies in universal health coverage for per capita for 2013: even the most basic health services and inadequate preparedness for health emergencies. • Low-income countries (LICs) = US$1,045 or less • Middle-income countries (MICs) are subdivided: Lower poverty levels and economic growth have (a) lower-middle-income = US$1,046 to US$4,125 moved many countries to the middle-income categories (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 and enabled an increasing proportion of countries to • High-income countries (HICs) = US$12,746 or more. become self-sufficient in health and even to become aid donors and health technology suppliers (Jamison and 1. Reduction of maternal mortality ratio to 70 per 100,000 others 2013). With enhanced investments to scale up live births; reduction of neonatal and under-age-5 mor- health systems toward universal health coverage and to tality rates to 12 and 25 per 1,000 live births, respectively; address major risk factors, continuing and accelerating 90 percent reduction in deaths from HIV/AIDS, tubercu- the convergence of death rates across country income losis, malaria, and neglected tropical diseases; 33 percent categories will be possible. At the same time, the chal- reduction in deaths from hepatitis, cancer, diabetes, lenges of population aging may be joined by additional cardiovascular disease, and chronic respiratory disease; challenges arising from climate change, political instabil- 50 percent reduction in road injury deaths; 50 percent reduction in diarrheal deaths (through achievement of the ity, and potential new epidemic outbreaks. target for water, sanitation, and hygiene); and 33 percent reduction (arbitrary interpretation of the SDG target of substantial reduction) in deaths from homicide, conflicts, ACKNOWLEDGMENTS and disasters. These estimated mortality reductions are The authors thank Doris Ma Fat for her assistance in conservative and do not include the effects of suicide, extracting death registration data from the WHO pollution, and drug and alcohol use on mortality targets Mortality Database and mapping them to the cause (beyond their contribution to NCD mortality). categories used in this chapter and Dean Jamison for his support and advice. The authors also drew heavily on advice and inputs from other WHO departments, REFERENCES collaborating UN agencies, and WHO expert advisory Avenir Consulting. 2016. “Provisional Updated Spectrum groups and academic collaborators. Modelled Estimates of HIV Mortality for Years 1985– Although we cannot name all those who provided 2015.” Unpublished results provided by John Stover, Avenir advice, assistance, or data, both inside and outside the Consulting, Sydney. WHO, we would particularly like to note the assistance Beltran-Sanchez, H., S. H. Preston, and V. Canudas-Romo. and inputs provided by Bob Black, Ties Boerma, 2008. “An Integrated Approach to Cause-of-Death Analysis: 102 Disease Control Priorities: Improving Health and Reducing Poverty Cause-Deleted Life Tables and Decompositions of Life Rudan, I., and K. Y. Chan. 2015. “Global Health Metrics Needs Expectancy.” Demographic Research 19 (35): 1323–50. Collaboration and Competition.” The Lancet 385 (9963): China CDC (Centre for Disease Control and Prevention). 2016. 92–94. “Death Registration System.” Unpublished tabulations for 2013. Torgerson, P. R., B. Devleesschauwer, N. Praet, N. Speybroeck, Ferlay, J., I. Soerjomataram, M. Ervik, R. Dikshit, S. Eser, A. L. Willingham, and others. 2015. “World Health and others. 2013. Globocan 2012 v1.0, Cancer Incidence Organization Estimates of the Global and Regional Disease and Mortality Worldwide: IARC CancerBase No. 11. Burden of 11 Foodborne Parasitic Diseases, 2010: A Data International Agency for Research on Cancer, Lyon, France. Synthesis.” PLoS Medicine 12 (12): e1001920. http://globocan.iarc.fr. UN (United Nations). 2015. World Population Prospects: 2015 GBD 2015 (Global Burden of Disease 2015 Study) Mortality Revision. New York: UN, Department of Economic and and Causes of Death Collaborators. 2016. “Global, Regional, Social Affairs, Population Division. and National Life Expectancy, All-Cause Mortality, and UNAIDS (Joint United Nations Programme on HIV/AIDS). Cause-Specific Mortality for 249 Causes of Death, 1980– 2016. “HIV Estimates with Uncertainty Bounds 1990–2015.” 2015: A Systematic Analysis for the Global Burden of UNAIDS, Geneva. http://www.unaids.org/en/resources Disease Study 2015.” The Lancet 388 (10053): 1459–544. / documents/2016/HIV_estimates_with_uncertainty IARC (International Agency for Research on Cancer). 2013. _bounds_1990-2015. “Globocan 2012: Data Sources and Methods.” Lyon, UN-IGME (United Nations Inter-agency Group for Child France: IARC. http://globocan.iarc.fr/Pages/DataSource Mortality Estimation). 2015. Levels and Trends in Child _and_methods.aspx. Mortality: Report 2015. New York: United Nations Children’s IHME (Institute for Health Metrics and Evaluation). 2016. Fund on behalf of UN-IGME. “Global Health Data Exchange: GBD Results Tool.” Seattle, UN Statistics Division. 2017. “Revised List of Global Sustainable WA: IHME. http://ghdx.healthdata.org/gbd-results-tool. Development Goal Indicators.” In Report of the Inter- Jamison, D. T., L. H. Summers, G. A. Alleyne, K. J. Arrow, agency and Expert Group on Sustainable Development Goal S. Berkley, and others. 2013. “Global Health 2035: A World Indicators, annex III. New York: UN. https://unstats.un.org Converging within a Generation.” The Lancet 382 (9908): /sdgs/indicators/Official%20Revised%20List%20of%20 1898–955. global%20SDG%20indicators.pdf. Liu, L., S. Oza, D. Hogan, J. Perin, I. Rudan, and others. WHO (World Health Organization). 1990. International 2015. “Global, Regional, and National Causes of Child Statistical Classification of Diseases and Related Health Mortality in 2000–2013 with Projections to Inform Post- Problems: 10th Revision. Geneva: WHO. 2015 Priorities: An Updated Systematic Analysis.” The ———. 2014a. Global Status Report on Violence Prevention Lancet 385 (9966): 430–40. 2014. Geneva: WHO. MMEIG (Maternal Mortality Estimation Inter-Agency ———. 2014b. “An Overarching Health Indicator for the Group). 2015. Trends in Maternal Mortality: 1990 to Post-2015 Development Agenda: Brief Summary of 2015. Geneva: World Health Organization (WHO) on Some Proposed Candidate Indicators.” Background behalf of WHO, United Nations Children’s Fund, United paper for Expert Consultation, December 11–12, WHO, Nations Population Fund, World Bank, and United Nations Geneva. http://www.who.int/healthinfo/indicators /hsi Population Division. _indicators_SDG_TechnicalMeeting_December2015 Norheim, O. F., P. Jha, K. Admasu, T. Godal, R. J. Hum, _BackgroundPaper.pdf?ua=1. and others. 2015. “Avoiding 40% of the Premature ———. 2015a. Global Status Report on Road Safety 2015. Deaths in Each Country, 2010–30: Review of National Geneva: WHO. Mortality Trends to Help Quantify the UN Sustainable ———. 2015b. Health in 2015: From MDGs to SDGs. Development Goal for Health.” The Lancet 385 (9964): Geneva: WHO. http://www.who.int/gho/publications 239–52. http://www.thelancet.com/journals/lancet/article /mdgs-sdgs/en/. /PIIS0140-6736%2814%2961591-9/fulltext. ———. 2016a. Global Tuberculosis Report 2016. Geneva: WHO. Oeppen, J., and J. W. Vaupel. 2002. “Demography: Broken http://www.who.int/tb/publications/global_report/en/. Limits to Life Expectancy.” Science 296 (5570): 1029–31. ———. 2016b. “WHO Methods and Data Sources for Life Patel, M. K., M. Gacic-Dobo, P. M. Strebel, A. Dabbagh, M. M. Tables 1990–2015.” Global Health Estimates Technical Mulders, and others. 2016. “Progress towards Regional Paper WHO/HIS/IER/GHE/2016.2, WHO, Geneva. http:// Measles Elimination—Worldwide, 2000–2015.” Weekly www.who.int/healthinfo/statistics/LT_method.pdf. Epidemiological Record 65 (44): 1228–33. ———. 2016c. WHO Mortality Database. WHO, Geneva. Registrar General of India. 2009. “Causes of Death in India in http://www.who.int/healthinfo/mortality_data/en/. 2001–2003.” Registrar General of India, New Delhi. ———. 2016d. World Health Statistics 2016: Monitoring Registrar General of India and CGHR (Centre for Global Health for the SDGs. Geneva: WHO. http://who.int/gho Health Research). 2015. “Causes of Death Statistics 2010– /publications/world_health_statistics/2016/en. 2013.” Joint report, India Ministry of Home Affairs, Sample ———. 2016e. World Malaria Report 2016. Geneva: Registration System, New Delhi. http://www.censusindia.gov WHO. http://www.who.int/malaria/publications/world .in/2011-common/Sample_Registration_System.html. -malaria-report-2016/en/. Global and Regional Causes of Death: Patterns and Trends, 2000–15 103 ———. 2017a. Global Health Estimates. http://www.who.int WHO Global Health Observatory. 2016. “Life Tables by WHO /healthinfo/global_burden_disease/en/. Region.” WHO, Geneva. http://apps.who.int/gho/data ———. 2017b. “Immunization Surveillance, Assessment, /node.main.LIFEREGION?lang=en. and Monitoring.” WHO, Geneva. http://www.who.int WHO and MCEE (Maternal and Child Epidemiology /immunization/monitoring_surveillance/en/. Estimation). 2016. “MCEE-WHO Methods and Data ———. 2017c. World Health Statistics 2017: Monitoring Sources for Child Causes of Death 2000–2015.” WHO, Health for the SDGs. Geneva: WHO. http://who.int/gho Geneva. http://www.who.int/healthinfo/global_burden /publications/world_health_statistics/2017/en/. _disease/ChildCOD_method_2000_2015.pdf?ua=1. 104 Disease Control Priorities: Improving Health and Reducing Poverty Chapter 5 Annual Rates of Decline in Child, Maternal, Tuberculosis, and Noncommunicable Disease Mortality across 109 Low- and Middle-Income Countries from 1990 to 2015 Osondu Ogbuoji, Jinyuan Qi, Zachary D. Olson, Gavin Yamey, Rachel Nugent, Ole Frithjof Norheim, Stéphane Verguet, and Dean T. Jamison INTRODUCTION how a country’s health performance might relate to A country’s performance in health is typically defined by adjustments in policy. Most published work on country how much better or worse it performs with respect to a performance is based on estimates of mortality levels, particular outcome (for example, life expectancy) com- but some studies investigate rates of change (Bhutta and pared with what would be expected in light of certain others 2010; Croghan, Beatty, and Ron 2006; Kassebaum contextual attributes (for example, income and educa- and others 2014; Lozano and others 2011; Muennig and tion) (Jamison and Sandbu 2001). In Good Health at Low Glied 2010; Munshi, Yamey, and Verguet 2016; Verguet Cost, Halstead, Walsh, and Warren (1985) used a case and Jamison 2013a, 2013b, 2014; Wang and others 2014). study approach to assess country performance in levels of To the extent that rates of change respond to the intro- mortality, examining why three countries and one Indian duction of health policies (for example, a new immuni- state had low levels of mortality despite scant resources. zation program), rates of decline in mortality offer a Later analyses also quantified performance with respect to dependent variable with which to understand the effect levels of mortality and fertility (Wang and others 1999). on performance of social and system determinants. The number of deaths is affected strongly by long- Nevertheless, the measure—like any one-dimensional standing country-level determinants. Essentially, a coun- metric—still has weaknesses. Notably, large declines try that starts with a low level of mortality is likely to from high levels of mortality may still leave an unaccept- continue to have lower mortality, whereas a country that ably large number of deaths. Therefore, rates of change begins with a high level of mortality might improve complement rather than replace the important informa- substantially but still have comparatively high mortality. tion conveyed by estimates of mortality levels. Examining alterations in the number of deaths or annual The need to measure progress in health was especially rate of change in mortality is useful for understanding apparent when assessing whether countries were on track Corresponding author: Osondu Ogbuoji, Department of Global Health and Population, Harvard T. H. Chan School of Public Health, Boston, Massachusetts, United States, oogbuoji@mail.harvard.edu. 105 to achieve the Millennium Development Goals (Bhutta maternal mortality ratios, tuberculosis mortality rates, and others 2010; Kassebaum and others 2014; Lozano and and NCD mortality rates in persons between ages 50 and others 2011; Wang and others 2014). Measuring progress is 69 years. These indicators feature prominently in SDG 3, also crucial to determining whether countries can achieve and updated data became available since the last analysis. the next set of post-2015 Sustainable Development Goals We assessed the annual rates of decline in the chosen (SDGs) that were adopted by United Nations (UN) mem- mortality indicators for 109 LMICs, as defined in the ber states in 2015. The SDGs include health goals with an World Bank income classifications for 2014, with popu- associated set of targets; the Lancet Commission on lations greater than 1 million people (Zeileis 2015). We Investing in Health proposed a target of achieving a “grand used the 2013 World Bank income group classification convergence in global health” by 2035, defined as reducing to ensure that all of the countries in the original paper infectious, maternal, and child deaths to universally low were covered. Annex 5B presents the countries and levels, similar to today’s rates in the best-performing mid- regional groupings included in the analysis. dle-income countries, such as Chile and Turkey (Jamison We estimated rates of decline in under-five mortality and others 2013). Other targets were proposed by the rates (number of children who die after birth and before Global Investment Framework for Women’s and Children’s age five years per 1,000 live births), maternal mortality Health (Stenberg and others 2014), the United Nations ratios (number of pregnant women who die per 100,000 Children’s Fund (UNICEF 2013), the Sustainable live births), tuberculosis mortality ratios (number of Development Solutions Network (SDSN 2013), and the deaths from tuberculosis per 100,000 population per High-Level Panel on the post-2015 development agenda year), and NCD mortality rates (probability of dying (Norheim and others 2015; Peto, Lopez, and Norheim between ages 50 and 69 years from an NCD in the pres- 2014; UN 2013). All of these proposals were debated before ence of other causes). Depending on the availability of adoption of the SDGs by all UN member states. data, we used a 1990–2015 time series for under-five Studying historical rates of change (rates of decline) mortality rates (UNICEF and others 2015), a 1990–2015 in mortality across countries over recent decades can be time series for maternal mortality ratios (WHO 2016b), helpful for testing the feasibility of these different pro- a 1990–2014 time series for tuberculosis mortality rates posals and the SDGs, which include ambitious targets (WHO 2016a), and a 1993–2013 time series for NCD for child, maternal, tuberculosis, human immunodefi- mortality rates from UN-DESA (2015) life tables and ciency virus/acquired immune deficiency syndrome IHME (2015) cause-of-death data. We used several time (HIV/AIDS), and noncommunicable disease (NCD) anchor points for every indicator: 1990, 1995, 2000, 2005, mortality that would require high rates of decline from 2010, and 2015 for under-five mortality rates and mater- 2015 to 2030. Such targets for mortality can be tested nal mortality ratios; 1990, 1995, 2000, 2005, 2010, and for their feasibility by looking at whether high rates of 2014 for tuberculosis mortality rates; and 1993, 1998, decline in mortality have ever been achieved by any low- 2003, 2008, and 2013 for NCD mortality rates. Thus, our or middle-income country (LMIC) and whether similar calculations differ from annualized rates of reduction declines could be achieved in 2016–30. computed using different time frames. We calculated Assessing a country’s health performance with respect 95 percent uncertainty intervals around the estimates and to changes in rates of decline in mortality is, therefore, used R software for all analyses. valuable for studying the effects of policy and for testing We calculated the average annual rates of decline the feasibility of proposed post-2015 health goals. This from levels of the first three indicators for every five-year chapter updates a previous study (Verguet and others interval from 1990 to 2015 and the average annual rates 2014) that examined changes in the annual rate of decline of decline in NCD mortality rates for every five-year of key mortality indicators for 109 LMICs by expanding interval from 1993 to 2013 (equations 5.1 to 5.4). In the period to cover 1990–2015. In addition, we examine total, we have five estimates for the annual rate of decline annual rates of decline in NCD mortality (the probability in under-five mortality rates, maternal mortality ratios, of dying between ages 50 and 69 years from NCDs in the and tuberculosis mortality rates, and for estimates for presence of other causes) over 1993–2013. the annual rate of decline in NCD mortality rates for every country included in the study: 1990–94, 1995–99, 2000–04, 2005–09, and 2010–15 (six-year interval) for METHODS under-five mortality rates and maternal mortality ratios; Verguet and others (2014) analyzed the rates of decline 1990–94, 1995–99, 2000–04, 2005–09, and 2010–14 for for under-five, maternal, tuberculosis, and HIV/AIDS- tuberculosis mortality rates; and 1993–98, 1998–2003, related mortality. The analysis in this chapter is restricted 2003–08, and 2008–13 (mid-year estimate) for NCD to four indicators—under-five mortality rates (5q0), mortality rates. 106 Disease Control Priorities: Improving Health and Reducing Poverty Equations (5.1)–(5.4) are used to perform the We calculated the annual rate of change in the decline estimates: (either an acceleration or a deceleration) for every tran- sition from one five-year period to the next between Lt +1 − Lt 1990 and 2015 (equations 5.3 and 5.4). In total, we have R(t ) = , (5.1) Lt four values for the rate of change in decline for each 1 n country using equation 5.3 for the first three mortality R( p) = ∑ R(t ), n t =1 (5.2) indicators, three values using equation 5.3 for NCD mortality rates, and five values using equation 5.4 for R p +1 − R p under-five and maternal mortality ratios. For accelera- RCR( p) = , (5.3) Rp tion between periods in equation 5.3, we use the rates of decline from two consecutive five-year periods (for 1 n R p +1 − R p RCR( p)t = ∑ R , n p =1 (5.4) example, 1995–99 and 2000–05) to estimate the rate of p change in decline for the transition between those two periods. For simplicity, we present results obtained using where R(t) is the annual rate of decline; L represents levels equations 5.1, 5.2, and 5.4. of under-five mortality rates, maternal mortality ratios, For every mortality indicator, we estimated the year tuberculosis mortality rates, and NCD mortality rates; by which the Lancet Commission on Investing in Health R(p) is the average R(t) for each period; RCR(p) is the rate target (Jamison and others 2013) and SDG target of change in the rate of decline (acceleration or decelera- (UN 2016) would be achieved (figures 5.1–5.4). tion) from one period to the next; RCR(p)t is the period We obtained estimates for every country’s aspirational average of annual rate of change (acceleration or decel- best-performer rate of decline (90th percentile for all eration) in the rate of decline; t represents time intervals; countries) and every region’s aspirational rate of decline and n represents the number of time intervals in a period. (90th percentile for each region). Figure 5.1 Year by Which the Global Targets for Under-Five Mortality Rates Will Be Reached at Aspirational Rates of Decline, Disaggregated by Geographic Region, 2015–50 2050 AGO 2045 TCD SOM CAF SLE Year by which 20 per 1,000 target is achieved MLI NGA 2040 BEN COD GIN CIV NER AFG BFA CMR GNB LSO BDI MRT PAK MOZ TGO 2035 GMB HTI LBR SDN ZWE LAO MWI ZMB ETH GHA PNG SWZ UGA 2030 GAB MMR TKM ERI IND KEN MDG SEN TZA BWA COG NAM TJK RWA ZAF YEM UZB 2025 BGDBOL NPL AZE IRQ DOM KHM GTM PHL 2020 IDN MAR DZA EGY ECU MNG NIC VNM 2015 Target years for under-five mortality rates (5q0) at aspirational rates of decline Region Asia (East and South) Eastern Europe and Central Asia Latin America and the Caribbean Middle East and North Africa Sub-Saharan Africa Annual Rates of Decline in Child, Maternal, Tuberculosis 107 Figure 5.2 Year by Which the Global Targets for Maternal Mortality Ratios Will Be Reached at Aspirational Rates of Decline, Disaggregated by Geographic Region, 2015–50 2060 2055 SLE Year by which 94 per 100,000 target is achieved 2050 CAF TCD NGA BDI COD GMB LBR SOM 2045 CIV GIN MWIMRT CMR MLI GNB NER ERI KEN LSO MOZ 2040 AGO COG ZWE AFG BEN TZA BFA HTI SWZTGO YEM 2035 ETH MDG UGA GHA SEN SDN GAB RWA 2030 NAM NPL PNG ZMB BOL LAO 2025 BGD IND MMR PAK KHM SUR NIC DZA ZAF 2020 BWA HNDIDN PRY MAR PHL 2015 Target years for maternal mortality ratios at aspirational rates of decline Region Asia (East and South) Europe and Central Asia Latin America and the Caribbean Middle East and North Africa Sub-Saharan Africa Figure 5.3 Year by Which the Global Targets for Tuberculosis Mortality Rates Will Be Reached at Aspirational Rates of Decline, Disaggregated by Geographic Region, 2015–50 2050 NGA 2045 COD LBRLSO MOZ NAM SOM GNB KHM Year by which 4 per 100,000 target is achieved GAB TZA AGO BGD LAO MDOMMR SWZ CAF COG AFG SLE ZAF 2040 IDN PNG GHA CMR ETH ZMB BWA GIN 2035 PAK BDI CIV MRT TCD HTI KEN SDNSEN GMB IND NER VNM 2030 MWI NPL ZWE ERI UKR UGA DZA KGZ MLI THA 2025 BEN LBY PHL BFA KAZ TGO UZB MAR MYS BLR MDA PER GEO RWA 2020 LKA PAN ROU ARM YEM 2015 Target years for tuberculosis mortality rates at aspirational rates of decline Region Asia (East and South) Europe and Central Asia Latin America and the Caribbean Middle East and North Africa Sub-Saharan Africa 108 Disease Control Priorities: Improving Health and Reducing Poverty Figure 5.4 Year by Which the Global Targets for NCD Mortality Rates Will Be Reached at Aspirational Rates of Decline, Disaggregated by Geographic Region, 2015–40 2040 SYR KGZ Year by which reducing NCD mortality by 2038 MDA ZAF UKR IRN 2036 ALB KAZ MUS one-thrid is achieved BLR TUR MAR NIC LBN HUN SRB BOL TJK 2034 SLV MKD PAN CRI SEN BRA MNG BGR JAM ARM DZA ROU ERI BIH ECU ARG AZE KHM CHN COL COG CUBDOM EGY ETH GAB GEO HND IND JOR LAO LBY MDG MYS MEX MOZ NPL NER PAK PRY PER SDN SUR THA TGO TKM TZA UZB VEN VNM YEM 2032 AFG BGD CMR COD GMB GIN GNB IDN LSO MRT MMR NAM NGA PNG PHL SLE SOM TUN GTM AGO BEN BFA BDI CAF TCD CIV GHA LBR MWI MLI RWA LKA 2030 UGA IRQ ZWE SWZ ZMB KEN 2028 2026 BWA HTI Target years for NCD mortality rates at aspirational rates of decline Region Asia (East and South) Europe and Central Asia Middle East and North Africa Sub-Saharan Africa Latin America and the Caribbean Note: NCD = noncommunicable disease. RESULTS respectively (table 5.1). Between 1990 and 2004, countries Tables 5.1–5.4 show the rates of decline in mortality indi- with the worst performance for under-five mortality rate cators and highlight the best and worst performers had zero or negative rates of decline (that is, mortality (top-five and bottom-five rates of decline). For under-five remained the same or increased) and, with the exception mortality and NCD mortality rates, the distribution of Sri Lanka, were largely in Southern Africa. Some coun- of rates of decline among the 109 LMICs is narrow tries (for example, FYR Macedonia, Peru, and Serbia in (annex 5C) and becomes narrower in the most recent 1990–99; Cambodia and Rwanda in 2001–15) maintained 10-year period (2005–15 and 2003–13, respectively), while very high rates of decline in under-five mortality rates, the distribution of rates of decline in maternal mortality above 6.0 percent per year. ratios and tuberculosis mortality rates starts out wide and For maternal mortality ratio, in 2010–15, the mean becomes more narrow in recent periods; notably, several rate of decline was 2.7 percent per year; the aspirational countries had very high or very low rates of decline in rate was 6.6 percent per year, with some variation across maternal mortality ratios. For under-five mortality rate, in regions (4.3 percent for South-East Asia, 2.7 percent for 2010–15, the mean rate of decline was 3.5 percent per Sub-Saharan Africa, 1.6 percent for North Africa and the year; the aspirational rate was 6.5 percent per year, with Middle East, 2.1 percent for Eastern Europe and Central some variation across regions (3.9 percent for South-East Asia, and 1.9 percent for Latin America and the Caribbean). Asia, 4.2 percent for Sub-Saharan Africa, 3.8 percent for The top performers in 2010–15 were Kazakhstan, the Lao Middle East and North Africa, 4.8 percent for Europe and People’s Democratic Republic, and Ethiopia, with rates of Central Asia, and 3.5 percent for Latin America and the 10.0, 7.6, and 7.6 percent per year, respectively (table 5.2). Caribbean). The top two performers between 2010 and In all periods assessed, the five worst performers had neg- 2015 were Haiti and the former Yugoslav Republic of ative rates of decline, while the five top performers had Macedonia, with rates of 14.8 and 11.1 percent per year, high rates, greater than 7.0 percent per year. Annual Rates of Decline in Child, Maternal, Tuberculosis 109 110 Disease Control Priorities: Improving Health and Reducing Poverty Table 5.1 Top-Five and Bottom-Five Country Performers in Rate of Decline for Under-Five Mortality Rate (5q0), 1990–2015 1990–94 1995–99 2000–04 2005–09 2010–15 Rate of decine Rate of decline Rate of decline Rate of decline Rate of decline Country per year (%) Country per year (%) Country per year (%) Country per year (%) Country per year (%) Best performers 1 Macedonia, FYR 7.6 Bosnia and 9.5 Rwanda 9.6 Rwanda 10.3 Haiti 14.8 Herzegovina 2 Serbia 7.0 Serbia 8.5 Cambodia 9.6 Congo, Rep. 8.7 Macedonia, FYR 11.1 3 Peru 6.3 Macedonia, FYR 8.2 Moldova 8.8 Belarus 8.3 Rwanda 8.2 3 Hungary 6.3 Peru 7.7 China 8.2 China 8.1 Kazakhstan 8.2 4 Turkey 5.8 Brazil 7.1 Belarus 8.0 Cambodia 8.0 Cambodia 7.8 Worst performers 1 Rwanda −14.4 Swaziland −5.8 Sri Lanka −6.5 Haiti −29.4 Brazil 0.2 2 Swaziland −5.4 South Africa −3.9 Lesotho −1.1 Costa Rica 0.4 Costa Rica 1.0 3 Botswana −5.2 Botswana −3.5 Swaziland −0.3 Malaysia 0.5 Algeria 1.4 4 Zimbabwe −4.7 Lesotho −3.4 Somalia 0.0 Mauritius 0.6 Moldova 1.7 5 Moldova −2.9 Congo, Rep. −2.5 South Africa 0.0 Myanmar 1.5 Dominican Republic 1.8 Table 5.2 Top-Five and Bottom-Five Country Performers in Rate of Decline for Maternal Mortality Ratios, 1990–2015 1990–94 1995–99 2000–04 2005–09 2010–15 Rate of decline Rate of decline Rate of decline Rate of decline Rate of decline Country per year (%) Country per year (%) Country per year (%) Country per year (%) Country per year (%) Best performers 1 South Africa 10.4 Dominican 15.2 Belarus 13.0 Belarus 16.8 Kazakhstan 10.0 Republic 2 Thailand 9.9 Tajikistan 11.9 Rwanda 11.0 Turkey 16.5 Lao PDR 7.6 3 Uzbekistan 9.7 Azerbaijan 10.8 Mongolia 10.1 Kazakhstan 13.9 Ethiopia 7.6 4 Honduras 9.4 Iran, Islamic Rep. 8.7 Lebanon 8.3 Botswana 9.3 Afghanistan 7.5 5 Romania 9.1 Ukraine 7.9 Libya 8.3 Cambodia 8.5 Brazil 7.3 Worst performers 1 Suriname −7.1 Suriname −7.9 South Africa −5.8 Dominican −8.9 Dominican −12.5 Republic Republic Annual Rates of Decline in Child, Maternal, Tuberculosis 2 Azerbaijan −6.8 South Africa −6.5 Uzbekistan −4.2 Mauritius −8.7 Syrian Arab −6.5 Republic 3 Moldova −5.5 Zimbabwe −5.6 Kyrgyz Republic −3.2 South Africa −6.5 Hungary −1.8 4 Tajikistan −4.8 Botswana −5.6 Lesotho −2.8 Panama −3.2 Libya −1.8 5 Nicaragua −4.3 Lesotho −4.4 Honduras −2.5 Georgia −1.6 Serbia −1.6 111 112 Disease Control Priorities: Improving Health and Reducing Poverty Table 5.3 Top-Five and Bottom-Five Country Performers in Rate of Decline for Tuberculosis Mortality Rates, 1990–2014 1990–94 1995–99 2000–04 2005–09 2010–14 Rate of decline Rate of decline Rate of decline Rate of decline Rate of decline Country per year (%) Country per year (%) Country per year (%) Country per year (%) Country per year (%) Best performers 1 Zimbabwe 16.5 Syrian Arab 14.0 Azerbaijan 12.7 Azerbaijan 34.4 Azerbaijan 24.1 Republic 2 Mauritius 15.3 Morocco 13.6 Mongolia 11.6 Tajikistan 16.4 Turkmenistan 22.8 3 Kenya 14.2 Lebanon 13.2 Georgia 11.1 Turkmenistan 15.2 Philippines 20.9 4 Lesotho 13.2 Cuba 11.6 Ecuador 11.0 Honduras 13.7 Egypt, Arab Rep. 18.1 5 Libya 12.9 Mongolia 11.5 Turkey 11.0 Kazakhstan 13.5 Syrian Arab Republic 16.7 Worst performers 1 Cameroon −23.8 Mauritius −19.3 Suriname −26.2 Lebanon −16.4 Albania −17.7 2 Kazakhstan −20.7 Lesotho −15.3 Mauritius −19.7 Suriname −16.0 Libya −15.9 3 Burundi −19.3 Albania −14.1 Jamaica −10.8 Cuba −8.8 Mauritius −11.1 4 Azerbaijan −16.6 Tajikistan −11.9 Lebanon −8.7 Libya −8.6 Lebanon −9.8 5 Moldova −16.2 Thailand −11.1 Congo, Rep. −7.6 Georgia −7.9 Kenya −8.9 Table 5.4 Top-Five and Bottom-Five Performers in Rate of Decline for Noncommunicable Disease Mortality Rates, 1993–2013 1993–98 1998–2003 2003–08 2008–13 1993–13 Rate of decline Rate of decline Rate of decline Rate of decline Rate of decline Country per year (%) Country per year (%) Country per year (%) Country per year (%) Country per year (%) Best performers 1 Rwanda 10.3 Botswana 7.1 Haiti 8.0 Syrian Arab 4.5 Rwanda 3.6 Republic 2 Malawi 4.9 Zimbabwe 6.0 Lebanon 3.6 Kyrgyz Republic 4.0 Malawi 2.5 3 Eritrea 4.4 Sri Lanka 4.6 South Africa 3.4 Moldova 3.7 South Africa 2.5 4 Uganda 3.3 Albania 4.1 Lesotho 3.0 Iran, Islamic Rep. 3.3 Syrian Arab 2.3 Republic 5 Burundi 3.2 Kenya 3.6 Mongolia 2.7 South Africa 3.2 Algeria 2.2 Worst performers 1 Kazakhstan −3.8 Liberia −3.8 Eritrea −3.0 Haiti −7.3 Burkina Faso −0.7 Annual Rates of Decline in Child, Maternal, Tuberculosis 2 Belarus −2.9 Guinea −1.7 Zambia −1.2 Botswana −5.1 Guinea −0.6 3 Sri Lanka −2.6 Bosnia and −1.3 Central African −1.2 Kenya −3.1 Côte d’Ivoire −0.6 Herzegovina Republic 4 Kyrgyz Republic −2.6 Burkina Faso −0.7 Albania −1.1 Zambia −2.7 Ghana −0.4 5 Lesotho −2.0 Senegal −0.6 Burkina Faso −1.1 Zimbabwe −2.5 Central African −0.3 Republic 113 In contrast to under-five and maternal mortality A country’s performance with respect to the rate of ratios, rates of decline for tuberculosis mortality rates change in mortality differs greatly from its perfor- were distributed more widely and showed little change mance with respect to death rate. Examining rates of over time (annex 5C). During 2010–14, the mean rate of decline versus number of deaths for under-five and decline was 3.5 percent per year; the aspirational rate was maternal mortality from 1990 to 2015, we found little 6.5 percent per year, with substantial variation across correlation between the two indicators (annex 5D, regions (4.6 percent for South-East Asia, 1.0 percent for figure 5D.1). Our findings show that high rates of Sub-Saharan Africa, 3.7 percent for North Africa and the decline in mortality can be achieved even at low levels Middle East, 6.9 percent for Eastern Europe and Central of mortality. Asia, and 5.0 percent for Latin America and the For under-five mortality rates, 36 of 109 countries Caribbean). The top performers in 2010–14 were (33 percent) have already achieved the interim 2030 tar- Azerbaijan, Turkmenistan, and the Philippines, with rates get of 20 deaths per 1,000 live births and 73 have not. of 24.1, 22.8, and 20.9 percent per year, respectively At current rates of mortality decline, none of these (table 5.3). In all periods assessed, the worst performers 73 countries will achieve the target between 2030 and had high negative rates, with more than half of them 2050. With an aspirational best-performer rate of decline having rates of less than −15 percent per year. In the last (at the 90th percentile), 38 (35 percent) of the 73 coun- three periods Azerbaijan ranked as the best performer, tries will achieve the target by 2030 and the remaining with rates above 10 percent (12.7 percent in 2000–04, 35 countries (32 percent) will achieve it over 2030–50 34.4 percent in 2005–09, and 24.1 percent in 2010–14). (figure 5.1). With regional aspirational rates, 37 of the For NCD mortality rates, the distribution of rates of 73 countries (34 percent) will achieve the target by 2030, decline varied greatly across World Bank income groups and the remaining 36 countries (33 percent) will achieve (annex 5C). From 1993 to 2013, the mean rate of it between 2030 and 2050 (annex 5E). decline was 0.51 percent per year for low-income coun- For maternal mortality ratios, 46 of 109 countries tries and 0.48 percent per year for lower-middle-income (42 percent) have already achieved the interim 2030 target countries. For upper-middle-income and high-income of 94 deaths per 100,000 live births and 63 have not. At countries, the mean rate of decline over 20 years was current rates, none of these 63 countries will achieve the much higher, at 1.43 and 1.71 percent per year, respec- target by 2050. At the aspirational rate, 21 countries tively. Low- and lower-middle-income countries are (19 percent) will achieve the target by 2030, 41 countries off-track to achieve the SDG target of reducing prema- (38 percent) will achieve it between 2030 and 2050, and ture mortality from NCDs by one-third by 2030 (UN one country (Sierra Leone) will achieve it after 2050 2016). LMICs exhibit wide distribution in the rates of (figure 5.2). At regional aspirational rates, 21 (19 percent) decline, with NCD mortality rates rising in some coun- of these 63 countries will achieve the target by 2030, tries. Over the periods assessed, the worst performers 28 countries (26 percent) will achieve it between 2030 and were Burkina Faso and Guinea, with mean rates of 2050, and 14 countries (13 percent) will achieve the target decline per year of −0.7 and −0.6 percent, respectively, after 2050 (annex 5E). and the best performers were Rwanda (3.6 percent), For tuberculosis mortality rates, 36 (33 percent) of Malawi (2.5 percent), and South Africa (2.5 percent), 108 countries have already achieved the Lancet with mean annual rates of decline of more than 2 Commission’s target of 4 deaths per 100,000 population percent (table 5.4). per year and 72 have not. At current rates, none of these Based on the change in the rate of decline, it is possi- 72 countries will achieve the target by 2050. At the aspira- ble to identify rapid transitions in performance over tional rate, 27 countries (25 percent) will achieve the tar- time (annex 5D, tables 5D.1 to 5D.3). For under-five get by 2030, and the remaining 45 countries (42 percent) mortality rates, most countries had small rates of accel- will achieve it between 2030 and 2050 (figure 5.3). At eration or deceleration (0 percent ± 3 percent) for all regional aspirational rates, 25 countries (23 percent) will periods; when the estimates were larger, they were not achieve the target by 2030, 46 countries (43 percent) will significant, with uncertainty intervals spanning zero. achieve it between 2030 and 2050, and the remaining Likewise, for tuberculosis mortality rates, the point esti- country (Nigeria) will achieve it in 2054 (annex 5E). mates were small, ranging from 2 percent per year to For NCD mortality rates between age 50 and 69, we −3.4 percent per year. However, unlike for under-five estimated the 2016 (January) NCD mortality level as the mortality rates, many of the point estimates for rates of starting point to achieve the SDG target of one-third change in tuberculosis mortality rates were significant. lower NCD mortality in 2030. At current rates, 30 coun- For maternal mortality ratio, although many of the point tries have increasing rates of NCD mortality; only 6 coun- estimates were large, none was found to be significant. tries will achieve the target by 2030, and 27 countries will 114 Disease Control Priorities: Improving Health and Reducing Poverty achieve it by 2050. At the aspirational rate, all countries A few countries have sustained high rates of decline— will achieve the target by 2040 (figure 5.4). At regional for example, under-five mortality rates in Turkey from aspirational rates, 30 countries (28 percent) will achieve 1990 to 2015, maternal mortality ratios in Cambodia the target by 2030, and 24 countries (22 percent) will from 1990 to 2015, and NCD mortality rates in Rwanda achieve it between 2030 and 2050 (Annex 5C, figure 5C.4). from 1993 to 2003. Did unusual circumstances or specific Countries in South-East Asia and Sub-Saharan Africa policies account for these changes in mortality? Indeed, have much lower rates of decline. subsequent assessments could control for contextual deter- minants (for example, income) and exceptional events (for example, natural disasters, political instability) and try to identify the contributions of specific policies implemented. DISCUSSION For instance, Turkey’s high rates of decline in under-five We studied the historical rates of decline in rates of mortality rates coincide with substantial economic growth, under-five, maternal, tuberculosis, and NCD mortality political stability, and the introduction of the Health for 109 LMICs. Annex 5A of this chapter provides a Transformation Program, which rapidly expanded access graphical overview of our findings by country income to health care services (Atun and others 2013). Cambodia’s group. We also identified countries with the best and progress in maternal mortality can probably be attributed worst performance and regions in which performance to socioeconomic improvements, better primary educa- had changed rapidly, either improving or deteriorating. tion, and specific policies leading to increases in skilled Analysis of rates of change in health is useful because birth attendance (Liljestrand and Sambath 2012). rapid alterations in rates of decline—whether accelera- We used the rates of decline in mortality to test the tions or decelerations—can point to a potential effect of feasibility of achieving SDGs, with a particular focus on policy changes and provide a mechanism for under- the 2030 targets proposed by the Lancet Commission on standing what constitutes good policy. We noted almost Investing in Health. Because post-2015 goals present no correlation between number of deaths and rate of ambitious targets for levels of mortality, meeting them decline in mortality indicators (annex 5D, figure 5D.1), will require high (aspirational) rates of mortality decline which suggests that rates of change augment the infor- from 2015 to 2030. Hence, we used historical rates of mation conveyed by mortality estimates but cannot decline—including best-performer aspirational rates—to replace the examination of number of deaths, particu- identify how many countries will achieve these ambitious larly with regard to capturing the underlying intensity of targets if they achieve similar rates of decline over 2015–30. country-level mortality. If all LMICs are able to achieve aspirational best-performer As in our original analysis (Verguet and others rates of decline in mortality, some countries will meet the 2014), this update reveals some interesting patterns. targets for under-five, maternal, and tuberculosis mortal- Rates of decline in child mortality indicate the severe ity by 2030, but the majority will reach their targets by effect of the HIV/AIDS epidemic in Southern Africa. 2050. However, meeting the SDG target of reducing pre- In this region, large increases were recorded in child mature mortality from NCDs by one-third by 2030 mortality over 1995–99, but the number of deaths fell requires a 2.7 percent annual rate of decline. Only rapidly beginning in 2000, reaching a peak rate of Lebanon and South Africa had average annual rates of decline of 6.3 percent per year in 2005–09. This is prob- decline greater than 2.7 percent during most of the ably linked to the rollout of antiretroviral therapy for 15 years between 1998 and 2013, and a few countries the prevention of mother-to-child transmission of maintained rates greater than 2 percent in the same HIV/AIDS (UNAIDS 2013; WHO 2011). Likewise, period, including Algeria, the Islamic Republic of Iran, rates of decline in maternal and tuberculosis mortality Malawi, Rwanda, and the Syrian Arab Republic. The rates deteriorated during 1990–99 in many Central majority of LMICs will not reach the NCD target by 2030. Asian countries after the collapse of the Soviet Union in Similar methods have been used to assess the feasibil- 1991, and rates of decline in under-five mortality rates ity of other post-2015 targets. Norheim and others (2015) dropped abruptly in Rwanda during 1990–99, probably have suggested setting (in addition to specific subtargets because of the genocide in 1994. Low rates of decline in for under-five mortality) an overarching goal of reducing NCD mortality rates between ages 50 and 69 years for premature (under age 70) deaths by 40 percent in 2030 low- and lower-middle-income countries over the from what they were in 2010. 20 years between 1993 and 2013 suggest lack of effec- Our analysis has three key limitations. First, for some tive health interventions (screening, prevention, treat- countries with poor data, the mortality estimates were ment) and rising risk factors (smoking, alcohol predicted largely from past trends. Many countries, consumption, high-calorie processed food). particularly those with high mortality, do not have strong Annual Rates of Decline in Child, Maternal, Tuberculosis 115 registration systems for vital statistics, so mortality esti- • Annex 5C: Distribution of Country-Level Rates of mates are not always reliable. In view of the large number Decline in Mortality Indicators, by Period of countries and distinct mortality indicators analyzed, • Annex 5D: Rate of Change in Decline for Mortality some findings might also be attributable to poor quality Indicators of data. We used mortality estimates from the UN, • Annex 5E: Reaching Global Targets for Mortality UNICEF, World Health Organization (WHO), and Indicators under Regional Best-Performer Rates of Institute for Health Metrics and Evaluation to draw gen- Decline. eral lessons, but our findings could be strengthened fur- ther by incorporating additional sources of data (IHME 2015; Jamison, Murphy, and Sandbu 2016; Kassebaum NOTES and others 2014; Liu and others 2012; Lozano and others Large portions of this chapter have been reproduced from: 2013; Murray and others 2014; UN-DESA 2015; Wang Verguet, S., O. F. Norheim, Z. D. Olson, G. Yamey, and D. T. and others 2014). Jamison. 2014. “Annual Rates of Decline in Child, Maternal, Second, in contrast to our original analysis, where we HIV, and Tuberculosis Mortality across 109 Countries of Low used five-year intervals, we used annual estimates for this and Middle Income from 1990 to 2013: An Assessment of update. Although this may improve the accuracy of the the Feasibility of Post-2015 Goals.” The Lancet Global Health estimates, it may also produce too much noise and mask 2 (12): e698–709. changes or reveal only small changes that may not be World Bank Income Classifications as of July 2014 are as relevant for policy. Despite this noise, annual outcomes follows, based on estimates of gross national income (GNI) could isolate inflection points that capture times when per capita for 2013: countries make performance transitions and help iden- tify seasonal variations or cyclical patterns that longer • Low-income countries (LICs) = US$1,045 or less intervals (for example, every five years) might not flag. • Middle-income countries (MICs) are subdivided: The final limitation is that other modeling tech- (a) lower-middle-income = US$1,046 to US$4,125 (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 niques could be used to forecast rates of decline in • High-income countries (HICs) = US$12,746 or more. mortality and to ascertain whether countries would achieve targets by 2030. For instance, specific explana- tory variables related to declines in mortality could be used, and regression models could be fitted to mortality ANNEX 5A: CROSS-COUNTRY VARIATION time series to make future predictions. However, it is the IN RATES OF DECLINE FOR MORTALITY purpose of our analyses to provide specific performance INDICATORS, 1998–2013 indicators to be explained, rather than explanations. For under-five mortality rates, tuberculosis mortality As such, they provide a starting point. Further research rates, and maternal mortality ratios, we calculated the focusing on individual countries can elucidate the rea- average annual rate of decline over a 15-year period sons for these differences in the rates of change. (1998–2013). We also calculated separate average rates of decline for the World Bank’s low-income, lower- ANNEXES middle-income, and upper-middle-income countries. For NCD mortality, we calculated the mean rate of This chapter has one accompanying print annex: decline over the same 15-year period and average rates of change for all four World Bank income groups, including • Annex 5A: Cross-Country Variation in Rates of high income. Decline for Mortality Indicators, 1998–2013 For each of the four mortality indicators, we graph the distribution of rates of decline separately The online annexes to this chapter are as follows. They for the three income groups (four income groups are available at http://www.dcp-3.org/DCP. for NCDs). Each graph also displays the mean for its income group and the rate of decline for a popu- • Annex 5B: Countries and Regional Groupings in the lous country in the group (China, Ethiopia, India, Analysis United States). 116 Disease Control Priorities: Improving Health and Reducing Poverty Annex Figure 5A.1 Cross-country variation in rates of Annex Figure 5A.2 Cross-country variation in rates of decline of under-five mortality rates (5q0) decline of maternal mortality ratios Panel a. Low-income countries (1998–2013) Panel a. Low-income countries (1998–2013) 0.3 0.3 0.2 0.2 Density Density 0.1 0.1 0 0 –20 –15 –10 –5 0 5 10 15 20 25 30 –20 –15 –10 –5 0 5 10 15 20 25 30 LIC mean = 4.2 percent/year Ethiopia = 5.6 percent/year LIC mean = 3.4 percent/year Ethiopia = 5.6 percent/year Panel b. Lower-middle-income countries (1998–2013) Panel b. Lower-middle-income countries (1998–2013) 0.3 0.3 0.2 0.2 Density Density 0.1 0.1 0 0 –20 –15 –10 –5 0 5 10 15 20 25 30 –20 –15 –10 –5 0 5 10 15 20 25 30 LMIC mean = 4.0 percent/year India = 4.2 percent/year LMIC mean = 2.7 percent/year India = 5.0 percent/year Panel c. Upper-middle-income countries (1998–2013) Panel c. Upper-middle-income countries (1998–2013) 0.3 0.3 0.2 0.2 Density Density 0.1 0.1 0 0 –20 –15 –10 –5 0 5 10 15 20 25 30 –20 –15 –10 –5 0 5 10 15 20 25 30 UMIC mean = 4.1 percent/year China = 7.8 percent/year UMIC mean = 2.8 percent/year China = 4.9 percent/year Average annual rate of decline in under-five mortality rates (5q0): Percent per year Average annual rate of decline in maternal mortality ratios: Percent per year Note: LIC = low-income countries; LIMC = lower-middle-income countries; Note: LIC = low-income countries; LIMC = lower-middle-income countries; UMIC = upper-middle-income countries. UMIC = upper-middle-income countries. Annual Rates of Decline in Child, Maternal, Tuberculosis 117 Annex Figure 5A.3 Cross-country variation in rates of Annex Figure 5A.4 Cross-country variation in rates of decline of tuberculosis mortality rates decline in mortality rates age 50-69 from noncommunicable diseases Panel a. Low-income countries (1998–2013) 0.20 Panel a. Low-income countries (1998–2013) 0.50 0.15 Density 0.40 0.10 Density 0.30 0.05 0.20 0 0.10 –20 –15 –10 –5 0 5 10 15 20 25 30 LIC mean = 3.0 percent/year Ethiopia = 7.1 percent/year 0 Panel b. Lower-middle-income countries (1998–2013) –5 –4 –3 –2 –1 0 1 2 3 4 5 0.20 LIC mean = 0.5 percent/year Ethiopia = 1.0 percent/year Panel b. Lower-middle-income countries (1998–2013) 0.15 Density 0.50 0.10 0.40 0.05 Density 0.30 0 0.20 –20 –15 –10 –5 0 5 10 15 20 25 30 LMIC mean = 3.8 percent/year India = 4.4 percent/year 0.10 Panel c. Upper-middle-income countries (1998–2013) 0.20 0 –5 –4 –3 –2 –1 0 1 2 3 4 5 0.15 LMIC mean = 0.7 percent/year India = 0.8 percent/year Average annual rate of decline in NCD mortality rates: Percent per year Density 0.10 figure continues next page 0.05 0 –20 –15 –10 –5 0 5 10 15 20 25 30 UMIC mean = 3.0 percent/year China = 7.8 percent/year Average annual rate of decline in tuberculosis mortality rates: Percent per year Note: LIC = low-income countries; LIMC = lower-middle-income countries; UMIC = upper-middle-income countries. 118 Disease Control Priorities: Improving Health and Reducing Poverty Annex Figure 5A.4 (continued) Jacobs, R., P. C. Smith, and A. Street. 2006. Measuring Efficiency in Health Care: Analytic Techniques and Health Policy. Panel c. Upper-middle-income countries (1998–2013) Cambridge, U.K.: Cambridge University Press. Jamison, D. T., S. M. Murphy, and M. E. Sandbu. 2016. “Why 0.80 Has Under-5 Mortality Decreased at Such Different Rates in Different Countries?” Journal of Health Economics 48: 16-25. 0.60 Jamison, D. T., and M. E. Sandbu. 2001. “WHO Ranking of Health System Performance.” Science 293 (5535): 1595–96. Density Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, S. 0.40 Berkley, and others. 2013. “Global Health 2035: A World Converging within a Generation.” The Lancet 382 (9908): 0.20 1898–955. Kassebaum, N. J., A. Bertozzi-Villa, M. S. Coggeshall, K. A. Shackelford, C. Steiner, and others. 2014. “Global, Regional, 0 and National Levels and Causes of Maternal Mortality dur- –5 –4 –3 –2 –1 0 1 2 3 4 5 ing 1990–2013: A Systematic Analysis for the Global Burden UMIC mean = 1.2 percent/year China = 1.6 percent/year of Disease Study 2013.” The Lancet 384 (9947): 980–1004. Liljestrand, J., and M. R. Sambath. 2012. “Socio-Economic Panel d. High-income countries (1998–2013) Improvements and Health System Strengthening of Maternity Care Are Contributing to Maternal Mortality Reduction in 0.50 Cambodia.” Reproductive Health Matters 20 (39): 62–72. Liu, L., H. L. Johnson, S. Cousens, J. Perin, S. Scott, and others. 0.40 2012. “Global, Regional, and National Causes of Child Mortality: An Updated Systematic Analysis for 2010 with Density 0.30 Time Trends since 2000.” The Lancet 379 (9832): 2151–61. Lozano, R., M. Naghavi, K. Foreman, S. Lim, K. Shibuya, and 0.20 others. 2013. “Global and Regional Mortality from 235 Causes of Death for 20 Age Groups in 1990 and 2010: A 0.10 Systematic Analysis for the Global Burden of Disease Study 2010.” The Lancet 380 (9859): 2095–128. 0 Lozano, R., H. Wang, K. J. Foreman, J. K. Rajaratnam, M. Naghavi, –5 –4 –3 –2 –1 0 1 2 3 4 5 and others. 2011.“Progress towards Millennium Development HIC mean = 1.9 percent/year Goals 4 and 5 on Maternal and Child Mortality: An Updated United States = 1.4 percent/year Systematic Analysis.” The Lancet 378 (9797): 1139–65. Muennig, P. A., and S. A. Glied. 2010. “What Changes in Average annual rate of decline in NCD mortality rates: Percent per year Survival Rates Tell Us about US Health Care.” Health Affairs Note: HIC = high-income countries; LIC = low-income countries; LIMC = lower-middle- 29 (11): 2105–13. income countries; NCD = noncommunicable diseases; UMIC = upper-middle-income Murray, C. J. L., K. F. Ortblad, C. Guinovart, S. S. Lim, countries. T. M. Wolock, and others. 2014. “Global, Regional, and National Incidence and Mortality for HIV, Tuberculosis, and Malaria during 1990–2013: A Systematic Analysis for the Global Burden of Disease Study 2013.” The Lancet REFERENCES 384 (9947): 1005–70. Atun, R., S. Aydın, S. Chakraborty, S. Sümer, M. Aran, and Munshi, V., G. Yamey, and S. Verguet. 2016. “Trends In State- others. 2013. “Universal Health Coverage in Turkey: Level Child Mortality, Maternal Mortality, and Fertility Enhancement of Equity.” The Lancet 382 (9886): 65–99. Rates in India.” Health Affairs 35 (10): 1759–63. Bhutta, Z. A., M. Chopra, H. Axelson, P. Berman, T. Boerma, Norheim, O. F., P. Jha, K. Admasu, T. Godal, R. J. Hum, and and others. 2010. “Countdown to 2015 Decade Report others. 2015. “Avoiding 40% of the Premature Deaths in (2000–10): Taking Stock of Maternal, Newborn, and Child Each Country, 2010–30: Review of National Mortality Survival.” The Lancet 375 (9730): 2032–44. Trends to Help Quantify the UN Sustainable Development Croghan, T. W., A. Beatty, and A. Ron. 2006. “Routes to Goal for Health.” The Lancet 385 (9964): 239–52. Better Health for Children in Four Developing Countries.” Peto, R., A. D. Lopez, and O. F. Norheim. 2014. “Halving Milbank Quarterly 84 (2): 333–58. Premature Death.” Science 345 (6202): 1272–72. Halstead, S. B., J. A. Walsh, and K. S. Warren. 1985. Good Health SDSN (Sustainable Development Solutions Network). 2013. “An at Low Cost. New York: Rockefeller Foundation. Action Agenda for Sustainable Development.” SDSN, Paris. IHME (Institute for Health Metrics and Evaluation). 2015. “GBD Stenberg, K., H. Axelson, P. Sheehan, I. Anderson, A. M. Compare.” IHME, University of Washington, Seattle, WA. Gülmezoglu, and others. 2014. “Advancing Social and Annual Rates of Decline in Child, Maternal, Tuberculosis 119 Economic Development by Investing in Women’s and ———. 2014. “Estimates of Performance in the Rate of Children’s Health: A New Global Investment Framework.” Decline of Under-Five Mortality for 113 Low- and Middle- The Lancet 383 (9925): 1333–54. Income Countries, 1970–2010.” Health Policy and Planning UN (United Nations). 2013. “A New Global Partnership: 29 (2): 151–63. Eradicate Poverty and Transform Economies through Verguet, S., O. F. Norheim, Z. D. Olson, G. Yamey, and Sustainable Development.” UN High-Level Panel, New York. D. T. Jamison. 2014. “Annual Rates of Decline in Child, ———. 2016. “Sustainable Development Knowledge Platform.” Maternal, HIV, and Tuberculosis Mortality across 109 UN, New York. https://sustainabledevelopment.un.org/sdg3. Countries of Low and Middle Income from 1990 to 2013: UNAIDS (Joint United Nations Programme on HIV/AIDS). An Assessment of the Feasibility of Post-2015 Goals.” The 2013. UNAIDS Report on the Global AIDS Epidemic 2013. Lancet Global Health 2 (12): e698–709. Geneva: UNAIDS. Wang, H., C. A. Liddell, M. M. Coates, M. D. Mooney, C. E. Levitz, UN-DESA (United Nations Department of Economic and Social and others. 2014. “Global, Regional, and National Levels Affairs). 2015. World Population Prospects: The 2015 Revision. of Neonatal, Infant, and Under-5 Mortality during DVD ed. New York: UN-DESA, Population Division. 1990–2013: A Systematic Analysis for the Global Burden UNICEF (United Nations Children’s Fund). 2013. “UNICEF of Disease Study 2013.” The Lancet 384 (9947): 957–79. Key Asks on the Post-2015 Development Agenda.” UNICEF, Wang, J., D. T. Jamison, E. Bos, A. Preker, and J. Peabody. New York. 1999. Measuring Country Performance on Health: Selected UNICEF (United Nations Children’s Fund), WHO (World Indicators for 115 Countries. Washington, DC: World Health Organization), World Bank, and UN-DESA (United Bank. Nations Department of Economic and Social Affairs), WHO (World Health Organization). 2011. Global HIV/AIDS Population Division. 2015. Levels and Trends in Child Response: Epidemic Update and Health Sector Progress Mortality Report 2015. New York: UNICEF. towards Universal Access. Progress report, WHO, Geneva. Verguet, S., and D. T. Jamison. 2013a. “Improving Life Expectancy: ———. 2016a. “Tuberculosis Burden Estimates.” WHO, Geneva. How Many Years Behind Has the USA Fallen? A Cross- ———. 2016b. “WHO GHO Maternal Mortality Data.” National Comparison among High-Income Countries from WHO, Geneva. http://apps.who.int/gho/data/node.main.15 1958 to 2007.” British Medical Journal Open 3 (7): e002814. ?lang=en. ———. 2013b. “Performance in Rate of Decline of Adult Zeileis, A. 2015. “pwt8: Penn World Table (Version 8.x).” Mortality in the OECD, 1970–2010.” Health Policy 109 (2): R package version 8.1-0. http://CRAN.R-project.org/ 137–42. package=pwt8. 120 Disease Control Priorities: Improving Health and Reducing Poverty Chapter 6 Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries Beverley M. Essue, Tracey-Lea Laba, Felicia Knaul, Annie Chu, Hoang Van Minh, Thi Kim Phuong Nguyen, and Stephen Jan INTRODUCTION Development Solutions Network 2014; WHO 2015). The World Health Report 2000: Health Systems: Improving This imperative is reflected in the third Sustainable Performance (WHO 2000); the World Health Organization Development Goal, which sets a target for achieving uni- (WHO) resolution on sustainable health financing, uni- versal health coverage, including financial risk protection; versal health coverage, and social health insurance (WHO access to high-quality essential health care services; and 2005); and the World Health Report: Health Systems access to safe, effective, high-quality, and affordable essen- Financing: The Path to Universal Coverage (WHO 2010) all tial medicines and vaccines for all (UN General Assembly highlighted the substantial economic burden faced by 2015). This commitment is echoed in the World Bank’s individuals with no access to affordable, high-quality recent call to eradicate impoverishment owing to health health care. These reports placed the need to address the care expenditures by 2030 (Kim 2014). economic effect of illness—in particular, catastrophic and A lack of both prepayment mechanisms and the means impoverishing health expenditure—on the global health and resources to pool risks has limited the capacity of policy agenda. many health care systems to provide access to high-quality Financial protection—a core element of universal health care services. As a result, for decades, many health health coverage—aims to ensure that people receive the systems, particularly in low- and middle-income coun- health care services they require without facing finan- tries (LMICs), have relied heavily on private payments in cial ruin (WHO 2010). Devising strategies to protect the form of out-of-pocket costs to fund health care. populations from financial risk has become a major In 2014, 18 percent of total health expenditure globally focus of global health policy development (WHO and came from out-of-pocket payments (WHO 2014). The World Bank 2014). burden is even greater in LMICs. In 2014, out-of-pocket Affordable access to high-quality health care is now payments equaled approximately 39 percent of total considered a basic human right and a critical step to health expenditure for low-income countries, 56 percent the achievement of sustainable economic and social for lower-middle-income countries, and 30 percent for development and the elimination of poverty (Sustainable upper-middle-income countries (WHO 2016). Corresponding author: Beverley Essue, University of Sydney, Sydney, Australia; beverley.essue@sydney.edu.au. 121 Relying on out-of-pocket costs to finance health care We begin by estimating the population-level burden of is both inefficient and inequitable and places a major CHE—the most common indicator of the household financial strain on individuals and households (WHO economic burden of health expenditure—and draw on 2010). Out-of-pocket costs can perpetuate poverty and empirical research of specific chronic diseases and injuries lead many individuals to delay or forgo necessary care to estimate the prevalence of CHE associated with seven (Peters and others 2008; van Doorslaer and others 2006). categories of conditions: cancers, CVDs, chronic infec- This link, where the household’s investment in health tious diseases, endocrine diseases, injuries, renal diseases, further impoverishes that household, can lead to a con- and respiratory diseases. We then draw on a review of tinuous cycle of poor health and poverty (Knaul, Wong, NCDs in LMICs to describe the broader household eco- and Arreola-Ornelas 2012). nomic effects associated with ill health, including impov- This burden is of particular concern for persons with erishing health expenditure, productivity effects, distressed chronic diseases, for whom repeated and lifelong costs financing, and treatment discontinuation. We discuss are associated with the management and treatment of implications of the results for improving financial protec- illness (Kankeu and others 2013). For example, in some tion and offer directions for future research. countries, a household may have to pay as much as eight days’ worth of wages to purchase one month’s supply of only one of the multiple medicines required for the opti- POPULATION-LEVEL ESTIMATES OF mal treatment of cardiovascular disease (CVD) or diabe- CATASTROPHIC AND IMPOVERISHING tes (Cameron and others 2009; Gelders and others 2006). HEALTH EXPENDITURES In more extreme cases, the costs of treatment for chronic and long-term conditions such as human immunodefi- Catastrophic and impoverishing health expenditures, ciency virus/acquired immune deficiency syndrome also referred to as medical impoverishment, continue to (HIV/AIDS) and surgery for some cancers have kept challenge health systems around the world and pose a patients confined to hospitals indefinitely pending pay- key barrier to improving economic and social well-being ment to the hospitals or forced them to stop treatment (Knaul, Wong, and Arreola-Ornelas 2012). Very conser- altogether (Human Rights Watch 2006). Although vative estimates suggest that, globally, at least 150 million households, even those that are already impoverished, people a year face financial catastrophe and 100 million may be able to manage a one-time shock and recover in are driven into poverty by expenditure on health care the short run (for example, over a period of a week or a (Xu and others 2007). month), they may not be able to withstand the ongoing CHE and impoverishing health expenditure are inter- costs of treatment for chronic diseases. related, but distinct, concepts (figure 6.1). Consensus Furthermore, LMICs are undergoing a protracted is lacking on the definition of what constitutes a epidemiological transition (Frenk and others 1989). Underfunded and weak health systems continue to face Figure 6.1 Definition of Catastrophic and Impoverishing a backlog of acute diseases and conditions associated Health Expenditures with poverty, together with the onslaught of costly and chronic noncommunicable diseases (NCDs), conditions that affect the entire population at all income levels. This situation inevitably results in competing priorities about which services to include in essential packages of care A. Catastrophic B. Impoverishing and which to cover through national insurance funds C. health expenditure health expenditure (Beaglehole and others 2011). However, evidence is lack- ing on the household-level economic burden associated with certain categories of disease, particularly chronic diseases. Such evidence would inform global health pol- icy development by highlighting where the greatest gains in financial protection might be realized (Shrime and A. Health care expenditure is defined as catastrophic using any of the conventional definitions. others 2015) and help governments prioritize the mea- B. Impoverishing health expenditure results at any level of expenditure: sures needed to move toward universal health coverage. • Darker shaded area: for the population already in poverty, This chapter estimates the burden of catastrophic any level of spending further entrenches social disadvantage, health expenditure (CHE) associated with chronic ill and there is a high likelihood of forgoing care. C. Health care expenditure is catastrophic and impoverishes health and injuries in LMICs and describes the broader the household. economic effects on households. It is organized as follows. 122 Disease Control Priorities: Improving Health and Reducing Poverty catastrophic level of expenditure for households and medical expenditure (Essue and others 2015; the most appropriate denominator for measuring CHE: Knaul, Arreola-Ornelas, and Méndez-Carniado 2016). expenditure, income, or consumption (Knaul, Wong, Although progress has been made at a population and Arreola-Ornelas 2012; O’Donnell and others 2007). level, research shows variations in the financial Box 6.1 distinguishes between these two concepts. protection afforded to different subgroups (box 6.2). The economic burden associated with ill health extends beyond paying for care (table 6.1). Household members cope with the onset of illness in various ways, Box 6.1 and the response can influence their treatment-seeking behavior (McIntyre and others 2006; Okoli and Cleary 2011; Sauerborn, Adams, and Hien 1996; Xu and others Conceptual Relationship between Catastrophic 2007). When faced with ill health, particularly unex- Health Expenditure and Impoverishing Health pected events, the household must mobilize resources to Expenditure pay for health care, often by borrowing money, using limited savings, and selling assets—all of which can neg- Conceptually, catastrophic health expenditure is a measure atively affect the long-term economic well-being of the of the burden of health care expenditure (that is, out-of- household, including its ability to deal with ongoing pocket costs) on a household’s available resources. It can health care needs and future health shocks (Kruk, result from sizable and unpredictable one-off payments Goldmann, and Galea 2009; McIntyre and others 2006; and from a steady flow of unbudgeted medical bills, includ- Peters and others 2008; Russell 2004). Ill health can also ing relatively small payments (Knaul and others 2006; affect the productivity of both the sick individual and a Schoenberg and others 2007; Thuan and others 2006). family caregiver, leading to loss of paid employment or educational opportunities. All these factors severely Impoverishing health expenditure is defined as expendi- impair the family’s capacity to earn income in both ture on health care that results in a household falling below temporary and longer-term ways. the prevailing poverty line or deepening its impoverish- Financial protection through tax-financed social ment if it is already poor (Knaul, Wong, and Arreola- health insurance programs is a major pillar of efforts Ornelas 2012; Xu 2005). Such impoverishment is also by national governments to achieve universal health linked to employment, because loss of income owing to ill coverage. Indeed, there is evidence of the extent to health can drive households into poverty (Gertler and which health insurance–based measures effectively Gruber 2002). provide financial protection by curbing the burden of Table 6.1 Indicators Used to Measure the Household Economic Burden of Ill Health Indicator Definition Advantages Limitations Catastrophic health Total health care expenditure (out-of-pocket • Provides objective measure • Has wide variation in the threshold and expenditure costs) as a percentage of household of the drain on available denominator used and the categories of resources (O’Donnell and others 2007; Xu and household resources caused health care expenditure included, which others 2003). The denominator, household by health care expenditure makes it difficult to use as a benchmark resources, is measured as discretionary • Is the most commonly across studies expenditure (also referred to as capacity to used indicator and widely • Does not capture forgone care owing to pay or nonfood expenditure), total expenditure, endorsed unaffordable health care costs or household income. • Arbitrary threshold: implicitly assumes that the given level of expenditure will impose the same burden across the population Impoverishing The outcome when total health care • Provides a measure of • Does not account well for the health expenditure expenditure subtracted from baseline income the effect of illness on poorest households, for whom any (also referred results in the household’s income falling below the household’s economic level of expenditure further entrenches to as medical the prevailing poverty line (Wagstaff and van well-being and potentially their poverty impoverishment) Doorslaer 2003) the national economy table continues next page Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 123 Table 6.1 Indicators Used to Measure the Household Economic Burden of Ill Health (continued) Indicator Definition Advantages Limitations Economic hardship A measure of the potential consequences for • Takes account of the • Has wide variation in the definition and or financial stress the household of health care expenditure. It opportunity costs categories of expenses included, which captures instances in which the household is associated with health care limits its generalizability unable to meet the costs of essential payments expenditure and potential • Does not account well for instances in (housing, food, heating, child care, transport, economic consequences for which households were unable to meet health care). It is most commonly defined as households essential bills before the onset of illness an instance of missing any one of the specified payments (Essue and others 2011). • Tends to be measured in cross-sectional studies, which are unable to assess the effect and recurrence of these consequences over time Distressed financing A measure of the strategies used by the • Accounts for the economic • Has wide variation in the distressed household to pay for health care expenses, consequences of health care financing categories included, which limits often including savings, borrowed funds expenditure for household its generalizability (either through formal or informal loan or economies • Tends to be measured in cross-sectional through credit schemes), or sale of assets. It • Offers insights into studies, which are unable to assess the is a descriptive measure that accounts for the potentially effective informal effect of using these strategies over time percentage of households using each of the strategies for dealing with financing strategies (Kruk, Goldmann, and Galea health care costs 2009; McIntyre and others 2006). Box 6.2 Monitoring Universal Health Coverage: Achieving Financial Protection in Asia Universal health coverage entails everyone having health expenditure followed the WHO methodol- access to needed health services without financial ogy in all four countries (Xu 2005). hardship. In the Western Pacific region, several coun- tries have made progress toward achieving universal Annual household out-of-pocket health expendi- health coverage and protecting their populations tures ranged from US$144 in Mongolia to US$190 from financial risk. in Vietnam. Medicines were a major component of out-of-pocket health expenditures in Mongolia and Country-specific studies on the equity of health the Philippines. The average proportion of house- service use and financial protection have been holds that incurred catastrophic health expendi- conducted in Mongolia (Tsilaajav, Nanzad, and ture (CHE) ranged from 0.9 percent in Mongolia Ichinnorov 2015), the Philippines (Ulep and dela to 2.3 percent in Vietnam (figure B6.2.1). Across Cruz 2013), and Vietnam (Minh and Phuong expenditure quintiles, the proportion of house- 2016). These studies examined health service use, holds that incurred CHE increased in Mongolia out-of-pocket health expenditures, catastrophic and the Philippines but decreased in Vietnam as health expenditure, impoverishing health expendi- the expenditure quintile increased. Over time, ture, and their determinants over time. Data were the proportion of households incurring CHEs from nationally representative surveys—socio- increased in the Philippines, but it fell in Mongolia economic or income and expenditure surveys— and Vietnam. containing information on health service use and health expenditure. The method used to calculate Impoverishment resulting from health expenditures out-of-pocket, catastrophic, and impoverishing was highest in the lowest and second-to-lowest box continues next page 124 Disease Control Priorities: Improving Health and Reducing Poverty Box 6.2 (continued) Figure B6.2.1 Proportion of Households with Figure B6.2.2 Proportion of Households Impoverished Catastrophic Health Expenditure in Selected Asian Owing to Health Expenditures in Selected Asian Countries, by Expenditure Quintile, Various Years Countries, by Expenditure Quintile, Various Years 3 7 Percentage of households with CHE Percentage of households incurring 6 5 impoverishment 2 4 3 1 2 1 0 0 e 1 2 3 e 4 5 1 3 2 4 5 ag ag ile ile ile ile ile ile ile ile ile ile er er int int int int int int int int int int Av Av Qu Qu Qu Qu Qu Qu Qu Qu Qu Qu Expenditure quintile Expenditure quintile Mongolia Philippines Vietnam Mongolia Vietnam Sources: Tsilaajav, Nanzad, and Ichinnorov 2015, based on data from the 2012 Sources: Tsilaajav, Nanzad, and Ichinnorov 2015, based on data from the 2012 Mongolia Household Socio-Economic Survey; Ulep and dela Cruz 2013, based Mongolia Household Socio-Economic Survey; Ulep and dela Cruz 2013, based on on data from the 2012 Philippines Family and Income Expenditure Survey; data from the 2012 Philippines Family and Income Expenditure Survey; Minh and Minh and Phuong 2016, based on data from the 2014 Vietnam Living Standards Phuong 2016, based on data from the 2014 Vietnam Living Standards Survey. Survey. Note: For the Philippines, the national average proportion of impoverishment owing to health expenditures was 1.0 percent. Analyses by quintile are not available. expenditure quintiles, with Vietnam at 6.4 percent effects of universal health coverage, including health and Mongolia at 2.3 percent in the lowest expendi- service use and financial protection. Further research ture quintile (figure B6.2.2). and cross-country comparisons should focus on examining the shock and cumulative effects of the Given differences in the data sources, methods, recall burden of health payments, particularly for poor periods, and survey years, there are limitations com- and vulnerable populations and for households with paring results across countries. However, these coun- members who are aging or have chronic diseases, try-specific studies offer evidence for monitoring the where the effect of these outcomes is likely greater. The poorest quintile of populations and older adults those that are chronic, affect the economic well-being continue to be at greater risk than the general popula- of households. tion (Goeppel and others 2016). Population-based estimates of CHE using data Much of the work in this field has focused on des- from household surveys have been found to vary sub- cribing the burden associated with catastrophic and stantially from research in populations with chronic impoverishing health expenditure at the population diseases. For instance, in Vietnam, population-level level, illuminating the problem, and mobilizing support surveys found that 2.3 percent of all households had for population-wide initiatives such as universal health CHE in 2014 (box 6.2), whereas studies of individuals coverage. A limitation of the research to date is its use with diabetes (Smith-Spangler, Bhattacharya, and of population-based data that lack detailed indicators of Goldhaber-Fiebert 2012), acute myocardial infarction the health status, including specific diseases, of individu- (Jan and others 2016), and HIV/AIDS (Tran and als in the households under study. Research on the eco- others 2013) found that 8 percent, 38 percent, nomic burden associated with particular diseases is and 35 percent, respectively, had CHE. In China, needed to understand how specific diseases, especially population-level surveys found that 13 percent of all Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 125 households had CHE in 2008 (Y. Li and others 2012), Table 6.2 Global Burden of Disease, by Category of whereas studies of individuals with stroke (Heeley and Disease, 2012 others 2009), diabetes (Smith-Spangler, Bhattacharya, Percentage of total global and Goldhaber-Fiebert 2012), and acute myocardial Disease category burden of diseasea infarction (Jan and others 2016) found that 71 percent, 80 percent, and 15 percent, respectively, had CHE. This Infectious diseases 15.8 difference between population-level and disease- Cardiovascular diseases 14.4 related estimates of CHE has also been found in both Injuries 11.1 high-income countries (Essue and others 2011; Essue Cancers 8.2 and others 2014; Schoen and others 2010) and other LMICs (Huffman and others 2011; Saito and others Respiratory diseases 5.0 2014; Xu and others 2003). Endocrine diseases 2.2 The household economic burden of ill health is not Renal diseases 1.1 simply a population-level problem; it is also highly Total 57.8 influenced by the disease course of individual condi- Source: WHO 2014. tions. Understanding variations in outcomes within a. Measured using disability-adjusted life year. populations can help decision makers identify the highest-risk populations, account for the ways in which We initially included maternal, infant, and childhood different conditions affect patients and their house- conditions and mental illnesses in the search, but holds, and generate economic incentives for preventing excluded them from the analysis, because too few studies and managing disease. reported rates of CHE for these conditions. From a broader perspective, the remaining seven categories of disease constitute almost 60 percent of the total global PREVALENCE ESTIMATES OF CATASTROPHIC burden of disease, as shown in table 6.2. HEALTH EXPENDITURE ASSOCIATED Methodology WITH CHRONIC ILL HEALTH AND INJURIES This discussion is based on a systematic search of studies IN LMICS that reported rates of CHE associated with the treatment This section analyzes the prevalence of CHE related to and management of chronic ill health and injuries. The chronic ill health and injuries in LMICs and the way it detailed search strategy and the equations used for the differs among regions. The analysis is based on a system- calculations are described in online annex 6A, along with atic search of studies that reported rates of CHE associ- the characteristics of the studies identified in the search. ated with the treatment and management of seven One issue that arose is the lack of consensus in the conditions: measurement of CHE. A commonly used approach is to measure the household’s total annual expenditure on • Cancers: Breast, uterine, cervical, colorectal, mouth, health care or health-related expenses (for example, pharynx, ovarian, stomach and tracheal, and bron- transport) as a proportion of the household’s resources, chial or lung measured in terms of income, expenditure, or consump- • CVDs: CVD (undefined), angina, heart disease, acute tion (O’Donnell and others 2007). Household resources coronary syndrome, acute myocardial infarction, as the denominator in this equation may involve a stroke, cerebrovascular disease (undefined), and measure of either nondiscretionary expenditure ischemic heart disease (Wagstaff and van Doorslaer 2003) or capacity to pay • Chronic infectious diseases: HIV/AIDS, malaria, (Xu and others 2003), both of which define CHE in tuberculosis, and hepatitis B terms of nonfood expenditure. In this analysis, we note • Endocrine diseases: Diabetes and endocrine disease the CHE definitions and thresholds used in each study (undefined, but not diabetes) but nonetheless include each as essentially the same out- • Injuries: Injuries caused by assault, blunt objects, come when calculating the prevalence of CHE associated burns, falls, road traffic accidents, and sharp objects with each condition. • Renal diseases: Chronic kidney disease and kidney disease (undefined). Summary of Findings • Respiratory diseases: Asthma, chronic obstructive pul- The systematic search identified 41 studies (42 published monary disease, and pulmonary disease (undefined). papers) that reported rates of disease-related CHE. 126 Disease Control Priorities: Improving Health and Reducing Poverty Most studies used a cross-sectional design (30), recruit- Most of the studies were conducted in middle-income ing either a convenience sample (22) or a random countries, clustered in South and East Asia; the greatest sample (18) from either a health care facility or a numbers were conducted in China (8) and India (6) hospital (26) or from households in the community (map 6.1). Endocrine diseases and CVDs were the most (14); 1 study used administrative data. The studies studied conditions (table 6.3), which is reasonably con- were conducted between 1997 and 2013, with 14 con- sistent with the 20 leading causes of disease burden ducted between 2010 and 2013. Of these 41 studies, (Global Burden of Disease Study 2013 Collaborators 7 were conducted in high-income countries (2 in 2015). Data coverage from the systematic search was best Australia, 1 in Greece, 2 in the Republic of Korea, for countries in the upper-middle-income group; the and 2 in the United States). This analysis focuses only greatest gaps were for research on renal and respiratory on LMICs. diseases (see online annex 6A, table 6A.4). Map 6.1 Density of Studies on Disease-Related Catastrophic Health Expenditure Number of records 1 20 Note: The map includes studies found for all country income categories. For multicountry studies, each country is represented in the figure so the total number of studies depicted exceeds the number of studies identified in the systematic search. Table 6.3 Density of Conditions for the Study of Disease-Related Catastrophic Health Expenditure, by Country Income Group Country income group Disease Low-income Lower-middle-income Upper-middle-income Endocrine diseases 7 17 10 Cardiovascular diseases 5 9 7 Cancers 1 5 5 table continues next page Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 127 Table 6.3 Density of Conditions for the Study of Disease-Related Catastrophic Health Expenditure, by Country Income Group (continued) Country income group Disease Low-income Lower-middle-income Upper-middle-income Chronic infectious diseases 3 4 6 Injuries 1 2 — Maternal, infant, and childhood — 2 1 conditions Renal diseases — — 1 Respiratory diseases — 1 — Mental illnesses — 1 — Multiple conditions — — 1 Note: The number in each cell is the count of studies of each condition identified in the review. Some studies included multiple conditions and different countries, and thus the total count in this table exceeds the total number of articles reviewed. — = none. All studies collected data on out-of-pocket payments Rates of CHE from studies based on samples from for direct medical expenses, although the categories of hospitals or health care facilities were significantly higher expenses collected varied somewhat. Where specified, than those from studies based on samples from house- most studies collected data on medicines (30), and more holds or communities for each World Bank income than half collected data on hospitalizations (24) and category (low-income: − x diff , 56.2; t = 5.00, p = 0.007; medical consultations (27). Nonmedical costs (travel, lower-middle-income: − x diff , 27.1; t = 4.97, p < 0.0001; accommodation, care expenses) were taken into account upper-middle-income: − x diff , 26.5; t = 3.75, p < 0.0001). in 19 studies and lost productivity in 4 studies. This difference is not surprising, because hospitals are CHE was most commonly measured in terms of a not an unbiased source of population data on health household’s capacity to pay, defined as total expenditure net expenditure. of food expenses (Xu and others 2003), followed by income Overall, across all LMICs, the largest population thresholds and total expenditure (figure 6.2). By condition experiencing CHE comprised persons with renal dis- category, the ranges in CHE rates were as follows: eases (187.7 million), followed by CVDs (138.4 million), chronic infectious diseases (101.9 million), endocrine • Cancers: 6.2 percent (cancer, undefined, Republic of diseases (46.0 million), cancers (14.3 million), respira- Korea) to 67.9 percent (cancer, undefined, the Islamic tory diseases (9.6 million), and injuries (0.9 million). In Republic of Iran) upper-middle-income countries, the largest population • CVDs: 0.05 percent (heart disease, Nepal) to 84.3 experiencing CHE comprised persons with renal diseases percent (CVD, Tanzania) (100.6 million), followed by CVDs (78.2 million), • Chronic infectious diseases: 7.1 percent (malaria, chronic infectious diseases (74.2 million), endocrine South Africa) to 90.0 percent (HIV/AIDS, the Lao diseases (22.4 million), cancers (11.9 million), respira- People’s Democratic Republic) tory diseases (8.2 million), and injuries (0.5 million). • Endocrine diseases: 1.0 percent (diabetes, Nepal) to In lower-middle-income countries, the largest popula- 26.6 percent (diabetes, Ecuador) tion experiencing CHE comprised persons with renal • Injuries: 0.8 percent (injury, undefined, Nepal) to diseases (83.3 million), followed by CVDs (59.9 million), 46 percent (road traffic injury, India). endocrine diseases (23.3 million), and chronic infectious • Maternal, infant, and childhood conditions: 1.0 percent diseases (6.2 million). In low-income countries, chronic (rotavirus, Malaysia) to 44.8 percent (rotavirus, Bolivia) infectious diseases were associated with the greatest bur- • Mental illnesses: 5.5 percent (depressive disorders, India) den of CHE (21.4 million), followed by renal diseases • Renal diseases: 9.8 percent (kidney disease, the (3.8 million), CVDs (0.4 million), and endocrine dis- United States) to 71.0 percent (chronic kidney dis- eases (0.3 million) (figure 6.3). ease, Australia) In a sensitivity analysis, we calculated the populations • Respiratory diseases: 3.0 percent (asthma, Myanmar) with CVD-related CHE using only studies that measured to 46.0 percent (chronic obstructive pulmonary dis- CHE defined as health care expenditures in excess of ease, Australia). 40 percent of the household’s capacity to pay. We found 128 Disease Control Priorities: Improving Health and Reducing Poverty Figure 6.2 Catastrophic Health Expenditure Rates, by Source and Disease Category Banthin and Bernard 2006 Che and others 2016 Choi and others 2014 Cancers Choi and others 2015 Davidoff and others 2013 Jan and others 2015 Kavosi and others 2014 Rocha-Garcia and others 2003 Alam and Mahel 2014 Banthin and Bernard 2006 Choi and others 2015 Daivadanam and others 2012 Cardiovascular diseases Heeley and others 2009 Htet, Alam, and Mahal 2015 Huffman and others 2011 Jan and others 2015 Murphy and others 2013 Saito and others 2014 Skroumpelos and others 2014 Sun and others 2015 Zhao and others 2012 Barennes and others 2015 Beauliere and others 2010 Boyer and others 2011 Chronic infectious diseases Castillo-Riquelme, McIntyre, and Barnes 2008 Disease category Che and others 2016 Chen and others 2015 Cleary and others 2012 Ilunga-Ilunga and others 2015 Laokri and others 2014 Moshabela and others 2012 Tran and others 2013 Wingfield and others 2014 Banthin and Bernard 2006 Choi and others 2015 Endocrine diseases Saito and others 2014 Skroumpelos and others 2014 Smith-Spangler, Bhattacharya, and Goldhaber-Fiebert 2012 Kumar and others 2012 Injuries Nguyen and others 2013 Saito and others 2014 Burke and others 2014 infant, and conditions Maternal, childhood Dalaba and others 2015 Loganathan and others 2015 Mental illnesses Patel and others 2007 Banthin and Bernard 2006 diseases Choi and others 2015 Renal Essue and others 2013 Prakongsai and others 2009 Banthin and Bernard 2006 Respiratory diseases Essue and others 2011 Htet, Alam, and Mahal 2015 Skroumpelos and others 2014 0 10 20 30 40 50 60 70 80 90 100 Percentage of households Capacity to pay (>30%) Income (threshold range: 1–40%) Total expenditure (threshold range: 5–15%) Note: For most studies, capacity to pay was defined as in Xu and others (2003). Different data were used to calculate the denominator for each catastrophic health expenditure (CHE) outcome (capacity to pay, income, total expenditure), so standardizing the estimates to a common benchmark was not possible. Each threshold of CHE was used to denote an event of catastrophic significance for the individual patient or household under investigation. Because they are linked through a common conceptual construct and as a way to allow for comparisons of the burden of CHE across the range of diseases, the varying thresholds used in each study are noted here but are treated as essentially the same outcome in this analysis. The CHE rate of 100 percent, reported for renal replacement therapy in Thailand (Prakongsai and others 2009), was excluded from the calculation of the case catastrophe rate for renal diseases. Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 129 no significant difference in case catastrophe rates and the diseases and CVDs. The case catastrophe rate for injuries prevalence of CVD-related CHE for all regions when is lower in low-income countries than in the other the analysis was limited to studies using this common country income groups, despite the high prevalence of definition (table 6.4). injuries. This variation is in contrast to cancers, where Figure 6.4 summarizes the case catastrophe rate rela- the prevalence of disease is relatively lower, so the main tive to the prevalence of each category of condition. driver of the prevalence of cancer-related CHE is the The case catastrophe rate is the population-weighted high case catastrophe rate associated with the treatment average CHE rate for each condition and World Bank and management of these conditions in all national income category. The large estimated burden of CHE income groups. predicted to be associated with renal diseases is explained by the high prevalence of disease and the high case catastrophe rate in populations with prevalent disease; OTHER MEASURES OF HOUSEHOLD-LEVEL renal diseases affect many individuals and are associated ECONOMIC EFFECT OF CHRONIC ILL HEALTH with a high burden because of the type of care required. AND INJURIES IN LMICS Those circumstances also apply to chronic infectious In this section, we report data from a review of the disease-related burden associated with indicators other Figure 6.3 Estimated Population with Catastrophic Health than CHE: impoverishing health expenditure, produc- Expenditures Related to Chronic Ill Health and Injuries, by Disease tivity effects, distressed financing, and treatment discon- Category and Country Income Group tinuation (table 6.1). These indicators supplement and complement the measurement of CHE, because they help describe the effect of ill health on a household’s Upper-middle economic well-being (Moreno-Serra, Millett, and Smith 2011; Ruger 2012), including the way households Income group respond, opportunity costs, and the effect of forgone Lower-middle income. The indicators also tend to focus on the effect of ill health on the poorest of the poor, who may be omitted from other measures, including CHE, because their income is so low. Low We did not estimate the disease-related prevalence associated with each indicator, as done for CHE, given 0 20 40 60 80 100 insufficient data. We thus restrict this discussion to a Number (millions) descriptive analysis. The populations affected by these other measures are not mutually exclusive, so there is Injuries Respiratory diseases significant overlap with the population estimates of Cancers Endocrine diseases Cardiovascular diseases Renal diseases disease-related CHE reported in the previous section. Chronic infectious diseases A systematic review of 47 LMIC studies was conducted to evaluate the household economic effect of NCDs. Table 6.4 Sensitivity Analysis: Comparison of Case Catastrophe Rates and the Projected Population with Cardiovascular Disease–Related Catastrophic Health Expenditure Definition limited to CHE as > 40% of All definitions of CHEa household’s capacity to pay Country Case catastrophe Population with Case catastrophe Population with CVD- income level rate (%) CVD-related CHE rate (%) related CHE Low 8.1 162,163 6.6 131,398 Lower-middle 21.2 22,065,683 21.0 21,829,842 Upper-middle 51.9 78,153,956 46.9 70,665,614 Note: CHE = catastrophic health expenditure; CVD = cardiovascular disease. a. Catastrophic health expenditure was defined as (a) more than 40 percent of household capacity to pay (or nonfood expenditure); (b) more than 10 percent of household expenditure; (c) more than 40 percent of effective income; or (d) more than 30 percent of household income in the published studies. 130 Disease Control Priorities: Improving Health and Reducing Poverty Figure 6.4 Rate of Catastrophic Health Expenditure Relative to Average Prevalence of Each Condition, by Country Income Group a. Low income 100 34 32 90 30 28 80 74.0 26 24 Case catastrophe rate (%) 70 22 Prevalence (%) 60 20 18 50 16 38.0 14 40 35.6 12 30 10 19.8 8 20 6 10 6.5 4 3.0 0.8 2 0 0 Chronic Renal diseases Cardiovascular Endocrine Cancers Respiratory Injuries infectious diseases diseases diseases diseases b. Lower-middle income 100 18 90 16 80 14 70 Case catastrophe rate (%) 12 57.7 Prevalence (%) 60 10 48.0 50 8 40 35.6 31.3 6 30 20.3 4 20 11.3 10 2 3.0 0 0 Chronic Renal diseases Cardiovascular Endocrine Cancers Respiratory Injuries infectious diseases diseases diseases diseases figure continues next page Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 131 Figure 6.4 (continued) c. Upper-middle income 100 18 90 16 80 14 70 Case catastrophe rate (%) 12 Prevalence (%) 60 51.9 10 48.8 50 40.7 8 40 35.6 6 30 4 20 13.7 11.3 10 7.8 2 0 0 Chronic Renal diseases Cardiovascular Endocrine Cancers Respiratory Injuries infectious diseases diseases diseases diseases Note: The blue stacks correspond to the left-hand axis and illustrate the weighted case catastrophe rate (%) for each condition, in each country income category. The blue line corresponds to the right-hand axis and illustrates the average weighted prevalence of each condition, in each country income category. The t-bars illustrate the 95 percent confidence intervals for the weighted case catastrophe rates, in cases where they could be calculated. The methods are described in annex 6B. The systematic Productivity Changes review synthesized evidence from studies in populations Six studies examined the effect of chronic diseases, of patients with NCDs. Of the 47 studies identified, particularly CVDs, on an individual’s capacity to 11 overlapped with the studies identified in the previously maintain usual working status. In some settings, more described systematic search. CHE was the most com- than 80 percent of patients affected by CVDs reported monly measured outcome. However, several studies also having to limit their usual work activities and more incorporated additional indicators of the economic bur- than 60 percent reported having to work less. In addi- den of NCDs on households. tion to the effect on individuals’ productivity, one study conducted across four countries also found that family members had to increase their work activities Impoverishing Health Expenditure or find new work. Whether such changes in productiv- Although impoverishing health expenditure is now rou- ity are different for households that are experiencing tinely investigated in many population-based studies, disease than for those that are not is unclear. For including alongside CHE, few studies have investigated instance, a study conducted in India found that the the disease-related burden. In the review of NCD studies decreases in workforce participation of individuals in LMICs, seven studies measured the rate of NCD-related experiencing angina were not significantly different impoverishing health expenditure. Across the studies, the from those of households not experiencing disease rate of impoverishment was below 15 percent. However, (Alam and Mahal 2014). in a study conducted among Chinese people experienc- By contrast, a study by Zhang, Chongsuvivatwong, ing hypertension, stroke, or coronary heart disease, the and Geater (2006) found that the presence of major incidence of impoverishment hovered around 50 percent chronic illness resulted in a 6.5 percent decrease in the and was not statistically different after implementation probability of remaining in paid work in China. Similarly, of the national health insurance scheme (J. Wang and although the workforce participation rates of cancer- others 2012; figure 6.5). affected households were significantly lower than those 132 Disease Control Priorities: Improving Health and Reducing Poverty Figure 6.5 Proportion of Households with Noncommunicable Diseases Experiencing Impoverishing Health Expenditure, by Disease Category and Country Income Group Van Minh and Tran 2012—Vietnama NCD, general Hamid, Ahsan, and Begum 2014—Bangladesh Bhojani and others 2012—India Alvarez-Hernandez and Rheumatoid arthritis others 2012—Mexico (OECD) Disease category Alvarez-Hernandez and others 2012—Mexico (BFB) Q. Wang and others 2014—Chinab; CHD, Stroke, Hypertension Alam and Mahal 2014—Nepala -angina CVD Alam and Mahal 2014—Sri Lankaa -angina Alam and Mahal 2014—Indiaa -angina Alam and Mahal 2014—Bangladesha -angina 0 10 20 30 40 50 60 70 80 Percentage of households Low Lower-middle Upper-middle Note: BFB = the proportion of households was calculated using the basic food basket method as the threshold; CVD = cardiovascular disease; NCD = noncommunicable disease; OECD = the proportion of households was calculated using the Organisation for Economic Co-operation and Development definition for poverty as the threshold. The t-bars illustrate the 95 percent confidence intervals for the estimate (percentage of households), in cases where they could be calculated. a. Statistically significant difference was found between those with and those without disease. b. No statistically significantly difference was found between those with and those without disease of non-cancer-affected households, when an individual Treatment Discontinuation with cancer was removed from consideration, there were An obvious consequence of unaffordable health care is no discernible differences between households with and treatment attrition or abandonment (Arora, Eden, and without disease. In spite of this finding, although the Pizer 2007; Israels and others 2008; Jan and others 2015). incidence of work-related changes was captured, very For example, in a study of CVD patients in Argentina, few studies valued these changes in monetary terms China, India, and Tanzania, up to 99 percent of households (figure 6.6). reported not taking CVD medications because of the cost (Huffman and others 2011). Similarly, in a study con- ducted among diabetes-affected households across 35 Distressed Financing LMICs, less than 30 percent of individuals were in posses- Six studies attempted to quantify the financing strate- sion of medications in 71 percent of countries (Smith- gies used to pay for health care for NCDs, including Spangler, Bhattacharya, and Goldhaber-Fiebert 2012). This CVDs and cancers. Whereas in one study, almost all outcome was not routinely examined within studies of households relied on savings to finance their health care NCD-related CHE. The relationship between CHE and (Bhojani and others 2013), more commonly, households treatment discontinuation is important for discerning reported selling assets or calling on family and friends. whether trends in health care expenditure, and CHE in This circumstance was especially evident in the most particular, have been affected by the discontinuation or socioeconomically disadvantaged households (Huffman avoidance of necessary health care by households or indi- and others 2011). The few studies that compared house- viduals when faced with unaffordable costs. This is highly holds with and without disease found that these strate- relevant for the treatment of chronic conditions in cases gies were needed more often in households confronted where treatment attrition or abandonment can lead to with chronic disease (Alam and Mahal 2014; figure 6.7). further deterioration of health and higher health care costs. Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 133 Figure 6.6 Proportion of Households with Noncommunicable Diseases Reporting Productivity Effects, by Disease Category and Country Income Group initiated work, family Huffman and others 2011—Tanzania Increased or Huffman and others 2011—India CVD Huffman and others 2011—China Huffman and others 2011—Argentina Huffman and others 2011—Tanzania ceased work, family decreased or Huffman and others 2011—India CVD Huffman and others 2011—China Huffman and others 2011—Argentina Huffman and others 2011—Tanzania activities, patient Limited work Huffman and others 2011—India CVD Huffman and others 2011—China Productivity effect Huffman and others 2011—Argentina Huffman and others 2011—Tanzania Decreased work time, patient Huffman and others 2011—India CVD Huffman and others 2011—China Huffman and others 2011—Argentina Alam and Mahal 2014—Nepal Employment status affected, patient Alam and Mahal 2014—Sri Lanka CVD Alam and Mahal 2014—India Alam and Mahal 2014—Bangladesh Alam and Mahal 2014—Nepal owing to illness Unemployment Alam and Mahal 2014—Sri Lanka CVD Alam and Mahal 2014—India Alam and Mahal 2014—Bangladesh 0 10 20 30 40 50 60 70 80 90 100 Percentage of households Low Lower-middle Upper-middle High Note: CVD = cardiovascular disease. The t-bars illustrate the 95 percent confidence intervals for the estimate (percentage of households), in cases where they could be calculated. DISCUSSION others 2016; Lange and others 2004; W. Li and others 2016), Patients with chronic conditions and injuries in LMICs and the costs associated with treating renal disease are face a substantial economic burden as a result of paying high, including the costs of medicines and dialysis for health care. Chronic conditions such as renal, cardio- (Teerawattananon and others 2016; White and others 2008). vascular, and endocrine diseases account for the largest The high costs of treatment for different conditions populations with CHE. However, in low-income coun- are due to factors such as place of treatment and out-of- tries individuals with chronic infectious diseases such pocket costs for different types of treatment. For exam- as HIV/AIDS, tuberculosis, and malaria are the largest ple, out-of-pocket costs associated with hospitalization populations with CHE. for an acute event may be high, as for conditions such The factors underlying these estimates are both preva- as stroke in China (Heeley and others 2009) and acute lence of disease and rates of CHE associated with each myocardial infarction in both China and India (Jan and category of conditions. For example, the comparatively others 2016). However, paying for treatment that is higher burden associated with renal conditions in all set- required on an ongoing basis can also lead to a high tings is likely explained by the fact that renal disease is an cost burden, whether the payments are marginal, such end product of other NCDs, notably diabetes and CVDs. as paying for medicines or, at a more extreme end, the These precursory NCDs are undertreated (Khatib and cost of regular dialysis for managing chronic kidney 134 Disease Control Priorities: Improving Health and Reducing Poverty Figure 6.7 Proportion of Households with Noncommunicable Disease Using Distressed Financing Strategies, by Disease Category and Country Income Group savings Engelgau, Karan, and Mahal 2012—lndia NCD Use Bhojani and others 2012—lndia Engelgau, Karan, and Mahal, 2012—lndia NCD Bhojani and others 2012—lndia Huffman and others 2011—Tanzania assets Sell Huffman and others 2011—lndia CVD Huffman and others 2011—China Huffman and others 2011—Argentina Huffman and others 2011—Tanzania Borrow money money lender from bank or Distressed financing strategy Huffman and others 2011—lndia CVD Huffman and others 2011—China Huffman and others 2011—Argentina Huffman and others 2011—Tanzania Borrow money from family or Huffman and others 2011—lndia friends CVD Huffman and others 2011—China Huffman and others 2011—Argentina Mahal and others 2013—India Cancer Mahal and others 2013—lndia, inpatient carea Borrow, sell assets, Engelgau, Karan, and Mahal 2012—lndia general NCD, Bhojani and others 2012—lndia or both Alam and Mahal 2014—Nepalb Alam and Mahal 2014—Sri Lankaa CVD Alam and Mahal 2014—lndiaa Alam and Mahal 2014—Bangladesha 0 20 40 60 80 100 Percentage of households Low Lower-middle Upper-middle High Note: CVD = cardiovascular disease; NCD = noncommunicable disease. The t-bars illustrate the 95 percent confidence intervals for the estimate (percentage of households), in cases where they could be calculated. a. Statistically significant difference was found between those with and those without disease. b. No statistically significantly difference was found between those with and those without disease. disease (Prakongsai and others 2009; Ramachandran from traffic accidents are lower (Dalal and others 2013), and Jha 2013). so the risk of incurring CHE is lower. Endocrine diseases and injuries in low-income set- HIV/AIDS, like other long-term illnesses, is associated tings both have relatively high prevalence but compara- with a relatively higher rate of CHE, likely because of the tively lower rates of CHE. For injuries, although the costs ongoing costs of medicines in settings where access to associated with treating an acute episode in either a free antiretroviral treatment is suboptimal. For cancers, hospital or a community health setting may be high, the prevalence of disease is relatively lower, both overall ongoing health care costs after recovery may be minimal. and in each country income category, but the cost burden However, if the severity of the injury affects the is comparatively high because of treatment costs associ- individual’s ability to continue in paid work, the house- ated with chemotherapy, radiation, and surgery (Aggarwal hold may still experience negative economic conse- and Sullivan 2014; Pramesh and others 2014). quences from this loss of income, which is not captured In the context of an increasing prevalence of multiple in the CHE measures. In addition, in low-income coun- morbidity, estimated at 7.8 percent in LMICs (Afshar and tries, survival rates from injuries such as those resulting others 2015), such high levels of expenditure associated Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 135 with one condition would potentially compromise an mitigating variations in health care practice among set- individual’s ability to afford the range of care that is tings, the disease-specific treatment options that are avail- required when faced with multiple morbidity. This cir- able and that constitute best practice may vary among cumstance could lead to trade-offs, including a prioriti- (and within) countries. These differences ultimately zation of treatment for acute conditions over chronic influence the generalizability and interpretation of the care, especially in cases where conditions are asymptotic. individual estimates. There is substantial variation in the cost burden and risk of CHE associated with chronic conditions and injuries in cases where expenditures are often repeated Differences in Measurement of CHE and continuous. Curbing the rates of CHE will require The studies consulted measured CHE using different targeting financial risk protection to cover elements of definitions, thresholds, and categories of expenditure treatment for conditions with high risk of CHE and high included as out-of-pocket costs, different data sources, prevalence, such as renal diseases and CVDs. In low- and different recall periods, which potentially intro- income settings, additional protection might be required duced measurement error. However, the findings from a for major infectious diseases. Identifying the elements of sensitivity analysis indicated that our results were robust treatment that impose the greatest cost burden, which despite the combining of varied estimates. may be common across various disease categories, will help achieve the greatest gains in mitigating the risk of CHE at a population level. Differences in Quality and Breadth of Evidence Global work, especially from the WHO, has high- Given the lack of comprehensive evidence on the level lighted the significant household economic burden that is of CHE in different populations, estimates for one set- associated with accessing and using health care services, ting sometimes were based on data extrapolated from particularly in LMICs. In addition, it has been a driving studies conducted in other settings. In cases where data force in efforts to implement effective financial protection on the prevalence of CHE for any particular country mechanisms to mitigate this burden. Comparability of income category were missing, we applied a conserva- our results with WHO global estimates of the prevalence tive strategy of using the estimate from the next-highest of CHE depends on the relative distribution of chronic dis- income category. In addition, the results describe the eases, injuries, and comorbidity within the population-level relative burden of disease-related CHE between condi- data used to generate the estimates. The rates of CHE are tions and country-income categories but not the poten- much higher when measured in the population with dis- tial distributional burden within the populations in ease than in the population as a whole. Our analysis, each category. which uses samples of persons with disease, shows that Much of the evidence on the disease-related burden many more people in LMICs and globally are at risk of of CHE is from cross-sectional studies that lack a control CHE than previously estimated (Xu and others 2007). group and cannot capture repeat expenditures, so they Furthermore, the estimates reported here for each cate- are limited in their ability to attribute CHE directly to gory of conditions are not cumulative, given the high the disease or injuries. In addition, the smaller, clinic- prevalence of multiple morbidity overall and the overlap- based studies may not be fully representative of the ping of comorbid conditions between disease categories population with disease in each country. Despite their included in this analysis. limitations, these studies are the sole source of evidence and provide a starting point from which to investigate differences in the burden of CHE among different LIMITATIONS OF THE RESEARCH categories of chronic conditions. The evidence also tends to come from smaller studies Comparability among Countries and Health Care of cohorts recruited from hospitals or health care facili- System Contexts ties, which can lead to higher estimates of health care The economic burden associated with health care expenses expenditure than those based on community or house- is context specific. Differences in the financing and service hold samples (Lavado, Brooks, and Hanlon 2013; Raban, provision arrangements among health care systems in Dandona, and Dandona 2013). Hospital expenses may each country may influence the populations and the explain some of this difference, because the samples in breadth of services covered, the mix of private and pub- hospitals are a biased (nonrandom) sample of the licly funded services, and the out-of-pocket costs associ- population. Moreover, household samples were asked to ated with health care use. In addition, despite advances in report costs associated with previous hospitalizations, evidence-based medicine and its contribution toward which suggests that recall bias may be stronger in 136 Disease Control Priorities: Improving Health and Reducing Poverty the community-based studies than in the clinic- or illness should not be overlooked when developing strat- hospital-based studies. egies to improve financial risk protection. This study has important implications for the design of benefit packages. The conventional approach has POLICY DIRECTIONS FOR IMPROVING been to place cost-effectiveness or best buys as the over- riding consideration in designing benefit packages FINANCIAL PROTECTION (Chisholm and others 2012; Evans and Etienne 2010; As the epidemiological transition progresses over the WHO and World Economic Forum 2011). The rationale next few decades, the double burden of infectious dis- for this approach is strong: given severe resource con- eases and NCDs will continue to challenge health care straints, priority needs to be given to funding programs systems in LMICs, which will be confronted with caring that deliver the greatest health outcomes for the dollar. for older and more costly populations. Catastrophic and However, although this approach promotes the objec- impoverishing health expenditure will increase globally tive of health maximization, it does not directly address unless action is taken to offer deeper packages of finan- the problem that such benefit packages are designed to cial protection that include the treatment of chronic address—that is, financial protection. This study pro- disease and injury. In formulating measures to address vides evidence to guide policy makers in the design of this issue, policy makers focus on universal health cover- benefit packages and entitlements. It demonstrates the age, which aims to provide population-wide protection need to prioritize the relative financial burden across through various social health protection mechanisms. disease areas and in different settings to ensure cover- However, given severe resource constraints, such pro- age of the disease-specific health care and health- grams are often able to provide only limited protection related services that are most associated with of certain diseases and treatments; achieving compre- catastrophic and impoverishing health expenditure hensive financial protection will inevitably be a long- (Jamison and others 2013). term goal. The design of the package of entitlements and This research also highlights the need for an ongoing covered services should take into account both the pop- focus on and investment in prevention. The most effec- ulations most at risk and the diseases and conditions that tive way to reduce disease-related CHE is to prevent such drive catastrophic and impoverishing health expenditure. conditions. This prevention is particularly critical in Country examples exist of how to implement this LMICs, where the double burden of infectious diseases through progressive universalism (Gwatkin and Ergo and NCDs continues to place a major strain on 2011; Jamison and others 2013); one example, about health care systems. Evidence from the extended cost- which much has been written, is the catastrophic expen- effectiveness literature has demonstrated the gains to be diture fund of Mexico’s Seguro Popular (Knaul, Arreola- made in strengthening financial protection through Ornelas, and Méndez-Carniado 2016). investment in prevention. Public financing of programs In this study, we identify significant variation in the such as vaccination for human papillomavirus infection household economic burden by condition. The high and management of risk factors, such as obesity for burden observed for many chronic conditions such as diabetes and hypertension for CVD, have been shown to renal diseases indicates potential areas where targeted have the potential to curb catastrophic and impoverish- programs could be developed to address the populations ing health expenditure significantly, thereby enhancing currently experiencing the greatest financial burden. financial protection across populations (Levin and These results suggest that universal health coverage others 2015; Verguet and others 2015). should be developed as part of a multipronged strategy Addressing the factors that lead to and perpetuate that addresses not only system-level drivers of the entrenched poverty will also produce the greatest gains household economic burden but also disease-specific in mitigating the economic burden of chronic ill health drivers. For individual diseases, basic packages should experienced by households. Rates of catastrophic and include specific interventions that are shown to be impoverishing health expenditure should decline over effective—for example, low-cost dialysis (Liyanage and time as universal health coverage is implemented along- others 2015) and polypill treatments for CVD (Webster side other poverty reduction strategies, including efforts and Rodgers 2016) as well as disease management and to meet the Sustainable Development Goals. These prevention strategies. efforts should reduce the burden of disease overall and The research on disease-related CHE tends to be clus- improve the capacity of households to access and use tered in areas that do not necessarily reflect the diseases required health care services. In monitoring progress, that have the greatest burden and largest household including the effect of efforts to reach the Sustainable economic effect. Under-researched areas such as mental Development Goals, priority should be given to Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 137 evaluating changes in financial protection among the CONCLUSIONS population as a whole as well as within subgroups most at risk of catastrophic and impoverishing health In this chapter, we estimate the economic burden associ- expenditure. ated with seven categories of chronic conditions as well as injuries. We find that most CHE is due to renal, cardiovas- cular, and chronic infectious diseases and that the global FUTURE RESEARCH DIRECTIONS burden of CHE is much higher than previously estimated. More prospective longitudinal studies are needed to Meeting the global commitment to enhance financial examine the extent to which households can recover protection of populations, including the World Bank’s from the burden of catastrophic and impoverishing goal of eliminating impoverishing health expenditure health expenditure. These types of studies, although few, by 2030, requires a concerted effort to address the main have helped identify the determinants of recovery from drivers of CHE in all settings. In designing financial an illness shock as well as factors that potentially protection programs, policy makers need to give prior- enhance resilience to such shocks (Essue and others ity to covering populations and conditions associated 2012; Heeley and others 2009; Jan and others 2015; Jan with the greatest economic burden. Furthermore, and others 2016; Kimman and others 2015). Prospective needed health care services still remain out of reach for studies will also help distinguish between the effect and millions with disease who live in poverty. Strategies to consequences of one shock versus cumulative expendi- enhance financial protection need to be implemented ture as well as the potential for health interventions to alongside broader poverty alleviation efforts, which improve household economic circumstances (Essue and collectively will generate the greatest gains in mitigating others 2014; Kuper and others 2010). the household-level economic burden of chronic ill Longitudinal research is also needed to monitor health globally. progress in mitigating CHE and impoverishing health expenditure. Monitoring progress using different cross- sections of population data over time cannot account ANNEXES well for the fact that new households may encounter The annexes to this chapter are as follows. They are avail- CHE, while others may become nonspenders because able at http://www.dcp-3.org/DCP. they are no longer able to pay for care. Therefore, declines over time do not necessarily mean that health • Annex 6A. Description of Data Sources and Search care has become more affordable for all. Strategy. Furthermore, the long-term effect on households of • Annex 6B. Search Strategy for Prospectively Designed impoverishing health expenditure, distressed financing Studies of Household Economic Effect of Chronic arrangements, changes in workforce participation, and Disease. treatment discontinuation are poorly understood. More multidimensional assessments of the household eco- nomic burden of chronic ill health are needed using NOTE routinely measured indicators along with CHE and World Bank Income Classifications as of July 2014 are as impoverishing health expenditure (Moreno-Serra, Millet, follows, based on estimates of gross national income and Smith 2011; Ruger 2012). Such studies would sup- (GNI) per capita for 2013: port the design of financial protection programs and improve the targeting of interventions, because these • Low-income countries (LICs) = US$1,045 or less indicators provide greater insights into the effect of illness • Middle-income countries (MICs) are subdivided: and health care expenditure on the household economy. (a) lower-middle-income = US$1,046 to US$4,125 More research is needed to understand the link back (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 to health. Although the effect of the social determinants • High-income countries (HICs) = US$12,746 or more. of health is well understood (Friel and Marmot 2011), longer-term cohort studies are needed to assess how these economic consequences perpetuate the cycle of ACKNOWLEDGMENTS chronic ill health and social disadvantage (van Doorslaer The authors are grateful to Ke Xu for the input received and others 2006). Evidence on the link between the eco- on the structure and scope of this chapter early on as nomic burden of disease, health outcomes, and social well as to the teams in Mongolia, the Philippines, and disadvantage would strengthen the economic case for Vietnam for their contribution to the case studies pre- improving access to affordable care. sented in box 6.2. Financial protection information 138 Disease Control Priorities: Improving Health and Reducing Poverty from Mongolia was taken from a WHO-commissioned Payments on Chronic Conditions Impoverish Urban Poor report on catastrophic health payments and benefit in Bangalore, India.” BMC Public Health 12 (990): 1–13. incidence of government expenditure in Mongolia Boyer, S., M. Abu-Zaineh, J. Blanche, S. Loubiere, R. C. Bonono, conducted by Tsolmongerel Tsilaajav, Oyungerel and others. 2011. “Does HIV Services Decentralization Protect Against the Risk of Catastrophic Health Nanzad, and Enkhbaatar Ichinnorov, under the coordi- Expenditures? Some Lessons from Cameroon.” Health nation of Erdenechimeg Enkhee of the Office of the Services Research 46 (6): 2029–56. WHO Representative in Mongolia. We also acknowl- Burke, R. M., E. R. Smith, R. M. Dahl, P. A. Rebolledo, edge the contribution of Melanie Bisnauth, who pro- C. Calderón Mdel, and others. 2014. “The Economic vided research assistance to support the analysis in Burden of Pediatric Gastroenteritis to Bolivian Families: this chapter. A Cross-Sectional Study of Correlates of Catastrophic Cost and Overall Cost Burden.” BMC Public Health 14 (June): 642. Cameron, A., M. Ewen, D. Ross-Degnan, D. Ball, and R. Laing. REFERENCES 2009. “Medicine Prices, Availability, and Affordability in 36 Afshar, S., P. J. Roderick, P. Kowal, B. D. Dimitrov, and Developing and Middle-Income Countries: A Secondary A. G. Hill. 2015. “Multimorbidity and the Inequalities of Analysis.” The Lancet 373 (9659): 240–49. Global Ageing: A Cross-Sectional Study of 28 Countries Castillo-Riquelme, M., D. McIntyre, and K. Barnes. 2008. Using the World Health Surveys.” BMC Public Health “Household Burden of Malaria in South Africa and 5 (776): 1–10. doi:10.1186/s12889-015-2008-7. Mozambique: Is there a Catastrophic Impact?” Tropical Aggarwal, A., and R. Sullivan. 2014. “Achieving Value in Cancer Medicine and International Health 13 (1): 108–22. Care: The Case of Low- and Middle-Income Countries.” Che, Y. H., V. Chongsuvivatwong, L. Li, H. Sriplung, Y. Y. Wang, American Journal of Managed Care 20 (12): 292–94. and others. 2016. “Financial Burden on the Families of Alam, K., and A. Mahal. 2014. “The Economic Burden of Patients with Hepatitis B Virus–Related Liver Diseases and Angina on Households in South Asia.” BMC Public Health the Role of Public Health Insurance in Yunnan Province of 14 (February): 179. China.” Public Health 130 (January): 13–20. Alvarez-Hernandez, E., I. Pelaez-Ballestas, A. Boonen, Chen, S., H. Zhang, Y. Pan, Q. Long, L. Xiang, and others. 2015. J. Vázquez-Mellado, A. Hernández-Garduño, and others. “Are Free Anti-Tuberculosis Drugs Enough? An Empirical 2012. “Catastrophic Health Expenses and Impoverishment Study from Three Cities in China.” Infectious Diseases of of Households of Patients with Rheumatoid Arthritis.” Poverty 4 (October): 47. Reumatología Clínica 8 (4): 168–73. Chisholm, D., R. Baltussen, D. B. Evans, G. Ginsberg, J. A. Lauer, Arora, R. S., T. Eden, and B. Pizer. 2007. “The Problem of and others. 2012. “What Are the Priorities for Prevention Treatment Abandonment in Children from Developing and Control of Non-Communicable Diseases and Injuries Countries with Cancer.” Pediatric Blood and Cancer 49 (7): in Sub-Saharan Africa and South East Asia?” British Medical 941–46. Journal 344 (March): e586. Banthin, J. S., and D. M. Bernard. 2006. “Changes in Financial Choi, J.-W., K. H. Cho, Y. Choi, K. T. Han, J. A. Kwon, and Burdens for Health Care: National Estimates for the others. 2014. “Changes in Economic Status of Households Population Younger than 65 Years, 1996 to 2003.” Journal of Associated with Catastrophic Health Expenditures for the American Medical Association 296 (22): 2712–19. Cancer in South Korea.” Asian Pacific Journal of Cancer Barennes, H., A. Frichittavong, M. Gripenberg, and P. Koffi. Prevention 15 (6): 2713–17. 2015. “Evidence of High Out-of-Pocket Spending for HIV Choi, J.-W., J.-W. Choi, J.-H. Kim, K.-B. Yoo, and E.-C. Park. 2015. Care Leading to Catastrophic Expenditure for Affected “Association between Chronic Disease and Catastrophic Patients in Lao People’s Democratic Republic.” PLoS One Health Expenditure in Korea.” BMC Health Services Research 10 (9): e0136664. 15 (1): 464–78. Beaglehole, R., R. Bonita, R. Horton, C. Adams, G. Alleyne, and Cleary, S. M., S. Birch, M. Moshabela, and H. Schneider. 2012. others. 2011. “Priority Actions for the Non-Communicable “Unequal Access to ART: Exploratory Results from Rural Disease Crisis.” The Lancet 377 (9775): 1438–47. and Urban Case Studies of ART Use.” Sexually Transmitted Beauliere, A., S. Toure, P. K. Alexandre, K. Koné, A. Pouhé, and Infections 88 (2): 141–46. others. 2010. “The Financial Burden of Morbidity in HIV- Daivadanam, M., K. R. Thankappan, P. S. Sarma, and Infected Adults on Antiretroviral Therapy in Côte d’Ivoire.” S. Harikrishnan. 2012. “Catastrophic Health Expenditure PLoS One 5 (6): e11213. and Coping Strategies Associated with Acute Coronary Bhojani, U., T. S. Beerenahalli, R. Devadasan, C. M. Munegowda, Syndrome in Kerala, India.” Indian Journal of Medical N. Devadasan, and others. 2013. “No Longer Diseases of the Research 136 (4): 585–92. Wealthy: Prevalence and Health-Seeking for Self-Reported Dalaba, M. A., P. Akweongo, R. A. Aborigo, H. P. Saronga, Chronic Conditions among Urban Poor in Southern India.” J. Williams, and others. 2015. “Cost to Households in BMC Health Services Research 13 (August): 1–10. Treating Maternal Complications in Northern Ghana: Bhojani, U., B. Thriveni, R. Devadasan, C. M. Munegowda, A Cross-Sectional Study.” BMC Health Services Research N. Devadasan, and others. 2012. “Out-of-Pocket Healthcare 15 (January): 34. Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 139 Dalal, K., Z. Lin, M. Gifford, and L. Svanström. 2013. Goeppel, C., P. Frenz, L. Grabenhenrich, T. Keil, and “Economics of Global Burden of Road Traffic Injuries P. Tinnemann. 2016. “Assessment of Universal Health and Their Relationship with Health System Variables.” Coverage for Adults Aged 50 Years or Older with Chronic International Journal of Preventive Medicine 4 (12): 1442–50. Illness in Six Middle-Income Countries.” Bulletin of the Davidoff, A. J., M. Erten, T. Shaffer, J. S. Shoemaker, World Health Organization 94 (4): 276–85C. I. H. Zuckerman, and others. 2013. “Out-of-Pocket Health Gwatkin, D. R., and A. Ergo. 2011. “Universal Health Coverage: Care Expenditure Burden for Medicare Beneficiaries with Friend or Foe of Health Equity?” The Lancet 377 (9784): Cancer.” Cancer 119 (6): 1257–65. 2160–61. Engelgau, M. M., A. Karan, and A. Mahal. 2012. “The Economic Hamid, S. A., S. M. Ahsan, and A. Begum. 2014. “Disease-Specific Impact of Non-Communicable Diseases on Households in Impoverishment Impact of Out-of-Pocket Payments for India.” Globalization and Health 8 (1): 9. Health Care: Evidence from Rural Bangladesh.” Applied Essue, B. M., M. Hackett, Q. Li, N. Glozier, R. Lindley, Health Economics and Health Policy 12 (4): 421–33. and others. 2012. “How Are Household Economic Heeley, E., C. A. Anderson, Y. Huang, S. Jan, Y. Li, and others. Circumstances Affected after a Stroke? The Psychosocial 2009. “Role of Health Insurance in Averting Economic Outcomes In StrokE (POISE) Study.” International Journal Hardship in Families after Acute Stroke in China.” Stroke of Stroke 7 (September): 36. 40 (6): 2149–56. Essue, B. M., P. Kelly, M. Roberts, S. Leeder, and S. Jan. 2011. Htet, S., K. Alam, and A. Mahal. 2015. “Economic Burden of “We Can’t Afford My Chronic Illness! The Out-of-Pocket Chronic Conditions among Households in Myanmar: The Burden Associated with Managing Chronic Obstructive Case of Angina and Asthma.” Health Policy and Planning Pulmonary Disease in Western Sydney, Australia.” Journal of 30 (9): 1173–83. Health Services Research and Policy 16 (4): 226–31. Huffman, M. D., K. D. Rao, A. Pichon-Riviere, D. Zhao, Essue, B. M., M. Kimman, N. Svenstrup, K. L. Kjoege, T. L. Laba, S. Harikrishnan, and others. 2011. “A Cross-Sectional and others. 2015. “The Effectiveness of Interventions to Study of the Microeconomic Impact of Cardiovascular Reduce the Household Economic Burden of Illness and Disease Hospitalization in Four Low- and Middle-Income Injury: A Systematic Review.” Bulletin of the World Health Countries.” PLoS One 6 (6): e20821. Organization 93 (2): 102–12B. Human Rights Watch. 2006. “A High Price to Pay: Detention Essue, B. M., Q. Li, M. L. Hackett, L. Keay, B. Lezzi, and others. of Poor Patients in Burundian Hospitals.” Human Rights 2014. “A Multicenter Prospective Cohort Study of Quality Watch, September 7. of Life and Economic Outcomes after Cataract Surgery in Ilunga-Ilunga, F., A. Leveque, S. Laokri, and M. Dramaix. Vietnam: The VISIONARY Study.” Ophthalmology 121 (11): 2015. “Incidence of Catastrophic Health Expenditures 2138–46. for Households: An Example of Medical Attention for Essue, B. M., G. Wong, J. Chapman, Q. Li, and S. Jan. 2013. the Treatment of Severe Childhood Malaria in Kinshasa “How Are Patients Managing with the Costs of Care for Reference Hospitals, Democratic Republic of Congo.” Chronic Kidney Disease in Australia? A Cross-Sectional Journal of Infection and Public Health 8 (2): 136–44. Study.” BMC Nephrology 14 (January): 5. Israels, T., C. Chirambo, H. Caron, J. de Kraker, E. Molyneux, Evans, D. B., and C. Etienne. 2010. “Health Systems Financing and others. 2008. “The Guardians’ Perspective on Paediatric and the Path to Universal Coverage.” Bulletin of the World Cancer Treatment in Malawi and Factors Affecting Health Organization 88 (6): 402. Adherence.” Pediatric Blood and Cancer 51 (5): 639–42. Frenk, J., J. L. Bobadilla, J. Sepuúlveda, and M. Cervantes. 1989. Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, “Health Transition in Middle-Income Countries: New S. Berkley, and others. 2013. “Global Health 2035: A World Challenges for Health Care.” Health Policy and Planning Converging within a Generation.” The Lancet 382 (9908): 4 (1): 29–39. 1898–955. Friel, S., and M. G. Marmot. 2011. “Action on the Social Jan, S., M. Kimman, S. A. E. Peters, and M. Woodward. Determinants of Health and Health Inequities Goes 2015. “Financial Catastrophe, Treatment Discontinuation, Global.” Annual Review of Public Health 32 (April): 225–36. and Death Associated with Surgically Operable Cancer in Gelders, S., M. Ewen, N. Noguchi, and R. Laing. 2006. Price, South-East Asia: Results from the ACTION Study.” Surgery Availability, and Affordability: An International Comparison 157 (6): 971–82. of Chronic Disease Medicines. Cairo: World Health Jan, S., S. W. L. Lee, J. P. S. Sawhney, T. K. Ong, C. T. Chin, and Organization Regional Office for the Eastern Mediterranean others. 2016. “Catastrophic Health Expenditure on Acute and Health Action International. Coronary Events in Asia: A Prospective Study.” Bulletin of Gertler, P., and J. Gruber. 2002. “Insuring Consumption against the World Health Organization 94 (3): 193–200. Illness.” American Economic Review 92 (March): 51–70. Kankeu, H. T., P. Saksena, K. Xu, and D. B. Evans. 2013. “The Global Burden of Disease Study 2013 Collaborators. 2015. Financial Burden from Non-Communicable Diseases in “Global, Regional, and National Incidence, Prevalence, Low- and Middle-Income Countries: A Literature Review.” and Years Lived with Disability for 301 Acute and Chronic Health Research Policy and Systems 11 (August): 1–12. Diseases and Injuries in 188 Countries, 1990–2013: Kavosi, Z., H. Delavari, A. Keshtkaran, and F. Setoudehzadeh. A Systematic Analysis for the Global Burden of Disease 2014. “Catastrophic Health Expenditures and Coping Study 2013.” The Lancet 386 (9995): 743–800. Strategies in Households with Cancer Patients in 140 Disease Control Priorities: Improving Health and Reducing Poverty Shiraz Namazi Hospital.” Middle East Journal of Cancer Levin, C. E., M. Sharma, Z. Olson, S. Verguet, J. F. Shi, and 5 (1): 13–22. others. 2015. “An Extended Cost-Effectiveness Analysis of Khatib, R., M. McKee, H. Shannon, C. Chow, S. Rangarajan, Publicly Financed HPV Vaccination to Prevent Cervical and others. 2016. “Availability and Affordability of Cancer in China.” Vaccine 33 (24): 2830–41. Cardiovascular Disease Medicines and their Effect on Li, W., H. Gu, K. K. Teo, J. Bo, Y. Wang, and others. 2016. Use in High-Income, Middle-Income, and Low-Income “Hypertension Prevalence, Awareness, Treatment, and Countries: An Analysis of the PURE Study Data.” The Lancet Control in 115 Rural and Urban Communities Involving 387 (10013): 61–69. 47, 000 People from China.” Journal of Hypertension 34 (1): Kim, J. Y. 2014. “Speech by World Bank Group President Jim 39–46. Yong Kim on Universal Health Coverage in Emerging Li, Y., Q. Wu, L. Xu, D. Legge, Y. Hao, and others. 2012. Economies.” Conference on Universal Health Coverage in “Factors Affecting Catastrophic Health Expenditure and Emerging Economies, Center for Strategic and International Impoverishment from Medical Expenses in China: Policy Studies, Washington, DC, January 14. Implications of Universal Health Insurance.” Bulletin of the Kimman, M., S. Jan, C. H. Yip, H. Thabrany, S. A. Peters, World Health Organization 90 (9): 664–71. and others. 2015. “Catastrophic Health Expenditure and Liyanage, T., T. Ninomiya, V. Jha, B. Neal, H. M. Patrice, and 12-Month Mortality Associated with Cancer in Southeast others. 2015. “Worldwide Access to Treatment for End- Asia: Results from a Longitudinal Study in Eight Countries.” Stage Kidney Disease: A Systematic Review.” The Lancet 385 BMC Medicine 13 (August): 190. (9981): 1975–82. Knaul, F. M., H. Arreola-Ornelas, and O. Méndez-Carniado. Loganathan, T., W. S. Lee, K. F. Lee, M. Jit, and C. W. Ng. 2016. “Protección financiera en salud: Actualizaciones 2015. “Household Catastrophic Healthcare Expenditure para México a 2014.” Salud Pública de México 58 (3): and Impoverishment Due to Rotavirus Gastroenteritis 341–50. Requiring Hospitalization in Malaysia.” PLoS One 10 (5): Knaul, F. M., H. Arreola-Ornelas, O. Méndez-Carniado, e0125878. C. Bryson-Cahn, J. Barofsky, and others. 2006. “Evidence Mahal, A., A. Karan, V. Y. Fan, and M. Engelgau. 2013. Is Good for Your Health System: Policy Reform to Remedy “The Economic Burden of Cancers on Indian Households.” Catastrophic and Impoverishing Health Spending in PLoS One 8 (8): e71853. Mexico.” The Lancet 368 (9549): 1828–41. McIntyre, D., M. Thiede, G. Dahlgren, and M. Whitehead. 2006. Knaul, F. M., R. Wong, and H. Arreola-Ornelas, eds. 2012. “What Are the Economic Consequences for Households of Household Spending and Impoverishment. Vol. 1, Financing Illness and of Paying for Health Care in Low- and Middle- Health in Latin America. Cambridge, MA: Harvard Income Country Contexts?” Social Science and Medicine Global Equity Initiative, Mexican Health Foundation, and 62 (4): 858–65. International Development Research Centre. Minh, H. V., and N. T. P. Phuong. 2016. “Burden of Household Kruk, M. E., E. Goldmann, and S. Galea. 2009. “Borrowing and Out-of-Pocket Health Expenditures in Vietnam.” Working Selling to Pay for Health Care in Low- and Middle-Income Paper, Hanoi School of Public Health, Hanoi. Countries.” Health Affairs 28 (4): 1056–66. Moreno-Serra, R., C. Millett, and P. C. Smith. 2011. “Towards Kumar, G. A., T. R. Dilip, L. Dandona, and R. Dandona. 2012. Improved Measurement of Financial Protection in Health.” “Burden of Out-of-Pocket Expenditure for Road Traffic PLoS Medicine 8 (9): e1001087. Injuries in Urban India.” BMC Health Services Research Moshabela, M., H. Schneider, S. P. Silal, and S. M. Cleary. 2012. 12 (August): 285. “Factors Associated with Patterns of Plural Healthcare Kuper, H., S. Polack, W. Mathenge, C. Eusebio, Z. Wadud, Utilization among Patients Taking Antiretroviral Therapy and others. 2010. “Does Cataract Surgery Alleviate in Rural and Urban South Africa: A Cross-Sectional Study.” Poverty? Evidence from a Multi-Centre Intervention Study BMC Health Services Research 12 (July): 182. Conducted in Kenya, the Philippines, and Bangladesh.” Murphy, A., A. Mahal, E. Richardson, and A. E. Moran. 2013. PLoS One 5 (11): e15431. “The Economic Burden of Chronic Disease Care Faced by Lange, S., C. Diehm, H. Darius, R. Haberl, J. R. Allenberg, Households in Ukraine: A Cross-Sectional Matching Study and others. 2004. “High Prevalence of Peripheral Arterial of Angina Patients.” International Journal for Equity in Disease and Low Treatment Rates in Elderly Primary Health 12 (May): 38. Care Patients with Diabetes.” Experimental and Clinical Nguyen, H., R. Ivers, S. Jan, A. Martiniuk, and C. Pham. 2013. Endocrinology and Diabetes 112 (10): 566–73. “Catastrophic Household Costs Due to Injury in Vietnam.” Laokri, S., M. Dramaix-Wilmet, F. Kassa, S. Anagonou, and Injury 44 (5): 684–90. B. Dujardin. 2014. “Assessing the Economic Burden of O’Donnell, O., E. van Doorslaer, A. Wagstaff, and M. Lindelow, Illness for Tuberculosis Patients in Benin: Determinants eds. 2007. Analyzing Health Equity Using Household Survey and Consequences of Catastrophic Health Expenditures Data: A Guide to Techniques and Their Implementation. and Inequities.” Tropical Medicine and International Health Washington, DC: World Bank Group. 19 (10): 1249–58. Okoli, C. I., and S. M. Cleary. 2011. “Socioeconomic Status Lavado, R. F., B. P. Brooks, and M. Hanlon. 2013. “Estimating and Barriers to the Use of Free Antiretroviral Treatment for Health Expenditure Shares from Household Surveys.” HIV/AIDS in Enugu State, South-Eastern Nigeria.” African Bulletin of the World Health Organization 91 (7): 519–24. Journal of AIDS Research 10 (2): 149–55. Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 141 Patel, V., D. Chisholm, B. R. Kirkwood, and D. Mabey. 2007. Shrime, M. G., A. J. Dare, B. C. Alkire, K. O’Neill, and “Prioritizing Health Problems in Women in Developing J. G. Meara. 2015. “Catastrophic Expenditure to Pay for Countries: Comparing the Financial Burden of Reproductive Surgery Worldwide: A Modelling Study.” The Lancet Global Tract Infections, Anaemia, and Depressive Disorders in Health 3 (Suppl 2): S38–44. a Community Survey in India.” Tropical Medicine and Skroumpelos, A., E. Pavi, S. Pasaloglou, and J. Kyriopoulos. International Health 12 (1): 130–39. 2014. “Catastrophic Health Expenditures and Chronic Peters, D. H., A. Garg, G. Bloom, D. G. Walker, W. R. Brieger, Condition Patients in Greece.” Value in Health 17 (7): and others. 2008. “Poverty and Access to Health Care in A501–02. Developing Countries.” Annals of the New York Academy of Smith-Spangler, C. M., J. Bhattacharya, and J. D. Goldhaber- Sciences 1136 (October): 161–71. Fiebert. 2012. “Diabetes, Its Treatment, and Catastrophic Prakongsai, P., N. Palmer, P. Uay-Trakul, V. Tangcharoensathien, Medical Spending in 35 Developing Countries.” Diabetes and A. Mills. 2009. “The Implications of Benefit Package Care 35 (2): 319–26. Design: The Impact on Poor Thai Households of Excluding Sun, J., T. Liabsuetrakul, Y. Fan, and E. McNeil. 2015. “Protecting Renal Replacement Therapy.” Journal of International Patients with Cardiovascular Diseases from Catastrophic Development 21 (2): 291–308. Health Expenditure and Impoverishment by Health Finance Pramesh, C. S., R. A. Badwe, B. B. Borthakur, M. Chandra, Reform.” Tropical Medicine and International Health 20 (12): E. Hemanth Raj, and others. 2014. “Delivery of Affordable 1846–54. and Equitable Cancer Care in India.” The Lancet Oncology Sustainable Development Solutions Network. 2014. “Health 15 (6): e223–33. in the Framework of Sustainable Development: Technical Raban, M. Z., R. Dandona, and L. Dandona. 2013. “Variations Report for the Post-2015 Development Agenda.” Thematic in Catastrophic Health Expenditure Estimates from Group on Health for All, Sustainable Development Household Surveys in India.” Bulletin of the World Health Solutions Network, New York. Organization 91 (10): 726–35. Teerawattananon, Y., A. Luz, S. Pilasant, S. Tangsathitkulchai, Ramachandran, R., and V. Jha. 2013. “Kidney Transplantation S. Chootipongchaivat, and others. 2016. “How to Meet Is Associated with Catastrophic Out-of-Pocket Expenditure the Demand for Good Quality Renal Dialysis as Part of in India.” PLoS One 8 (7): e67812. Universal Health Coverage in Resource-Limited Settings?” Rocha-Garcia, A., P. Hernandez-Pena, S. Ruiz-Velazco, L. Avila- Health Research Policy and Systems 14 (March): 21. Burgos, T. Marin-Palomares, and others. 2003. “Out-of- Thuan, N. T., C. Lofgren, N. T. Chuc, U. Janlert, and Pocket Expenditures during Hospitalization of Young L. Lindholm. 2006. “Household Out-of-Pocket Payments Leukemia Patients with State Medical Insurance in Two for Illness: Evidence from Vietnam.” BMC Public Health Mexican Hospitals [Spanish].” Salud Pública de México 6 (November): 283. 45 (4): 285–92. Tran, B. X., A. T. Duong, L. T. Nguyen, J. Hwang, B. T. Nguyen, Ruger, J. P. 2012. “An Alternative Framework for Analyzing and others. 2013. “Financial Burden of Health Care for Financial Protection in Health.” PLoS Medicine 9 (8): HIV/AIDS Patients in Vietnam.” Tropical Medicine and e1001294. International Health 18 (2): 212–18. Russell, S. 2004. “The Economic Burden of Illness for Tsilaajav, T., O. Nanzad, and E. Ichinnorov. 2015. “Analysis Households in Developing Countries: A Review of of Catastrophic Health Payments and Benefit Incidence Studies Focusing on Malaria, Tuberculosis, and Human of Government Spending for Health in Mongolia.” Paper Immunodeficiency Virus/Acquired Immunodeficiency commissioned for the World Health Organization, Western Syndrome.” American Journal of Tropical Medicine and Pacific Regional Office, Manila. Hygiene 71 (Suppl 2): 147–55. Ulep, G. T., and N. A. O. dela Cruz. 2013. “Analysis of Out-of- Saito, E., S. Gilmour, M. M. Rahman, G. S. Gautam, P. K. Pocket Expenditures in the Philippines.” Philippine Journal Shrestha, and others. 2014. “Catastrophic Household of Development 40 (1–2d). Expenditure on Health in Nepal: A Cross-Sectional Survey.” UN (United Nations) General Assembly. 2015. “Seventieth Bulletin of the World Health Organization 92 (10): 760–67. Session. Resolution A/RES/70/1: Transforming Our World: Sauerborn, R., A. Adams, and M. Hien. 1996. “Household The 2030 Agenda for Sustainable Development.” UN, Strategies to Cope with the Economic Costs of Illness.” New York, October 21. Social Science and Medicine 43 (3): 291–301. van Doorslaer, E., O. O’Donnell, R. P. Rannan-Eliya, Schoen, C., R. Osborn, D. Squires, M. M. Doty, R. Pierson, and A. Somanathan, S. R. Adhikari, and others. 2006. “Effect others. 2010. “How Health Insurance Design Affects Access of Payments for Health Care on Poverty Estimates in to Care and Costs, by Income, in Eleven Countries.” Health 11 Countries in Asia: An Analysis of Household Survey Affairs 29 (12): 2323–34. Data.” The Lancet 368 (9544): 1357–64. Schoenberg, N. E., H. Kim, W. Edwards, and S. T. Fleming. 2007. Van Minh, H., and B. X. Tran. 2012. “Assessing the Household “Burden of Common Multiple-Morbidity Constellations Financial Burden Associated with the Chronic Non- on Out-of-Pocket Medical Expenditures among Older Communicable Diseases in a Rural District of Vietnam.” Adults.” Gerontologist 47 (4): 423–37. Global Health Action 5 (December): 1–7. 142 Disease Control Priorities: Improving Health and Reducing Poverty Verguet, S., Z. D. Olson, J. B. Babigumira, D. Desalegn, ———. 2016. Global Health Expenditure Database. WHO, K. A. Johansson, and others. 2015. “Health Gains and Financial Geneva (accessed August 29, 2016), http://apps.who.int Risk Protection Afforded by Public Financing of Selected /nha/database. Interventions in Ethiopia: An Extended Cost-Effectiveness WHO (World Health Organization) and World Bank. 2014. Analysis.” The Lancet Global Health 3 (5): e288–96. “Monitoring Progress towards Universal Health Coverage Wagstaff, A., and E. van Doorslaer. 2003. “Catastrophe and at Country and Global Levels: Framework, Measures, and Impoverishment in Paying for Health Care: With Applications Targets.” WHO, Geneva; World Bank, Washington, DC. to Vietnam 1993–1998.” Health Economics 12 (11): 921–34. WHO (World Health Organization) and World Economic Wang, J., H. W. Zhou, Y. X. Lei, and X. W. Wang. 2012. Forum. 2011. From Burden to “Best Buys”: Reducing the “Financial Protection under the New Rural Cooperative Economic Impact of Non-Communicable Diseases in Low- Medical Schemes in China.” Medical Care 50 (8): 700–04. and Middle-Income Countries. Geneva: WHO. Wang, Q., H. Liu, Z. X. Lu, Q. Luo, and J. A. Liu. 2014. “Role of Wingfield, T., D. Boccia, M. Tovar, A. Gavino, K. Zevallos, and the New Rural Cooperative Medical System in Alleviating others. 2014. “Defining Catastrophic Costs and Comparing Catastrophic Medical Payments for Hypertension, Stroke, Their Importance for Adverse Tuberculosis Outcome with and Coronary Heart Disease in Poor Rural Areas of China.” Multi-Drug Resistance: A Prospective Cohort Study, Peru.” BMC Public Health 14 (1): 907. PLoS Medicine 11 (7): e1001675. Webster, R., and A. Rodgers. 2016. “Polypill Treatments for Xu, K. 2005. “Distribution of Health Payments and Catastrophic Cardiovascular Diseases.” Expert Opinion on Drug Delivery Expenditures: Methodology.” Discussion Paper 2-2005, 13 (1): 1–6. WHO, Geneva. White, S. L., S. J. Chadban, S. Jan, J. R. Chapman, and A. Cass. Xu, K., D. B. Evans, G. Carrin, A. M. Aguilar-Rivera, 2008. “How Can We Achieve Global Equity in Provision of P. Musgrove, and others. 2007. “Protecting Households Renal Replacement Therapy?” Bulletin of the World Health from Catastrophic Health Spending.” Health Affairs 26 (4): Organization 86 (3): 229–37. 972–83. WHO (World Health Organization). 2000. World Health Xu, K., D. B. Evans, K. Kawabata, R. Zeramdini, J. Klavus, and Report 2000: Health Systems: Improving Performance. others. 2003. “Household Catastrophic Health Expenditure: Geneva: WHO. A Multicountry Analysis.” The Lancet 362 (9378): 111–17. ———. 2005. “World Health Assembly: Sustainable Health Zhang, J., V. Chongsuvivatwong, and A. Geater. 2006. “Clinical Financing, Universal Coverage, and Social Health Insurance.” Severity and Financial Burden among Road Traffic Injury WHA58.33, WHO, Geneva. Patients in Kunming, China.” Southeast Asian Journal of ———. 2010. The World Health Report: Health Systems Tropical Medicine and Public Health 37 (5): 1034–39. Financing: The Path to Universal Coverage. Geneva: WHO. Zhao, D., W. Wang, J. Liu, M. Wang, J. Sun, and others. 2012. ———. 2014. “Global Health Estimates.” WHO, Geneva. “The Impact of Cardiovascular Disease on Household ———. 2015. “Health and Human Rights.” Fact Sheet 323, Economic Well-Being in Chinese Population.” Circulation WHO, Geneva. 125 (19): e677–78. Economic Burden of Chronic Ill Health and Injuries for Households in Low- and Middle-Income Countries 143 Part 3 Economic Evaluation Results from Disease Control Priorities, Third Edition Chapter 7 Cost-Effectiveness Analysis in Disease Control Priorities, Third Edition Susan Horton INTRODUCTION METHODS League tables, which rank the cost-effectiveness of health A database of cost and cost-effectiveness results was con- interventions, are a useful input for prioritizing health structed for the first six volumes of the Disease Control expenditures, especially for national health budgets. Priorities, third edition (DCP3) (Black and others They have been used as policy tools for high-income 2016; Debas and others 2015; Gelband and others 2015; countries (HICs), including a comprehensive analysis Holmes and others 2017a; Patel and others 2015; for Australia (Vos and others 2010) and a similar analysis Prabhakaran and others 2017). Systematic searches were for cancer across HICs (Greenberg and others 2010). conducted in six major health areas, supplemented by Some low- and middle-income countries (LMICs), such expert surveys and existing published systematic surveys as Mexico, have also used league tables in their policy- and reviews (Gaziano and others 2017; Holmes and making process (Salomon and others 2012). others 2017b; Horton and Gauvreau 2015; Horton and For LMICs as a group, two major reviews of cost- Levin 2016; Levin and Chisholm 2015; Prinja and others effectiveness have informed strategies to achieve the 2015). The surveys covered literature from 2000 to mid- Millennium Development Goals (MDGs) (Evans and 2013 published in English, because the literature before others 2005; Laxminarayan, Chow, and Shahid-Salles 2000 had been reviewed previously (Laxminarayan, 2006). However, cost-effectiveness is not the only impor- Chow, and Shahid-Salles 2006). tant criterion for policy choice; sustainability, equity, and The searches undertaken employed keywords associ- affordability, among others, also matter. Nevertheless, ated with economic outcomes, the names of all LMICs cost-effectiveness provides a useful and comprehensible and regions, and the main disease conditions relevant for reference point. each major health area. In this chapter, we report the As strategies and priorities are set for the Sustainable results per disability-adjusted life year (DALY) averted. Development Goals and countries consider the transi- In most DCP3 volumes, studies were also graded accord- tion to universal health coverage, updating the previous ing to the Drummond checklist to assess the quality of reviews for LMICs is appropriate. This chapter synthe- the economic analysis (Drummond and others 2005). sizes the results from recent analyses in six different Further details of the searches and summaries of disease areas to provide a comprehensive, updated com- the findings for the six major health areas are available parison across a broad range of conditions; to examine (Gaziano and others 2017; Holmes and others 2017b; changes during the past 10–12 years; and to highlight Horton and Gauvreau 2015; Horton and Levin 2016; research gaps. Levin and Chisholm 2015; Prinja and others 2015). 147 Summary information about each of the 93 health inter- or the individual study provided a range of estimates. ventions analyzed and full references for the 149 pub- In the figures, the geometric mean of the endpoints of lished studies are provided in annex 7A. the range was the point estimate used. This approach All costs were converted to 2012 U.S. dollars by works better for a natural log scale axis and is more adjusting prices to 2012 values in the original currency appropriate when the ranges are very different. of the relevant country and then converting those The WHO has issued guidelines on thresholds for amounts to U.S. dollars using the exchange rate for 2012. acceptable costs per DALY averted. The recommenda- The costs for one group of studies were expressed tion is that anything costing less than the per capita gross in international dollars of a World Health Organization national income (GNI) per DALY averted is “very (WHO) region (Evans and others 2005) and could not cost-effective” (WHO 2001); anything costing less than be readily converted, because consumer price indices three times per capita GNI is “cost-effective.” Recent and exchange rates with the U.S. dollar are not publicly research suggests that health budget constraints are too available for those regional aggregates. Although meth- tight to be able to afford everything, even those items ods exist to make an approximate conversion, the addi- that are very cost-effective according to the WHO tional information required is not always readily available threshold. Accordingly, thresholds should be lower from the original study, namely, the proportion of all (Claxton and others 2015). Deriving a more appropriate costs (both of the intervention itself and, where relevant, threshold—for example, using the marginal health gain of those costs averted by the intervention) accounted for with the existing health budget—requires country- by tradable and nontradable inputs. specific data. A recent analysis suggests that a threshold We opted to use exchange rate conversions rather of approximately one-half of GNI per capita would be than purchasing power parity (PPP) conversions. Studies more appropriate for LMICs than the WHO-suggested using the Choosing Interventions that are Cost-Effective threshold and better reflects funds that taxpayers in (WHO-CHOICE) methodology (Evans and others those countries are able and willing to spend from the 2005) have often used PPP conversions, which assume public budget (Ochalek, Claxton, and Lomas 2016). that health interventions have the same mix of tradable In our review, a lower threshold of US$200 per DALY and nontradable inputs as the economy does overall. is used to identify priority interventions for consider- However, health interventions vary considerably, from ation in low-income countries (LICs); all but three those involving behavior change communication by countries in the World Bank database had per capita community health workers (relying heavily on nontrad- income above US$400 in 2014. A higher threshold of able inputs) to vaccine delivery or use of rapid diagnos- US$500 is used to identify priority interventions for tic tests (relying heavily on tradable inputs); no single consideration in lower-middle-income countries, all of conversion method is perfect. We opted for the exchange which had per capita GNI above US$1,045 in 2014. rate method because it is more readily understood by Other considerations, such as equity, affordability, and noneconomists, and it allows comparison with the ear- feasibility will also be important in priority setting for lier Disease Control Priorities work (Laxminarayan, individual countries, depending on the context. Chow, and Shahid-Salles 2006). Using market exchange rates, however, can be problematic if they do not respond RESULTS immediately to differential rates of inflation between countries. We identified cost-effectiveness estimates for 93 inter- The cost-effectiveness rankings from individual vol- ventions and contexts (figures 7.1–7.4), drawn from umes were aggregated to provide two sets of league 149 studies. We excluded cost-effectiveness studies of tax tables—one for adults and one for children. In a few and subsidy policies. Although broad national policy cases where no study using DALYs was available for an changes are very important, estimating their costs is important intervention—for example, human papillo- more difficult, and their cost-effectiveness is not readily mavirus (HPV) vaccination—a study using quality- compared with that of individual health interventions. adjusted life years (QALYs) was used instead, and this In a few cases, the same intervention appears more substitution is indicated. A natural logarithmic scale was than once in different contexts, with different costs per used for cost in the figures because small differences in DALY averted. For example, the cost-effectiveness of cost per outcome are less important for the least cost- HPV vaccination has been estimated at two different effective interventions, that is, those with the highest cost prices per vaccinated girl: the lower price from Gavi— per outcome. For some interventions, a single study the Vaccine Alliance (Gavi) is available to some lower- provided a point estimate for cost-effectiveness; for middle-income countries—and the usually higher other interventions, multiple studies were available, price applies to countries ineligible for Gavi support. 148 Disease Control Priorities: Improving Health and Reducing Poverty Figure 7.1 Interventions Costing Less than US$100 per DALY Averted for Adults Home presumptive treatment malaria, Africa Rural trauma hospital Supply ITNs for malaria, Africa Add Xpert to smear to diagnose TB, lower-middle-income countries Hepatitis B vaccination, LICs Treat smear negative TB with first-line drugs, LICs Comprehensive management of malaria (spray, nets, treat), Africa IRS for malaria, Africa Detect and treat leprosy IPTM in pregnancy, Africa Preventive chemotherapy for trachoma IPTM in infants, Africa Hernia repair Cleft lip and palate repair ACE inhibitor, heart failure, no treatment access PMTCT Option B versus no treatment, Africa Treat malaria with ACT, Africa Detect and treat visceral leishmaniasis Cataract surgery Treat smear positive TB with first-line drugs, LICs Detect and treat human African trypanosomiasis Screen and treat for syphilis, LICs Prehospital ECG versus none, MICs Emergency obstetric care Add syphilis screen to HIV screen and treat, LICs Voluntary male circumcision Salt reduction policy in food Treat severe malaria with artesunate Preventive chemotherapy for onchocerciasis Give female condom to sex workers, S Af ACE inhibitor, heart failure, treatment access Polypill for high absolute risk CVD, UMICs Blood pressure management, UMICs 1 10 100 1,000 10,000 Cost per DALY averted (2012 US$) Range Note: ACE = angiotensin converting enzyme; ACT = artemisinin-based combination therapy; CVD = cardiovascular disease; ECG = electrocardiogram; IPTM = intermittent preventive treatment for malaria; IRS = indoor residual spraying; ITNs = insecticide-treated nets; LICs = low-income countries; mgt = management; MICs = middle-income countries; Option B = use of two-drug regime for pregnancy for PMTCT; PMTCT = Prevention of Mother-to-Child Transmission of HIV; S Af = South Africa; TB = tuberculosis; UMICs = upper-middle-income countries. Gavi has used its ability to undertake bulk purchases and The results from southern Africa, which faces a general- multiyear commitments for vaccines to obtain favorable ized epidemic in a few countries, differ from those of prices. However, only those countries eligible for Gavi other countries, where the epidemic is more concen- support have access to these prices; other countries must trated in certain population groups. If no context is negotiate prices with manufacturers. identified, the results are expected to be generally appli- Where relevant, the economic level of the country cable in LMICs. where the study was conducted is identified (for exam- Of the 93 cost-effectiveness estimates, 37 percent ple, LICs as compared to lower-middle-income coun- relate to interventions for reproductive, maternal, tries and UMICs) because human resource costs vary newborn, and child health interventions and 24 percent significantly and disease patterns are different. In other relate to interventions for major infectious diseases— cases, particularly for the human immunodefi- HIV/AIDS, tuberculosis, malaria, and neglected tropi- ciency virus/acquired immune deficiency syndrome cal diseases (NTDs). This finding is not surprising, (HIV/AIDS), the epidemiologic context is identified. given that the MDGs focused on these areas of health. Cost-Effectiveness Analysis in Disease Control Priorities, Third Edition 149 Figure 7.2 Interventions Costing between US$100 and US$999 per DALY Averted for Adults Episodic psychosocial care for depression, primary care, UMICs Secondary prevention (medication) for CVD versus none BCC plus regulation, sex establishments, LAC Nonemergency orthopedic conditions Maintenance psychosocial care for depression, primary care, UMICs Treat CRC, LICs Nonprice interventions for tobacco PMTCT Option B+ versus Option A, Africa PMTCT Option A versus no treatment, SE Asia Eradicate yaws (detect and treat) Intrapartum care Older anti-epileptic drug in primary care, MICs β-blocker and ACE inhibitor, heart failure, no access to treatment Screen and treat for syphilis, UMICs Treat TB with second line drugs, MICs Trauma center HPV vaccination of US$50 per girl, MICsa Treat breast cancer, MICs Scale up ART to all with CD4 counts < 350 cells/mm2, or all infected, S Af β-blocker and ACE inhibitor, heart failure, access to treatment Add syphilis screen to HIV screen and treatment, UMICs PMTCT Option A versus no treatment, Africa Primary prevention of ARF/RHD, children with GAS pharyngitis PMTCT Option B versus Option A, Africa Preventive chemotherapy for schistosomiasis and STH 1 10 100 1,000 10,000 Cost per DALY averted (2012 US$) Range Note: ACE = angiotensin converting enzyme; ARF/RHD = acute respiratory failure/rheumatic heart disease; ART = antiretroviral therapy; BCC = behavior change communication; CRC = colorectal cancer; CVD = cardiovascular disease; HPV = human papillomavirus; LAC = Latin America and the Caribbean; LICs = low-income countries; MICs = middle-income countries; Option A = use of single-drug regime for pregnancy for PMTCT; Option B = use of two-drug regime for pregnancy for PMTCT; Option B+ = use of two-drug regime during pregnancy and then lifelong for PMTCT; PMTCT = Elimination of Mother-to-Child Transmission of HIV; STH = soil-transmitted helminths; TB = tuberculosis; UMICs = upper-middle- income countries. a. Denotes outcome in QALYs (quality-adjusted life years). International organizations, such as Gavi and the new guidelines for treatment, and new diagnostic Global Fund to Fight AIDS, Tuberculosis, and Malaria, tools. Hence, no new studies were found for well- mobilized significant resources, leading to consider- established interventions, such as the original able interest in, and funding for, cost-effectiveness Expanded Program of Immunization with six vac- studies in these health areas. Far fewer economic stud- cines. Pre-2000 studies of some of these established ies are available for each of the other four areas consid- interventions exist. In other cases, for example, emer- ered: cancer, cardiovascular disease, mental health, and gency appendectomy, the importance of the interven- surgery. tion was established long before cost-effectiveness Studies are typically conducted where new policy estimates became common for LMICs, and thus, no measures are being considered, such as new vaccines, studies were found. 150 Disease Control Priorities: Improving Health and Reducing Poverty Figure 7.3 Interventions Costing US$1,000 or More per DALY Averted for Adults PrEP - ART for noninfected partner, serodiscordant couples Regulate food ads and labels, MICs PMTCT Option A (with mass screen) versus no treatment, LAC Screen and treat breast cancer, LICs Online sex education to prevent STIs, LAC Vector control for dengue fever Primary prevention CVD with 4 drugs, MICs Screen and treat breast cancer, MICs Treatment of depression in primary care with drugs, MICs Telemedicine diabetic retinopathy screening, 1–2 times per lifetime, MICs Facility-based treatment of schizophrenia with drugs, MICs Primary prevention of CVD absolute risk > 40%, UMICs BCC alone, sex establishments, LAC Use Xpert to diagnose TB, MICs HPV vaccination of US$240+ per girla 1 10 100 1,000 10,000 Cost per DALY averted (2012 US$) Range Note: ART = antiretroviral therapy; BCC = behavior change communication; CVD = cardiovascular disease; HPV = human papillomavirus; LAC = Latin America and the Caribbean; LICs = low-income countries; MICs = middle-income countries; Option A = use of single-drug regime for pregnancy for EMTCT; PrEP = pre-exposure prophylaxis; PMTCT = Prevention of Mother-to-Child Transmission of HIV; STIs = sexually transmitted infections; TB= tuberculosis; UMICs = upper-middle-income countries. a. Denotes outcome in QALYs (quality-adjusted life years). More than half of the interventions in figures 7.1–7.4 against malaria, as well as insecticide-treated cost less than US$200 per DALY averted. These interven- nets and indoor residual spraying; antiretroviral tions could be considered for publicly funded health care therapy for pregnant women; hepatitis B vacci- in LICs and include the following: nations; and HPV vaccination at US$50 per fully vaccinated girl • Treatment of various, primarily infectious diseases: • Pneumococcus, rotavirus, and Haemophilus influ- Treatment for malaria, tuberculosis (including enza type b (Hib) vaccines in LICs tuberculosis that is resistant to first-line drugs), • Selected basic surgical interventions: Basic trauma HIV/AIDS, syphilis, and four of the NTDs; basic surgery and emergency obstetric care; surgery for treatment using medication for heart failure cataracts, hernia, and cleft lip and palate • Prevention of various, primarily infectious • Other miscellaneous interventions: Training tradi- diseases: Male circumcision; intermittent preven- tional birth attendants and general practitioners for tive treatment in pregnant women and in infants births; community-based neonatal care. Cost-Effectiveness Analysis in Disease Control Priorities, Third Edition 151 Figure 7.4 Interventions for Children Microfinance and gender training IPV Urban water supply and sanition, LICs Rural water supply and sanitation, LICs C-section, all lower-middle-income countries Pneumococcus and rotavirus, UMICs Cholera and typhoid vaccines Pneumococcus, rotavirus, lower-middle-income countries Yellow fever, Japanese encephalitis, meningitis A vaccines Hib and rubella added to EPI, LICs Mother’s groups to improve healtha Comprehensive nutrition package Intrapartum care, LICsa Intrapartum care, LAC QI protocol for newborns in hospital Access to modern contraceptives Household water treatment, LICs Oral rehydration therapy Handwashing (BCC) Pneumococcus and rotavirus, LICs Original EPI-6 plus hepatitis B Home management of fever with antimalarials Education programs on nutrition and WASH Clean delivery kit and training of TBAs Management of obstructed labor Micronutrient interventions Maternal and neonatal care at home Community management of severe malnutrition Zinc added to oral rehydration therapy Treatment of severe malaria with artesunate 1 10 100 1,000 10,000 Cost per DALY averted (2012 US$) Range Note: BCC = behavior change communication; EPI = expanded program of immunization; Hib = Haemophilus influenza type b; IPV = intimate partner violence; LAC = Latin America and the Caribbean; LICs = low-income countries; QI = quality improvement; TBAs = traditional birth attendants; UMICs = upper-middle income countries; WASH = water, sanitation, and hygiene. a. Denotes outcome in QALYs (quality-adjusted life years). Those interventions costing US$200–US$500 per • Secondary and primary prevention of cardiovascu- DALY averted could be considered for lower-middle- lar disease with medication income countries in addition to the items listed. These • Additional mental health interventions include the following: • Pre-exposure prophylaxis as antiretroviral treat- ment of uninfected partners of HIV-infected • Surgery for selected nonemergency orthopedic individuals conditions • Selected behavior-change interventions • Selected interventions for mental health in primary • Provision of balanced protein–energy supplements care settings in pregnancy. • Treatment of one additional NTD • Various nutrition interventions. DISCUSSION Examples of interventions costing more than US$500 per DALY averted and potentially appropriate A similar analysis to the one reported here was con- for consideration in upper-middle-income countries ducted for Disease Control Priorities in Developing include the following: Countries (second edition; Jamison and others 2006). 152 Disease Control Priorities: Improving Health and Reducing Poverty It covered studies through the year 2000 (Laxminarayan, Changes in Prices Chow, and Shahid-Salles 2006) and provided an infor- Reduced prices of pneumococcal and rotavirus vaccines mative source of comparison for the current results that are examples of changes in costs that dramatically date from 2000 through part of 2013. The differences change the cost-effectiveness of the interventions. These are not only in the results of cost-effectiveness studies interventions were high cost per DALY averted in the but are also—tellingly—in the topics studied. pre-2000 review, but at current Gavi prices for LICs, the About half of the interventions appear in both the cost is now less than US$100 per DALY averted. Another pre- and post-2000 compilations; the remainders rep- major example is the NTDs. Following the 2012 London resent some significant changes. Some new interven- Declaration (Uniting to Combat NTDs Coalition 2016), tions that were not in widespread use before the key drugs to combat NTDs have been donated by the 2000—many of them related to substantial investments manufacturers, which has moved the elimination of in new technologies and new methods to change NTDs by prevention and treatment substantially higher behavior over the MDG period—have been evaluated. up the priority list in terms of cost-effectiveness in the For some interventions, substantial reductions in prices past decade. have occurred that have made previously unaffordable interventions less costly and more cost-effective. This is particularly true for vaccines, in cases where efforts by New Health Areas Gavi and others have led to lower vaccine prices, and Efforts by the surgical community (for example, the for malaria and AIDS treatments, in cases where efforts Lancet Commission on Global Surgery and the DCP3 by the Global Fund to Fight AIDS, Tuberculosis, and volume 1 on surgery [Debas and others 2015]) have Malaria and Médecins sans Frontières, among others, increased the interest in and emphasis on cost- have similarly led to reduced drug prices. Some new effectiveness of surgery. Several surgical interventions areas of health care, particularly those not involving cost less than US$200 per DALY averted. In urgent MDG targets, have been studied, making more detailed cases, these same interventions can be implemented in cost-effectiveness data available beyond the areas of a first-level hospital with a general surgeon (for exam- maternal and child health and major infectious dis- ple, emergency obstetric care and basic trauma care); in eases. Some interventions have changed priorities, nonurgent cases, they can be implemented in a special- either as the disease context has changed or as experi- ized facility with high volume and modest cost (for ence has led to a revision of what was expected, based example, cataract surgery or repair of cleft lip and cleft on pilot programs. palate). Similar efforts are underway in the global can- Finally, some interventions no longer appear on the cer community. One study suggests that treatment of list, despite being found to be cost-effective in the previ- early-stage breast cancer falls in the category of less ous study. This may be because they have been main- than US$200 per DALY averted for middle-income streamed and either no further need exists to estimate or countries (although not in LICs, where screen-and- update cost-effectiveness or they have been superseded treat approaches cost more than US$200 per DALY by other more effective or more cost-effective interven- averted). tions. Examples in each of these categories are given in the following sections. Interventions That Have Changed Priority New Technologies and Methods School-based adolescent health and nutrition programs New interventions for which cost-effectiveness data have appeared as a high priority because of their low cost per become available for LMICs include treating severe DALY averted in 2006. This was not the case in 2016, malaria with rectal or injected artesunate, which can be because more recent studies are much more cautious done before hospital arrival; adding GeneXpert testing about whether these programs will have long-term pos- to sputum-smear testing to diagnose disease and deter- itive effects. mine antibiotic susceptibility; and HPV vaccination for girls to prevent cervical cancer. These all fall into the range of less than US$200 per DALY averted in the Interventions That Are No Longer on the List appropriate contexts. However, other new technologies, Changing technology also means that some previously such as pre-exposure prophylaxis, have a relatively high cost-effective interventions have been superseded or cost per DALY averted in most cases. have become usual care. This is particularly evident Cost-Effectiveness Analysis in Disease Control Priorities, Third Edition 153 for HIV/AIDS. In the pre-2000 compilation, eight Recent concerns about the DALY relate to the issue of interventions appeared in the highest-priority list. Peer discounting costs and health benefits further in the and education programs for high-risk groups; condom future. Although this issue is very much accepted by promotion and distribution; voluntary counseling and economists, some health specialists find it more prob- testing without treatment; diagnosis and treatment of lematic. The Institute for Health Metrics and Evaluation sexually transmitted infections; blood and needle safety; has begun using undiscounted DALYs to measure tuberculosis coinfection prevention and treatment; global burden of disease (Murray and others 2012) but opportunistic infection treatment; and prevention of without using a new term to differentiate these undis- mother-to-child transmission were included among the counted DALYs. This approach is already causing most cost-effective interventions (using less than confusion. US$150 per DALY averted in 2001 U.S. dollars, roughly The DALY measure itself has limitations. Using the comparable to less than US$200 per DALY averted in DALY measure tends to underrepresent interventions 2012 U.S. dollars). A decade later, with treatment with where outcomes are not readily measured in this metric, antiretroviral agents on the highest priority list, all but such as family planning, and interventions in nutrition two of the other interventions fell off the list; the where the outcomes are improved cognition rather than remaining two are prevention of mother-to-child trans- improved health, more readily measured with bene- mission and testing for and treatment of other sexually fit:cost analysis ratios. transmitted infections. Most of the interventions had On the cost side, studies predominantly use market become usual care, but voluntary counseling and test- exchange rates to compare across different currencies. ing without treatment had been superseded by test- However, an influential body of work from the WHO, and-treat approaches. the WHO-CHOICE study, used international dollars for A major limitation of the cost-effectiveness litera- WHO subregions rather than countries. International ture, particularly acute in LMICs, is its bias toward the dollars make cross-country comparisons somewhat eas- diseases of greatest interest during the period under ier to understand by adjusting for salary differences as a study. In the current study, the literature overrep- component of costs. The downside is that international resents infectious conditions and childbirth, because dollars make comparison more difficult with other stud- these have been prioritized by international donors. ies not using international dollars. One does not simply Drugs and vaccines tend to be overrepresented relative use the US$/PPP exchange rate, because having informa- to behavior change interventions, because manufac- tion about cost structure is necessary. A further compli- turers use cost-effectiveness data as part of the adop- cation is the lack of published indices for PPP exchange tion process. rates of regions. The advantage of WHO-CHOICE was the ability to compare many interventions at one time, when the Measurement Issues MDG strategies were being evaluated, and to compare The ability to conduct a large comparative study such the outcome of combinations of interventions. The dis- as this relies on use of common methodologies by advantage is that funding to replicate such a large com- individual study authors. For effectiveness studies, prehensive evaluation is difficult to attain. The use of progress has been made applying standard guidelines simpler methods, such as market exchange rates, allows for systematic reviews and using explicit criteria for the synthesis of many smaller, individually directed evaluating evidence. For economics studies, the fairly studies. recent adoption of a common set of reporting stan- dards (Husereau and others 2013) and the develop- CONCLUSIONS ment of a reference case for conducting economic evaluations in LMICs (NICE International 2014) are Cost-effectiveness is not the only criterion by which moves in the same direction. to choose health priorities, but it is useful for identi- A larger issue is the common metric for cost- fying what is given up when a less cost-effective inter- effectiveness. The DALY has been the predominant vention is prioritized. It is also a useful tool for health outcome metric used for studies of LMICs over advocacy for increased health budgets. This review the past decade or more. It has the advantage over the has used cost-effectiveness measures from several QALY for work in multiple countries in that a single set hundred studies for LMICs to help identify candidates of disability weights is used across countries, whereas for priority health packages, which may assist policy QALY weightings are, in theory, country specific, and makers considering how to move to universal health generating QALY weights can be a costly process. coverage. 154 Disease Control Priorities: Improving Health and Reducing Poverty This review has identified some of the gaps where NOTE future research on cost-effectiveness is needed: World Bank Income Classifications as of July 2014 are as follows, based on estimates of gross national income (GNI) • Given the ongoing decline in infectious disease bur- per capita for 2013: den and the growing burden of NCDs, more analy- ses for NCDs are needed for LMICs. Achieving the • Low-income countries (LICs) = US$1,045 or less goal of health convergence within a generation will • Middle-income countries (MICs) are subdivided: not be possible without initiating interventions to (a) lower-middle-income = US$1,046 to US$4,125 reduce NCDs, where the lag between intervention (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 and outcomes is often much longer than for infec- • High-income countries (HICs) = US$12,746 or more. tious diseases. • The review highlights the lack of any study of cost- effectiveness for childhood cancer and the dearth of REFERENCES information on cost-effective interventions for men- Black, R. E., R. Laxminarayan, M. Temmerman, and N. Walker, tal health in LMICs. eds. 2016. Disease Control Priorities (third edition): Volume • Another area for future work includes the cost- 2, Reproductive, Maternal, Newborn, and Child Health. effectiveness of resource-appropriate treatment Washington, DC: World Bank. of early-stage cancers, such as breast and cervical Claxton, K., S. Martin, S. Soares, N. Rice, E. Spackman, and cancers. others. 2015. “Methods for the Estimation of the National • Given the growth of obesity worldwide, cost- Institute for Health and Care Excellence Cost-Effectiveness effectiveness studies of interventions to change Threshold.” Health Technology Assessment 19 (14): 1–503. patterns of diet and inactivity in urban areas are doi:10.3310/hta19140. needed. Debas, H. T., P. Donkor, A. Gawande, D. T. Jamison, M. E. Kruk, and C. N. Mock, eds. 2015. Disease Control Priorities (third edition): Volume 1, Essential Surgery. Washington, A publicly available online global database of cost- DC: World Bank. effectiveness studies using DALY outcomes will make Drummond, M. F., M. J. Schulpher, G. W. Torrance, D. J. O’Brien, future updates easier (Tufts University 2016). and G. L. Stoddart. 2005. Methods for the Economic The major changes in ranking of health priorities Evaluation of Health Care Programmes. 3rd ed. New York: over the past decade underscore the need for periodic Oxford University Press. repetition of league table exercises such as this one. Evans, D. B., S. S. Lim, T. Adam, and T. Tan-Torres Edejer. 2005. “Achieving the Millennium Development Goals for Health: Evaluation of Current Strategies and Future Priorities for ACKNOWLEDGMENTS Improving Health in Developing Countries.” BMJ 331 (7530): 1457–61. The authors wish to thank the following individuals Gaziano, T., M. Suhrcke, E. Brouwer, C. Levin, I. Nikolic, for their excellent assistance: Elizabeth Brouwer at the and R. Nugent. 2017. “Summary of Costs and Cost- University of Washington for her work in organizing Effectiveness of Interventions and Policies to Prevent the cost and cost-effectiveness databases; Vittoria and Treat Cardiometabolic Diseases.” In Disease Control Lutje, from the Cochrane Infectious Diseases Group, Priorities (third edition): Volume 5, Cardiovascular, Respiratory, and Related Diseases, edited by D. Prabhakaran, for conducting extensive literature searches; Emily S. Anand, T. Gaziano, Y. F. Wu, J. C. Mbanya, and R. Nugent. Thacher and Julian Frare-Davis from the University of Washington, DC: World Bank. Washington for extracting the literature and creating Gelband, H., P. Jha, R. Sankaranarayanan, and S. Horton, eds. the databases; and Daphne Wu at the University of 2015. Disease Control Priorities (third edition): Volume 3, Waterloo for working on the databases and providing Cancer. Washington, DC: World Bank. research support for the final analysis presented here. Greenberg, D., C. Earle, C. H. Fang, A. Eldar-Lissai, and P. J. Neumann. 2010. “When Is Cancer Care Cost-Effective? A Systematic Overview of Cost-Utility Analyses in Oncology.” ANNEX Journal of the National Cancer Institute 102 (2): 82–88. Holmes, K. K., S. Bertozzi, B. Bloom, and P. Jha, eds. 2017a. The annex to this chapter is as follows. It is available at Disease Control Priorities (third edition): Volume 6, Major http://www.dcp-3.org/DCP. Infectious Diseases. Washington, DC: World Bank. • Annex 7A. Details of Interventions Included in Holmes, K. K., S. Bertozzi, B. Bloom, P. Jha, R. Nugent, figures 7.1, 7.2, 7.3, and 7.4, by Increasing Cost per H. Gelband, and others. 2017b. “Major Infectious Diseases: DALY Averted. Key Messages from Disease Control Priorities, Third Edition.” Cost-Effectiveness Analysis in Disease Control Priorities, Third Edition 155 In Disease Control Priorities (third edition): Volume 6, Major NICE International. 2014. “The Gates Reference Case: What Infectious Diseases, edited by K. K. Holmes, S. Bertozzi, It Is, Why It’s Important, and How to Use It.” Nice B. Bloom, and P. Jha. Washington, DC: World Bank. International, London. https://www.nice.org.uk/Media Horton, S., and C. L. Gauvreau. 2015. “Cancer in Low and /Default/About/what-we-do/NICE-International/projects Middle-Income Countries: Economic Overview.” In Disease /Gates-Reference-case-what-it-is-how-to-use-it.pdf. Control Priorities (third edition): Volume 3, Cancer, edited Ochalek, J., K. Claxton, and J. Lomas. 2016. “Cost per DALY by H. Gelband, P. Jha, R. Sankaranarayanan, and S. Horton. Averted Thresholds for Low- and Middle-Income Countries: Washington, DC: World Bank. Evidence from Cross Country Data.” CHE Research Paper Horton, S., and C. Levin. 2016. “Cost-Effectiveness of 122, Centre for Health Economics, University of York, Interventions for Reproductive, Maternal, Neonatal, and York, U.K. http://www.idsihealth.org/knowledge_base/cost Child Health.” In Disease Control Priorities (third edi- -per-daly-averted-thresholds-for-low-and-middle-income tion): Volume 2, Reproductive, Maternal, Newborn, and -countries-evidence-from-cross-country-data/. Child Health, edited by R. E. Black, R. Laxminarayan, Patel, V., D. Chisholm, T. Dua, R. Laxminarayan, and M. Temmerman, and N. Walker. Washington, DC: World M. E. Medina-Mora, eds. 2015. Disease Control Priorities Bank. (third edition): Volume 4, Mental, Neurological, and Husereau, D., M. Drummond, S. Petrou, C. Carswell, D. Moher, Substance Use Disorders. Washington, DC: World Bank. and others. 2013. “Consolidated Health Economic Prabhakaran, D., S. Anand, T. Gaziano, Y. F. Wu, J. C. Mbanya, Evaluation Reporting Standards (CHEERS): Explanation and R. Nugent, eds. 2017. Disease Control Priorities (third and Elaboration: A Report of the ISPOR Health Economic edition): Volume 5, Cardiovascular, Respiratory, and Related Evaluation Publication Guidelines Good Reporting Disorders. Washington, DC: World Bank. Practices Task Force.” Value Health 1 (2): 231–50. Prinja, S., A. Nandi, S. Horton, C. Levin, and R. Laxminarayan. Jamison, D. T., J. G. Breman, A. R. Measham, G. Alleyne, 2015. “Costs, Effectiveness, and Cost-Effectiveness of M. Claeson, D. B. Evans, P. Jha, A. Mills, and P. Musgrove, Selected Surgical Procedures and Platforms.” In Disease eds. 2006. Disease Control Priorities in Developing Control Priorities (3rd edition): Volume 1: Essential Countries, Second edition. Washington, DC: World Surgery, edited by H. T. Debas, P. Donkor, A. Gawande, Bank and Oxford University Press. Laxminarayan, D. T. Jamison, M. E. Kruk, and C. N. Mock. Washington, R., J. Chow, and S. A. Shahid-Salles. 2006. “Intervention DC: World Bank. Cost-Effectiveness: Overview of Main Messages.” In Salomon, J. A., N. Carvalho, C. Gutiérrez-Delgado, R. Orozco, Disease Control Priorities in Developing Countries (sec- A. Mancuso, and others. 2012. “Intervention Strategies to ond edition), edited by D. T. Jamison, J. G. Breman, Reduce the Burden of Non-Communicable Diseases in A. R. Measham, G. Alleyne, M. Claeson, D. B. Evans, Mexico: Cost Effectiveness Analysis.” BMJ 344: e355. P. Jha, A. Mills, and P. Musgrove. Washington, DC: World Tufts University. 2016. Global Health Cost-Effectiveness Bank and Oxford University Press. Analysis Registry database. http://healtheconomics Levin, C., and D. Chisholm. 2015. “Cost-Effectiveness and .tuftsmedicalcenter.org/orchard. Affordability of Interventions, Policies, and Platforms for Uniting to Combat NTDs Coalition. 2016. “Reaching the Prevention and Treatment of Mental, Neurological, and the Unreached: Fourth Progress Report of the London Substance Use Disorders.” In Disease Control Priorities (third Declaration.” http://unitingtocombatntds.org/report/fourth edition): Volume 4, Mental, Neurological, and Substance -report-reaching-unreached. Use Disorders, edited by V. Patel, D. Chisholm, T. Dua, Vos, T., R. Carter, J. Barendregt, C. Mihalopoulos, L. Veerman, R. Laxminarayan, and M. E. Medina-Mora. Washington, and others. 2010. Assessing Cost-Effectiveness in Prevention. DC: World Bank. University of Queensland, Brisbane, and Deakin University, Murray, C. J. L., T. Vos, R. Lozano, M. Naghavi, A. D. Flaxman, Melbourne. https://public-health.uq.edu.au/files/571/ACE and others. 2012. “Disability-Adjusted Life Years (DALYs) -Prevention_final_report.pdf. for 291 Diseases and Injuries in 21 Regions, 1991–2010: WHO (World Health Organization). 2001. Macroeconomics A Systematic Analysis for the Global Burden of Disease and Health: Investing in Health for Economic Development. Study 2010.” The Lancet 380 (9859): 2197–233. Geneva: WHO. 156 Disease Control Priorities: Improving Health and Reducing Poverty Chapter 8 Health Policy Analysis: Applications of Extended Cost-Effectiveness Analysis Methodology in Disease Control Priorities, Third Edition Stéphane Verguet and Dean T. Jamison INTRODUCTION objective of health policy instruments such as universal Multiple criteria are involved in making decisions and public finance (UPF), that is, full public finance irrespec- prioritizing health policies (Baltussen and Niessen 2006). tive of whether services are provided privately or publicly. Potential trade-offs between efficiency and equity are Indeed, out-of-pocket (OOP) medical payments can lead among these criteria and have long been emphasized in to impoverishment in many countries, with households the treatment and prevention of human immunodefi- choosing from among many coping strategies (borrowing ciency virus/acquired immune deficiency syndrome from friends and relatives, selling assets) to manage (HIV/AIDS) (for example, Cleary 2010; Kaplan and health-related expenses (Kruk, Goldmann, and Galea Merson 2002; Verguet 2013). Notably, several mathematical 2009; van Doorslaer and others 2006; Xu and others frameworks, including mathematical programming, have 2003). Absent other financing mechanisms, household proposed incorporating equity into resource allocation medical expenditures can often be catastrophic (Wagstaff decisions in the public sector (Birch and Gafni 1992; 2010; Wagstaff and van Doorslaer 2003), defined as Bleichrodt, Diecidue, and Quiggin 2004; Epstein and exceeding a certain fraction of total household expendi- others 2007; Segall 1989; Stinnett and Paltiel 1996). The tures. A large literature documents the significance of worldwide application of benefit-cost analysis provided medical impoverishment, but far less is known about the for “distributional weights” as early as the 1970s. medical conditions responsible for it. Essue and others Protection from financial risks associated with health (2017), in chapter 6 of this volume, review and extend that care expenses is emerging as a critical component of literature, and Verguet, Memirie, and Norheim (2016) national health strategies in many low- and middle- provide a framework for assessing the global burden of income countries (LMICs). The World Health medical impoverishment by cause, applying it to a case Organization’s World Health Reports of 1999 and 2000 study of a systematic categorization by disease in Ethiopia. included the provision of financial risk protection (FRP) In the literature on medical impoverishment, attenuating as one criterion of good performance for health systems such impoverishment is considered a significant objective (WHO 1999, 2000). Reducing these financial risks is one of health policy, but surprisingly little analysis has been Corresponding author: Stéphane Verguet, Department of Global Health and Population, Harvard T. H. Chan School of Public Health, Boston, Massachusetts, United States; verguet@hsph.harvard.edu 157 performed of efficient ways to address the problem. The others 2013; McClellan and Skinner 2006; Sassi, Archard, method of Extended cost-effectiveness analysis (ECEA) and Le Grand 2001; Smith 2007, 2013). It enables the was initially developed for DCP3 by Verguet, Laxminarayan, design of benefits packages that quantify both health and and Jamison (2015). nonhealth benefits for a given expenditure on specific Traditionally, economic evaluations of health interven- health policies, based on the quantitative inclusion of tions (cost-effectiveness analyses [CEAs]) have focused on how much nonhealth benefits are being bought as well improvements in health and estimated an intervention as how much health benefits are being bought with a cost per health gain in dollar per death averted or dollar given investment in an intervention or policy. In this per disability-adjusted life year (DALY) averted (Jamison respect, ECEA can answer some of the policy questions and others 2006). However, arguments have been devel- raised by the World Health Reports for 2010 and 2013 oped for some time that CEA in health should be extended (WHO 2010, 2013) regarding how to select and sequence to explicitly consider the multiple dimensions of outcome. the health services to be provided on the path toward Jamison (2009), for example, argued that CEAs can be universal health coverage. This chapter first describes the extended to include FRP on the outcome side and use of ECEA approach and then summarizes findings of ECEAs scarce health system capacity on the cost side (figure 8.1). undertaken in the context of the third edition of Disease Specific methods for advancing this agenda were first pro- Control Priorities (DCP3; http://www.dcp-3.org). posed and applied in assessments of the consequences of two alternative policies—public finance and improved access to credit—for extending coverage of tuberculosis APPROACH treatment in India (Verguet, Laxminarayan, and Jamison Consider the implementation of a given health policy 2015). That study and other early ECEAs (Verguet 2013; (HP) in a given population (P). Policy examples include Verguet, Gauvreau, and others 2015; Verguet, Olson, and public finance for a package of vaccines, taxation on others 2015) supplemented traditional economic evalua- tobacco products, legislation to enforce the mandatory tion with evaluation of nonhealth benefits (such as FRP use of helmets, and so forth. P can be divided into sub- and equity), with the broad objective of providing valu- groups, which can be denoted Pk (with 1 ≤ k ≤ n) per able guidance in the design of health policies.1 socioeconomic status according to five income quintiles, ECEA in this respect builds on the existing frame- per region according to geographic location (state, works of cost-benefit analysis and cost-consequence region, county), and per gender. analysis that tabulate disaggregated results (Mauskopf HP entails a given coverage (Cov) and given effective- and others 1998) and on analytical frameworks that ness (Eff) for preventing disease burden (D) in the pop- incorporate equity and FRP concerns into economic ulation as well as a net cost (C). The ECEA methodology evaluations (Asaria and others 2015; Brown and quantifies both health benefits (BH) and nonhealth ben- Finkelstein 2008; Cookson, Drummond, and Weatherly efits (BNH) in P for a given increment in public (or pri- 2009; Finkelstein and McKnight 2008; Fleurbaey and vate) expenditure (figure 8.2). Figure 8.1 Intervention Costs and Effects: A More General View Health Benefits With the introduction of HP, health benefits (BH) are procured—for example, quantified by the sum of the Improved burden of disease averted in each subgroup (Pk)—with a health specific effectiveness of the policy (Effk) assumed to be Outcomes constant per subgroup. In this respect, ECEA estimates the distributional health consequences—in particular, benefits (mortality, morbid- Financial ity averted, disability-adjusted life years averted, quality- protection adjusted life years gained)—per population strata, whether socioeconomic group or geographic setting (figure 8.3). Financial Health system capacity Nonhealth Benefits Costs With HP, nonhealth benefits (BNH,j) are procured, with Source: Jamison 2009, by permission of Oxford University Press. 1 ≤ j ≤ m, where j indicates the type of nonhealth benefits Note: The shaded box represents the domain of traditional cost-effectiveness analysis. (FRP, number of school days gained). For example, if 158 Disease Control Priorities: Improving Health and Reducing Poverty we consider FRP, given a preexisting burden of illness- Figure 8.2 Objective of Extended Cost-Effectiveness Analysis: related impoverishment due to medical expenses, direct Efficient Purchase of Health and Nonhealth Benefits nonmedical costs such as transportation costs, and indi- rect costs such as wages lost, the related nonhealth bene- Equity (for example, deaths fits could be expressed by the sum of the burden of Financial risk averted among bottom 20%) illness-related impoverishment averted in each popula- protection tion subgroup. (FRP) (for example, poverty Specifically, the ECEA approach goes beyond the socie- Health benefits cases averted) tal perspective in traditional economic evaluations (for example, (Drummond and others 2015) to examine the perspective deaths averted) of households in estimating the amount of OOP expendi- Outcomes per US$1 million tures (direct medical costs, direct nonmedical costs, indirect costs) that could be affected by a specific policy (figure 8.4). Subsequently, once the amount of OOP private expendi- Note: Similar to CEA measures in, say, US$ per death averted, estimate the efficient purchase of FRP in, say, US$ per FRP provided. CEA = cost-effectiveness analysis; FRP = financial risk protection. tures borne by households that may be “crowded out” has been estimated, ECEA can be used to scale the amount of Figure 8.3 Distribution of Under-Five Deaths Averted with Universal OOP household expenditures by households’ disposable Public Finance (UPF) of Pneumonia Treatment at a Coverage Level income to estimate FRP—in other words, to account for 20 Percent Higher Than the Current Level and UPF of Combined the fact that a household with annual income of US$100,000 Pneumonia Treatment and Pneumococcal Vaccination at 20 Percent and OOP expenditures of US$10 is much less severely Coverage Level in Ethiopia affected than a household with annual income of US$100. The crowding out of private health expenditures will often 2,500 be an objective as well as a consequence of health policy. Several metrics can be used to estimate FRP (Flores and others 2008; Wagstaff 2010; Verguet, Laxminarayan, 2,000 and Jamison 2015), including the following: Number of deaths averted • Number of catastrophic health expenditures averted, 1,500 estimating the number of households no longer crossing a catastrophic threshold (for example, 10 percent, 20 percent, 40 percent of income or capacity to pay) 1,000 from OOP expenditures • Number of poverty cases averted, estimating the num- ber of households no longer crossing a poverty line (for 500 example, US$1.25 per day) because of OOP expenditures • Number of instances of forced asset sales or forced borrowing averted 0 • A money-metric value of insurance provided, quan- I II III IV V tifying the willingness to pay or risk premium associ- Income quintile (poorest to richest) ated with the policy (figure 8.5). Pneumonia treatment Pneumonia treatment and pneumococcal vaccine Equity Benefits Source: Verguet and others 2016. With HP, equity benefits (BEq), estimated here in terms of health distribution, can be procured. For example, if HP provides more health benefits to poorer than to “Efficient Purchase” of Health and Nonhealth Benefits richer segments of the population, the policy could be The net cost of the policy is C. For that net cost, HP “effi- deemed equity enhancing (figure 8.3). There are several ciently” purchases health benefits (BH) but also nonhealth B benefits (BNH)—for example, BFRP. As in CEA, we can then ways to quantify BEq, including H ,w , where BH,w and BH BH define a usual incremental cost-effectiveness ratio (ICER)— are the health benefits procured by HP among the worst- ICER = C/BH—but we can also define an ICER for each of off group and the total sum of health benefits in all the nonhealth benefits: for FRP, ICERFRP = C/BFRP. In this groups, respectively. respect, ECEA can help quantify the efficient purchase of Health Policy Analysis: Applications of Extended Cost-Effectiveness Analysis Methodology in DCP3 159 Figure 8.4 Distribution of Household Private Expenditures Averted both equity and FRP in addition to health. It also can help with Universal Public Finance (UPF) of Pneumonia Treatment at a generate the evidence base to support informed trade-offs Coverage Level 20 Percent Higher Than the Current Level and UPF of among the partially competing objectives of improved Combined Pneumonia Treatment and Pneumococcal Vaccination at health, improved FRP, and improved equity. Figure 8.6 20 Percent Coverage Level in Ethiopia provides an illustration from Ethiopia. Private expenditures averted (US$, thousands) 1,400 1,200 APPLICATIONS ECEAs Completed to Date 1,000 ECEA was developed for DCP3 and has been used in 800 health policy assessments for a variety of both policies and settings (table 8.1). The policies include public 600 finance, excise taxes, legislation, regulation, conditional 400 cash transfers, task shifting, and education. ECEAs are context specific and depend substantially 200 on the epidemiology of the setting (endemicity, distribu- tion of specific diseases), local health system infrastruc- 0 ture (presence and distribution of health facilities), wealth I II III IV V of the location (low-income, lower-middle-income, Income quintile (poorest to richest) upper-middle-income country), and financial arrange- Pneumonia treatment ments (presence of social health insurance, community- Pneumonia treatment and pneumococcal vaccine based insurance). In total, more than 20 ECEAs have been published (or accepted for publication) as of May 2017. Source: Verguet and others 2016. Of these, nine are included in one of DCP3’s nine volumes. Figure 8.5 Distribution of Financial Risk Protection (Measured by a Money-Metric Value of Insurance Provided) with Universal Example: Use of Dashboard Public Finance (UPF) of Pneumonia Treatment at a Coverage Level We now illustrate ECEA in considering the example of UPF 20 Percent Higher Than the Current Level and UPF of Combined for tuberculosis treatment in India in a population composed Pneumonia Treatment and Pneumococcal Vaccination at 20 Percent Coverage Level in Ethiopia of five income quintiles totaling 1 million people (200,000 people per income quintile), drawing on the first completed 180 ECEA (Verguet, Laxminarayan, and Jamison 2015). Money-metric value of insurance (US$, thousands) Notably, we assume an average incidence of tuberculosis 160 of p0 = 100 per 100,000 per year, with incidence highest in 140 the lowest income quintile. The cost of tuberculosis treat- ment (that is, directly observed treatment, short course) is 120 US$100 per person. We also assume income in the popu- 100 lation is distributed following a Gamma distribution based on a mean income of US$1,500 and a Gini coeffi- 80 cient of 0.33, as produced by an algorithm given by Salem 60 and Mount (1974; see also Kemp-Benedict 2001). The total number of deaths averted would be about 80 40 a year. The health benefits would be concentrated among the bottom income quintile (50 percent) because tubercu- 20 losis has a higher incidence among this subgroup. The total 0 amount of private OOP expenditures averted by universal I II III IV V public funding would be about US$29,000. The bottom Income quintile (poorest to richest) income quintile would benefit from about 20 percent of Pneumonia treatment the private expenditures averted. The total incremental Pneumonia treatment and pneumococcal vaccine treatment costs incurred by the public sector would be about US$65,000. The total FRP afforded by UPF, esti- Source: Verguet and others 2016. mated here using a money-metric value of insurance, 160 Disease Control Priorities: Improving Health and Reducing Poverty would be about US$9,000, 60 percent of which would be Figure 8.6 Financial Risk Protection Afforded (Poverty Cases among the bottom quintile (table 8.2). Averted) Versus Health Gains (Deaths Averted) per US$100,000 Spent Examining the efficient purchase of health and non- (in 2011 U.S. Dollars) for Interventions Provided through Universal health benefits, we find the following: ICER = US$800 Public Finance in Ethiopia per death averted, and ICERFRP = US$7 per dollar of 100 7 insurance value provided. For each US$1 million spent, 8 about 1,200 deaths are averted, 600 of which are in the 9 Number of poverty cases averted bottom income quintile, and the money-metric value of 80 4 insurance is US$140,000, of which 60 percent is in the 5 6 bottom income quintile. 60 1 (US$1 per dose) In addition to examining UPF, the ECEA study for India examined the consequences of improving access to borrowing to cover treatment costs. It found that it was 40 plausible that such policies substantially reduce TB mor- tality among the poor but—relative to UPF—it would 2 (US$1 per dose) 20 1 (US$2.5 per dose) generate high burdens of lingering debt. 2 (US$3.5 per dose) 3 Poverty Reduction Benefits of Health Policies and 0 100 200 300 400 Design of the Benefits Package Number of deaths averted ECEA stresses the potential poverty reduction benefits of Rotavirus vaccine (1) Pneumococcal conjugate vaccine (2) health policies. Specifically, ECEA explicitly quantifies the Measles vaccine (3) Diarrhea treatment (4) FRP benefits or the poverty reduction benefits of policies. Pneumonia treatment (5) Malaria treatment (6) In this respect, it fulfills two major objectives. First, it Cesarean section (7) Tuberculosis treatment (8) provides a quantitative tool that enables intersectoral Hypertension treatment (9) comparison of health policies with other sectors (educa- tion and transport), which is of particular relevance for Source: Verguet, Olson, and others 2015. Table 8.1 Extended Cost-Effectiveness Analyses for Disease Control Priorities a. ECEAs in DCP3 DCP3 Authors and other relevant Volume Chapter and topic Policy instrument Country publications (if any) 1 19. Expanding surgical access Task sharing, public finance Ethiopia Shrime and others 2015; Shrime and others 2016 2 18. Universal home-based Public finance India Ashok, Nandi, and Laxminarayan 2015; neonatal care package in Nandi, Colson, and others 2016 rural India 19. Diarrhea and pneumonia Public finance Ethiopia Verguet, Pecenka, and others 2016; treatment Johansson, Pecenka, and others 2015; Pecenka and others 2015; Verguet, Murphy, and others 2013 3 18. Human papillomavirus Public finance China Levin and others 2015a; Levin and others vaccination to prevent 2015b cervical cancer 4 13. Universal coverage for Public finance Ethiopia, India Chisholm and others 2015; Johansson, mental, neurological, and Bjerkreim Strand, and others 2016; substance use disorders Megiddo and others 2016; Raykar and others 2016 table continues next page Health Policy Analysis: Applications of Extended Cost-Effectiveness Analysis Methodology in DCP3 161 Table 8.1 Extended Cost-Effectiveness Analyses for Disease Control Priorities (continued) DCP3 Authors and other relevant Volume Chapter and topic Policy instrument Country publications (if any) 5 20. Selected ECEAs for Public finance of China, Ethiopia, Watkins, Nugent, and Verguet 2017; cardiovascular diseases interventions, tobacco South Africa Verguet, Gauvreau, and others 2015; taxation, regulation of salt Verguet, Olson, and others 2015; Watkins and others 2015 7 11. Motorcycle helmet laws Regulation Vietnam Olson and others 2016; Olson and others 2017 12. Use of liquefied Commodity subsidy India Pillarisetti, Jamison, and Smith 2017 petroleum gas and other clean energy sources in household 8 28. Postponing adolescent Education India, Niger Verguet, Nandi, and Bundy 2016; Verguet, parity Nandi, and others 2017 b. Other published ECEAs (including those accepted for publication) Topic Policy instrument Country Reference Tuberculosis treatment Universal public finance; India Verguet, Laxminarayan, and Jamison 2015 policies to improve ease of borrowing for treatment costs Measles vaccine Conditional cash transfers Ethiopia Driessen and others 2015 Universal immunization Public finance India Megiddo and others 2014 Water and sanitation Clean piped water and India Nandi, Megiddo, and others 2016 improved sanitation Tobacco Taxation Lebanon/Armenia Verguet, Gauvreau, and others 2015; Salti, Brouwer, and Verguet 2016; Postolovska and others 2017 Palliative care Public finance Vietnam Krakauer and others 2017 Tutorial Not applicable Verguet, Kim, and Jamison 2016 Rotavirus vaccine Public finance Malaysia Loganathan and others 2016 Malaria vaccine Public finance Zambia Liu, True, and others, forthcoming Note: ECEA = extended cost-effectiveness analysis. These two papers reference the same study. Table 8.2 Extended Cost-Effectiveness Analysis Results for Universal Public Finance of Tuberculosis Treatment in India to 90 Percent Current Coverage (per Million Population) Income Quintile Outcome Total I II III IV V Tuberculosis deaths averted 80 40 25 12 3 0 Private expenditures averted (US$) 29,000 6,000 6,000 7,000 6,000 4,000 Insurance value (US$) 9,000 5,000 2,000 1,000 1,000 0 Source: Reproduced from table III of Verguet, Laxminarayan, and Jamison 2015. Note: Financial risk protection is measured as a money-metric value of insurance. 162 Disease Control Priorities: Improving Health and Reducing Poverty Figure 8.7 Use of Extended Cost-Effectiveness Analysis in the ECEA approach enables the design of benefits packages, Decision Making with the Inclusion of One Health Domain such as essential universal health coverage and the highest- (Deaths Averted by Policy) and One Nonhealth Domain priority package discussed in chapter 3 in this volume (Financial Risk Protection Provided by Policy) per Dollar (Watkins and others 2018), based on the quantitative Expenditure inclusion of information about how much nonhealth benefits can be bought, in addition to how much health can be bought, per dollar expenditure on health care (figures 8.6 and 8.7). Some health policies will rank higher on one metric Low health benefits, High health benefits, relative to another. ECEA allows policy makers to take both high FRP high FRP health and nonhealth outcomes into account when mak- ing decisions and thus to target scarce health care resources more effectively toward specific policy objectives. FRP NOTES Large parts of this chapter have been reproduced and Low health benefits, High health benefits, adapted from the following PharmacoEconomics publication: low FRP low FRP Verguet, S., J. J. Kim, and D. T. Jamison. 2016. “Extended Cost-Effectiveness Analysis for Health Policy Assessment: A Tutorial.” PharmacoEconomics 34 (9): 913–23. Licensed under Creative Commons Attribution (CC BY 4.0) available at: https:// creativecommons.org/licenses/by/4.0/ Deaths averted World Bank Income Classifications as of July 2014 are as follows, based on estimates of gross national income (GNI) Note: FRP = financial risk protection. As a simplification, the decision-making space can be divided into four quadrants: high health benefits and high FRP, high health benefits per capita for 2013: and low FRP, low health benefits and high FRP, and low health benefits and low FRP. • Low-income countries (LICs) = US$1,045 or less • Middle-income countries (MICs) are subdivided: ministries of finance in LMICs (figure 8.7). In this con- (a) lower-middle-income = US$1,046 to US$4,125 text, ECEAs may yield surprising results. Salti and others (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 (2016) found that tobacco taxation not only differentially • High-income countries (HICs) = US$12,746 or more. benefited the health of the poor, but it protected them from financial consequences of illness and thereby con- 1. Kim and others (2006) analyzed the effects of health system stituted a progressive tax. Second, it enables policy mak- constraints on optimal resource allocation, and Rheingans, ers to assemble a basic benefits package that takes into Atherly, and Anderson (2012) examined the distributional impact of rotavirus immunization. account how much health and how much FRP they can buy when designing the package. Depending on the pref- erences of policy makers and users, they can directly choose and optimize the benefits packages. REFERENCES Asaria, M., S. Griffin, R. Cookson, S. Whyte, and P. Tappenden. 2015. “Distributional Cost-Effectiveness Analysis of Health DISCUSSION Care Programmes: A Methodological Study of the UK Bowel Cancer Screening Programme.” Health Economics 24 (6): This chapter presents detailed methods for the broader 742–54. economic evaluation of health policies. ECEAs build on Ashok, A., A. Nandi, and R. Laxminarayan. 2015. “The Benefits CEAs by assessing consequences in both the health and of a Universal Home-Based Neonatal Care Package in nonhealth domains. Rural India: An Extended Cost-Effectiveness Analysis.” The ECEA approach is novel in that it includes equity In Disease Control Priorities (third edition): Volume 2, and nonhealth benefits (FRP) in the economic evaluation Reproductive, Maternal, Newborn, and Child Health, edited of health policies, which enables multiple criteria to be by R. E. Black, R. Laxminarayan, M. Temmerman, and included in the decision-making process. More important, N. Walker. Washington, DC: World Bank. Health Policy Analysis: Applications of Extended Cost-Effectiveness Analysis Methodology in DCP3 163 Baltussen, R., and L. Niessen. 2006. “Priority Setting of Health Measurement of Catastrophic Expenditures and Poverty.” Interventions: The Need for Multi-Criteria Decision Analysis.” Health Economics 17 (12): 1393–412. Cost Effectiveness and Resource Allocation 4 (August 21): 14. Jamison, D. T. 2009. “Cost Effectiveness Analysis: Concepts and Birch, S., and A. Gafni. 1992. “Cost-Effectiveness/Utility Applications.” In Oxford Textbook of Public Health, Vol. 2, Analyses: Do Current Decisions Lead Us to Where We Want edited by R. Detels, J. McEwen, R. Beaglehole, and H. Tanaka. to Be?” Journal of Health Economics 11 (3): 279–96. 5th ed. Oxford: Oxford University Press. Bleichrodt, H., E. Diecidue, and J. Quiggin. 2004. “Equity Weights Jamison, D. T. 2018. “Disease Control Priorities: Improving in the Allocation of Health Care: The Rank-Dependent Health and Reducing Poverty.” In Disease Control Priorities QALY Model.” Journal of Health Economics 23 (1): 157–71. (third edition): Volume 9, Disease Control Priorities: Brown, J. R., and A. Finkelstein. 2008. “The Interaction of Improving Health and Reducing Poverty, edited by D. T. Public and Private Insurance: Medicaid and the Long-Term Jamison, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, Care Insurance Market.” American Economic Review 98 (3): C. N. Mock, and R. Nugent. Washington, DC: World Bank. 1083–102. Jamison, D. T., J. G. Breman, A. R. Measham, G. Alleyne, Chisholm, D., K. A. Johansson, N. Raykar, I. Meggido, A. Nigam, M. Claeson, D. B. Evans, P. Jha, A. Mills, and P. Musgrove, and others. 2015. “Universal Health Coverage for Mental, eds. 2006. Disease Control Priorities in Developing Countries, Neurological, and Substance Use Disorders: An Extended second edition. Washington, DC: World Bank. Cost-Effectiveness Analysis.” In Disease Control Priorities Johansson, K. A., K. Bjerkreim Strand, A. Fekadu, and D. Chisholm. (third edition): Volume 4, Mental, Neurological, and 2016. “Health Gains and Financial Protection Provided by Substance Use Disorders, edited by V. Patel, D. Chisholm, the Ethiopian Mental Health Strategy: An Extended Cost- T. Dua, R. Laxminarayan, and M. Medina-Mora. Washington, Effectiveness Analysis.” Health Policy and Planning 1–8. DC: World Bank. doi:10.1093/heapol/czw134. Cleary, S. 2010. “Equity and Efficiency in Scaling Up Access to Johansson, K. A., C. Pecenka S. T. Memirie, D. T. Jamison, and HIV-Related Interventions in Resource-Limited Settings.” S. Verguet. 2015. “Health Gains and Financial Protection Current Opinion in HIV and AIDS 5 (3): 210–14. from Pneumococcal Vaccination and Pneumonia Treatment Cookson, R., M. Drummond, and H. Weatherly. 2009. “Explicit in Ethiopia: Results from an Extended Cost-Effectiveness Incorporation of Equity Considerations into Economic Analysis.” PLoS One 10 (12): e0142691. Evaluation of Public Health Interventions.” Health Economics, Kaplan, E. H., and M. H. Merson. 2002. “Allocating HIV- Policy, and Law 4 (2): 231–45. Prevention Resources: Balancing Efficiency and Equity.” Driessen, J. R., Z. D. Olson, D. T. Jamison, and S. Verguet. 2015. American Journal of Public Health 92 (12): 1905–7. “Comparing the Health and Social Protection Effects of Kemp-Benedict, E. 2001. “Income Distribution and Poverty Measles Vaccination Strategies in Ethiopia: An Extended Methods for Using Available Data in Global Analysis.” Cost-Effectiveness Analysis.” Social Science and Medicine PoleStar Technical Note 4, Stockholm Environment Institute 139 (August): 115–22. of the Global Scenario Group. http://gdrs.sourceforge.net Drummond, M. F., M. J. Sculpher, K. Claxton, G. L. Stoddart, /docs/PoleStar_TechNote_4.pdf. and G. W. Torrance. 2015. Methods for the Economic Kim, J. J., J. A. Salomon, M. C. Weinstein, and S. J. Goldie. 2006. Evaluation of Health Care Programmes. 4th ed. Oxford: “Packaging Health Services When Resources Are Limited: Oxford University Press. The Example of a Cervical Cancer Screening Visit.” PLoS Epstein, D., Z. Chalabi, K. Claxton, and M. Sculpher. 2007. Medicine 3 (11): e34. “Efficiency, Equity, and Budgetary Policies: Informing Krakauer, E., Z. Ali, H. Arreola, A. Bhadelia, S. Connor and others. Decisions Using Mathematical Programming.” Medical 2018. “Palliative Care in response to the Global Burden of Decision Making 27 (2): 128–37. Remediable Suffering.” In Disease Control Priorities (third Essue, B., T. Laba, F. M. Knaul, A. Chu, H. Van Minh, and edition): Volume 9, Disease Control Priorities: Improving others. 2018. “The Economic Burden of Chronic Ill-Health Health and Reducing Poverty, edited by D. T. Jamison, and Injuries for Households in Low- and Middle-Income R. Nugent, H. Gelband, S. Horton, P. Jha, R. Laxminarayan, Countries.” In Disease Control Priorities (third edition): and C. N. Mock. Washington, DC: World Bank. Volume 9, Disease Control Priorities: Improving Health and Kruk, M. E., E. Goldmann, and S. Galea. 2009. “Borrowing and Reducing Poverty, edited by D. T. Jamison, H. Gelband, Selling to Pay for Health Care in Low- and Middle-Income S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and Countries.” Health Affairs 28 (4): 1056–66. R. Nugent. Washington, DC: World Bank. Levin, C. L., M. Sharma, Z. Olson, S. Verguet, J. F. Shi, and others. Finkelstein, A., and R. McKnight. 2008. “What Did Medicare 2015a. “An Extended Cost-Effectiveness Analysis of Publicly Do? The Initial Impact of Medicare on Mortality and Out Financed HPV Vaccination to Prevent Cervical Cancer in of Pocket Medical Spending.” Journal of Public Economics China.” In Disease Control Priorities (third edition): Volume 3, 92 (7): 1644–68. Cancer, edited by H. Gelband, P. Jha, R. Sankaranarayanan, Fleurbaey, M., S. Luchini, C. Muller, and E. Schokkaert. 2013. and S. Horton. Washington, DC: World Bank. “Equivalent Income and Fair Evaluation of Healthcare.” Levin, C. L., M. Sharma, Z. Olson, S. Verguet, J. F. Shi, and Health Economics 22 (6): 711–29. others. 2015a. “An Extended Cost-Effectiveness Analysis Flores, G., J. Krishnakumar, O. O’Donnell, and E. van Doorslaer. of Publicly Financed HPV Vaccination to Prevent Cervical 2008. “Coping with Health-Care Costs: Implications for the Cancer in China.” Vaccine 33 (24): 2830–41. 164 Disease Control Priorities: Improving Health and Reducing Poverty Liu, L., Z. True, G. Fink, and S. Verguet. Forthcoming. “Estimating India under Universal Public Finance.” Cost Effectiveness and the Household Health and Financial Benefits of Malaria Resource Allocation 14: 9. Vaccine in Zambia: An Extended Cost-Effectiveness Analysis.” Rheingans, R. D., D. Atherly, and J. Anderson. 2012. Loganathan, T., M. Jit, R. Hutubessy, C. W. Ng, W. S. Lee, and “Distributional Impact of Rotavirus Vaccination in 25 GAVI others. 2016. “Rotavirus Vaccines Contribute towards Universal Countries: Estimating Disparities in Benefits and Cost- Health Coverage in a Mixed Public-Private Healthcare System.” Effectiveness.” Vaccine 30 (Suppl 1): A15–23. Tropical Medicine and International Health (August 24). Salem, A. B. Z., and T. D. Mount. 1974. “A Convenient Mauskopf, J. A., J. E. Paul, D. M. Grant, and A. Stergachis. 1998. Descriptive Model of Income Distribution: The Gamma “The Role of Cost-Consequence Analysis in Healthcare Density.” Econometrica 42 (6): 1115–27. Decision-Making.” PharmacoEconomics 13 (3): 277–88. Salti, N., E. D. Brouwer, and S. Verguet. 2016. “The Health, McClellan, M., and J. Skinner. 2006. “The Incidence of Financial, and Distributional Consequences of Increases in Medicare.” Journal of Public Economics 90 (1–2): 257–76. the Tobacco Excise Tax among Smokers in Lebanon.” Social Megiddo, I., A. R. Colson, D. Chisholm, T. Dua, A. Nandi, Science and Medicine 170: 161–69. and others. 2016. “Health and Economic Benefits of Public Sassi, F., L. Archard, and J. Le Grand. 2001. “Equity and the Financing of Epilepsy Treatment in India: An Agent-Based Economic Evaluation of Healthcare.” Health Technology Simulation Model.” Epilepsia 57 (3): 464–74. Assessment 5 (3): 1–138. Megiddo, I., A. R. Colson, A. Nandi, S. Chatterjee, S. Prinja, Segall, R. S. 1989. “Some Nonlinear Optimization Modeling for and others. 2014. “Analysis of the Universal Immunization Planning Objectives of Large Market-Oriented Systems: With Programme and Introduction of a Rotavirus Vaccine in an Application to Real Health Data.” Applied Mathematical India with IndiaSim.” Vaccine 32 (S1): A151–61. Modelling 13 (4): 203–14. Nandi, A., A. R. Colson, A. Verma, I. Megiddo, A. Ashok, and Shrime, M. G., S. Verguet, K. A. Johansson, D. Desalegn, others. 2016. “Health and Economic Benefits of Scaling Up D. T. Jamison, and M. E. Kruk. 2015. “Task-Sharing or Public a Home-Based Neonatal Care Package in Rural India: A Finance for Expanding Surgical Access in Rural Ethiopia: An Modeling Analysis.” Health Policy and Planning 31 (5): 634–44. Extended-Cost Effectiveness Analysis.” In Disease Control Nandi, A., I. Megiddo, A. Ashok, A. Verma, and R. Laxminarayan. Priorities (third edition): Volume 1, Essential Surgery, edited 2016. “Reduced Burden of Childhood Diarrheal Diseases by H. T. Debas, P. Donkor, A. Gawande, D. T. Jamison, M. E. through Increased Access to Water and Sanitation in India: A Kruk, and C. N. Mock. Washington, DC: World Bank. Modeling Analysis.” Social Science and Medicine. http://dx Shrime, M. G., S. Verguet, K. A. Johansson, D. Desalegn, .doi.org/10.1016/j.socscimed.2016.08.049. D. T. Jamison, and others. 2016. “Task-Sharing or Public Olson, Z. D., J. A. Staples, C. Mock, N. P. Nguyen, A. M. Bachani, Finance for the Expansion of Surgical Access in Rural and others. 2016. “Helmet Regulation in Vietnam: Impact Ethiopia: An Extended Cost-Effectiveness Analysis.” Health on Health, Equity, and Medical Impoverishment.” Injury Policy and Planning 31 (6): 706–16. Prevention 22 (4): 233–38. Smith, P. C. 2007. “Provision of a Public Benefit Package Olson, Z. D., J. A. Staples, C. N. Mock, N. P. Nguyen, alongside Private Voluntary Health Insurance.” In Private A. M. Bachani, and others. 2017. “Helmet Regulation Voluntary Health Insurance in Development: Friend or in Vietnam: Impact on Health, Equity, and Medical Foe? edited by A. Preker, R. M. Scheffer, and M. C. Bassett. Impoverishment.” In Disease Control Priorities (third Washington, DC: World Bank. edition): Volume 7, Injury Prevention and Environmental ———. 2013. “Incorporating Financial Protection into Decision Health, edited by C. N. Mock, O. Kobusingye, R. Nugent, Rules for Publicly Financed Healthcare Treatments.” Health and K. R. Smith. Washington, DC: World Bank. Economics 22 (2): 180–93. Pecenka, C., K. A. Johansson, S. T. Memirie, D. T. Jamison, Stinnett, A. A., and D. A. Paltiel. 1996. “Mathematical and S. Verguet. 2015. “Health Gains and Financial Risk Programming for the Efficient Allocation of Health Care Protection in Ethiopia: Assessing the Benefits of Diarrhea Resources.” Journal of Health Economics 15 (5): 641–53. Treatment and Prevention.” PLoS One 10 (12): e0142691. van Doorslaer, E., O. O’Donnell, R. P. Rannan-Eliya, A. Somanathan, Pillarisetti, A., D. T. Jamison, and K. R. Smith. 2017. “Effect of S. R. Adhikari, and others. 2006. “Effect of Payments for Health Household Energy Interventions on Health and Finances in Care on Poverty Estimates in 11 Countries in Asia: An Analysis Haryana, India: An Extended Cost-Effectiveness Analysis.” of Household Survey Data.” The Lancet 368 (9544): 1357–64. In Disease Control Priorities (third edition): Volume 7, Verguet, S. 2013. “Efficient and Equitable HIV Prevention: Injury Prevention and Environmental Health, edited by A Case Study of Male Circumcision in South Africa.” Cost C. N. Mock, O. Kobusingye, R. Nugent, and K. R. Smith. Effectiveness and Resource Allocation 11 (December): 1. Washington, DC: World Bank. Verguet, S., C. L. Gauvreau, S. Mishra, M. MacLennan, Postolovska, I., R. Lavado, G. Tarr, and S. Verguet. 2017. S. M. Murphy, and others. 2015. “The Consequences of “Estimating the Distributional Impact of Increasing Taxes Tobacco Tax on Household Health and Finances in Rich and on Tobacco Productions in Armenia: Results from an Poor Smokers in China: An Extended Cost-Effectiveness Extended Cost-Effectiveness Analysis.” Report. World Bank, Analysis.” The Lancet Global Health 3 (4): e206–16. Washington, DC. Verguet, S., J. J. Kim, and D. T. Jamison. 2016. “Extended Raykar N., A. Nigam, and D. Chisholm. 2016. “An Extended Cost-Effectiveness Analysis for Health Policy Assessment: Cost-Effectiveness Analysis of Schizophrenia Treatment in A Tutorial.” PharmacoEconomics 34 (9): 913–23. Health Policy Analysis: Applications of Extended Cost-Effectiveness Analysis Methodology in DCP3 165 Verguet, S., R. Laxminarayan, and D. T. Jamison. 2015. Wagstaff, A. 2010. “Measuring Financial Protection in Health.” “Universal Public Finance of Tuberculosis Treatment in In Performance Measurement for Health System Improvement, India: An Extended Cost-Effectiveness Analysis.” Health edited by P. C. Smith, E. Mossialos, I. Papanicolas, and Economics 24 (3): 318–32. S. Leatherman. Cambridge, U.K.: Cambridge University Press. Verguet, S., S. T. Memirie, and O. F. Norheim. 2016. “Assessing Wagstaff, A., and E. van Doorslaer. 2003. “Catastrophe the Burden of Medical Impoverishment by Cause: A and Impoverishment in Paying for Health Care: Its Systematic Breakdown by Disease in Ethiopia.” BMC Applications to Vietnam 1993–1998.” Health Economics Medicine 14 (1): 164. 12 (11): 921–34. Verguet, S., S. M. Murphy, B. Anderson, K. A. Johansson, Watkins, D. A., R. Nugent, and S. Verguet. 2017. “Extended R. Glass, and others. 2013. “Public Finance of Rotavirus Cost-Effectiveness Analyses of Cardiovascular Risk Factor Vaccination in India and Ethiopia: An Extended Cost- Reduction Policies.” In Disease Control Priorities (third Effectiveness Analysis.” Vaccine 31 (42): 4902–10. edition): Volume 5, Cardiovascular, Respiratory and Related Verguet, S., A. Nandi, V. Filippi, and D. A. P. Bundy. 2016. Disorders, edited by D. Prabhakaran, S. Anand, T. Gaziano, “Maternal-Related Deaths and Impoverishment among J. C. Mbanya, Y. Wu, and R. Nugent. Washington, DC: Adolescent girls in India and Niger: Findings from a World Bank. Modelling Study.” BMJ Open 6: e011586. Watkins, D. A., D. T. Jamison, A. Mills, C. N. Mock, R. Nugent, Verguet, S., A. Nandi, V. Filippi, and D. A. P. Bundy. 2017. and others. 2018. “Essential Universal Health Coverage.” In “Maternal Deaths and Impoverishment Averted by Disease Control Priorities (third edition): Volume 9, Disease Postponing Adolescent Parity in Developing Countries Control Priorities: Improving Health and Reducing Poverty, through Education: An Extended Cost-Effectiveness Analysis.” edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, In Disease Control Priorities (third edition): Volume 8, Child R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, and Adolescent Health and Development, edited by D. A. P. DC: World Bank. Bundy, N. de Silva, S. Horton, D. T. Jamison, and G. C. Patton. Watkins, D. A., Z. D. Olson, S. Verguet, R. Nugent, and D. T. Washington, DC: World Bank. Jamison. 2015. “Cardiovascular Disease and Impoverishment Verguet, S., Z. D. Olson, J. B. Babigumira, D. Desalegn, Averted due to a Salt Reduction Policy in South Africa: An K. A. Johansson, and others. 2015. “Health Gains and Extended Cost-Effectiveness Analysis.” Health Policy and Financial Risk Protection Afforded by Public Financing Planning (April 3). doi:10.1093/heapol/czv023. of Selected Interventions in Ethiopia: An Extended Cost- WHO (World Health Organization). 1999. World Health Effectiveness Analysis.” The Lancet Global Health 3 (5): Report 1999: Making a Difference. Geneva: WHO. e288–96. ———. 2000. World Health Report 2000: Health Systems: Verguet, S., C. Pecenka, K. A. Johansson, S. T. Memirie, I. K. Friberg, Improving Performance. Geneva: WHO. and others. 2016. “Health Gains and Financial Risk Protection ———. 2010. World Health Report 2010: Health Systems Afforded by Treatment and Prevention of Diarrhea and Financing, the Path to Universal Coverage. Geneva: WHO. Pneumonia in Ethiopia: An Extended Cost-Effectiveness ———. 2013. World Health Report 2013: Research for Universal Analysis.” In Disease Control Priorities (third edition): Health Coverage. Geneva: WHO. Volume 2, Reproductive, Maternal, Newborn, and Child Health, Xu, K., D. B. Evans, K. Kawabata, R. Zeramdini, J. Klavus, and edited by R. Black, R. Laxminarayan, M. Temmerman, and others. 2003. “Household Catastrophic Health Expenditure: N. Walker. Washington, DC: World Bank. A Multicountry Analysis.” The Lancet 362 (9378): 111–17. 166 Disease Control Priorities: Improving Health and Reducing Poverty Chapter 9 Benefit-Cost Analysis in Disease Control Priorities, Third Edition Angela Y. Chang, Susan Horton, and Dean T. Jamison INTRODUCTION: ROLE OF BENEFIT-COST We observe that BCA and CEA in the health sector represent two distinct cultures. The metric for ANALYSIS IN THE HEALTH SECTOR value in CEA can accommodate real health outcomes, A variety of economic methods is used for analysis in the such as child deaths averted, and aggregate measures, health sector. Other chapters in this volume summarize such as quality-adjusted life years (QALYs) or disability- the findings from Disease Control Priorities (third edi- adjusted life years (DALYs), as well as more granular tion) (DCP3) concerning cost-effectiveness analysis measures, such as malaria cases correctly treated. (CEA) and extended cost-effectiveness analysis (ECEA) When health benefits are measured in life years, both (Horton 2018; Verguet and Jamison 2018). This chapter the ages of the individuals and their remaining life summarizes the findings concerning benefit-cost analy- expectancies are implicitly factored into the analysis. sis (BCA). In contrast, in BCA, health benefits are often measured BCA has long been used for the analysis of public in terms of the number of statistical lives; ages and policy. The U.S. Secretary of the Treasury first used it remaining life expectancy of individuals are often not in 1808, and its use became mandatory for the U.S. considered. BCA involves an additional step of assign- Army Corps of Engineers in 1936. The U.S. Bureau of ing monetary value to health benefits; analysts are the Budget first issued guidelines for its use in 1952. required to explicitly assume a certain relationship Mills, Lubell, and Hanson (2008) suggest that BCA between the proportional change in this monetary became less well used for analysis of malaria eradica- value and the differences in countries’ income levels, tion around 1980, when CEA methods were becoming namely, income elasticity. This factor is often not con- well developed. More recently, there has been a resur- sidered in CEA. gence of interest in applying BCA to assess the viability The choice of applying CEA or BCA to evaluate of public investment programs and to set priorities economic benefits depends on the type of outcomes among a list of interventions (Jha and others 2015; produced by the health interventions. For some inter- Ozawa and others 2016). ventions, the main benefits include reduced mortality, BCA tends to be relatively readily understood by the improved quality of life, or reduced morbidity or general public, because the private sector uses analogous disability. For these outcomes, CEA works well and concepts. However, BCA also tends to raise controversies allows comparisons with other health interventions. because it assigns monetary values to outcomes (such as Many health interventions also affect future health care small changes in annual mortality probabilities) that requirements; preventive interventions, in particular, cannot be monetized according to many individuals. can reduce future health care costs. In CEA, these future Corresponding author: Angela Y. Chang, Harvard T. H. Chan School of Public Health, Boston, Massachusetts, United States; angela.chang@mail.harvard.edu. 167 cost reductions can be subtracted from current costs below 1 (likely owing to publication bias), a small of the intervention before comparing net costs to the number are in the 11–30 range, and a few outliers have health benefits. higher ratios. In part, this variation in results may stem Other interventions may improve health, but their from variations in the methodologies adopted. Some key outcomes are more easily expressed in monetary studies use methods of value per statistical life (VSL) terms. For example, supplementation or food fortifi- based on willingness to pay (for example, Alkire, cation with iron or iodine produces modest health Vincent, and Meara 2015; Cropper and others 2017). benefits in the form of reduced anemia and cretinism. Others assign dollar values to morbidity and mortality However, the most pervasive benefits accrue via averted (for example, Jamison and others 2013; Jha improved human capital—in this case, cognition and and others 2013; Stenberg and others 2016) or to mor- education—and thus BCA is more appropriate. The tality risk reduction (Fan, Jamison, and Summers eradication of a disease, such as smallpox, improves 2018; Jamison, Summers, and others 2013), using pro- health but can also save a substantial amount of ductivity or cost of illness averted to value years of life money through elimination of future prevention and lost. Of those assigning a value to mortality averted, treatment costs. Hence, BCA may be the most effec- only Stenberg and others (2016) include an explicit tive way to provide evidence of and advocate for this intrinsic value to life in excess of an assumed contribu- as a policy intervention. tion to, or share of, GDP. These methods are described A third group of interventions undertaken in sectors in more detail in the next section. outside health (for example, improvements in road safety, Several studies examine health interventions that safety regulations for vehicles, or water and sanitation) improve human capital and value the outcome are more naturally assessed by BCA methods. The invest- according to higher wages. These include interven- ment decisions are made in sectors that are accustomed tions in early child development and preschool to using BCA, and the investments with health benefits (Horton and Black 2017), school feeding and deworm- are being compared to other investments with outcomes ing (Fernandes and Aurino 2017) and programs to edu- that are assessed by BCA. CEA is more frequently used for cate school-age children and adolescents in health comparisons within the health sector; it has been refined prevention (Horton and others 2017). Other studies for specific policy purposes, such as the decision whether include future wages and averted future health care to allow insurance coverage of a particular new drug, costs in regard to malaria elimination (for example, technique, health technology, or diagnostic test within a Mills, Lubell, and Hanson 2008) and improvements in country, or for the prioritization of the use of donor sanitation (Hutton 2013; Whittington and others funds when international assistance is involved. (For alter- 2009). native approaches incorporating noneconomic consider- BCA findings were not surveyed and analyzed sys- ations, see also Norheim and others 2017.) tematically in all volumes (unlike CEAs), and thus we BCA, CEA, and ECEA are complementary techniques; can draw only tentative conclusions as to the areas each has value in addressing specific circumstances or where BCA is used most often. It is widely used in specific policy questions. This chapter summarizes the injury prevention and environmental health areas, and BCA findings from DCP3. It then examines the existing volume 7 (Mock and others 2017) has very few exam- methods for valuing life and considers possible improve- ples of CEA. Similarly, the analyses of pandemics and ments and ends with concluding comments. elimination or eradication of infectious diseases lend themselves to BCA. BCA is underrepresented in volume 2, because space did not permit the inclusion CONTRIBUTION OF DISEASE CONTROL of BCAs on nutrition, an area with many BCAs already PRIORITIES (THIRD EDITION) TO BCA IN THE (Black and others 2016). BCAs are scarcely visible in volume 3 (Gelband and others 2015) and volume 5 HEALTH SECTOR (Prabhakaran and others 2017). The focus of these The approaches in the DCP3 chapters and DCP- particular areas of noncommunicable diseases is on supported literature take many forms. Some directly health interventions more relevant to individuals than report benefit-cost ratios from existing literature, while populations and on treatment and screening of those others conduct their own BCA using primary data. Key individuals, which may make CEA methods more BCA findings and the methods applied are summarized appropriate. in tables 9.1 and 9.2. The next section considers the issues around the vari- Most of the benefit-cost ratios reported in tables 9.1 ation in methodology and associated effects on the mag- and 9.2 range from 1 to 10. Only one reported ratio is nitudes of BCA reported. 168 Disease Control Priorities: Improving Health and Reducing Poverty Table 9.1 Economic Burden of Disease, BCA, and Investment Cases in DCP3 Subject DCP3 reference Summary of key findings Method of valuing health or changes in mortality Essential Surgery Volume 1, chapter 21 • B/C of cleft lip and palate repair were 42 (income elasticity = • The base VSL was set at $7.4 million (2006 US$), and income elasticities (Alkire, Vincent, and 1.0) and 12 (income elasticity = 1.5), respectively. of 1.0 and 1.5 were applied when extrapolating to other countries. Age Meara 2015) • The median B/C of cesarean-section delivery for obstructed adjustment was applied, with the highest value of VSLY occurring at two- labor across countries is 4.0 (income elasticity = 1.5), ranging thirds of life expectancy. A 3 percent discount rate was applied. from 0.3 for the Democratic Republic of Congo to 76 for Gabon. Reproductive, Maternal, Volume 2, chapter 16 • Additional investments of $5 (2011 US$) per person per year Values for changes in mortality and morbidity and in consequences of decline Newborn, and Child (Stenberg and others in 74 countries with 95 percent of the global maternal and in fertility and unintended pregnancies were estimated using human capital Health 2016) child mortality burden would yield a B/C of 8.7 by 2035. methods. No age adjustment was applied. • B/C in low-, lower-middle-, and upper-middle-income • Mortality averted: The authors assigned an average benefit of 1.0 times the (excluding China) countries are 7.2, 11.3, and 6.1, GDP per capita for the direct economic benefits in terms of increased labor respectively, at 3 percent discount rate supply and productivity and an additional 0.5 times the GDP per capita for the social value of a life year. • Morbidity averted: A morbidity-to-mortality ratio of disability weights (namely, severity) was applied to estimate the social value of morbidity averted. • Positive economic and social consequences of decreases in fertility and reductions in unintended pregnancies: The economic benefit (expressed as percentage of GDP per capita) of this category was calculated by assuming different levels of decline in total fertility rate (TFR) and applying the model by Ashraf, Weil, and Wilde (2013) to calculate the effect of TFR reduction on Benefit-Cost Analysis in Disease Control Priorities, Third Edition GDP per capita. Major Infectious Volume 6, chapter 12 • B/C of malaria elimination programs surveyed by Mills, Various methods are used to value benefits (varies by study): Diseases: Malaria (Shretta and others Lubell and Hanson (2008) range from 2.4 in the Philippines to • Elimination of costs required to control malaria 2017) 4.1 and 9.2 for control in India, 17.1 for elimination in Greece to almost 150 in Sri Lanka. • Productivity gains (labor, land, or both) • B/C of global malaria reduction and elimination between • Modeled macroeconomic growth benefits 2013 and 2015 is estimated at 6.1 (Purdy and others 2013) • B/C of malaria eradication efforts between 2015 and 2040 is estimated to be 17 (Gates and Chambers 2015). Major Infectious Volume 6, chapter • B/C of interventions to end NTDs is 25 between 1990 • The benefits of the interventions include only health expenditure and lost Diseases: NTDs 17 (Fitzpatrick and and 2030. The benefits include health expenditure and wages averted. No value was assigned to the intrinsic value of mortality risk others 2017) lost wages averted, estimated at around $657 billion reduction. (international dollars) between 2011 and 2030. Total cost of the investment is estimated at US$27 billion. A discount rate of 3 percent per annum was applied for both benefits and costs table continues next page 169 170 Disease Control Priorities: Improving Health and Reducing Poverty Table 9.1 Economic Burden of Disease, BCA, and Investment Cases in DCP3 (continued) Subject DCP3 reference Summary of key findings Method of valuing health or changes in mortality Injury Prevention and Volume 7, chapter 9 B/C from Hutton (2013) and Whittington and others (2009): • Health estimates based on direct health costs (treatment of water- and Environmental Health: (Hutton and Chase • Networked water and sewerage services: 0.7 sanitation-related disease), productivity losses during illness, and mortality Environment 2017) losses were measured using human capital. • Deep borehole with public hand pump: 4.6 • Estimates also include reduced travel and access time for water and • Total sanitation campaign (South Asia): 3.0 sanitation owing to improvements. • Household water treatment (biosand filters): 2.5 • Improved water supply: 2.0 • Improved sanitation: 5.5 Injury Prevention and Volume 7, chapter 13 B/C of installing flue-gas desulfurization units at every coal-fired • Empirical estimates of the VSL in India range widely, from US$50,600 Environmental Health: (Cropper and others power plant in India is greater than 1, for all reasonable VSL (Bhattacharya, Alberini, and Cropper 2007) to US$362,000 (Madheswaran Environment 2017) estimates applied 2007) (2007 US$). • Transferring the U.S. VSL to India at current exchange rates, using an income elasticity of 1, suggests a VSL of US$250,000 (2006 US$). Child and Adolescent Volume 8, chapter 24 B/C for the following interventions: • Benefits include improved cognition and greater school grade attainment, Health and (Horton and Black • Videos on early childhood development shown to parents which translate into higher wages and employment. Same pathway exists for Development: Early 2017) with children age 2 years and younger waiting in health all interventions (except sprinkles, which reduce anemia and then also has childhood centers, followed by group discussion: 5.3 (Walker and same effects). others 2015) • Psacharopoulos (2014) study does not fully incorporate the cost of all • Responsive stimulation and nutrition intervention (sprinkles) interventions, hence the incredibly high B/C ratio. for children age 2 years and younger: 1.5 (López Boo, Palloni, and Urzua 2014) • Home visiting program that educates mothers with children age 2 years and younger in child development: 2.6–3.6 (Berlinski and Schady 2015) • Preschool programs for children ages 3 to 5 years: generally exceed 3 (Berlinski and Schady 2015) • Nutritional add-on to preschool: 77 (Psacharopoulos 2014) • Overall, B/C of a well-designed and well-implemented early childhood program is in the range of 2 to 5. table continues next page Table 9.1 Economic Burden of Disease, BCA, and Investment Cases in DCP3 (continued) Subject DCP3 reference Summary of key findings Method of valuing health or changes in mortality Child and Adolescent Volume 8, chapter • School feeding programs with micronutrient fortification had • Benefits are assumed to be gained through improved education outcomes over Health and 25 (Fernandes and estimated B/C of 3 and 7 for low- and lower-middle-income the lifetime of targeted children and to translate into improved productivity Development: school- Aurino 2017) countries, respectively (2012 US$, discount rate 3 percent). and contributions to GDP. No intrinsic value of health improvements was age children The average cost of school feeding is US$56 in low- and included. lower-middle-income countries Child and Volume 8, chapter 26 B/C for adolescent health in high-income countries is as follows: • Benefits included health care costs averted, human capital gains (via Adolescent Health (Horton and others • Education sessions with children ages 11–12 years and education, reduced mortality), and reduced costs of crime (for alcohol and drug and Development: 2017) parents and other interventions for alcohol use in the United interventions). adolescents States: range of 5 to 100 (McDaid and others 2014) • School-based smoking programs in Germany: 3.6 (McDaid and others 2014) • Programs to promote mental well-being in the United States: range of 5 to 28 (McDaid and others 2014) • Programs for reduced drug dependency, smoking, and delinquency in the United States: 25 (McDaid and others 2014) Disease Control Volume 9, chapter 18 • The total cost of a pandemic is presented as a sum of its • The values of a 1-in-10,000 mortality risk reduction for one year for a person Priorities: Improving (Fan, Jamison, and effect on income and the intrinsic value of lives prematurely age 35 years were set at 0.7, 1.0, 1.3, and 1.6 percent of income per capita for Health and Reducing Summers 2018) lost and illness suffered (Fan, Jamison, and Summers, 2018). low-, lower-middle-, upper-middle-, and high-income countries, respectively. Benefit-Cost Analysis in Disease Control Priorities, Third Edition Poverty: Pandemic flu • For the first dimension, the authors estimated the expected This amount was then adjusted for ages other than age 35 years in proportion annual income losses globally of US$16 billion for to the ratio of life expectancies at those ages to life expectancy at age moderately severe pandemics and US$64 billion for severe 35 years. pandemics. • For the second dimension, they estimated the expected annual loss for the whole world from the intrinsic cost as 0.6 percent of global income and variation by income group, from 0.3 percent in high-income countries to 1.6 percent in lower-middle-income countries. • In total, the expected annual inclusive cost, reflecting both dimensions above, amounts to about 0.7 percent (US$570 billion per year) of global income, with income losses accounting for a small fraction of inclusive costs (12 percent) for severe pandemics, but a larger fraction (40 percent) for moderately severe pandemics. Note: B/C = benefit/cost; GDP = gross domestic product; NTDs = neglected tropical diseases, VSL = value per statistical life; VSLY = value per statistical life year. 171 172 Disease Control Priorities: Improving Health and Reducing Poverty Table 9.2 Economic Burden of Disease, BCA, and Investment Cases Supported by DCP3 Subject Reference Summary of key findings Method of valuing health or changes in mortality Global Health 2035 grand Jamison, Summers, and • The recommended set of investments to • The value of a 1-in-10,000 mortality risk reduction for one year for a 35-year- convergence others (2013) scale up health technologies and systems old person was set at 1.8 percent of income per capita, assuming an income in LMICs, compared to a scenario of elasticity of 1.0. This was then adjusted for ages other than age 35 years in stagnant investment and no improvements proportion to the ratio of life expectancies at those ages to life expectancy at age in technology, would yield a B/C of 9 in 35 years, using the historical Japanese life table. lower-income countries and 20 in lower- • Four different age adjustment scenarios were applied: no adjustment, reducing middle-income countries over a 20-year progress in children under age 4 years by 50 percent, excluding all children under period. age 10 years from the calculation, and excluding over-70 mortality. Under the second age adjustment scenario, the value of a life year is 2.3 times the per person income. Infectious disease and maternal Jamison, Jha, and others Recommended investment solutions and B/Cs • US$1,000 per DALY was applied to value the health benefits gained; it roughly health (2013) are as follows: equals the lower end of the proposed value of a statistical life year of 2 to 4 1. Tuberculosis: Appropriate case finding and times per capita income of low-income countries. US$5,000 per DALY was used treatment, including dealing with MDR for sensitivity analysis. TB—15 • The DALYs were discounted at 3 percent, and the DALY cost of a typical death 2. Malaria: Subsidy for appropriate under age 5 years was reduced by 50 percent. For DALYs accrued near the time treatment via Affordable Medicines of birth, a smoothing formula using the concept of acquisition of life potential Facility–malaria—35 was applied to assign greater weights to DALYs resulting from deaths of a fetus. 3. Childhood diseases: Expanded immunization coverage—20 4. HIV: Accelerated vaccine development—11 5. Essential surgery: Management of difficult childbirth, trauma, and other—10 6. Deworming of schoolchildren—10 table continues next page Table 9.2 Economic Burden of Disease, BCA, and Investment Cases Supported by DCP3 (continued) Subject Reference Summary of key findings Method of valuing health or changes in mortality NCDs Jha and others (2013) Key investment priorities and B/Cs are as • Same method as the Copenhagen Consensus on infectious disease (Jamison and follows: others 2013b) was applied. 1. Tobacco taxation: 40 2. Acute management of heart attacks with low-cost drugs: 25 3. Salt reduction: 20 4. Hepatitis B immunization: 10 5. Secondary prevention of heart attacks and strokes with 3–4 drugs in a generic risk pill: 4 Rheumatic heart disease Watkins and Chang (2017) Economic burden of RHD found to be • The value of a 1-in-10,000 mortality risk reduction for one year for a 35-year-old approximately US$64.8 billion, or an average of person in the United States was set at $900. These were adjusted downward US$ 360,000 per preventable death in low- and for low- and middle-income countries based on average GDP per capita in middle-income countries each region, assuming an income elasticity of 1.0. This was then adjusted for ages other than age 35 years in proportion to the ratio of region-specific life expectancies at those ages to life expectancy at age 35 years. • Sensitivity analyses conducted for income elasticity (0.6 and 1.5), anchoring age (from age 35 years to ages with remaining life expectancy of 45 years). Note: DALY = disability-adjusted life year; HIV = human immunodeficiency disease; MDR = multidrug-resistant; NCDs = noncommunicable diseases; RHD = rheumatic heart disease; TB = tuberculosis. Benefit-Cost Analysis in Disease Control Priorities, Third Edition 173 USE OF THE VALUE PER STATISTICAL LIFE IN Development (OECD) as two reasonable starting points. ESTIMATING BCA IN THE HEALTH SECTOR In the United States, a simple average of the values applied by three regulatory agencies is US$9.3 million Several of the DCP3 chapters and related articles build on (Robinson and Hammitt 2015b; U.S. DOT 2015; U.S. the concept of the VSL to estimate the intrinsic value of EPA 2014), which translates into a VSLr of roughly 180. health improvements. The VSL is defined as the marginal OECD (2012, 2014) proposed a VSL of US$3.6 million rate of substitution between money and mortality risk in and a VSLr of roughly 100, which is much lower than the a defined time period. It is typically calculated by divid- U.S. estimates. ing individuals’ willingness to pay for a small change in Several considerations need to be made when extrap- their own risks in a defined time period by the risk olating existing estimates to other populations. The VSL change. For example, individuals have a VSL of US$9 is expected to vary, depending on the characteristics of million if they are willing to pay US$900 for a 10−4 reduc- those affected (for example, health status, age, life expec- tion in mortality risk in the current year. Note that tancy, and income) and the characteristics of the risks money is used as a measure to reflect the trade-offs indi- (for example, latency, morbidity before death, voluntar- viduals are willing to make, and it is not itself important. iness, and controllability). However, the effects of many Jamison, Summers, and others (2013) argue that termi- of these characteristics need further research. There are nology should be used in cases where the risk change significant inconsistencies and gaps in the available liter- units are close to those actually measured so that one ature, even for HICs (Hammitt 2017, Robinson and avoids the occasionally contentious interpretations of Hammitt 2015b; Viscusi and Masterman 2017b). The value of life (Chang and others 2017). They propose that most commonly adjusted characteristic is income, possi- risk be measured in source measure units (SMUs), or bly because both theoretical and empirical evidence are units of 10−4. Rather than referring to the value of a statis- readily available (although consensus on the magnitude tical life, they propose referring to the value of an SMU of adjustments one should make between countries with (VSMU). In the example just provided, the risk change varying income levels is still lacking). Other important was 1 SMU and the associated VSMU was US$900. Most characteristics, such as the average age or remaining life published VSL studies focus on the risks of accidental expectancy of those affected, are often ignored. deaths, mainly among adult populations in high-income settings (Lindhjem and others 2011; Robinson and Relationship to Income Hammitt 2015b; Viscusi 2015; Viscusi and Aldy 2003). Research on the relationship between income and the Far fewer VSL studies are conducted in low- and VSL generally indicates that the VSL increases as income middle-income countries (LMICs), and the quality of the increases. However, the proportional change in the VSL papers varies widely (Bhattacharya, Alberini, and Cropper in response to a change in real income—its income 2007; Guo and Hammitt 2009; Hammitt and Zhou 2006; elasticity—is uncertain (Robinson and Hammitt 2015a). Hoffmann and others 2012; Shanmugam 2001; Simon Income elasticity is of particular importance in estimat- and others 1999; Tekes¸ in and Ara 2014; Vassanadumrongdee ing the VSL for lower-income countries because changing and Matsuoka 2005; Viscusi and Masterman 2017a). the elasticity can affect the resulting VSL by orders of Under this limitation, analyses that value health improve- magnitude (equations 9.1 and 9.2) (Hammitt and ments in LMICs often rely on studies from high-income Robinson 2011). (In equations 9.1 and 9.2, r = ratio of countries (HICs) as their base VSL estimates, and these are VSL to GDP per capita and pc = per capita.) adapted on the basis of some characteristics of the popu- elasticity lation of interest. This section discusses the common ⎛ GDP per capita x ⎞ VSL country x = VSL US × ⎜ (9.1) practices, as well as the challenges, that analysts face in ⎝ GDP per capita US ⎟ ⎠ using previously established values for another setting of interest (also known as benefit transfer) and provides an VSL country x alternative to existing methods. VSL r = GDP per capita country x elasticity ⎛ GDPpc country x ⎞ Current Practice of Benefit Transfer in Global Health VSL US × ⎜ ⎝ GDPpc US ⎟ ⎠ Selection of Base VSL or VSL-to-Income Ratio = (9.2) Benefit transfer often begins with selecting a base VSL or GDPpc country x a VSL-to-income ratio (VSLr). We consider the VSL −1) GDPpc(elasticity country x estimates produced by major U.S. regulatory agencies = VSL US × and the Organisation for Economic Co-operation and GDPpc elasticity US 174 Disease Control Priorities: Improving Health and Reducing Poverty Empirical studies comparing VSL estimates from Figure 9.1 VSLr, with VSL Extrapolated from the U.S. VSL with HICs and middle-income countries (MICs), as well as Income Elasticity of 1.2 between higher- and lower-income groups in the United 240 States, support the use of elasticity greater than 1.0 when applying VSL across income levels (Biausque 2012; Costa 220 and Kahn 2004; Hammitt and Ibarrarán 2006; Kniesner, 200 Viscusi, and Ziliak 2010). However, similar research has 180 not been conducted in low-income countries (LICs). Nevertheless, the global meta-analysis in Lindhjem and 160 VSLr others (2011) and OECD (2012) for OECD countries 140 United States 180 yielded the estimate of 0.8 (range 0.7–0.9). Figure 9.1 120 illustrates the relationship between VSLr and income when an income elasticity of 1.2 is applied across coun- 100 tries, using the U.S. VSLr of 180 as the base. If elasticity 80 of 1 were applied, all countries would face the same VSLr 60 of 180. With greater income elasticity, countries with 0 25,000 50,000 75,000 100,000 125,000 150,000 greater GDP per capita will behave a higher VSLr, with 2013 GDP per capita, US$ (PPP) the highest occurring in Qatar at 217. For LMICs, the VSLr drops exponentially, with the lowest VSLr occur- Note: GDP = gross domestic product; PPP = purchasing power parity; VSLr = value per statistical ring in the Central African Republic at 73. life-to-income ratio. One issue with extrapolating the VSL from a higher- to a lower-income setting is that the VSL may fall below the expected income or consumption in the relevant willingness to pay to reduce risks to children is likely to period in the lower-income country. Theory suggests be larger than the value adults place on reducing risks to that the VSL will exceed the present value of future earn- themselves, although the magnitude of the difference ings and of future consumption, both of which vary by varies across studies. For example, Hammitt and age, because it reflects the intrinsic value of living in Haninger (2010) found that willingness to pay for risk addition to an individual’s productivity or consumption. reduction is nearly twice as large for children than for Accordingly, the VSL is expected to at least equal the adults. To date, we are unaware of a general consensus in present value of future income, as well as consumption, the BCA community on how to adjust the value of risk discounted to the age at which the risk reduction occurs change for differences in age. (Hammitt and Robinson 2011). Age and life expectancy are related but distinct con- cepts. As Sanderson and Scherbov (2007) stated, a per- Relationship to Age and Life Expectancy son has two different ages: the retrospective age, which is Because the VSL cannot be directly estimated from mar- a measure of how many years one has already lived, and ket measures such as earnings or consumption, research- the prospective age—remaining life expectancy—which ers instead rely on revealed or stated preference studies. reflects how many years a person will live. For example, The former estimates the value of risk reductions based a person age 35 years in 1960 and a person age 35 years on related market transactions or behavior, often on the in 2015 likely would have different levels of willingness relationship between wages and occupational risks in the to pay for mortality risk reduction, because they would case of the VSL. Some of these studies found an inverse have had different perceptions of how much longer they U-shaped relationship; the VSL increased in young will live. This distinction is important in transferring adulthood, peaked in middle age, and then declined, base VSL from an HIC to an LIC. Comparing the consistent with the patterns of income and consumption remaining life expectancies of persons at age 35 years in predicted under the lifecycle models (Rosen 1988; 2015 in Lesotho (the lowest life expectancy at birth), the Shepard and Zeckhauser 1982, 1984). Others found that United States, and Japan (the highest life expectancy at the values for older adults decrease or remain constant birth), one finds that the average person in Lesotho faced (Evans and Smith 2006; Krupnick 2007). One limitation a 26-year life expectancy, while a person in the United of the revealed preference method is that it addresses States and Japan faced 45 years and 49 years, respectively only working age populations. Stated preference meth- (UNDP 2015). Intuitively, all else equal, we would expect ods instead involve surveying respondents to determine lower willingness to pay among people in Lesotho, given their willingness to pay for risk reductions of various the lower number of years remaining. However, no types. Some stated preference studies suggest that adult empirical data support this claim. Benefit-Cost Analysis in Disease Control Priorities, Third Edition 175 As an illustration, in figure 9.2 we estimate the VSLr ⎛ remaining life expectancy ⎞ for all countries, based on the ratio of the remaining ⎜ (35)country x ⎟ life expectancy at age 35 years of persons of a selected VSL r country x = VSL r US × ⎜ ⎟ remaining life expectancy country and of the United States (equation 9.3). (In ⎜ ⎝ (35)US ⎟ ⎠ equation 9.3, r = ratio of VSL to GDP per capita.) The figure shows a narrower range of the VSLr across (9.3) countries, because the differences among remaining life expectancies are smaller than among income lev- els. The lowest VSLr occurs in Lesotho, the country Alternative Approaches with the lowest life expectancy, at a VSLr of 101, and Given the limited theoretical and empirical evidence on highest in Japan, at 194. the appropriate framework to account for transferring the value of mortality risk reduction to populations with Figure 9.2 VSLr Extrapolated with the Ratio of Remaining Life different characteristics, we propose five simple and Expectancies at Age 35 Years for Persons in Selected Countries and defensible alternative approaches to incorporate these the United States key characteristics. We start with the two VSLr described earlier as the starting point (VSLr = 180 and 100), and United States 180 we estimate the VSLr for each World Bank income group 200 in table 9.3.1 The first approach ([1] in table 9.3) is to not apply 180 any adjustments based on income or age and to assume that the VSLr remains the same across all 160 populations. The second approach ([2] in table 9.3) makes 140 income adjustments by applying an elasticity of 0.8 VSLr for HICs and 1.2 for all other countries, based on 120 equation 9.2, to the VSLr. We use 2013 GDP per capita in U.S. dollars (PPP) for each income group. 100 The third approach applies age and life expectancy adjustment ([3] in table 9.3) by assuming that the 80 value decreases proportional to remaining life expec- tancy. This method reflects common practices in the 60 health economics literature, and specifically in CEA in 0 20,000 40,000 60,000 80,000 the health sector, in which the units of health benefits 2013 GDP per capita, US$ (PPP) are in life years, rather than, for example, lives saved. Note: GDP = gross domestic product; PPP = purchasing power parity; VSLr = value per statistical These analyses implicitly assume that the VSL decreases life–to–income ratio. in proportion to remaining life expectancy and that Table 9.3 Estimated VSLr for Four Alternative Approaches, World Bank Income Anchor VSL Alternative options HICs UMICs LMICs LICs US 180 [1] No adjustment 180 180 180 180 US 180 [2] Income adjustment 191 137 115 88 US 180 [3] Age adjustment 80 81 104 117 US 180 [4] Income and age adjustment 85 62 66 57 OECD 100 [1] No adjustment 100 100 100 100 OECD 100 [2] Income adjustment 101 80 67 51 OECD 100 [3] Age adjustment 44 45 58 65 OECD 100 [4] Income and age adjustment 45 36 39 33 Note: HICs = high-income countries; LICs = low-income countries; LMICs = lower-middle-income countries; OECD = Organisation for Economic Co-operation and Development; UMICs = upper-middle-income countries; VSL = value per statistical life; VSLr = VSL-to-income ratio. 176 Disease Control Priorities: Improving Health and Reducing Poverty saving the life of a younger person with higher remain- functional form to represent the relationship between ing life expectancy has a greater yield than saving the VSLr and income. For example, one function form that life of an older person. To estimate the changes in meets the criteria is as follows: VSLr, we first collected the most recent (2010–15) age-specific death rates for all four income groups VSLr(y ) = 115 + 70 sin(y n ) (9.5) (UNDP 2015) and used the 2015 world population distribution to create age-standardization for the dis- tribution of deaths. Assuming that the value of risk where yn is the normalized 2013 GDP per capita in reduction decreases proportional to remaining life U.S. dollars (PPP). expectancy, we then applied a ratio of the remaining life expectancy at that age and at age 35 years for each x −a (9.6) yn = age group (equation 9.4). b−a where x is the country’s income level. We set a (where Age-adjusted VSLrj = Base VSLr sin(yn) = 0) as the average income of upper-middle- e (a )ij income countries and b as the average income of = 0 world population size i × death rate ij × ∑ i21 e ( 35 ) j non-OECD HICs. We excluded the following small × countries with very high income levels to simplify the ∑ 21 i =0 world population size i × death rate ij analysis: Qatar, Luxembourg, Kuwait, and Singapore. (9.4) We present this relationship between VSLr and income level under the scenario in figure 9.3 and the implied VSL as a function of income under this analysis in where j is income group, i is age group (0, 1–4, 5–9, and figure 9.4. We constrained the U.S. VSLr to be approx- so on up to 95+), e(a)ij is the remaining life expectancy imately at 180. The lowest VSLr occurs in the Central at age a in age group i in the jth income group, and African Republic, an LIC with a 2013 GDP per capita e(35)j is the remaining life expectancy of 35 year olds in of US$603, and the highest VSLr occurs in several the jth income group. HICs, including Iceland and the Netherlands, with the The fourth approach combines the second and third 2013 GDP per capita ranging from US$42,000 to approach to adjust for both differences in income and in US$46,000. Under this formulation, the income elas- age and life expectancy ([4] in table 9.3). ticities in LMICs and HICs are approximately 1.2 and The fifth and final approach involves using an alter- 0.9, respectively. native functional form that incorporates different char- acteristics. This varies substantially from the previous four approaches, which are all built on the same func- tional form commonly applied in the VSL literature Figure 9.3 VSLr Extrapolated with the Sine Function (equation 9.2). In searching for an appropriate func- 200 tional form to calculate the VSLr for countries, we set the following criteria that we consider important when 180 transferring VSLr from one country to another: 160 140 1. The base VSLr is set roughly at the U.S. average of 120 180 or the OECD’s estimate at 100 (for purpose of VSLr 100 illustration, we use the former in the calculation that follows equation 9.5). 80 2. Following the income elasticity literature, we apply 60 an elasticity of roughly 0.8 for HICs and 1.2 for 40 LMICs. 20 3. All VSLr should be above the income floor, namely, the VSLr should not be lower than the discounted 0 0 20,000 40,000 60,000 remaining life expectancy. 2013 GDP per capita, US$ (PPP) We found that the sine function can approximately Note: GDP = gross domestic product; PPP = purchasing power parity; VSLr = value per statistical meet these criteria and could therefore be an appropriate life–to–income ratio. Benefit-Cost Analysis in Disease Control Priorities, Third Edition 177 Figure 9.4 Implied VSL, Based on the Sine Function Extrapolation of A possible proposal is that each BCA (or economic bur- the VSLr den of disease assessment) would select its own values for key parameters while also reporting standardized 12 sensitivity analyses to enable accumulation of compar- ative knowledge. 10 Second, more empirical VSL estimates from low- and middle-income countries are needed. The current prac- 8 tice of benefit transfer does not adequately reflect the VSL (US$) different characteristics between populations, and we 6 believe this inadequacy leads to inaccurate estimations of the population’s willingness to pay. Having empirical 4 estimates of VSL from a diverse set of populations will fill an important research gap in this field. 2 Third, advances in BCA also need to be harmonized with the evolution in thinking about thresholds for 0 0 2 4 6 cost-effectiveness. We know that VSL methods tend to assign large values to health because they focus on willing- 2013 GDP per capita, US$ (PPP) x 10,000 ness to pay without specific reference to ability to pay. Note: GDP = gross domestic product; PPP = purchasing power parity; VSLr = value per statistical At the same time, recent studies (Claxton and others 2015; life–to–income ratio. Ochalek, Lomas, and Claxton 2015) have shown that the public tends to undervalue public dollars spent on health care, acting as if a DALY (one year of enjoyment of full CONCLUSIONS health) is worth only 50 percent of per capita at the mar- This chapter reviews estimates of B/C ratios from DCP3 gin. If this methodological issue is not resolved, health and illustrates the large number of applications of the policy makers will overspend on health interventions technique to the health sector. Two major streams of assessed by BCA (for example, environmental interven- methods are used within the health sector for B/C esti- tions, injury prevention, and human capital promotion) mation in DCP3. One uses willingness to pay and the and underspend on those assessed primarily by CEA VSL concept. The other uses a human capital measure, (used to decide between many curative interventions). analyzing costs of lost productivity because of morbidity This is an important area for future work. and mortality or improved productivity associated with Finally, both CEA and BCA entail implicit ethical improved cognition. The literature on VSL is evolving, judgments. An approach using BCA that incorporates and we have presented current thinking on how that considerations of future wages gives a larger weight to evolution might continue. The following research prior- individuals who are of working age, to those with ities are recommended for future examination. higher labor force participation rates (men compared First, standardization of the assumptions within to women), and to urban populations as compared to each methodology would be useful. Currently, actual rural populations. These same groups (working-age differences across alternative interventions are obscured population, men, urban residents) also tend to have by variations in methods and assumptions. Disagreements higher health-care expenditures and, hence, also about how the VSL should vary with population char- receive greater weight in benefit calculations of future acteristics are built on both empirical and normative health expenditures averted. Because benefits mea- arguments. The human capital side lacks consistency of sured in CEA are denominated in years of health, they rules for valuing future years of human life: Do we use are less subject to bias by gender, higher income, and current GDP? Do we use rates of actual growth per residence. However, they share similar ethical con- capita of countries? Do we use a common measure of cerns as do measures of the global burden of disease. expected growth, for example, 2 percent per capita per Years of life saved for someone who suffers from a annum? This lack of consistency makes the comparison disability or mental illness are valued less than those of estimates challenging. Estimates made in different for someone who is free of disability, for example. For sectors with different traditions is part of the problem. these reasons, a common compromise between CEA The development of a reference case would help. Such a and BCA methods is to assign the same VSL to every- reference case is being supported by the Bill & Melinda one within a country. These topics may be usefully Gates Foundation, in part as a follow-up to DCP3. examined in future research. 178 Disease Control Priorities: Improving Health and Reducing Poverty NOTES Analysis of the Value of a Life Year Estimates.” Journal of Global Health 7 (1). doi:10.7189/jogh.07.010401. World Bank Income Classifications as of July 2014 are as fol- Claxton, K., S. Martin, M. Soares, N. Rice, E. Spackman, and lows, based on estimates of gross national income (GNI) per others. 2015. “Methods for the Estimation of the National capita for 2013: Institute for Health and Care Excellence Cost-Effectiveness Threshold.” Health Technology Assessment 19 (14): 1–503. • Low-income countries (LICs) = US$1,045 or less doi:10.3310/hta19140. • Middle-income countries (MICs) are subdivided: Costa, D. L., and M. E. Kahn. 2004. “Changes in the Value of (a) lower-middle-income = US$1,046 to US$4,125 Life, 1940–1980.” Journal of Risk and Uncertainty 29 (2): (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 159–80. • High-income countries (HICs) = US$12,746 or more. Cropper, M., S. Guttikunda, P. Jawahar, K. Malik, and I. Partridge. 2017. “Costs and Benefits of Installing Scrubbers 1. These scenarios build on conversations among an informal at Coal-Fired Power Plants in India.” In Disease Control group of researchers interested in developing standardized Priorities (third edition): Volume 7, Injury Prevention and VSL sensitivity analyses to enhance the comparability Environmental Health, edited by C. N. Mock, O. Kobusingye, of assessments of global health and development issues. R. Nugent, and K. Smith. Washington, DC: World Bank. The group was initially convened by Dean Jamison and Evans, M. R., and V. K. Smith. 2006. “Do We Really Understand Maureen Cropper in February 2016 and ultimately grew the Age-VSL Relationship?” Resource and Energy Economics to include over 30 participants as of April 2016. Major 28 (3): 242–61. contributors included Kenneth Arrow, Nils Axel Braathen, Fan, V. Y., D. T. Jamison, and L. H. Summers. 2018. “The Angela Y. Chang, Rob Dellink, James K. Hammitt, Michael Inclusive Loss from Pandemic Influenza Risk.” In Disease Holland, Alan Krupnick, Elisa Lanzi, Urvashi Narain, Ståle Control Priorities (third edition): Volume 9, Disease Navrud, Lisa A. Robinson, Rana Roy, and Christopher Sall, Control Priorities: Improving Health and Reducing Poverty, among others. The analysis presented here uses these dis- edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, cussions as a starting point, but it has not been reviewed or R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, approved by that group. DC: World Bank. Fernandes, M., and E. Aurino. 2017. “Identifying an Essential Package for School-Age Child Health: Economic Analysis.” REFERENCES In Disease Control Priorities (third edition): Volume 8, Child and Adolescent Health and Development, edited by D. A. P. Alkire, B. C., J. R. Vincent, and J. G. Meara. 2015. “Benefit-Cost Bundy, N. de Silva, S. Horton, D. T. Jamison, and G. Patton. Analysis for Selected Surgical Interventions in Low- and Washington, DC: World Bank. Middle-Income Countries.” In Disease Control Priorities Fitzpatrick, C., U. Nwankwo, E. Lenk, S. J. de Vlas, and (third edition): Volume 1, Essential Surgery, edited by D. A. P. Bundy. 2017. “An Investment Case for Ending H. Debas, P. Donkor, A. Gawando, D. T, Jamison, M. Kruk, Neglected Tropical Diseases.” In Disease Control Priorities and C. N. Mock. Washington, DC: World Bank. (third edition): Volume 6, Major Infectious Diseases, edited Ashraf, Q., D. Weil, and J. Wilde. 2013. “The Effect of Fertility by K. K. Holmes, S. Bertozzi, B. Bloom, and P. Jha. Reduction on Economic Growth.” Population and Washington, DC: World Bank. Development Review 39 (1): 97–130. Gates, B., and R. Chambers. 2015. “From Aspiration to Action: Berlinski, S., and N. Schady, eds. 2015. The Early Years: Child What Will It Take to End Malaria?” United Nations, New Well-Being and the Role of Public Policy. Washington, DC: York, and the Bill & Melinda Gates Foundation, Seattle, Inter-American Development Bank. WA. https://www.mmv.org/sites/default/files/uploads/docs Bhattacharya, S., A. Alberini, and M. L. Cropper. 2007. “The /publications/Aspiration-to-Action.pdf. Value of Mortality Risk Reductions in Delhi, India.” Journal Gelband, H., P. Jha, R. Sankaranarayanan, and S. Horton, eds. of Risk and Uncertainty 34 (1): 21–47. 2015. Disease Control Priorities (third edition): Volume 3, Biausque, V. 2012. “The Value of Statistical Life: A Meta-Analysis.” Cancer. Washington, DC: World Bank. Report ENV/EPOC/WPNEP(2010)9/FINAL, Working Party Guo, X., and J. K. Hammitt. 2009. “Compensating Wage on National Environmental Policies, Organisation for Differentials with Unemployment: Evidence from China.” Economic Development and Co-operation. http://www Environmental and Resource Economics 42 (2): 187–209. .oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote- doi:10.1007/s10640-008-9217-9. =ENV/EPOC/WPNEP(2010)9/FINAL&doclanguage=en. Hammitt, J. K. 2017. “Extrapolating the Value per Statistical Life Black, R., R. Laxminarayan, M. Temmerman, and N. Walker. between Populations: Theoretical Implications.” Journal of 2016. Disease Control Priorities (third edition): Volume Benefit-Cost Analysis 8 (2), forthcoming. 2, Reproductive, Maternal, Newborn, and Child Health. Hammitt, J. K., and K. Haninger. 2010. “Valuing Fatal Risks Washington, DC: World Bank. to Children and Adults: Effects of Disease, Latency, and Chang, A. Y., L. A. Robinson, J. K. Hammitt, and S. C. Resch. Risk Aversion.” Journal of Risk and Uncertainty 40 (1): 2017. “Economics in ‘Global Health 2035’: A Sensitivity 57–83. Benefit-Cost Analysis in Disease Control Priorities, Third Edition 179 Hammitt, J. K., and M. E. Ibarrarán. 2006. “The Economic Costs and Benefits, edited by B. Lomborg, 137–69. Value of Fatal and Non-Fatal Occupational Risks in Mexico Cambridge, U.K.: Cambridge University Press. City Using Actuarial- and Perceived-Risk Estimates.” Health Kniesner, T., W. K. Viscusi, and J. P. Ziliak. 2010. “Policy Economics 15 (12): 1329–35. doi:10.1002/hec.1137. Relevant Heterogeneity in the Value of Statistical Life: Hammitt, J. K., and L. A. Robinson. 2011. “The Income New Evidence from Panel Data Quantile Regressions.” Elasticity of the Value per Statistical Life: Transferring Journal of Risk and Uncertainty 40 (1): 15–31. doi:10.1007 Estimates between High and Low Income Populations.” /s11166-009-9084-y. Journal of Benefit-Cost Analysis 2 (1): 1–29. Krupnick, A. 2007. “Mortality-Risk Valuation and Age: Stated Hammitt, J. K., and Y. Zhou. 2006. “The Economic Value of Preference Evidence.” Review of Environmental Economics Air-Pollution-Related Health Risks in China: A Contingent and Policy 1 (2): 261–82. Valuation Study.” Environmental and Resource Economics 33 Lindhjem, H., S. Navrud, N. A. Braathen, and V. Biausque. 2011. (3): 399–423. doi:10.1007/s10640-005-3606-0. “Valuing Mortality Risk Reductions from Environmental, Hoffmann, S., P. Qin, A. Krupnick, B. Badrakh, S. Batbaatar, Transport, and Health Policies: A Global Meta-Analysis of and others. 2012. “The Willingness to Pay for Mortality Risk Stated Preference Studies.” Risk Analysis 31 (9): 1381–407. Reductions in Mongolia.” Resource and Energy Economics 34 doi:10.1111/j.1539-6924.2011.01694. (4): 493–513. doi:10.1016/j.reseneeco.2012.04.005. López Boo, F. L., G. Palloni, and S. Urzua. 2014. “Cost-Benefit Horton, S. 2018. “Cost-Effectiveness in Disease Control Analysis of a Micronutrient Supplementation and Early Priorities, Third Edition.” In Disease Control Priorities (third Childhood Stimulation Program in Nicaragua.” Annals of edition): Volume 9, Disease Control Priorities: Improving the New York Academy of Sciences 1308: 139–48. Health and Reducing Poverty, edited by D. T. Jamison, Madheswaran, S. 2007. “Measuring the Value of Statistical H. Gelband, S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, Life: Estimating Compensating Wage Differentials among and R. Nugent. Washington, DC: World Bank. Workers in India.” Social Indicators Research 84 (1): 83–96. Horton, S., and M. Black. 2017. “Identifying an Essential McDaid, D., A.-L. Park, C. Currie, and C. Zanotti. 2014. Package for Early Child Development: Economic Analysis.” “Investing in the Wellbeing of Young People: Making the In Disease Control Priorities (third edition): Volume 8, Economic Case.” In The Economics of Wellbeing: Vol. 5, Child and Adolescent Health and Development, edited by Wellbeing: A Complete Reference Guide, edited by D. McDaid D. A. P. Bundy, N. de Silva, S. Horton, D. T. Jamison, and and C. L. Cooper, 181–214. Oxford: Wiley-Blackwell. G. Patton. Washington, DC: World Bank. Mills, A., Y. Lubell, and K. Hanson. 2008. “Malaria Eradication: Horton, S., J. Waldfogel, E. De la Cruz Toledo, J. Mahon, The Economic, Financial and Institutional Challenge.” and J. Santelli. 2017. “Economic Factors in Defining Malaria Journal 7 (Supp 1): S11. the Adolescence Package.” In Disease Control Priorities Mock, C. N., O. Kobusingye, R. Nugent, and K. R. Smith, eds. 2017. (third edition): Volume 8, Child and Adolescent Health Disease Control Priorities (third edition): Volume 7, Injury and Development, edited by D. A. P. Bundy, N. de Silva, Prevention and Environmental Health. Washington, DC: S. Horton, D. T. Jamison, and G. Patton. Washington, DC: World Bank. World Bank. Norheim, O. F., E. Emanuel, D. T. Jamison, K. A. Johansson, Hutton, G. 2013. “Global Costs and Benefits of Reaching J. Millum, and others. 2017. Global Health Priority-Setting: Universal Coverage of Sanitation and Drinking-Water Beyond Cost-Effectiveness. New York: Oxford University Supply.” Journal of Water and Health 11 (1): 1–12. Press. Hutton, G., and C. Chase 2017. “Water Supply, Sanitation, Ochalek, J., J. Lomas, and K. Claxton. 2015. “Cost per and Hygiene.” In Disease Control Priorities (third edition): DALY Averted Thresholds for Low- and Middle-Income Volume 7, Injury Prevention and Environmental Health, Countries: Evidence from Cross Country Data.” CHE edited by C. N. Mock, O. Kobusingye, R. Nugent, and Research Paper 122, Centre for Health Economics, K. Smith. Washington, DC: World Bank. University of York, York, U.K. Jamison D. T., P. Jha, V. Malhotra, and S. Verguet. 2013. OECD (Organisation for Economic Co-operation and “Human Health: The Twentieth-Century Transformation Development). 2012. Mortality Risk Valuation in of Human Health: Its Magnitude and Value.” In How Much Environment, Health and Transport Policies. Paris: OECD Have Global Problems Cost the World?, ed. by B. Lomborg. Publishing. http://www.oecd.org/env/tools-evaluation Cambridge University Press. /mortalityriskvalua tioninenvironmenthealthandtrans Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, portpolicies.htm. S. Berkley, and others. 2013. “Global Health 2035: A World ———. 2014. The Cost of Air Pollution: Health Impacts of Converging within a Generation.” The Lancet 382 (9908): Road Transport. OECD Publishing. http://www.oecd 1898–955. .org / env/the-cost-of-air-pollution-9789264210448-en Jha, P., K. Jordan, R. Hum, and C. Gauvreau. 2015. .htm. “Health.” Copenhagen Consensus Center, Copenhagen. Ottersen, T., J. Millum, J. Prah Ruger, S. Verguet, K. A. Johansson, http://www.copenhagenconsensus.com/publication/ and others. 2017. “The Future of Priority Setting in Global post-2015-consensus-health-assessment-jha-et-al. Health.” In Global Health Priority-Setting: Beyond Cost- Jha, P., R. Nugent, S. Verguet, D. Bloom, and R. Hum. 2013. Effectiveness, edited by O. F. Norheim and others. New York: “Chronic Disease.” In Global Problems, Smart Solutions: Oxford University Press. 180 Disease Control Priorities: Improving Health and Reducing Poverty Ozawa, S., S. Clark, A. Portnoy, S. Grewal, L. Brenzel, and edited by R. E. Black, R. Laxminarayan, M. Temmerman, D. G. Walker. 2016. “Return on Investment from Childhood and N. Walker. Washington, DC: World Bank. Immunization in Low- and Middle-Income Countries, Tekes¸in, C., and S. Ara. 2014. “Measuring the Value of Mortality 2011–20.” Health Affairs 35 (2): 199–207. Risk Reductions in Turkey.” International Journal of Prabhakaran, D., S. Anand, T. Gaziano, J.-C. Mbanya, Y. Wu, Environmental Research and Public Health 11 (7): 6890–922. and R. Nugent. 2017. Disease Control Priorities (third edi- doi:10.3390/ijerph110706890. tion): Volume 5, Cardiovascular, Respiratory, and Related UNDP (United Nations Population Division). 2015. 2015 Disorders.Washington, DC: World Bank. Revision of World Population Prospects. New York: UNDP. Psacharopoulos, G. 2014. “Education Assessment Paper: http://esa.un.org/unpd/wpp/. Benefits and Costs of the Education Targets for the Post- U.S. DOT (U.S. Department of Transportation). 2015. 2015 Development Agenda.” Copenhagen Consensus “Guidance on Treatment of the Economic Value of a Center, Tewksbury, MA. http://www.copenhagenconsensus Statistical Life (VSL) in Departmental Analyses—2015 .com/publication/post-2015-consensus-education-assess Adjustment. Memorandum to Secretarial Officers and ment-psacharopoulos. Modal Administrators from K. Thomson, General Counsel, Purdy, M., M. Robinson, K. Wei, and D. Rublin. 2013. “The and C. Monje, Assistant Secretary for Policy.” U.S. DOT, Economic Case for Combating Malaria.” American Journal Washington, DC. of Tropical Medicine and Hygiene 89 (5): 819–23. U.S. EPA (U.S. Environmental Protection Agency). 2014. Robinson, L. A., and J. K. Hammitt. 2015a. “The Effect of Guidelines for Preparing Economic Analyses. Washington, Income on the Value of Mortality and Morbidity Risk DC: U.S. EPA. Reductions.” Prepared for Industrial Economics, Vassanadumrongdee, S., and S. Matsuoka. 2005. “Risk Perceptions Inc., Cambridge, MA, on behalf of the U.S. and Value of a Statistical Life for Air Pollution and Traffic Environmental Protection Agency, Washington, Accidents: Evidence from Bangkok, Thailand.” Journal of DC. https://yosemite .epa.gov /sab/SABPRODUCT Risk and Uncertainty 30 (3): 261–87. .NSF/0/0CA9E925C9A702F285257F380050C842/$File/IEc_ Verguet, S., and D. T. Jamison. 2018. “Health Policy Income%20elasticity%20Report%20_final.pdf. Assessment: Applications of Extended Cost-Effectiveness ———. 2015b. “Valuing Reductions in Fatal Illness Risks: Analysis Methodology in Disease Control Priorities.” In Implications of Recent Research.” Health Economics 25 (8): Disease Control Priorities (third edition): Volume 9, Disease 1039–52. doi:10.1002/hec.3214. Control Priorities: Improving Health and Reducing Poverty, Rosen, S. 1988. “The Value of Changes in Life Expectancy.” edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, Journal of Risk and Uncertainty 1 (3): 285–304. R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, Sanderson, W. C., and S. Scherbov. 2007. “A New Perspective DC: World Bank. on Population Aging.” Demographic Research 16 (January): Viscusi, W. K. 2015. “The Role of Publication Selection Bias 27–58. doi:10.4054/DemRes.2007.16.2. in Estimates of the Value of a Statistical Life.” American Shanmugam, K. R. 2001. “Self Selection Bias in the Estimates of Journal of Health Economics 1 (1): 27–52. Compensating Differentials for Job Risks in India.” Journal Viscusi, W. K., and J. E. Aldy. 2003. “The Value of a Statistical of Risk and Uncertainty 22 (3): 263–75. Life: A Critical Review of Market Estimates throughout the Shepard, D. S., and R. Zeckhauser. 1982. “Life-Cycle World.” Journal of Risk and Uncertainty 27 (1): 5–76. Consumption and Willingness to Pay for Increased Viscusi, W. Kip, and Clayton J. Masterman. 2017a. “Income Survival.” Institute for Research on Poverty, University of Elasticities and Global Values of a Statistical Life.” Journal of Wisconsin-Madison, WI. Benefit-Cost Analysis: 1–25. ———. 1984. “Survival versus Consumption.” Management Viscusi, W. K., and C. J. Masterman. 2017b. “Anchoring Biases Science 30 (4): 423–39. in International Estimates of the Value of a Statistical Life.” Shretta, R., J. Liu, C. Cotter, J. Cohen, C. Dolenz, and others. Journal of Risk and Uncertainty 54 (2): 103–28. doi:10.1007 2017. “Malaria Elimination and Eradication.” In Disease /s11166-017-9255-1. Control Priorities (third edition): Volume 6, Major Infectious Walker, S. P., C. Powell, S. M. Chang, H. Baker-Henningham, Diseases, edited by K. K. Holmes, S. Bertozzi, B. Bloom, and S. Grantham-McGregor, and others. 2015. “Delivering P. Jha. Washington, DC: World Bank. Parenting Interventions through Health Services in the Simon, N. B., M. L. Cropper, A. Alberini, and S. Arora. 1999. Caribbean: Impact, Acceptability and Costs.” Working “Valuing Mortality Reductions in India: A Study of Paper IDB-WP-642, Inter-American Development Bank, Compensating Wage Differentials.” Policy Research Working Washington, DC. Paper 2078, World Bank, Washington, DC. http://documents Watkins, D. A., and A. Chang. 2017. “The Economic Impact .worldbank.org/curated/en/582261468751139383/pdf/multi- of Rheumatic Heart Disease in Low- and Middle-Income page.pdf. Countries.” Working Paper 19, DCP3 Working Paper Series, Stenberg, K., K. Sweeney, H. Axelson, M. Temmerman, and World Bank, Washington, DC. P. Sheehan. 2016. “Returns on Investment in the Continuum Whittington, D., W. M. Hanemann, C. Sadoff, and M. Jeuland. of Care for Reproductive, Maternal, Newborn, and Child 2009. “The Challenge of Improving Water and Sanitation Health.” In Disease Control Priorities (third edition): Volume Services in Less Developed Countries.” Foundations and 2, Reproductive, Maternal, Newborn and Child Health, Trends in Microeconomics 4 (6): 469–607. Benefit-Cost Analysis in Disease Control Priorities, Third Edition 181 Part 4 Health System Topics from Disease Control Priorities, Third Edition Chapter 10 Quality of Care John Peabody, Riti Shimkhada, Olusoji Adeyi, Huihui Wang, Edward Broughton, and Margaret E. Kruk INTRODUCTION health of the patient. Simply put, this chapter looks at Just after dawn, Vivej arrives at the hospital with her the decisions and actions of the provider when seeing a newborn under her arm to see you. She is 21 years old, patient. It is at this critical moment when we expect the two days postpartum, and exhausted after 36 hours of doctor or nurse, or whoever is caring for the patient, to protracted labor. She is worried because she cannot get provide the best possible care by skillfully combining the her firstborn, Esmile, to breastfeed. You learn that she available resources and technologies with the best clini- delivered at a birthing clinic near her home and tells cal evidence and professional judgment. you that, even after her water broke, it took more than a Esmile and Vivej received poor-quality care at the day before the birth attendant could deliver her son. time of delivery. Several clinical steps were not taken. Your examination reveals a dire clinical picture: Esmile The prolonged rupture of membranes was not diag- is lethargic and hypotonic, he has a poor suck reflex, his nosed in a timely manner. Vivej needed either to have temperature is 39.8°C, his pulse is 180, and his breath- her labor induced or, failing that, to be referred for a ing is labored. You check his white blood count, con- cesarean section. Prophylactic antibiotics should have firming leukocytosis. A spinal tap shows pleocytosis. been administered. Just as important, the provider at the You start him on fluids and antibiotics for neonatal birthing center needed support and professional over- sepsis with likely meningitis and quickly turn your sight, with guidelines, supervision, or default referral attention to Vivej. Her situation is easier to diagnose but systems in place to provide a path to the best care possi- no less urgent: she is febrile and tachycardic, her blood ble. The multiple failures in this case led to puerperal pressure is 85/50. You give her fluids and start her on and neonatal sepsis. At worst, these conditions have a antibiotics. Ultimately, despite your efforts, both mother fatality rate greater than one in four; at best, they lead to and child die. protracted care, recovery, and clinical expense that could What went wrong? This chapter looks narrowly at have been avoided. It is possible, however, to imagine these situations—the critical points after access and providers in a different setting, with the same physical availability (including affordability) are already accom- resources, giving better care and avoiding this tragic plished, when patients are in health care facilities that are scenario. staffed and equipped with appropriate technology. These In the next section, we answer the questions raised in are the situations in which the inputs are brought this scenario and in countless clinics and hospitals together and it is up to the provider to improve the around the world. How much variation is there in the Corresponding author: John Peabody, professor, University of California, San Francisco and University of California, Los Angeles; president, QURE Healthcare, San Francisco, California, United States; jpeabody@qurehealthcare.com. 185 quality of care? How do we measure clinical practice? Clinicians failed to provide even the most basic care— How and where has quality been systematically improved only 12 percent of standardized patients who reported a and practice variation reduced? What elements of care child with symptoms of dysentery were told to give their variation can be addressed by policy and what are the child oral rehydration therapy (Das and others 2012). A costs? Most important, what can be done to elevate the study of 296 providers in India found that a mere 6 care given by providers in developing country settings? percent followed the six diagnostic standards of the Our focus, therefore, is on the steps that can be taken to International Standards for Tuberculosis Care (Achanta optimize the quality of care for patients like Esmile in and others 2013). pediatrics, Vivej in obstetrics, and other patients receiv- Such deficits in quality of care can come from many ing care for the clinical conditions considered through- sources, including gaps in knowledge, inappropriate out the nine volumes of the third edition of Disease application of available technology, and inability of Control Priorities (DCP3). organizations to monitor and support care standardiza- tion. This striking variation in quality within countries occurs across facilities, among providers, and between PROBLEM OF VARIATIONS IN QUALITY specialists and nonspecialists (Beracochea and others 1995; Das and Hammer 2007; Das and others 2012; OF CARE Dumont and others 2002; Nolan and others 2001; Health policy makers, researchers, and clinicians recog- Peabody, Gertler, and Leibowitz 1998; Weinberg 2001; nize the wide variations in access to care (Peabody and Xu and others 2015). others, forthcoming). However, once individuals and Some cross-national comparisons have reached the populations avail themselves of health care services, vari- same conclusion. A 2007 DCP-sponsored study that ations in health outcomes raise disturbing questions about evaluated quality for three common clinical conditions the quality of care delivered, defined as “the degree to in five countries simultaneously found that the average which health services for individuals and populations quality of care was low in every country (61 percent) and increase the likelihood of desired outcomes and are con- the difference in average score between countries was sistent with professional knowledge” (IOM 2013, 21). small (ranging from 60.2 to 62.6 percent). However, the Variations in care entail policy challenges similar to those quality scores within every country varied widely, rang- associated with variations in access, including equity and ing from 30 to 93 percent (Peabody and Liu 2007). This efficiency (Saleh, Alameddine, and Natafgi 2014). In stud- wide variation was constant across type of facility, med- ies comparing clinical practice with evidence-based stan- ical condition, and domain of care. dards, researchers found that high-quality care is provided Poor health outcomes are the result of many factors, inconsistently to large segments of the population ranging from the nature and severity of disease to patient (McGlynn and others 2003). For example, a landmark behavior and structural elements of care (IOM and Institute of Medicine report found that, in the United National Academy of Engineering 2011; Steinwachs and States, medical errors kill more people than traffic acci- Hughes 2008; Xu and others 2015). Some factors are not dents (Kohn, Corrigan, and Donaldson 2000). amenable to change (genetic predisposition), while others Many subsequent studies have documented varia- are slow to affect outcomes (changes in payment incen- tions in quality of care in low- and middle-income tives). Discouragingly, better access, more infrastructure, countries (LMICs) (Barber, Bertozzi, and Gertler and structural measures of quality do not always translate 2007; Barber, Gertler, and Harimurti 2007; Hansen and into better health outcomes. Indeed, some structural others 2008; Loevinsohn, Guerrero, and Gregorio 1995; indexes can be inversely related to health (for example, National Academies of Sciences, Engineering, and number of hospital beds versus health status) (Ng and Medicine 2015; Peabody, Nordyke, and others 2006; others 2014). Thus, improving the quality of care may World Bank 2003). In India, studies have found alarm- well provide the greatest sectoral opportunity to improve ingly low rates of correct diagnosis, limited adherence to health outcomes (Peabody and others 2017). Care can be treatment guidelines, and frequent use of harmful or improved quickly and, if based on best evidence, improved unnecessary drugs. In one study, only 31 percent of care will improve outcomes and lower costs (Scott and Jha standardized patients who described symptoms of 2014). Reducing unwarranted variation and addressing unstable angina and 48 percent who reported symptoms poor-quality provider practices deserve the most urgent of asthma were given the correct drugs (Das and attention possible from policy makers (Kirkpatrick and Hammer 2014). Even more worrying, providers pre- Burkman 2010; Ransom, Pinsky, and Tropman 2000). scribed an incorrect or harmful treatment to more than Providers, health care systems, governments, and 60 percent of patients reporting asthma symptoms. payers are beginning to recognize this urgency and are 186 Disease Control Priorities: Improving Health and Reducing Poverty seeking innovative, effective ways to improve the quality QUALITY IMPROVEMENT INFRASTRUCTURE of care. Metrics and measurement, pathways, clinical REQUIREMENTS checklists, educational interventions, and payment incentives all raise awareness and offer opportunities to Clinical solutions are typically not generalizable because provide accountability and improve care. These they are disease-specific, vary by clinical condition, and approaches have been tried in many LMICs, but their rely on the training of health care providers and the con- effectiveness varies. Changing practice at the system level text of the health care system (Dayal and Hort 2015). is difficult and requires coordination, vision, planning, Policy, however, is designed to work at the group level— and consideration of how effective, high-impact inter- that is, at scales larger than the individual level. Effective ventions can be scaled up and applied across an entire quality improvement policies that work at the group system (Massoud and Abrampah 2015). At the level of level have several common features, specifically the individual providers, knowledge improvement and means to collect information and synthesize it and the acquisition of new skills need to be motivated by both means to encourage skills and technologies to be applied extrinsic and intrinsic factors, which are enabled through in a timely fashion. The following four common policy access to knowledge and measurement tools that change attributes, detailed below, improve quality: behavior and ideally are accompanied by peer support (Schuster, Terwoord, and Tasosa 2006; Woolf 2000). We • Measurement of the clinical activity (including mea- have learned that improved clinical practice requires surement tied to feedback) active participation (not passive learning), peer and • Standards for those measurements (based on scien- leadership support, and communication of relevant tific evidence for standardizing care) feedback (Kantrowitz 2014; Mostofian and others 2015). • Training of providers (including supervision) Multifaceted interventions seem more successful than • Incentives that align and motivate providers (includ- single interventions, underscoring the importance of ing financial incentives, but also incentives of profes- practice-level change that focuses on supporting the sionalism and reputation). individual provider (training) and creating a suitable environment for change (accountability). Even more challenging than finding disease-level Measurement interventions for individual providers is identifying Accurate, affordable, and valid measurements “are health care policies that improve the quality of care the basis for quality of care assessments” (Peabody and for populations. While clinical practice interventions, others 2004, 771). For too long, routine measures of such as checklists, for acute and chronic diseases quality in LMICs relied on structural elements (ros- work at the provider-patient level, policies need to ters, catalogs, and inventories of coverage and access), address group-level practice, for example, through giving little thought to how these elements improve incentives and indirect means. Preventing the deaths health. Such elements are relatively easy to count of Vivej and Esmile, for example, would have required and measure, but are only remotely linked to better the timely use of simple uterotonic commodities outcomes. Improving quality requires measurement and prophylactic antibiotics, which might happen of the care process—that is, what providers do when with better supervision. An effective policy, however, they see patients (Ansong-Tornui and others 2007; compels groups of providers to set up the supervi- Peabody, Taguiwalo, and others 2006; Peabody and sion or the training that leads to the use of oxytocin or others 2011). cephalosporins. Measurement of the care process is critical, creating In the second edition of DCP, the chapter on quality awareness of deficits in practice, gaps in care, and of care largely summarized the emerging policy evidence accountability at the individual and system levels, which that better quality could lead to better outcomes improves focus and motivation. To serve as an instru- (Peabody, Taguiwalo, and others 2006). Just a decade ment of change and accountability, provider-level mea- later, every volume in this edition discusses quality of surement needs to be ongoing and cyclical. Transparency care. We consider in this chapter the different policy of results can increase knowledge and change intentions, interventions that have been tried around the world. We but requires a supportive context to be effective (National begin with the quality infrastructure that is required for Patient Safety Foundation 2015). every policy intervention, then expand on the policy When coupled with useful feedback and done in a framework for changing clinical practice, and use this timely manner, measurement is the foundation for expanded framework to discuss the challenges, returns, improving quality. If the measures are reliable, afford- and costs of improved quality. able, and anchored in valid, evidence-based criteria, Quality of Care 187 quality of care can be followed over time and the impact The available methods for measuring performance of policy interventions can be assessed (Felt-Lisk and include provider self-reports, patient vignette simula- others 2012). Various quality measures have been devel- tions, patient self-reports, and reviews of medical oped, each with its own set of advantages and disadvan- records. These methods vary in their ability to capture tages. Although no measure is perfect, adequate measures improvement and account for differences in the type of exist, and every health system—from small clinics to patients treated (case-mix adjustment). They also vary in national governments—can benefit from measurement. their economic feasibility (Epstein 2006; Spertus and Feedback has the potential to promote improvement, others 2003), reliability (repeated measures), validity but studies are limited, tending to focus on health care (against a gold standard), and ability to be “gamed” report cards (Baker and Cebul 2002; Dranove and others (Petersen and others 2006). The policy challenge is that 2003; Kolstad 2013; Shaller and others 2003), which performance-measurement methods may need to be include public disclosure of quality scores that may not developed and adapted to low-resource settings provide the same motivation to improve scores as when (Engelgau and others 2010). Table 10.1 lists available feedback is provided privately. methods for measuring quality of the care process. Table 10.1 Methods for Measuring Quality of the Care Process Method Advantage Issues Chart abstraction or • Nearly ubiquitous and theoretically could be obtained after • May lack relevant clinical details, especially when written for review of medical the patient-provider encounter; in practice, record keeping other purposes, such as legal protection record in most LMICs is inadequate • Poor record keeping and documentation lead to incomplete and • Electronic medical record technology: improved uniformity, inaccurate content legibility, communication • Illegibility of handwritten notes • Records of clinical events • Inaccuracies in the process of abstracting to produce data suitable for analysis • High costs involved in training medical abstractors • Variation in documentation practices across providers, facilities, and countries Direct observation • Records of clinical events • Ethical considerations and recording of • First-hand observation of actual encounters • Need to inform providers and patients, which can induce the visits Hawthorne effect (bias when participant changes his or her behavior as a result of being evaluated) • High cost of training observers • Variations across observers Administrative data • Available in most facilities • Lack sufficient clinical detail • Ubiquitous and inexpensive to collect when data collection • Inaccuracies in content system is in place • Poor data collection or management systems, especially in LMICs Standardized • The gold standard for process measurement • Expensive patients • Captures technical and interpersonal elements of process • Not practical for routinely evaluating quality • Reliable over a range of conditions, providing valid • Limited range of applicability (works best for adult conditions and measurements that accurately capture variation in clinical conditions that can be simulated) practice among providers across patients Clinical vignettes • Can measure quality within a group of providers and • Potential resistance of providers to complete the vignettes evaluate quality at the population level • Different methods for administering vignettes • Responsive to variations in quality • Instrument validation • Cases simulate actual patient visit and evaluate • Link to patient-level data physician’s knowledge • Validated against other methods and criteria for standard- of-quality measurement • Useful for comparison studies • Easy and inexpensive to administer • Ability to collect data independently Sources: Bertelsen 1981; Peabody and others 2004; Peabody and others 2011; Peabody, Nordyke, and others 2006. Note: LMICs = low- and middle-income countries. 188 Disease Control Priorities: Improving Health and Reducing Poverty The usefulness of any method for measuring process are the need to staff projects and train evaluators, which depends on the completeness and accuracy of the data can be difficult to scale up. Ethical challenges must be collected—a ubiquitous problem with charts, medical addressed, and both providers and patients must be records, and administrative data. Another significant informed of the observation or recording. Although concern is patient case mix, given that different patient research performed in Tanzania showed that the characteristics may affect quality (Zaslavsky 2001). Hawthorne effect can disappear after 10 to 15 observa- Validity and comparability of results across measure- tions, this notification introduces participation bias ment units (individual patients, providers, facilities, and when providers change their behavior as a result of countries) are questionable unless these differences are being evaluated (Leonard and Masatu 2006). Perhaps a controlled for through complex instrument design and more salient problem is that trained observers are statistical techniques (Peabody and others 2004). costly, and variation between observers is difficult to Operational concerns, such as the need for highly remedy. These challenges have stimulated the search for trained staff, can increase the cost and complexity of other ways to measure and record what happens in implementing some methods. clinical visits. Data Derived from Medical Charts Administrative Data Chart abstraction, or review of the medical record, has Administrative data are available in all but the poorest long been used to measure quality of care. Clinical settings. A data collection system, once established, can audits, physician report cards, and profiles are based on provide information on charges and many cost inputs. chart abstraction. Reliable health records can provide However, administrative data are assembled for pur- credible evidence of the health status of patients and poses other than improving quality, such as document- assist policy makers with developing plans and making ing and processing medical claims (Calle and others decisions to improve health care delivery (Haux 2006). 2000; Goeree and others 2009), and often lack sufficient The core strength of the medical record is that it is ubiq- clinical detail to be useful in evaluating clinical pro- uitous and could potentially be obtained after each cesses. In a 2004 study, an incorrect diagnosis was encounter. recorded 30 percent of the time (although the actual Chart reviews, however, suffer from many problems. diagnosis was correct). The actual diagnosis was recorded First, the medical chart must be completed (and found) only 57 percent of the time (Peabody and others 2004). to proceed with an abstraction. Handwritten notes on As information systems advance, accuracy may improve, paper charts may be illegible. Medical charts may be but the lack of adequate clinical detail will continue to generated for reasons other than documenting the key limit the use of administrative data. Clinical databases clinical events of the visit (for legal protection or such as registries may be helpful but are primarily avail- obtaining payment) and thus may lack crucial clinical able only in high-income countries (HICs) and for details. Luck and others (2000) found that charts iden- commercial interests. tified only 70 percent of activities performed during the Globally, both administrative and clinical health clinical encounter. Even abstracting measures of quality databases are of poor quality, and administrative data- from electronic medical records is challenging given the bases are usually the only resource available in LMICs. heterogeneity in record-keeping practices (Ali, Shah, Even when available, health information is underused and Tandon 2011; Parsons and others 2012). The costs for planning and decision making (Corrao and others and logistical challenges of securing medical records, 2009), especially in resource-constrained settings training medical abstractors, and reviewing records can (Bosch-Capblanch and others 2009) and when data be significant. Throughout acquisition, verification, are paper based or decentralized to the district level and abstraction, a process is needed to ensure that the (LaFond and Siddiqi 2003). District-level informa- data collected are reliable (Koh and Tan 2005). Beyond tion systems often do not feed information back to these costs and challenges, chart review also suffers the local level (Lippeveld, Sauerborn, and Bodart 2000). from the inability to control for patient case mix and Paper-based information systems often generate difficulty of comparing physicians caring for different poor-quality data (Lium, Tjora, and Faxvaag 2008), patient populations. which weakens confidence in reported progress made toward health care system goals (Kerr and Direct Observation and Recording of Visits Fleming 2007) and toward the Sustainable Development Direct observation and recording of visits are common Goals and the Millennium Development Goals practices in LMICs (Nolan and others 2001). Some of (AbouZahr and Boerma 2005). In the absence of greater the most obvious challenges to using direct observation attention and resources from government or private Quality of Care 189 health insurance initiatives, using administrative data to and Aktas 2013; Li and others 2007; Tiemeier and others measure and track clinical performance should be done 2002; Veloski and others 2005). Vignettes are particularly cautiously. effective in comparative evaluations because the same case or type of case can be presented to many providers Standardized Patients simultaneously, and the results can be examined over Using standardized patients, when unannounced, is the time. Vignettes have been used in large cross-national gold standard for measuring process (Luck and Peabody studies, such as a six-country policy evaluation in 2002). Trained to simulate patients with a given illness, Central Asia and Eastern Europe (Peabody and others, standardized patients present themselves in a clinical forthcoming). This study, involving 1,039 facilities and setting to providers who have given their consent to par- 3,121 providers, evaluated quality of care in obstetrics, ticipate in the study. After the visit, the standardized newborns, and chronic disease. Because vignettes are patient reports on the technical and interpersonal ele- inexpensive to administer, they are especially well suited ments of the care process. Interest in using standardized for use in resource-poor settings (Peabody, Luck, and patients has been growing in LMICs, with most studies others 2014; Peabody, Shimkhada, and others 2014; done in China and India (Das and others 2012; Das and Peabody, Taguiwalo, and others 2006). others 2015; Mohanan and others 2015; Sylvia and oth- ers 2015). Well-trained standardized patients are not susceptible to observation bias (Glassman and others Standards 2000) and, when rigorously monitored, enable compari- Evidence-Based and Best-Practice Standards sons of quality within and between facilities. Much of the early disagreement about what to measure However, this method also has major drawbacks, has given way to a consensus that performance should be including high costs of training, significant difficulties in measured against evidence-based criteria. The scientific large-scale application (consistent training), and limited literature is replete with evidence-based quality metrics conditions that actors can reliably portray, for example, that describe processes as varied as whether a patient’s excluding surgical and pediatric cases (Felt-Lisk and blood pressure is under control, whether a patient is on others 2012). the correct medication to slow down renal failure, whether the timing of a specific surgery is correct, or Clinical Vignettes whether a diagnostic test is needed. Collectively, clinical The shortcomings of the previous methods have spurred care metrics are based on the evidence and the supposi- development of more facile methods. One of these, tion that meeting these metrics results in better out- developed in work starting in 1999, is the use of vali- comes. Critics point out that evidence-based practice has dated clinical performance vignettes (Peabody and oth- only been established for a limited number of care ele- ers 2000). Clinical performance vignettes use a full set of ments (Contreras and others 2007; Karolinski and others clinical care elements to assess the patient-provider 2009; Vogel and others 2014). However, clinicians interaction (Glassman and others 2000). routinely rely on best-practice standards, even as There are many types of vignettes from which to high-quality data from well-designed studies continue to choose—for example, multiple choice versus open- emerge and evolve. In practical terms, there will never be ended, or short case versus full clinical care delivery a complete set of evidence-based standards, and quality scenarios—producing variable results at predicting of care will always rely on the best available evidence and actual practice. Clinical performance and value vignettes local standards. have been validated in randomized evaluations against An important body of evidence-based, best-practice standardized patients in two large trials (Peabody and standards in LMICs comes from using surgical and others 2000; Peabody and others 2004). In these studies, childbirth safety checklists. Checklists have recently vignette scores for clinical performance and value con- been rapidly introduced into LMIC settings, and the sistently reflected quality as measured by standardized evidence indicates that using these evidence-based stan- patients better than abstracted medical records and dards in checklist form improves health outcomes, pri- worked across different health care systems, clinical con- marily by setting a quality standard for treatment and ditions, and levels of training among randomly sampled facilitating communication within provider teams physicians. (Ergo and others 2012). An intervention in Michigan Various types of vignettes have been used in diverse that used a surgical checklist to decrease catheter- settings around the world (Canchihuaman and others related bloodstream infections in hospital intensive 2011; Das and Hammer 2005a, 2005b; Holm and care units, for example, led the World Health Burkhartzmeyer 2015; Jörg and others 2006; Kaptanoğlu Organization (WHO) to create the Surgical Safety 190 Disease Control Priorities: Improving Health and Reducing Poverty Checklist (Pronovost and others 2006). As of 2012, the young children reduced under-five mortality within two WHO Surgical Safety Checklist has been adopted by years (Mtango and Neuvians 1986). Physician-reported 1,790 health care facilities worldwide (Treadwell, Lucas, continuing medical education has been linked to better and Tsou 2014), helping teams to manage crises, avoid quality and health status when accountability is included clinical errors, and minimize health risks. However, using clinical performance vignettes (Luck and others successful uptake of checklists requires “constant super- 2014). A six-nation study linked continuing education to vision and instruction until it becomes self-evident and evidence-based practice as measured with simulated accepted” (Sendlhofer and others 2015). patients (Peabody and others, forthcoming). Using a systematic database of quality improvement studies, Licensing, Certification, and Accreditation Rowe and colleagues at the U.S. Centers for Disease Provider certification and hospital accreditation were Control and Prevention (National Academies of Sciences, introduced into health care in the early twentieth cen- Engineering, and Medicine 2015) found that, in LMICs, tury and have been adopted globally as a cornerstone of training and supervision have modest positive effects on health care quality assurance. The number of health care provider performance and that strategies may work bet- accreditation programs, including national accreditation ter when used in combination than when used by them- systems, is doubling every few years, with as many as 70 selves. Work by Das and others (2016) on providers in programs around the world in 2013 (Saleh and others India suggests that better incentives can improve quality 2013). Accreditation has expanded beyond hospitals to without any additional provider training. include primary care, health systems, and laboratories. Despite its ubiquity, continuing education will not Additionally, many LMICs are replacing voluntary greatly improve the quality of clinical practice or health accreditation from independent organizations with outcomes (Davis and others 1999; Forsetlund and oth- national programs that, in some instances, link accredi- ers 2009). An analysis of 62 studies and 20 systematic tation to licensing (Greenfield and Braithwaite 2008; reviews found that the “continuing education ‘system,’ Jovanovic 2005). as it is structured today, is so deeply flawed that it can- However, national licensing and accreditation pro- not properly support the development of health profes- grams require political commitment, human and finan- sionals” (IOM, Committee on Planning a Continuing cial resources, and planning. This issue is further Health Professional Education Institute 2010, ix). Davis complicated in LMICs by the complexity of the develop- and others (2006) found that physicians cannot self- ment of the accreditation process and the dearth of assess their skills accurately and suggested that external resources for implementing and maintaining a strong assessment, scoring, and feedback would drive more accreditation process. Evidence on the effectiveness of effective professional development. Moreover, physi- accreditation for enhancing clinical outcomes or defin- cians are often “not trained” to evaluate or use published ing when accreditation is most effective is limited and guidelines and best practices. Passive dissemination of inconclusive: in a systematic review of the literature, information (publishing guidelines, reading peer- health sector accreditation was consistently associated reviewed articles) is generally ineffective at changing with professional development and promotion of practice and is unlikely to change group-wide practice change, but not consistently associated with quality when used alone. improvement or other organizational and financial Newer educational techniques—targeted education, impacts (Greenfield and Braithwaite 2008). One study in case-based learning, and interactive and multimodal the Philippines showed that licensing and accreditation teaching techniques—have had more success. independently and substantively improved clinical prac- Interventions that are multifaceted and include active tice and health outcomes, but with modest impact participation and targeted feedback are much more (Quimbo and others 2008). likely to be effective than single interventions. Engaging clinicians is the key to translating training into improved quality (Mostofian and others 2015). Physicians engaged Training in hospital initiatives, for example, are much more likely Clinical training starts in medical or other professional to report successful experiences with quality improve- schools and continues throughout a practitioner’s pro- ment programs. Methods that require active physi- fessional career. Continuing medical education is often a cian learning (one-on-one meetings, small-group requirement for licensing or certification and is part of workshops, and programs tailored to a specific clinic) almost every health care system. Continuing education are effective at aligning patterns of physician prac- has shown positive impacts on care. In Tanzania, train- tice with new clinical guidelines. In Guatemala, distance ing staff in the control of acute respiratory infections in education that targeted diarrhea and cholera case Quality of Care 191 management increased the accurate assessment and for example, attending school, having up-to-date vacci- classification of diarrhea cases by 25 percent (Flores, nations, or visiting a health center for prenatal care Robles, and Burkhalter 2002). (box 10.1). Although CCTs do not directly provide incentives to health care providers, they require quality health services, adding a supply- or provider-side com- Supervision ponent to demand-side interventions. There is also an Supervision is an established method for assessing qual- indirect supply-side incentive when consumers use cash ity. The power and influence of peer review supervision, incentives to pay for services. A systematic review of the often conducted through professional societies, vary evidence suggests that CCTs improve the uptake of pre- widely among countries (Heaton 2000). Large providers, ventive services by children and pregnant women such as hospitals or public health institutions, often have (Lagarde, Haines, and Palmer 2009). more resources for collecting information on provider However, in shorter time frames of months to a year, practices and patient outcomes and for using those data CCTs have difficulty driving lasting effects and affecting to guide, educate, supervise, discipline, or recognize pro- health (Beegle, Frankenberg, and Thomas 2001; World viders. Providers at clinics and primary care facilities also Bank 2003). From a policy perspective, it is also difficult benefit from supervision (Loevinsohn, Guerrero, and to distinguish the effects of the CCT incentive from the Gregorio 1995). Other studies point to the benefits of impact of the cash itself, that is, it is unclear whether the quality review committees and standing groups that behavioral change is associated with the conditional review all hospital deaths. However, oversight can also incentive or with an income effect (Fernald, Gertler, and create an antagonistic relationship between workers and Neufeld 2008). A systematic review of the impact of managers that may preclude cooperative problem solv- vouchers found modest evidence that the vouchers ing and continuous improvement (Berwick 2002). improved quality of care (Brody and others 2013). The question that remains is whether there are long-term Incentives effects because clinical practice was not improved. Demand Incentives Provider Payment Demand-side interventions, such as conditional cash In the past two decades, health care administrators and transfer (CCT) and voucher programs, pay partici- policy makers in both LMICs and HICs have been using pants (not providers) a stipend for specific behaviors, pay for performance (P4P) as a means to improve clini- cal practice. Although the details of programs vary, health care P4P programs link physician compensation Box 10.1 to measures of clinical quality (Epstein, Lee, and Hamel 2004). P4P and other forms of results-based compensa- tion have been used routinely in business settings. The Progresa/Oportunidades challenge in health, however, is to identify suitable met- rics that are under the control of the provider (Werner Progresa/Oportunidades is a major government initiative and Asch 2007). For example, care providers are hard that used demand-side interventions (conditional cash pressed to reduce infant mortality rates that are driven transfers) to reduce long-standing poverty and develop primarily by poverty and nutrition, but they can readily human capital within poor households in Mexico (Fernald, change the frequency of unnecessary cesarean sections. Gertler, and Neufeld 2009). The demand incentives were Even with suitable metrics, isolating and linking P4P payments to mothers for health behaviors, such as partici- changes in practice to better health has been challenging pation in programs like prenatal care, immunizations, and (Atkinson and others 2000; Derose and Petitti 2003). nutrition supplementation, as well as for children’s school P4P might be linked, at best, to modest improvements in attendance. The intervention had a broad positive impact quality of care (Epstein 2007; Lindenauer and others on many measures and improved patient outcomes such 2007; Petersen and others 2006; Rosenthal and others as stunting and anemia in preschool children (Fernald, 2005). However, most studies are not experimentally Gertler, and Neufeld 2009; Rivera and others 2004). The designed, and participation in P4P programs is volun- implication of this work is that, for certain health outcomes, tary, leading to selection bias. Although much of the lit- improving access was sufficient to improve outcomes. erature on the equivocal benefit of provider incentive Although there are no data, this improvement occurred systems comes from HICs, the Quality Improvement even though clinical practice was (certainly) varied. Demonstration Study (QIDS), carried out in the Philippines as a social policy experiment, provides 192 Disease Control Priorities: Improving Health and Reducing Poverty strong experimental evidence that P4P can be effective in outcomes beyond the performance seen in controls an LMIC (Quimbo and others 2008) (see box 10.2). (Glickman, Boulding, and others 2007; Grossbart 2006; Similar results were found in the work by Gertler and Lindenauer and others 2007). Although these studies Vermeersch (2013). had controls, the interventions were not randomly The large QIDS randomized community-level exper- assigned. iment found greater improvement in health outcomes than previous P4P studies (Peabody and others 2017). Results- and Performance-Based Financing This finding may have occurred because most other Results-based financing (RBF) encompasses various studies providing incentives to doctors have been con- types of interventions that provide demand-side incen- ducted in wealthier countries and been nonrandomized, tives (for example, CCTs), refine provider payments (for which introduces the possibility of selection bias wherein example, P4P), and trigger government reforms. providers who adopt the incentives may be the most The RBF lending projects financed by the Health likely to respond and improve their clinical practice Results Innovation Trust Fund and World Bank credits anyway (Petersen and others 2006). Three randomized or loans (World Bank 2014) operationalized the concept P4P studies conducted in the United States found that of RBF at a large scale in many LMICs and intended to rewarding physicians improved outpatient care, such as provide incentives to policy makers to build and leverage immunization rates (Fairbrother and others 1997; their quality infrastructure as a condition for financing. Fairbrother and others 2001; Kouides and others 1998). Since 2008, RBF projects like these have been widely However, other randomized studies found that physi- adopted in more than 30 countries, with interventions at cian P4P had no effect on mammography, other cancer the national, subnational, district, facility, and commu- screening, or adherence to pediatric preventive guide- nity levels. Operationally, funds are provided to govern- lines (Grady and others 1997; Hillman and others 1998; ments at the national and subnational level based on Hillman and others 1999). Three hospital-based studies agreed-on disbursement-linked indicators and their examining inpatient P4P programs in the United States established targets (often nation- and state-wide esti- also included control hospitals. These studies, which mates). At the facility level, payments to individual facil- focused on adult care in cardiovascular disease, commu- ities are based on their contracts with fund holders nity-acquired pneumonia, and joint replacement, found (often district or provincial health authorities). And, modest improvements of 2 to 4 percentage points in increasingly used at the community level, payments are Box 10.2 Impact of P4P on Quality: Results of the Quality Improvement Demonstration Study The Quality Improvement Demonstration Study wasted (underweight for height) increased 9 percent- (QIDS) is unique in that it was an explicit policy age points relative to control sites. The share of par- experiment that randomized communities into pay ents reporting at least good health for their children for performance (P4P) versus universal health cover- was 7 percentage points higher in P4P sites than in age versus a true control. P4P improved both quality controls (Peabody, Shimkhada, and others 2014). and outcomes. The introduction of P4P led to improvements in QIDS was a large policy experiment conducted in the quality of care as measured by clinical case vignettes Philippines among 119 doctors, 3,162 children, and (Peabody and others 2011). Difference-in-differences 30 communities, covering about one-third of the model estimations indicated that P4P improved not country. The communities were randomized into an only the measured quality of physician practice but incentives-based policy program rewarding physi- also health outcomes. The impact of policy can be cians financially for providing higher-quality care to measured in a relatively short (two-year) time frame children than provided by universal health coverage when evaluation is integrated into policy making and and controls (Quimbo and others 2008). In the com- planning before implementation (Peabody and oth- munities where doctors were eligible for the bonus ers 2017), making it possible to measure policy effec- payments, the number of children who were not tiveness and to identify ineffective polices early on. Quality of Care 193 provided to community organizations or community Impact evaluation studies show positive evidence health workers based on RBF contracts with fund hold- about the impact of RBF programs on certain dimen- ers (often districts or facilities). sions of quality. Several countries, including Argentina, A flexible approach, RBF focuses on results: Rwanda, and Zimbabwe, report improvement in qual- ity of prenatal care (Basinga and others 2010; Gertler • Payments linked to results (both demand and supply and Vermeersch 2013; World Bank 2014). Afghanistan side) based on context-specific health priorities demonstrated substantial improvements in quality of • Contracts or agreements that clarify the responsibili- examinations and counseling, as well as time spent ties of all stakeholders with patients (Engineer and others 2016). Under • Autonomy for those contracted to be able to use RBF Argentina’s Plan Nacer1 incentives-based program, the funds to attain the agreed-on results most effectively estimated probability of low birthweight was reduced • Verification of results to ensure that they are accurate by 19 percent among beneficiaries, and in-hospital and reliable neonatal mortality for babies of enrolled mothers was • Data sharing to enhance the results, which can be reduced by 74 percent (Gertler, Giovagnoli, and used for planning, design, and implementation Martinez 2014). • Community involvement to enhance accountability. RBF programs exercise interventions beyond pro- vider performance incentives, such as policy reform, RBF operational data show improvement of quality system strengthening, transparency improvement, and (especially structural quality) in the RBF programs. management and accountability enhancement. Because Facilities’ quarterly quality scores, calculated based on a of this, establishing the effectiveness of clinical interven- supervisory checklist, improved in almost all of these tions through randomized controlled trials becomes a programs. In Burundi, for example, quality scores challenge. How best to use operational data and experi- improved significantly during the first two years follow- ences remains important in disentangling the effects of ing rollout of a national RBF program (figure 10.1). In incentives and the key bottlenecks addressed by RBF. Ethiopia, where RBF was implemented at the national government level, the Ministry of Health undertook Service Availability and Readiness Assessment (SARA) of LINKING POLICY AND PRACTICE AT THE its primary care facilities on an annual basis to achieve PLATFORM LEVEL targets associated with disbursement-linked indicators and developed action plans to address weaknesses iden- How do quality infrastructure policies at the govern- tified through SARA. ment level translate into improved clinical care at the patient level? At its heart, quality improves only when providers deliver the right care to the patient at the right time, do so efficiently, and focus on the patient. However, Figure 10.1 Average Quality Score among All Health Centers in Burundi, less variation among a group of providers depends on 2010–12 individual providers treating their patients and their diseases the same way. This section examines how policy 100 and practice come together at the platform level. Technical quality score 80 Specifically, we review the policy elements described above that would be implemented for 11 clinical condi- 60 tions across four platforms. 40 We start by looking at where care services are deliv- ered. Delivery occurs through various platforms, from 20 community and public health settings to primary care clinics, first-level hospitals, and the most advanced facil- 0 ities in every country. 010 010 010 11 11 011 11 12 2 The quality of care will vary in each setting, which 1 –20 –20 –20 –20 –20 –2 –2 –2 –2 means that the policy elements discussed above are Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Quarter relevant to each setting. These policy elements are categorized as quality measurement, practice standards, Source: Calculations based on operational data from the Burundi Health Sector Development Support training management, and incentives. Project. Note: Quality score (“technical quality score”) is measured based on a comprehensive supervisory Table 10.2 shows how each policy element might checklist on a quarterly basis. Outlying values are not plotted on this graph. be implemented across the four delivery platforms. 194 Disease Control Priorities: Improving Health and Reducing Poverty Table 10.2 Infrastructure Elements for Improving Quality for 4 Delivery Platforms and 11 Representative Clinical Conditions Delivery platforms Referral and specialized Community-based services Primary health centers First-level hospitals hospitals Infrastructure Disease or elements condition Outcomes Metrics Outcomes Metrics Outcomes Metrics Outcomes Metrics Quality measurement Measurement Reproductive Fertility Coverage rates Prenatal, Referral rates; Management of Provider-level Treatment Provider-level health management, or service perinatal care; folic acid labor; vaginal data on practice of birth data on use contraception, use, provider recognition coverage; delivery or (vignettes, complications, (vignettes, family planning knowledge, of high-risk ability to cesarean section charts); patient- such as sepsis charts); patient- (information, patient behavior pregnancies recognize high level data level data condom (condom use); risk on outcomes on outcomes availability, birth unintended (charts, patient (charts, spacing) pregnancy rate reports) registries, patient reports; mortality rates; readmission rates) Feedback and Cardiovascular Use of nutritional Patient-level Blood Screening at Triage of acute Provider-level Arrhythmias, Provider-level accountability disease and exercise awareness pressure, the population myocardial data on practice; endovascular data on use; programs of programs; lipid, diabetes level; screening infarction and patient-level procedures, patient-level availability of screening; with patient treatment of data on valvular data on programs management data on blood congestive heart outcomes surgery outcomes pressure, lipids failure Practice standards Evidence- Sexually Patient Patient surveys Syphilis Direct Pelvic Provider-level Penile, Provider-level based practice transmitted knowledge, safe of knowledge, screening, observation inflammatory data on practice; cervical cancer data on use; infections sex practices behaviors; treatment of of successful disease patient-level patient-level provider surveys gonorrhea management data on data on of clinical and treatment outcomes outcomes knowledge regarding sexually transmitted infections; Quality of Care cultural competency in communication table continues next page 195 196 Disease Control Priorities: Improving Health and Reducing Poverty Table 10.2 Infrastructure Elements for Improving Quality for 4 Delivery Platforms and 11 Representative Clinical Conditions (continued) Delivery platforms Referral and specialized Community-based services Primary health centers First-level hospitals hospitals Infrastructure Disease or elements condition Outcomes Metrics Outcomes Metrics Outcomes Metrics Outcomes Metrics Checklists, Pediatric Preventive, Immunization Diarrhea Provider Pneumonia Provider-level Cancer, Provider-level clinical infectious evidence-based rates; incidence treatment, compliance (diagnosis, data on use; bacterial data on use; guidelines diseases measures to of disease referral with guidelines treatment of patient-level meningitis, patient-level prevent disease (charts, bacterial versus data on other serious data on vignettes); viral) outcomes infections outcomes provider’s ability to diagnose, referral rates Licensing, Infectious Provider hygiene, Direct observation Wound care Direct Tuberculosis, HIV/ Expert board Ebola, International certification, disease handwashing; of program with suturing; observation AIDS diagnosis review to SARS, other body review accreditation proper disposal of implementation aseptic of explicit and treatment determine if outbreaks to determine if needles technique; management explicit criteria diagnosis and explicit criteria instrument or treatment standards are treatment are met; sterilization criteria met retransmission rates Training, management Training Mental health Provider and Provider and Acute mental Institutional Emergency Presence of care Long-term Provider-level community community health first training care and coordination care for compliance awareness of attitudes, aid and triage outcomes hospitalization for and team dementia, according to mental health; knowledge (suicide (provider acute psychosis; practice with chronic evidence-based destigmatization using surveys; prevention, knowledge) for treatment and counseling and affective care; patient- of mental health incidence surveys crisis diagnosis and detoxification of drug therapies disorders, level data: use illness of mental illness intervention, counseling; substance abuse available; schizophrenia of procedures, by socioeconomic disaster provider’s use provider’s ability complications status; counseling); of screening to diagnose; destigmatization screening for for ASD referral rates; of mental health ASD treatment per at community institutional level guidelines table continues next page Table 10.2 Infrastructure Elements for Improving Quality for 4 Delivery Platforms and 11 Representative Clinical Conditions (continued) Delivery platforms Referral and specialized Community-based services Primary health centers First-level hospitals hospitals Infrastructure Disease or elements condition Outcomes Metrics Outcomes Metrics Outcomes Metrics Outcomes Metrics Continuing Diabetes Patient preventive Whether Diabetes Knowledge- Treatment of Knowledge- Transplant Provider use medical behaviors: continuing management based testing renal failure, based testing; surgery rates, provider’s education physical activity, medical education with cardiovascular team-based ability to healthy eating requirements behavioral disease, and practice identify are being interventions consumptive measures transplant met; provider and medication heart failure with candidates; knowledge medical therapies patient-level regarding and medication data on patient programs mortality, and ways to complications engage patient in behavioral change; patient knowledge Management Accidents, Provider’s Provider Provider’s Provider-level Successful Mortality rates, Successful Mortality rates, injury, trauma and patient’s and patient ability to data on ability surgical treatment wrong-site surgical wrong-site knowledge and knowledge recognize and to make correct of trauma (minor surgeries; proper treatment of surgeries; use of preventive surveys assess severity diagnosis surgery) use of surgery trauma (major proper use measures for of injury or (vignettes), or surgical surgery); of surgery injury (child complications time to techniques; treatment of or surgical safety, car seats, treatment, readmission burns techniques; water safety; referral rates readmission elder safety) rates Professional Cancer Smoking Patient-level data Screening Assessment Breast, skin Provider’s use Colorectal Provider-level oversight cessation, on immunization for breast, of provider’s cancer diagnosis; of biopsies and cancer care data on use, hepatitis B rates, cancer colon, cervical, knowledge clinical staging compliance screening compliance immunization incidence; lung, and skin of risk and with treatment (colonoscopy), with rates, school- hospital-based cancer referral protocol colectomy, guidelines; based human cancer registries standards chemotherapy patient-level papillomavirus (set and data on Quality of Care vaccination disseminated); outcomes patient screening rates table continues next page 197 198 Disease Control Priorities: Improving Health and Reducing Poverty Table 10.2 Infrastructure Elements for Improving Quality for 4 Delivery Platforms and 11 Representative Clinical Conditions (continued) Delivery platforms Referral and specialized Community-based services Primary health centers First-level hospitals hospitals Infrastructure Disease or elements condition Outcomes Metrics Outcomes Metrics Outcomes Metrics Outcomes Metrics Incentives Performance- Orthopedics Provider’s Provider’s Care for Provider’s Care for mid- Proper use of Use and Proper use of based knowledge and and patient’s low-level ability to level trauma and surgery, drugs, success surgery, drugs, remuneration communication knowledge trauma (simple diagnose fractures antibiotics; of joint antibiotics; of preventive and behavior broken bones), low-back pain mortality replacement mortality behaviors surveys; patient management (vignettes); rate; bleeding rate; bleeding (healthy eating, participation of low-back use of certain during surgery; during surgery; immunization, in preventive pain prescription complications complications healthy lifting); health programs; drugs for pain (thromboembolic (thromboembolic patient’s behavior monitoring of management disease) disease) of the same physical activity of patients Team-based, Malaria Provider’s Provider’s Provider’s Provider’s Provider’s ability Provider’s ability Treatment and Patient-level multidisciplin- knowledge and and patient’s ability to ability to to diagnose to diagnose management data: mortality ary care (global communication knowledge recognize diagnose type of malaria malaria, of severe rate, compli- payment) of vector control and behavior malaria malaria (drug resistant recognize drug malaria- cation rate; to patient (use of surveys; patient’s rapidly; use (vignettes), or not); proper resistance; associated provider-level insecticides) participation of ACT time to treat, treatment, long- proper use of complications data: treatment in preventive proper use term follow-up, antimalarial (cerebral of compli- health programs; of ACT management drugs; proper malaria, renal cations per community management dysfunction, evidence-based malaria rates of relapse hepatic dys- guidelines function, acute respiratory distress, ane- mia, thrombo- cytopenia) Note: Quality interventions provided at lower-level platforms are also provided at higher-level facilities. ACT = artemisinin-based combination therapy; ASD = autism spectrum disorder; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome; SARS = severe acute respiratory syndrome. For each element, the table details how quality outcomes level, is that the effectiveness of the improvement strategy and metrics could be operationalized for a given disease must be assessed regularly. Recommendations published or clinical condition. For example, community-based by the WHO and the International Association for services in reproductive health (a condition) would focus Trauma Surgery and Intensive Care on quality improve- on family planning and fertility management, which can ment strategies are broadly applicable to all levels of care be assessed by metrics of patient behavior (condom use); and types of settings and include strategies such as mor- primary clinics would focus on high-risk pregnancies, bidity and mortality conferences to review errors occur- which can be assessed using referral rates for women at ring during the care of patients, panel reviews of risk. Outcomes and metrics tend to become more con- preventable deaths, and tracking of complications, crete as care progresses across platforms. Primary clinics adverse events, sentinel events, and errors (WHO, and first-level hospitals, for example, might require Association for Trauma Surgery and Intensive Care, and chart-level data or provider-level assessments of skill, International Society of Surgery 2009). knowledge, and practice. Specialized hospitals, where care is more complex (treatment of birth complications) UPDATED QUALITY OF CARE FRAMEWORK and outcome metrics are more serious (mortality rates), are likely to have more readily available data and better As shown in figure 10.2, health actions take place in the outcomes. A key element of quality improvement, context of and are influenced by political (laws, govern- whether at the specialized hospital or community clinic mental stability), cultural and social (societal norms, Figure 10.2 DCP3 Approaches to Improving Quality of Care Framework Political Institutional factors factors Equity Policy levers: Access: Platform-level: Provider-level: Patient-level: Coverage, benefits Aggregate system Practice, behavior Engagement, experience, adherence Structure, Provider: Health care systems, Process, Clinical skill, adherence to Health access human clinical care guidelines, diagnostic outcomes resources accuracy, communication Payment systems: Incentives Environmental Social factors and cultural factors Note: Blue indicates items discussed in this chapter; DCP3 = Disease Control Priorities, third edition. Quality of Care 199 practices), environmental (natural disasters), and insti- The updated framework in figure 10.2 adds policy tutional (functioning health departments, corruption) levers for improving quality of care and showcases the factors. Demographic and socioeconomic makeup, provider’s practice and behavior as well as the unique including genetics and personal resources, also affect the perspectives of policy makers, physicians, and patients, health status of individuals seeking care. which are essential to establishing accountability. The The classic construct of structure, process, and out- early frameworks focused on the lack of structural come is at the core of the framework (Brook, McGlynn, inputs, whereas recent frameworks look at care processes and Cleary 1996; De Geyndt 1995; Donabedian 1980; (Kruk and others 2009). The Institute of Medicine was McGlynn 1997). These three elements are described in the first to include additional elements of care regarding table 10.3. safety and efficacy, patient focus, affordability and time- Structure refers to stable, material health care assets liness, and effectiveness (Berwick 2002; IOM, Committee (infrastructure, tools, technology, implements), the on Quality Health Care in America 2001). The remain- resources of the organizations providing care, and the der of this section discusses these elements. financing of that care (levels of funding, staffing, pay- ment schemes, incentives). These factors can be mea- sured inexpensively and data are typically readily Safety and Efficacy available (De Geyndt 1995). Patient safety has not received enough attention in Process captures the interaction between caregivers LMICs. Globally, up to 1 in 10 patients is harmed by an and patients, including appropriateness of the care deliv- adverse incident in a hospital not directly related to his ered, cognitive skill of the provider, and communication or her clinical care, with approximately US$6 billion in (Murray, Gakidou, and Frenk 1999). The private nature costs per year (WHO 2008). Even procedures that are of the doctor-patient consultation, lack of measurement not considered high risk in HICs have the potential to criteria, and absence of reliable measurement tools make lead to harm or poor outcomes in LMICs. For example, it difficult to assess process, especially in LMICs (Peabody up to 1 in 4 cataract surgeries in India results in poor and others 2004). However, new approaches to measur- visual acuity (Lindfield and others 2012). ing process have come a long way toward capturing A study on patient safety practices in low-income process measures across settings. countries suggests that improved staff-patient commu- Outcome includes direct measures of health status, nication, use of protocols, control of infections, and death, or disability-adjusted life years as well as patient standardization between providers can improve overall satisfaction or patient responsiveness to the health care safety (Lindfield, Knight, and Bwonya 2015). Efficacy of system. Outcome measurement has matured in the past care has an ascendant role in clinical practice as the prac- decade with the use of electronic medical records and tice of evidence-based medicine continues to expand. data registries. Many new and exciting studies of clinical efficacy are Table 10.3 Quality-of-Care Framework Elements of quality Description Subcomponents Structure Stable, material characteristics (infrastructure, tools, • Physical characteristics technology) and resources of the organizations that • Management (executive leadership, board responsibilities) provide care and the financing of care (levels of • Culture funding, staffing, payment schemes, incentives) • Organizational design • Information management • Incentives Process The interaction between caregivers and patients during • Making the diagnosis which structural inputs from the health care system are • Providing evidence-based treatment transformed into health outcomes Outcomes Measures of health status, deaths, or disability- • Morbidity adjusted life years (a measure that encompasses • Mortality the morbidity and mortality of patients or groups of • Patient satisfaction patients); outcomes such as patient satisfaction or patient responsiveness to the health care system Sources: Glickman, Boulding, and others 2007; Peabody, Taguiwalo, and others 2006. 200 Disease Control Priorities: Improving Health and Reducing Poverty driving better care, including the use of antibiotic pro- Affairs 2014). Estimates are not available for LMICs, but phylaxis before surgery and the elimination of antibiot- as much as one-third of health care costs may be due to ics for otitis media. unexplained variation in quality and unnecessary care in practice. A study in eight countries found that the intro- duction of surgical guidelines in hospitals led to less Patient Focus variation in quality, better health outcomes, and lower As with efficacy, focus on the patient and his or her per- costs (Haynes and others 2009). spective has become more prominent, leading evalua- tions of performance to include satisfaction as a necessary outcome. The availability and growing acceptance of Effectiveness patient satisfaction surveys are striking given that these Effectiveness refers to how well evidence-based practices tools were almost unheard of 20 years ago. are followed. Translating promising research findings The focus on the patient is important because and evidence, especially results that improve health or patients’ or users’ perspectives determine whether they lower health care costs, into scalable interventions is seek care and where they obtain services (demand). This challenging. The high stakes—and rare successes—have perspective is based on the individual’s own opinions, led to increasing calls for evidence-based policy making. previous experiences with the health system, and input Ideally, evidence-based policy making is based on evalu- received from others. ations of real-world economic effectiveness, allowing a Perception of low quality has been reported as a determination to be made of value as well as efficacy. major factor in the decision not to use or to bypass With this effort has come interest in determining the health services. For example, in a study in Tanzania, 42 comparative cost-effectiveness of clinical interventions. percent of women who delivered children in a health Few studies compare policy approaches to quality care facility in rural parts of the country bypassed the improvement. Peabody and others (2017) compared a local primary care clinic and delivered directly in a hos- demand-side intervention (universal health coverage for pital or health center (Kruk and others 2014). This find- children under age five years) with a supply-side inter- ing is striking because all of them lived near a functioning vention (P4P scheme for physicians) and found that clinic with delivery services and the sample excluded both interventions were effective, reducing wasting by women referred to a hospital. Primary care clinics tend about 9 percent (relative to controls). Costs were notably to have poor infrastructure, lack equipment, and are lower in the supply-side intervention than in the understaffed, and women may choose care based on demand-side intervention, suggesting that increasing their perception of specific factors, such as a provider’s quality is more cost-effective than expanding insurance attitude or competency and the availability of drugs and benefits in resource-constrained settings. medical equipment. Affordability and Timeliness CHALLENGES FOR ASSESSMENT Determining affordability is challenging given that there The conversation on quality needs to include issues is no recognized, consistent association between afford- related to equity, misdiagnosis, perceptions, accountabil- ability and quality. High-quality care is often assumed to ity, and learning from patients, all of which are challeng- mean more expensive care (Starfield and others 1994). ing to assess. Indeed, early quality improvement efforts were often costly because the quality interventions themselves had to be paid for, and new measures of performance had to Equity be introduced to calibrate the baseline quality and detect Equity is an increasingly recognized part of the quality subsequent change (U.S. Congress, Office of Technology equation. Inequality—a situation in which poor-quality Assessment 1994). care is disproportionately provided to people from a However, high-quality care is potentially more afford- particular disadvantaged group—is rampant world- able care because consistent, high-quality, standardized wide (Barber, Bertozzi, and Gertler 2007; Barber, Gertler, care entails fewer unnecessary tests, less time spent in the and Harimurti 2007; Hansen and others 2008). hospital, and fewer complications. In the United States, Socioeconomically disadvantaged groups have poorer as much as one-third of health care costs are unneces- access to services and, once they have access, are less sary, and as much as US$799 billion in costs is due to likely to receive effective treatment (Garrido-Cumbrera unexplained variation in practice and quality (Health and others 2010; Health Affairs 2011; Rogers 2004). Quality of Care 201 If they are lucky enough to obtain treatment, they Real-world practicalities make investigating misdiag- receive poorer-quality care than people from other noses a substantial challenge. Methodological problems groups. The impact of quality interventions on equity include the difficulty of aggregating patients with the has not received enough attention in the literature. same diagnosis to overcome the unobserved (and unre- corded) case-mix variation, legitimate disagreements on reference standards for practice, reliance on recorded Misdiagnosis retrospective data, and challenges of measuring a clini- Misdiagnosis, also referred to as diagnostic error, is a cian’s cognitive thought processes. Perhaps the biggest significant shortcoming, with worrisome, albeit poorly methodological challenge is to reach some agreement understood, consequences (box 10.3). For example, a regarding the correct diagnosis. Short of having a group study reported that 5 percent of adults are misdiagnosed of experts reexamine the patient, the correctness of diag- during outpatient visits, and about 50 percent of these noses is difficult to evaluate. errors could harm the patient (WHO 2000). Misdiagnosis in breast cancer is as high as 20 percent in some cases (Lozano and others 2006). Perceptions of Quality Misdiagnosis is likely to be especially high in LMICs Identifying a perspective—or multiple perspectives— (Galactionova and others 2015). A study in India found from which to assess quality is difficult (Strauss and that only one-third of primary care providers articulated Corbin 1998; Tafreshi, Pazargadi, and Abed Saeedi 2007; a diagnosis, either correct or incorrect, and when a diag- Van der Bij, Vollmar, and Weggeman 1998; Wisniewski nosis was given, close to 50 percent were wrong and Wisniewski 2005). Judging quality requires balanc- (Marchant and others 2015). In an observational study ing the competing viewpoints of many players in the of primary care providers in rural China, the misdiagno- system. For example, payers and purchasers typically sis rate was 74 percent, and clinicians provided medicine judge quality by how well insurance premium dollars are that was unnecessary or harmful to 64 percent of their spent for each covered life; patients typically judge qual- patients (WHO and World Bank 2014). Diagnostic ity by how well their individual needs are addressed; and errors occur around the world and in all types of set- physicians assess quality by their own clinical judgment tings, suggesting a need to include misdiagnosis in con- or training, patient demands, available resources, and ceptualizing quality-of-care deficiencies. cost-controlling mechanisms (Luck and others 2014). Box 10.3 Misdiagnosis as a Core Element of Poor Quality Diagnosis is a key determinant of a successful out- the prevalence and consequences of misdiagnosis come (Freedman and Kruk 2014). Yet the extent of among 103 obstetrical providers in an urban misdiagnosis has not been fully recognized (Jamison area of the Philippines using identical vignettes and others 2013; Ng and others 2014; OECD 2015; and reviewing each provider’s clinical records Rockers, Kruk, and Laugesen 2012; WHO 2000). A (Shimkhada and others 2016). The misdiagnosis wrong diagnosis will lead, at best, to unnecessary of three common obstetric conditions—obstructed evaluations and treatment and, at worst, to harmful labor, postpartum hemorrhage, and preeclampsia— tests and toxic treatment. Diagnostic errors result in was almost 30 percent overall. Providers who mis- potential delays in treatment, putting the patient at diagnosed these conditions were more likely to risk (WHO 2000) and leading to severe complications have patients with a complication. Patients with a and overtreatment. They are an important cause of complication were significantly less likely to be preventable morbidity and mortality (Freedman and referred to a hospital immediately and were Kruk 2014; Jamison and others 2013; Ng and others more likely to be readmitted to a hospital after 2014; Rockers, Kruk, and Laugesen 2012). delivery, to have significantly higher medical costs, and to lose more income than patients without a The field of obstetrics provides a rich opportunity complication. to study misdiagnosis in LMICs. A study examined 202 Disease Control Priorities: Improving Health and Reducing Poverty When different perspectives collide—for example, when ways by an individual’s health system experiences. physician performance metrics (penalties for high surgi- For example, in addition to health outcome data, the cal complication rates) are not in the best interest of the Organisation for Economic Co-Operation and patient (a diabetic who is a higher surgical risk and may Development now measures the patient experience, be turned down for surgery to keep complication rates including metrics on wait times, communication, and low)—the patient’s outcomes, including satisfaction, costs of care. should be given the greatest weight. Methods of obtaining data on the patient experience include exit surveys (in person or anonymous), mailed or online questionnaires, and, increasingly, phone sur- Accountability veys. The large and growing penetration of mobile Establishing accountability is challenging. It can be diffi- phones makes it more and more feasible to collect short cult to determine which platform is responsible for telephone or mobile Web assessments of the patient achieving certain measurement goals and which individ- experience in LMICs (Solon and others 2009). uals within each level should be held accountable for those measures (Emanuel and Emanuel 1996; Wachter 2013). The challenge of establishing accountability is IMPACT OF QUALITY IMPROVEMENT tied to the larger challenge of convincing all players that poor quality should not be attributed to an individual Global health goals and projections are predicated on clinician. Poor quality cuts across all types of care, facil- assumptions about achieving high coverage and improv- ities, providers, health insurance offerings, geographic ing the quality of care in high-mortality countries areas, and patient populations. Accountability must be (Jamison and others 2013). Given the lack of high- established at all levels (Brinkerhoff 2003). Holding phy- quality data from LMICs, data from high-income set- sicians accountable may be especially difficult in a fee- tings are used to predict health gains from expanded for-service environment where individuals are used to coverage in LMICs. These extrapolations do not reflect being independent, and there are significant methodo- the real-life impact of quality on use and eventual out- logical, political, and legal obstacles to measuring comes in LMICs. Diagnosis and treatment, for example, accountability (Quimbo and others 2008). are often egregiously poor in understaffed, under- A common trap is to let the availability of data deter- resourced and underregulated health systems. Yet it is mine which system-level metrics are tracked. System critical to understand whether health care visits translate accountability is analogous to provider accountabil- into quality health care—both for projecting better ity, and metrics must be relevant, reliable, valid, com- health and for estimating the health returns on initia- prehensive, and financially achievable; data availability tives such as universal health coverage. should not drive the selection of metrics (Hsia 2003). Accountability also means that those who judge quality have the opportunity to go beyond explicit, Influence on Demand for Services and Outcomes evidence-based measures of practice or even structure. Quality of care is a major driver of use. Various studies Recent work points to system- and platform-level have shown that perceived quality of care influences accountability for collaboration, local ownership, and patterns of use—for example, perceptions of poor qual- shared learning (Boucar and others 2014). ity can motivate patients to stay at home or to choose far-away providers perceived to be more competent (Bohren and others 2014; Kruk and others 2009; Leonard Learning from Patients 2014). Perceptions of poor quality are a strong factor A final, neglected area of quality assessment is health pushing patients to bypass care, as are users’ assessments system responsiveness to patients, specifically data on of the complexity of their health needs (Akin and the patient’s experience and satisfaction with care Hutchinson 1999; Kruk and others 2014; Leonard, (Bernhart and others 1999). Therefore, improving the Mliga, and Mariam 2002). In sum, patients in low-income patient experience is a stand-alone goal of health systems settings increasingly behave like their rich-country in the updated framework (Rockers, Kruk, and Laugesen counterparts: as active consumers making rational 2012; WHO 2000). choices about their care rather than as passive beneficia- Initiatives such as the current push for universal ries of health care. health coverage assume that people will value and want The demand for quality is likely to grow as coverage to fund health benefits, whether through taxes or premi- expands. Kruk and others (2015) found that, when ums. Public support, however, is shaped in important childbirth at a health facility (that is, in-facility delivery) Quality of Care 203 exceeds 80 percent of all births in a community, proxim- example of patient-centered reform (Groene 2011). For ity to hospitals, not primary care clinics, matters in pre- example, when the quality of obstetric care provided at dicting delivery of care, potentially because of growing first-level, low-volume facilities is of poor quality, refer- demand for high-quality care that is difficult for low- rals to higher levels of emergency care is inefficient, volume clinics to deliver. resulting in excessively high maternal and newborn How accurately do patients assess quality? Although mortality (Hsia and others 2012; Thorsen and others patients are well positioned to report on interpersonal or 2014). Women who deliver in the health system clearly nontechnical quality-of-care issues, such as clarity of prefer higher-volume, higher-quality facilities, as evi- communication, respect, confidentiality, and waiting denced by choice of provider. Thus, the answer to times, they do not have full information with which to improved quality and outcomes may be to establish gauge the technical quality of care. Doyle, Lennox, and high-volume maternity health centers or hospital units Bell (2013) found that the patient experience of care was and provide support for travel to these facilities, rather positively associated with clinical effectiveness and safety than to invest more in primary care obstetrics or in more than 75 percent of studies. For example, Glickman low-volume, first-level facilities. Focusing on customer and others (2010) found that higher patient satisfaction service and respect requires paying attention to staffing, was linked to lower mortality among patients with acute training, and supervision. myocardial infarction. Similarly, more satisfied patients Health systems that can satisfy people’s expectations had lower 30-day hospital readmission rates and higher may experience a double benefit: better health outcomes adherence to physician recommendations (Boulding and and greater support for the health system. For example, others 2011; Fenton and others 2012). Other research women who bypassed their first-level clinic and deliv- found little correlation between patient ratings of care ered in hospitals rated quality of care more highly than and chart-measured adherence to standards of care, use women who delivered in first-level clinics across a wide of inpatient care, or mortality (Chang and others 2006). range of indicators (Kruk and others 2014). Experiencing Whether accurate or not, perceptions drive behavior. responsive health services may enhance confidence in Patient ratings of quality and satisfaction are also associ- government. A multicountry study of LMICs found that ated with future care seeking, an important consider- a combination of high-quality care and financial risk ation given the rise of chronic diseases requiring ongoing protection raised the probability of having trust in gov- contact with the health system (Bohren and others 2014; ernment by 13 percent (Rockers, Kruk, and Laugesen Groene 2011; Kruk and others 2014; Sun and others 2012). More responsive, patient-centered health systems 2000). More work is needed to understand which patient should be a health and political priority. assessments are most reliable and the best ways to collect these data. Patient-reported quality and satisfaction are impor- COSTS OF IMPROVING QUALITY tant indicators of the responsiveness and accountability of health systems (Thaddeus and Maine 1994). Almost all deficits in the quality of care can be addressed Responsiveness, defined as meeting patients’ nonhealth if enough resources are made available for the purpose. expectations, should be a goal of every health system The question is not, “Can we improve the quality of (WHO 2000). Yet recent research has documented disre- health care services?” Instead, it is, “How can we use the spectful and abusive treatment of patients in health resources available to achieve that improvement?” Thus, facilities. For example, nearly 20 percent of women in when resource constraints are considered, policy makers two districts of Tanzania reported harsh treatment by will have to choose from a range of interventions, and health workers, including yelling and slapping (Freedman the question becomes, “What are the most efficient and and Kruk 2014). Such treatment leads to a loss of confi- feasible ways to improve health outcomes?” For example, dence (Kujawski and others 2015). Abusive treatment is nosocomial infections could be treated with costly anti- distressingly common in other settings as well (Asefa biotics, new facilities, and equipment. However, it is and Bekele 2015; Gourlay and others 2014; Okafor, likely to be far more efficient to introduce a handwash- Ugwu, and Obi 2015; Sando and others 2014). ing protocol, to ensure that providers comply with it, and to develop a rapid response team that can be deployed when infections occur. Fit between Services and Patient Needs The costs of improving quality are different from the One promising strategy is to improve the fit between costs of the intervention itself. For example, the cost of people’s expectations and health needs and the health delivering care to patients with closed fractures requiring services available to them. This tailoring of care is an internal fixation includes facility costs (patient room, 204 Disease Control Priorities: Improving Health and Reducing Poverty equipment, sterile supplies), personnel costs (clinicians, a quality improvement collaborative conducted in support staff), and patient costs (transportation to the childbirth facilities reduced the overall costs per birth facility, time costs). If a high proportion of patients an average of 20 percent (from US$35 to US$28); when develop nosocomial infections, the cost of quality would accounting for the decrease in average clinical costs due be the costs incurred to reduce the risk of facility- to improved efficiency and the reductions in post- associated infection through strategies such as providing partum hemorrhage, the authors determined that the training, supervising staff, procuring new cleaning and incremental cost of the improvement collaborative sterilization equipment, and developing care pathways was US$2.43 (Broughton, Boucar, and Alagane 2012). or checklists. The incremental cost-effectiveness was an impressive Cost-effectiveness analysis (CEA) can be used to US$147 per disability-adjusted life year averted, com- determine how cost-efficient a quality improvement pared with US$870 for the rotavirus vaccine, US$135 intervention is. CEAs compare the resources consumed for hypertension treatment, and US$1,480 for a tobacco and the effects on the desired outcome of an interven- tax (Tran and others 2014). Interventions to improve tion to improve the quality of care against a valid com- health care quality can also save money as shown in the parison, which is either business-as-usual or a different example of improving uptake of Kangaroo Mother Care intervention. Three results are possible. First, the inter- for premature and low birthweight infants in Nicaragua vention may fail to improve the outcome of interest and (Broughton and others 2013). In this case, the cost of is not cost-effective at any price. Second, the intervention the improvement intervention was less than the cost may achieve the intended improvements, but require savings realized from decreased treatment costs result- additional resources, in which case implementation is a ing from improved adherence to evidence-based stan- matter of willingness to pay for the level of improvement dards of care. achieved. Third, the intervention may improve health Despite the many difficulties in determining efficient outcomes as a result of better quality while also reducing ways to address deficits in the quality of health care, it is overall expenditure. Lower cost comes from spending a important to include these cost analyses in every quality lesser amount on care or avoiding an expensive compli- improvement intervention. Systematic accounting for cation or an adverse event. Economically, it is best to the resources and rigorous evaluation of the effects on implement all interventions matching the third result. the outcomes of interest are essential for prioritizing There is a dearth of literature on the cost-efficiency of decision making. Basic guidance on what costs to include quality of care interventions (IOM, Committee on in economic evaluations and how to analyze cost and Quality of Health Care in America 2001). Several diffi- effectiveness data is needed to move the field of health culties are involved in determining efficiency: care quality forward. • Inaccurate, incomplete, or unavailable routinely col- lected data CONCLUSIONS • Fidelity of the intervention to the outcome stated in research design In LMICs, quality of care is an emerging conversation. • The challenge of choosing comparison groups to iso- Mostly ignored a few decades ago, studies are now exam- late the variable of interest ining health system priorities once access to care has • The difficulty of capturing all of the effects of the been addressed. Conversations over the past 10 years intervention to account for positive or negative spill- have largely acknowledged the importance of quality of over effects care in resource-constrained LMICs. Quality of care is • The challenge of calibrating the extent to which discussed in all volumes of DCP3. the quality improvement can be attributed to the Quality of care matters because it relates directly intervention to outcomes and can be addressed in a shorter time • The perceived costs and economic consequences frame than other policy interventions. The updated meaningful to different audiences quality framework presented in this chapter describes • The difficulty of valuing in-kind contributions the urgency, connections, and responsibilities for cre- • The difficulty of capturing complexity of a system ating quality infrastructure that ties this responsibil- and the implications for economic evaluation. ity to individual providers through the diseases they address and the patients who access care via various health care platforms. The framework is applica- Nevertheless, CEA can show substantive returns from ble across country settings, emphasizing the funda- better quality. In one study in Niger, high quality from mental role that providers and patients play in Quality of Care 205 determining quality. With the growing evidence base Barber, S. L., S. M. Bertozzi, and P. J. Gertler. 2007. “Variations on quality improvement efforts around the globe, in Prenatal Care Quality for the Rural Poor in Mexico.” there is reason for renewed hope that quality can be Health Affairs (Millwood) 26 (3): 310–23. improved and done so rapidly. Successful policies will Barber, S. L., P. J. Gertler, and P. Harimurti. 2007. “Differences in Access to High-Quality Outpatient Care in Indonesia.” always be linked to practice on a disease-by-disease Health Affairs (Millwood) 26 (3): 352–66. basis and will only occur where access to health care Basinga, P., P. J. Gertler, A. Binagwaho, A. L. B. Soucat, is not in question. J. R. Sturdy, and others. 2010. “Paying Primary Health Care Centers for Performance in Rwanda.” Human Development Network Chief Economist’s Office and Africa Region NOTES Health, Nutrition and Population Unit, World Bank, World Bank Income Classifications as of July 2014 are as Washington, DC. follows, based on estimates of gross national income (GNI) Beegle, K., E. Frankenberg, and D. Thomas. 2001. “Bargaining per capita for 2013: Power within Couples and Use of Prenatal and Delivery Care in Indonesia.” Studies in Family Planning 32 (2): • Low-income countries (LICs) = US$1,045 or less 130–46. • Middle-income countries (MICs) are subdivided: Beracochea, E., R. Dickson, P. Freeman, and J. Thomason. 1995. (a) lower-middle-income = US$1,046 to US$4,125 “Case Management Quality Assessment in Rural Areas of (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 Papua New Guinea.” Tropical Doctor 25 (2): 69–74. • High-income countries (HICs) = US$12,746 or more. Bernhart, M. H., I. G. Wiadnyana, H. Wihardjo, and I. Pohan. 1999. “Patient Satisfaction in Developing Countries.” Social Science and Medicine 48 (8): 989–96. 1. A program that delivers insurance for maternal and child Bertelsen, J. 1981. “Who Should Abstract Medical Records? health services to uninsured families. A Study of Accuracy and Cost.” Evaluation and the Health Professions 4 (1): 79–92. Berwick, D. M. 2002. “A User’s Manual for the IOM’s ‘Quality REFERENCES Chasm’ Report.” Health Affairs (Millwood) 21 (3): 80–90. AbouZahr, C., and T. Boerma. 2005. “Health Information Bohren, M. A., E. C. Hunter, H. M. Munthe-Kaas, J. P. Souza, Systems: The Foundations of Public Health.” Bulletin of the J. P. Vogel, and others. 2014. “Facilitators and Barriers World Health Organization 83 (8): 578–83. to Facility-Based Delivery in Low- and Middle-Income Achanta, S., J. Jaju, A. M. Kumar, S. B. Nagaraja, S. R. Shamrao, Countries: A Qualitative Evidence Synthesis.” Reproductive and others. 2013. “Tuberculosis Management Practices by Health 11 (1): 71. Private Practitioners in Andhra Pradesh, India.” PLoS One Bosch-Capblanch, X., O. Ronveaux, V. Doyle, V. Remedios, and 8 (8): e71119. A. Bchir. 2009. “Accuracy and Quality of Immunization Akin, J. S., and P. Hutchinson. 1999. “Health-Care Facility Information Systems in Forty-One Low-Income Countries.” Choice and the Phenomenon of Bypassing.” Health Policy Tropical Medicine and International Health 14 (1): 2–10. and Planning 14 (2): 135–51. Boucar, M. L., K. Hill, A. Coly, S. Djibrina, Z. Saley, and Ali, M. K., S. Shah, and N. Tandon. 2011. “Review of Electronic others. 2014. “Improving Postpartum Care for Mothers Decision-Support Tools for Diabetes Care: A Viable Option and Newborns in Niger and Mali: A Case Study of for Low- and Middle-Income Countries?” Journal of an Integrated Maternal and Newborn Improvement Diabetes Science and Technology 5 (3): 553–70. Programme.” British Journal of Obstetrics and Gynaecology Ansong-Tornui, J., M. Armar-Klemesu, D. Arhinful, S. Penfold, 121 (Suppl 4): 127–33. and J. Hussein. 2007. “Hospital Based Maternity Care in Boulding, W., S. W. Glickman, M. P. Manary, K. A. Schulman, Ghana: Findings of a Confidential Enquiry into Maternal and R. Staelin. 2011. “Relationship between Patient Deaths.” Ghana Medical Journal 41 (3): 125–32. Satisfaction with Inpatient Care and Hospital Readmission Asefa, A., and D. Bekele. 2015. “Status of Respectful and Non- within 30 Days.” American Journal of Managed Care 17 (1): Abusive Care during Facility-Based Childbirth in a Hospital 41–48. and Health Centers in Addis Ababa, Ethiopia.” Reproductive Brinkerhoff, D. 2003. Accountability and Health Systems: Health 12 (1): 33. Overview, Framework, and Strategies. Bethesda, MD: Atkinson, S., R. L. R. Medeiros, P. H. L. Oliveira, and R. D. Almeida. Partners for Health Reformplus Project, Abt Associates. 2000. “Going Down to the Local: Incorporating Social Brody, C. M., N. Bellows, M. Campbell, and M. Potts. 2013. Organisation and Political Culture into Assessments of “The Impact of Vouchers on the Use and Quality of Health Decentralised Health Care.” Social Science and Medicine Care in Developing Countries: A Systematic Review.” Global 51 (4): 619–36. Public Health 8 (4): 363–88. Baker, D. W., and R. D. Cebul. 2002. “Should Consumers Trust Brook, R. H., E. A. McGlynn, and P. D. Cleary. 1996. “Measuring Hospital Quality Report Cards?” Journal of the American Quality of Care.” New England Journal of Medicine 335 (13): Medical Association 287 (24): 3206–07. 966–70. 206 Disease Control Priorities: Improving Health and Reducing Poverty Broughton, E. I., A. M. Boucar, and D. Alagane. 2012. “Cost- Davis, D. A., P. E. Mazmanian, M. Fordis, R. Van Harrison, Effectiveness of a Quality Improvement Collaborative K. E. Thorpe, and others. 2006. “Accuracy of Physician for Obstetric and Newborn Care in Niger.” International Self-Assessment Compared with Observed Measures of Journal of Health Care Quality Assurance 26 (3): 250–61. Competence: A Systematic Review.” Journal of the American Broughton, E. I., I. Gomez, N. Sanchez, and C. Vindell. 2013. Medical Association 296 (9): 1094–102. “The Cost-Savings of Implementing Kangaroo Mother Davis, D. A., M. A. O’Brien, N. Freemantle, F. M. Wolf, Care in Nicaragua.” Revista Panamericana de Salud Publica P. Mazmanian, and others. 1999. “Impact of Formal 34 (3): 176–82. Continuing Medical Education: Do Conferences, Calle, J. E., P. J. Saturno, P. Parra, J. Rodenas, M. J. Pérez, and Workshops, Rounds, and Other Traditional Continuing others. 2000. “Quality of the Information Contained in the Education Activities Change Physician Behavior or Minimum Basic Data Set: Results from an Evaluation in Health Care Outcomes?” Journal of the American Medical Eight Hospitals.” European Journal of Epidemiology 16 (11): Association 282 (9): 867–74. 1073–80. Dayal, P., and K. Hort. 2015. “Quality of Care: What Are Canchihuaman, F. A., P. J. Garcia, S. S. Gloyd, and K. K. Holmes. Effective Policy Options for Governments in Low- and 2011. “An Interactive Internet-Based Continuing Education Middle-Income Countries to Improve and Regulate the Course on Sexually Transmitted Diseases for Physicians and Quality of Ambulatory Care?” Policy Brief 4 (1), Asia Midwives in Peru.” PLoS One 6 (5): e19318. Pacific Observatory on Health Systems and Policies, WHO, Chang, J. T., R. D. Hays, P. G. Shekelle, C. H. Malean, Geneva. D. H. Solomon, and others. 2006. “Patients’ Global Ratings De Geyndt, W. 1995. “Managing the Quality of Health Care in of Their Health Care Are Not Associated with the Technical Developing Countries.” Technical Paper 258, World Bank, Quality of Their Care.” Annals of Internal Medicine 144 (9): Washington, DC. 665–72. Derose, S. F., and D. B. Petitti. 2003. “Measuring Quality of Contreras, G. Y., B. S. Olavaria, S. M. Pérez, D. P. Haemmerli, Care and Performance from a Population Health Care M. L. Cafferata, and others. 2007. “Practices in the Care Perspective.” Annual Review of Public Health 24 (1): 363–84. of the Low-Risk Delivery in Hospitals in Southern Chile.” Donabedian, A. 1980. “Methods for Deriving Criteria for Ginecología y Obstetricia de México 75 (1): 24–30. Assessing the Quality of Medical Care.” Medical Care Corrao, S., V. Arcoraci, S. Arnone, L. Calvo, R. Scaglione, and Review 37 (7): 653–98. others. 2009. “Evidence-Based Knowledge Management: Doyle, C., L. Lennox, and D. Bell. 2013. “A Systematic An Approach to Effectively Promote Good Health-Care Review of Evidence on the Links between Patient Experience Decision-Making in the Information Era.” Internal and and Clinical Safety and Effectiveness.” BMJ Open 3 (1): Emergency Medicine 4 (2): 99–106. e001570. Das, J., and J. Hammer. 2005a. “Which Doctor? Combining Dranove, D., D. Kessler, M. McClellan, and M. Satterthwaite. Vignettes and Item Response to Measure Clinical 2003. “Is More Information Better? The Effects of ‘Report Competence.” Journal of Development Economics 78 (2): Cards’ on Health Care Providers.” Journal of Political 348–83. Economy 111 (3): 555–88. ———. 2005b. “Which Doctor? Combining Vignettes and Dumont, A., L. De Bernis, D. Bouillin, A. Gueye, J. P. Dompnier, Item Response to Measure Doctor Quality.” Policy Research and others. 2002. “Maternal Morbidity and Qualification of Working Paper 3301, World Bank, Washington, DC. Health-Care Workers: Comparison between Two Different ———. 2007. “Location, Location, Location: Residence, Populations in Senegal.” Journal de Gynécologie, Obstétrique, Wealth, and the Quality of Medical Care in Delhi, India.” et Biologie de la Reproduction (Paris) 31 (1): 70–79. Health Affairs (Millwood) 26 (3): 338–51. Emanuel, E. J., and L. L. Emanuel. 1996. “What Is Accountability ———. 2014. “Quality of Primary Care in Low-Income in Health Care?” Annals of Internal Medicine 124 (2): Countries: Facts and Economics.” Annual Review of 229–39. Economics 6 (1): 525–53. Engelgau, M., K. Okamoto, K. V. Navaratne, and S. Gopalan. Das, J., A. Holla, A. Das, M. Mohanan, D. Tabak, and others. 2010. “Prevention and Control of Selected Chronic NCDs 2012. “In Urban and Rural India, A Standardized Patient in Sri Lanka: Policy Options and Action.” HNP Discussion Study Showed Low Levels of Provider Training and Paper 2010, World Bank, Washington, DC. Huge Quality Gaps.” Health Affairs (Millwood) 31 (12): Engineer, C. Y., E. Dale, A. Agarwal, A. Agarwal, O. Alonge, 2774–84. and others. 2016. “Effectiveness of a Pay-for-Performance Das, J., A. Holla, A. Mohpal, and K. Muralidharan. 2016. Intervention to Improve Maternal and Child Health “Quality and Accountability in Healthcare Delivery: Audit- Services in Afghanistan: A Cluster-Randomized Trial.” Study Evidence from Primary Care in India.” American International Journal of Epidemiology 45 (2): 451–59. Economic Review 106 (12): 3765–99. Epstein, A. M. 2006. “Pay for Performance in the United States Das, J., A. Kwan, B. Daniels, S. Satyanarayana, R. Subbaraman, and Abroad.” New England Journal of Medicine 355 (4): and others. 2015. “Use of Standardised Patients to Assess 406–08. Quality of Tuberculosis Care: A Pilot, Cross-Sectional ———. 2007. “Pay for Performance at the Tipping Point.” Study.” The Lancet Infectious Diseases 15 (11): 1305–13. New England Journal of Medicine 356 (5): 515–17. Quality of Care 207 Epstein, A. M., T. H. Lee, and M. B. Hamel. 2004. “Paying Gertler, P., P. Giovagnoli, and S. Martinez. 2014. “Rewarding Physicians for High-Quality Care.” New England Journal of Provider Performance to Enable a Healthy Start to Life: Medicine 350 (4): 406–10. Evidence from Argentina’s Plan Nacer.” Working Paper Ergo, A., L. Paina, L. Morgan, and R. Eichler. 2012. Creating 6884, World Bank Group, Washington, DC. Stronger Incentives for High-Quality Health Care in Low- Gertler, P., and C. Vermeersch. 2013. “Using Performance and Middle-Income Countries. Report for the Maternal and Initiatives to Improve Medical Care Productivity and Child Health Integrated Program (MCHIP). Washington, Health Outcomes.” Working Paper 19046, National Bureau DC: U.S. Agency for International Development (USAID). of Economic Research, Cambridge, MA. Fairbrother, G., S. Friedman, K. L. Hanson, and G. C. Butts. Glassman, P. A., J. Luck, E. M. O’Gara, and J. W. Peabody. 1997. “Effect of the Vaccines for Children Program on 2000. “Using Standardized Patients to Measure Quality: Inner-City Neighborhood Physicians.” Archives of Pediatrics Evidence from the Literature and a Prospective Study.” Joint and Adolescent Medicine 151 (12): 1229–35. Commission Journal on Quality Improvement 26 (11): 644–53. Fairbrother, G., M. J. Siegel, S. Friedman, P. D. Kory, and Glickman, S. W., W. Boulding, M. Manary, R. Staelin, M. T. Roe, G. C. Butts. 2001. “Impact of Financial Incentives on and others. 2010. “Patient Satisfaction and Its Relationship Documented Immunization Rates in the Inner City: Results with Clinical Quality and Inpatient Mortality in Acute of a Randomized Controlled Trial.” Ambulatory Pediatrics Myocardial Infarction.” Circulation: Cardiovascular Quality 1 (4): 206–12. and Outcomes 3 (2): 188–95. Felt-Lisk, S., G. Ferry, R. Roper, M. Au, J. Walker, and others. Glickman, S. W., W. Boulding, R. Staelin, J. Mulgund, 2012. “Sustainability, Partnerships, and Teamwork in M. T. Roe, and others. 2007. “A Framework for Quality Health IT Implementation: Essential Findings from the Improvement: An Analysis of Factors Responsible for Transforming Healthcare Quality through IT Grant.” Improvement at Hospitals Participating in the Can AHRQ Publication 12-0075-EF, Agency for Healthcare Rapid Risk Stratification of Unstable Angina Patients Research and Quality, Rockville, MD. Suppress Adverse Outcomes with Early Implementation Fenton, J. J., A. F. Jerant, K. D. Bertakis, and P. Franks. 2012. of the ACC/AHA Guidelines (CRUSADE) Quality “The Cost of Satisfaction: A National Study of Patient Improvement Initiative.” American Heart Journal 154 (6): Satisfaction, Health Care Utilization, Expenditures, and 1206–20. Mortality.” Archives of Internal Medicine 172 (5): 405–11. Goeree, R., L. Levin, K. Chandra, J. M. Bowen, G. Blackhouse, Fernald, L. C., P. J. Gertler, and L. M. Neufeld. 2008. “Role of and others. 2009. “Health Technology Assessment and Cash in Conditional Cash Transfer Programmes for Child Primary Data Collection for Reducing Uncertainty in Health, Growth, and Development: An Analysis of Mexico’s Decision Making.” Journal of the American College of Oportunidades.” The Lancet 371 (9615): 828–37. Radiologists 6 (5): 332–42. ———. 2009. “10-Year Effect of Oportunidades, Mexico’s Gourlay, A., A. Wringe, I. Birdthistle, G. Mshana, D. Michael, Conditional Cash Transfer Programme, on Child Growth, and others. 2014. “‘It Is Like That, We Didn’t Understand Cognition, Language, and Behaviour: A Longitudinal Each Other’: Exploring the Influence of Patient-Provider Follow-up Study.” The Lancet 374 (9706): 1997–2005. Interactions on Prevention of Mother-to-Child Transmission Flores, R., J. Robles, and B. R. Burkhalter. 2002. “Distance of HIV Service Use in Rural Tanzania.” PLoS One 9 (9): Education with Tutoring Improves Diarrhea Case e106325. Management in Guatemala.” International Journal of Grady, K. E., J. P. Lemkau, N. R. Lee, and C. Caddell. 1997. Quality Health Care 14 (Suppl 1): 47–56. “Enhancing Mammography Referral in Primary Care.” Forsetlund, L., A. Bjorndal, A. Rashidian, G. Jamtvedt, Preventive Medicine 26 (6): 791–800. M. A. O’Brien, and others. 2009. “Continuing Education Greenfield, D., and J. Braithwaite. 2008. “Health Sector Meetings and Workshops: Effects on Professional Practice Accreditation Research: A Systematic Review.” International and Health Care Outcomes.” Cochrane Database of Journal for Quality in Health Care 20 (3): 172–83. Systematic Reviews 2 (2): CD003030. Groene, O. 2011. “Patient Centredness and Quality Freedman, L. P., and M. E. Kruk. 2014. “Disrespect and Improvement Efforts in Hospitals: Rationale, Measurement, Abuse of Women in Childbirth: Challenging the Global Implementation.” International Journal for Quality in Health Quality and Accountability Agendas.” The Lancet 384 Care 23 (5): 531–37. (9948): e42–44. Grossbart, S. R. 2006. “What’s the Return? Assessing the Effect Galactionova, K., F. Tediosi, D. de Savigny, T. Smith, and of ‘Pay for Performance’ Initiatives on the Quality of M. Tanner. 2015. “Effective Coverage and Systems Care Delivery.” Medical Care Research and Review 63 (1): Effectiveness for Malaria Case Management in Sub-Saharan 29S–48S. African Countries.” PLoS One 10 (5): e0127818. Hansen, P. M., D. H. Peters, A. Edward, S. Gupta, A. Arur, and Garrido-Cumbrera, M., C. Borrell, L. Palència, A. Espelt, M. others. 2008. “Determinants of Primary Care Service in Rodríguez-Sanz, and others. 2010. “Social Class Inequalities Afghanistan.” International Journal for Quality in Health in the Utilization of Health Care and Preventive Services Care 20 (6): 375–83. in Spain, a Country with a National Health System.” Haux, R. 2006. “Individualization, Globalization, and Health: International Journal of Health Services 40 (3): 525–42. About Sustainable Information Technologies and the Aim 208 Disease Control Priorities: Improving Health and Reducing Poverty of Medical Informatics.” International Journal of Medical Jovanovic, B. 2005. “Hospital Accreditation as Method for Informatics 75 (12): 795–808. Assessing Quality in Healthcare.” Archive of Oncology Haynes, A. B., T. G. Weiser, W. R. Berry, S. R. Lipsitz, 13 (3–4): 156–57. A. H. Breizat, and others. 2009. “A Surgical Safety Checklist Kantrowitz, B. 2014. “The Science of Learning.” Scientific to Reduce Morbidity and Mortality in a Global Population.” American 311 (2): 68–73. New England Journal of Medicine 360 (5): 491–99. Kaptanoğlu, A. Y., and I. Aktas. 2013. “Measuring Quality of Health Affairs. 2011. “Achieving Equity in Health.” Health Care Using a Vignette-Based Survey in Turkish Hospitals.” Policy Brief, October 6. Korean Journal of Critical Care Medicine 1 (1): 5–7. ———. 2014. “Geographic Variation in Medicare Spending.” Karolinski, A., P. Micone, R. Mercer, L. Gibbons, F. Althabe, Health Policy Brief, March 6. and others. 2009. “Evidence-Based Maternal and Perinatal Heaton, C. 2000. “External Peer Review in Europe: An Overview Healthcare Practices in Public Hospitals in Argentina.” from the ExPeRT Project. External Peer Review Techniques.” International Journal of Gynaecology and Obstetrics 105 (2): International Journal for Quality in Health Care 12 (3): 177–82. 118–22. Hillman, A. L., K. Ripley, N. Goldfarb, I. Nuamah, J. Weiner, and Kerr, E. A., and B. Fleming. 2007. “Making Performance others. 1998. “Physician Financial Incentives and Feedback: Indicators Work: Experiences of U.S. Veterans Health Failure to Increase Cancer Screening in Medicaid Managed Administration.” British Medical Journal 335 (7627): Care.” American Journal of Public Health 88 (11): 1699–701. 971–73. Hillman, A. L., K. Ripley, N. Goldfarb, J. Weiner, I. Nuamah, Kirkpatrick, D. H., and R. T. Burkman. 2010. “Does and others. 1999. “The Use of Physician Financial Incentives Standardization of Care through Clinical Guidelines and Feedback to Improve Pediatric Preventive Care in Improve Outcomes and Reduce Medical Liability?” Medicaid Managed Care.” Pediatrics 104 (4, Pt 1): 931–35. Obstetrics and Gynecology 116 (5): 1022–26. Holm, M., and H. Burkhartzmeyer. 2015. “Implementation of Koh, H. C., and G. Tan. 2005. “Data Mining Applications in a Phased Medical Educational Approach in a Developing Healthcare.” Journal of Healthcare Information Management Country.” Global Health Action 8 (November): 29882. 19 (2): 64–72. Hsia, D. C. 2003. “Medicare Quality Improvement: Bad Kohn, L. T., J. M. Corrigan, and M. S. Donaldson, eds. 2000. To Apples or Bad Systems?” Journal of the American Medical Err Is Human: Building a Safer Health System. Washington, Association 289 (3): 354–56. DC: National Academies Press. Hsia, R. Y., N. A. Mbembati, S. Macfarlane, and M. E. Kruk. Kolstad, J. T. 2013. “Information and Quality When Motivation 2012. “Access to Emergency and Surgical Care in Sub- Is Intrinsic: Evidence from Surgeon Report Cards.” Saharan Africa: The Infrastructure Gap.” Health Policy and American Economic Review 103 (7): 2875–910. Planning 27 (3): 234–44. Kouides, R. W., N. M. Bennett, B. Lewis, J. D. Cappuccio, IOM (Institute of Medicine). 2013. Announcement. Crossing W. H. Barker, and others. 1998. “Performance-Based Physician the Quality Chasm: The IOM Health Care Quality Initiative. Reimbursement and Influenza Immunization Rates in the Washington, DC: Health and Medicine Division, National Elderly. The Primary-Care Physicians of Monroe County.” Academy of Sciences, Engineering, Medicine. American Journal of Preventive Medicine 14 (2): 89–95. IOM, Committee on Planning a Continuing Health Kruk, M. E., S. Hermosilla, E. Larson, and G. M. Mbaruku. Professional Education Institute. 2010. Redesigning 2014. “Bypassing Primary Care Clinics for Childbirth: A Continuing Education in the Health Professions. Washington, Cross-Sectional Study in the Pwani Region, United Republic DC: National Academies Press. of Tanzania.” Bulletin of the World Health Organization 92 IOM, Committee on Quality of Health Care in America. 2001. (4): 246–53. Crossing the Quality Chasm: A New Health System for the Kruk, M. E., S. Hermosilla, E. Larson, D. Vail, Q. Chen, and 21st Century. Washington, DC: National Academies Press. others. 2015. “Who Is Left Behind on the Road to Universal IOM and National Academy of Engineering. 2011. “Healthcare Facility Delivery? A Cross-Sectional Multilevel Analysis in System Complexities, Impediments, and Failures.” In Rural Tanzania.” Tropical Medicine and International Health Engineering a Learning Healthcare System: A Look at the 20 (8): 1057–66. Future, Workshop Summary, edited by C. Grossmann, Kruk, M. E., M. Paczkowski, G. Mbaruku, H. de Pinho, and W. A. Goolsby, L. Olsen, and J. M. McGinnis, 117–70. S. Galea. 2009. “Women’s Preferences for Place of Delivery Washington, DC: National Academies Press. in Rural Tanzania: A Population-Based Discrete Choice Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, Experiment.” American Journal of Public Health 99 (9): S. Berkley, and others. 2013. “Global Health 2035: A World 1666–72. Converging within a Generation.” The Lancet 382 (9908): Kujawski, S., G. Mbaruku, L. P. Freedman, K. Ramsey, W. Moyo, 1898–955. and others. 2015. “Association between Disrespect and Jörg, F., N. Borgers, A. J. Schrijvers, and J. J. Hox. 2006. Abuse during Childbirth and Women’s Confidence in “Variation in Long-Term Care Needs Assessors’ Willingness Health Facilities in Tanzania.” Maternal and Child Health to Support Clients’ Requests for Admission to a Residential Journal 19 (10): 2243–50. Home: A Vignette Study.” Journal of Aging and Health LaFond, A., and M. Siddiqi. 2003. “Second International 18 (6): 767–90. RHINO Workshop on Enhancing the Data Quality and Use Quality of Care 209 of Routine Health Information at District Level.” RHINO Luck, J., J. W. Peabody, T. R. Dresselhaus, M. Lee, and P. Glassman. (Routine Health Information Network), Mpekweni Sun, 2000. “How Well Does Chart Abstraction Measure Quality? Eastern Cape, South Africa, September 29–October 4. A Prospective Comparison of Standardized Patients with Lagarde, M., A. Haines, and N. Palmer. 2009. “The Impact of the Medical Record.” American Journal of Medicine 108 (8): Conditional Cash Transfers on Health Outcomes and Use 642–49. of Health Services in Low- and Middle-Income Countries.” Marchant, T., R. D. Tilley-Gyado, T. Tessema, K. Singh, Cochrane Database of Systematic Reviews 4 (October): M. Gautham, and others. 2015. “Adding Content to CD008137. Contacts: Measurement of High-Quality Contacts for Leonard, K. L. 2014. “Active Patients in Rural African Health Maternal and Newborn Health in Ethiopia, North East Care: Implications for Research and Policy.” Health Policy Nigeria, and Uttar Pradesh, India.” PLoS One 10 (5): and Planning 29 (1): 85–95. e0126840. Leonard, K. L., and M. C. Masatu. 2006. “Outpatient Process Massoud, R., and N. Abrampah. 2015. “Scaling-Up of High- Quality Evaluation and the Hawthorne Effect.” Social Impact Interventions.” In Improving Aid Effectiveness in Science and Medicine 63 (9): 2330–40. Global Health, edited by E. Beracochea, 139–47. New York: Leonard, K. L., G. R. Mliga, and D. H. Mariam. 2002. “Bypassing SpringerVerlag. Health Centres in Tanzania: Revealed Preferences for McGlynn, E. A. 1997. “Six Challenges in Measuring the Quality Quality.” Journal of African Economics 11 (4): 441–71. of Health Care.” Health Affairs (Millwood) 16 (3): 7–21. Li, L., Z. Wu, Y. Zhao, C. Lin, D. Detels, and others. 2007. “Using McGlynn, E. A., S. M. Asch, J. Adams, J. Keesey, J. Hicks, Case Vignettes to Measure HIV-Related Stigma among and others. 2003. “The Quality of Health Care Delivered Health Professionals in China.” International Journal of to Adults in the United States.” New England Journal of Epidemiology 36 (1): 178–84. Medicine 384 (26): 2635–45. Lindenauer, P. K., D. Remus, S. Roman, M. B. Rothberg, Mohanan, M., M. Vera-Hernandez, V. Das, S. Giardili, E. M. Benjamin, and others. 2007. “Public Reporting and J. D. Goldhaber-Fiebert, and others. 2015. “The Know-Do Pay for Performance in Hospital Quality Improvement.” Gap in Quality of Health Care for Childhood Diarrhea and New England Journal of Medicine 356 (5): 486–96. Pneumonia in Rural India.” Journal of the American Medical Lindfield, R., A. Knight, and D. Bwonya. 2015. “An Approach Association Pediatrics 169 (4): 349–57. to Assessing Patient Safety in Hospitals in Low-Income Mostofian, F., C. Ruban, N. Simunovic, and M. Bhandari. 2015. Countries.” PLoS One 10 (3): e0121628. “Changing Physician Behavior: What Works?” American Lindfield, R., K. Vishwanath, F. Ngounou, and R. C. Khanna. Journal of Managed Care 21 (1): 75–84. 2012. “The Challenges in Improving Outcome of Cataract Mtango, F. D., and D. Neuvians. 1986. “Acute Respiratory Surgery in Low- and Middle-Income Countries.” Indian Infections in Children under Five Years: Control Project Journal of Ophthalmology 60 (5): 464–69. in Bagamoyo District, Tanzania.” Transactions of the Royal Lippeveld, T., R. Sauerborn, and C. Bodart. 2000. Design Society of Tropical Medicine and Hygiene 80 (6): 851–58. and Implementation of Health Information Systems. Geneva: Murray, C. J., E. E. Gakidou, and J. Frenk. 1999. “Health WHO. Inequalities and Social Group Differences: What Should Lium, J. T., A. Tjora, and A. Faxvaag. 2008. “No Paper, but the We Measure?” Bulletin of the World Health Organization 77 Same Routines: A Qualitative Exploration of Experiences (7): 537–43. in Two Norwegian Hospitals Deprived of the Paper-Based National Academies of Sciences, Engineering, and Medicine. Medical Record.” BMC Medical Informatics and Decision 2015. Improving Quality of Care in Low- and Middle-Income Making 8 (2): PMC2245928. Countries. Washington, DC: National Academies Press. Loevinsohn, B. P., E. T. Guerrero, and S. P. Gregorio. 1995. National Patient Safety Foundation. 2015. Shining a Light: Safer “Improving Primary Health Care through Systematic Health Care through Transparency; Report of the Roundtable Supervision: A Controlled Field Trial.” Health Policy and on Transparency. Boston, MA: National Patient Safety Planning 10 (2): 144–53. Foundation. Lozano, R., P. Soliz, E. Gakidou, J. Abbott-Klafter, Ng, M., N. Fullman, J. L. Dieleman, A. D. Flaxman, C. J. Murray, D. M. Freehan, and others. 2006. “Benchmarking of and others. 2014. “Effective Coverage: A Metric for Performance of Mexican States with Effective Coverage.” Monitoring Universal Health Coverage.” PLoS Medicine The Lancet 368 (9548): 1729–41. 11 (9): e1001730. Luck, J., and J. W. Peabody. 2002. “Using Standardised Nolan, T., P. Angos, A. J. Cunha, L. Muhe, S. Qazi, and Patients to Measure Physicians’ Practice: Validation Study others. 2001. “Quality of Hospital Care for Seriously Using Audio Recordings.” British Medical Journal 325 Ill Children in Less-Developed Countries.” The Lancet (7366): 679. 357 (9250): 106–10. Luck, J., J. W. Peabody, L. M. DeMaria, C. S. Alvarado, and OECD (Organisation for Economic Co-Operation and R. Menon. 2014. “Patient and Provider Perspectives on Development). 2015. Health Care Quality Indicators. Paris: Quality and Health System Effectiveness in a Transition OECD. Economy: Evidence from Ukraine.” Social Science and Okafor, I. I., E. O. Ugwu, and S. N. Obi. 2015. “Disrespect Medicine 114 (August): 57–65. and Abuse during Facility-Based Childbirth in a 210 Disease Control Priorities: Improving Health and Reducing Poverty Low-Income Country.” International Journal of Gynecology the Quality of Health Care?” Annals of Internal Medicine and Obstetrics 128 (2): 110–13. 145 (4): 265–72. Parsons, A., C. McCullough, J. Wang, and S. Shih. 2012. Pronovost, P., D. Needham, S. Berenholtz, D. Sinopoli, H. Chu, “Validity of Electronic Health Record–Derived Quality and others. 2006. “An Intervention to Decrease Catheter- Measurement for Performance Monitoring.” Journal of Related Bloodstream Infections in the ICU.” New England the American Medical Informatics Association 19 (4): Journal of Medicine 355 (26): 2725–32. 604–9. Quimbo, S. A., J. W. Peabody, R. Shimkhada, K. Woo, Peabody, J. W., L. DeMaria, S. N. Nguyen, O. Smith, A. Hoth, and O. Solon. 2008. “Should We Have Confidence If a and others. Forthcoming. “Quality of Care in Six Eastern Physician Is Accredited? A Study of the Relative Impacts European Countries Using Clinical Performance and Value of Accreditation and Insurance Payments on Quality of Vignettes: A Cross-Sectional Study.” Bulletin of the World Care in the Philippines.” Social Science and Medicine 67 (4): Health Organization. 505–10. Peabody, J. W., P. J. Gertler, and A. Leibowitz. 1998. “The Policy Ransom, S. B., W. W. Pinsky, and J. E. Tropman, eds. 2000. Implications of Better Structure and Process on Birth Enhancing Physician Performance: Advanced Principles of Outcomes in Jamaica.” Health Policy 43 (1): 1–13. Medical Management. Tampa, FL: American College of Peabody, J. W., and A. Liu. 2007. “A Cross-National Comparison Physician Executives. of the Quality of Clinical Care Using Vignettes.” Health Rivera, J. A., D. Sotres-Alvarez, J. P. Habicht, T. Shamah, Policy and Planning 22 (5): 294–302. and S. Villalpando. 2004. “Impact of the Mexican Peabody, J. W., J. Luck, L. M. DeMaria, and R. Menon. 2014. Program for Education, Health, and Nutrition (Progresa) “Quality of Care and Health Status in Ukraine.” BMC on Rates of Growth and Anemia in Infants and Health Services Research 14: 446. Young Children: A Randomized Effectiveness Study.” Peabody, J. W., J. Luck, P. A. Glassman, T. R. Dresselhaus, and Journal of the American Medical Association 291 (21): M. Lee. 2000. “Comparison of Vignettes, Standardized 2563–70. Patients, and Chart Abstraction: A Prospective Validation Rockers, P. C., M. E. Kruk, and M. J. Laugesen. 2012. “Perceptions Study of 3 Methods for Measuring Quality.” Journal of the of the Health System and Public Trust in Government in American Medical Association 283 (13): 1715–22. Low- and Middle-Income Countries: Evidence from the Peabody, J. W., J. Luck, P. Glassman, S. Jain, M. Spell, and others. World Health Surveys.” Journal of Health Politics, Policy, and 2004. “Measuring the Quality of Physician Practice by Law 37 (3): 405–37. Using Clinical Vignettes: A Prospective Validation Study.” Rogers, W. A. 2004. “Evidence-Based Medicine and Justice: Annals of Internal Medicine 141 (10): 771–80. A Framework for Looking at the Impact of EBM upon Peabody, J. W., R. J. Nordyke, F. Tozija, J. Luck, J. A. Munoz, and Vulnerable or Disadvantaged Groups.” Journal of Medical others. 2006. “Quality of Care and Its Impact on Population Ethics 30 (2): 141–45. Health: A Cross-Sectional Study from Macedonia.” Social Rosenthal, M. B., R. G. Frank, Z. Li, and A. M. Epstein. Science and Medicine 62 (9): 2216–24. 2005. “Early Experience with Pay-for-Performance: From Peabody, J. W., S. Quimbo, J. Florentino, R. Shimkhada, Concept to Practice.” Journal of the American Medical X. Javier, and others. 2017. “Comparative Effectiveness Association 294 (14): 1788–93. of Two Disparate Policies on Child Health: Experimental Saleh, S. S., M. S. Alameddine, and N. M. Natafgi. 2014. Evidence from the Philippines.” Health Policy and Planning “Beyond Accreditation: A Multi-Track Quality-Enhancing czw179. doi:http://dx.doi.org/10.1093/heapol/czw179. Strategy for Primary Health Care in Low- and Middle- Peabody, J. W., R. Shimkhada, S. Quimbo, J. Florentino, Income Countries.” International Journal of Health Services M. Bacate, and others. 2011. “Financial Incentives and 44 (2): 355–72. Measurement Improved Physicians’ Quality of Care in the Saleh, S. S., J. Bou Sleiman, D. Dagher, H. Sbeit, and N. Natafgi. Philippines.” Health Affairs (Millwood) 30 (4): 773–81. 2013. “Accreditation of Hospitals in Lebanon: Is It a Worthy Peabody, J. W., S. Shimkhada, S. Quimbo, O. Solon, X. Javier, Investment?” International Journal for Quality in Health and others. 2014. “The Impact of Performance Incentives Care 25 (3): 284–90. on Health Outcomes: Results from a Cluster Randomized Sando, D., T. Kendall, G. Lyatuu, H. Ratcliffe, K. McDonald, and Controlled Trial in the Philippines.” Health Policy and others. 2014. “Disrespect and Abuse during Childbirth in Planning 29 (5): 615–21. Tanzania: Are Women Living with HIV More Vulnerable?” Peabody, J. W., M. M. Taguiwalo, D. A. Robalino, and J. Frenk. Journal of Acquired Immune Deficiency Syndromes 67 (Suppl 2006. “Improving the Quality of Care in Developing 4): S228–34. Countries.” In Disease Control Priorities in Developing Schuster, R. J., N. A. Terwoord, and J. Tasosa. 2006. “Changing Countries, second edition, edited by D. T. Jamison, J. G. Physician Practice Behavior to Measure and Improve Breman, A. R. Measham, G. Alleyne, M. Claeson, D. B. Clinical Outcomes.” American Journal of Medical Quality Evans, P. Jha, A. Mills, and P. Musgrove. Washington, DC: 21 (6): 394–400. World Bank. Scott, K. W., and A. K. Jha. 2014. “Putting Quality on the Petersen, L. A., L. D. Woodard, T. Urech, C. Daw, and Global Health Agenda.” New England Journal of Medicine S. Sookanan. 2006. “Does Pay-for-Performance Improve 371 (1): 3–5. Quality of Care 211 Sendlhofer, G., N. Mosbacher, L. Karina, B. Kober, L. Jantscher, Cross-Sectional Study.” African Journal of Reproductive and others. 2015. “Implementation of a Surgical Safety Health 18 (1): 16–26. Checklist: Interventions to Optimize the Process and Hints Tiemeier, H., W. J. de Vries, M. van Het Loo, J. P. Kahan, to Increase Compliance.” PLoS One 10 (2): e0116926. N. Klazinga, and others. 2002. “Guideline Adherence Rates Shaller, D., S. Sofaer, S. Findlay, J. Hibbard, D. Lansky, and Interprofessional Variation in a Vignette Study of and others. 2003. “Consumers and Quality-Driven Health Depression.” Quality and Safety in Health Care 11 (3): Care: A Call to Action.” Health Affairs (Millwood) 22 (2): 214–18. 95–101. Tran, B. X., V. M. Nong, R. M. Maher, P. K. Nguyen, and Shimkhada, R., O. Solon, D. Tamondong-Lachica, J. W. Peabody. H. N. Luu. 2014. “A Systematic Review of Scope and Quality 2016. “Misdiagnosis of Obstetrical Cases and the Clinical of Health Economic Evaluation Studies in Vietnam.” PLoS Cost and Consequences to Patients: A Cross-Sectional One 9 (8): e103825. Study of Urban Providers in the Philippines.” Global Health Treadwell, J. R., S. Lucas, and A. Y. Tsou. 2014. “Surgical Action 9 (1): 32672. Checklists: A Systematic Review of Impacts and Solon, O., K. Woo, S. A. Quimbo, R. Shimkhada, J. Florentino, Implementation.” BMJ Quality and Safety 23 (4): 299–318. and others. 2009. “A Novel Method for Measuring Health U.S. Congress, Office of Technology Assessment. 1994. Care System Performance: Experience from QIDS in the Identifying Health Technologies that Work: Searching for Philippines.” Health Policy and Planning 24 (3): 167–74. Evidence. OTA-H-608. Washington, DC: U.S. Government Spertus, J. A., M. J. Radford, N. R. Every, E. F. Ellerbeck, Printing Office. E. D. Peterson, and others. 2003. “Challenges and Van der Bij, J. D., T. Vollmar, and M. C. Weggeman. 1998. Opportunities in Quantifying the Quality of Care for “Quality Systems in Health Care: A Situational Approach.” Acute Myocardial Infarction: Summary from the Acute International Journal of Health Care Quality Assurance Myocardial Infarction Working Group of the American 11 (2): 65–70. Heart Association/American College of Cardiology First Veloski, J., S. Tai, A. S. Evans, and D. B. Nash. 2005. “Clinical Scientific Forum on Quality of Care and Outcomes Vignette-Based Surveys: A Tool for Assessing Physician Research in Cardiovascular Disease and Stroke.” Circulation Practice Variation.” American Journal of Medical Quality 107 (12): 1681–91. 20 (3): 151–57. Starfield, B., N. R. Powe, J. R. Weiner, M. Stuart, D. Steinwachs, Vogel, J. P., A. M. Gülmezoglu, G. J. Hofmeyr, and and others. 1994. “Costs vs. Quality in Different Types of M. Temmerman. 2014. “Perspectives on Elective Induction Primary Care Settings.” Journal of the American Medical of Labor.” Clinical Obstetrics and Gynecology 7 (2): 331–42. Association 272 (24): 1903–08. Wachter, R. 2013. “Personal Accountability in Healthcare: Steinwachs, D. M., and R. G. Hughes. 2008. “Health Services Searching for the Right Balance.” BMJ Quality and Safety Research: Scope and Significance.” In Patient Safety and 22 (2): 176–80. Quality: An Evidence-Based Handbook for Nurses, edited by Weinberg, N. 2001. “Using Performance Measures to Identify R. G. Hughes, 8–43. Rockville, MD: Agency for Healthcare Plans of Action to Improve Care.” Joint Commission Journal Research and Quality. on Quality and Patient Safety 27 (12): 683–88. Strauss, A., and J. Corbin. 1998. Basics of Qualitative Research: Werner, R. M., and D. A. Asch. 2007. “Clinical Concerns about Techniques and Procedures for Developing Grounded Theory. Clinical Performance Measurement.” Annals of Family 2nd ed. Newbury Park, CA: Sage. Medicine 5 (2): 159–63. Sun, B. C., J. Adams, E. J. Orav, D. W. Rucker, T. A. Brennan, WHO (World Health Organization). 2000. World Health and others. 2000. “Determinants of Patient Satisfaction Report 2000: Health Systems, Improving Performance. and Willingness to Return with Emergency Care.” Annals of Geneva: WHO. Emergency Medicine 35 (5): 426–34. ———. 2008. Summary of the Evidence on Patient Safety: Sylvia, S., Y. Shi, H. Xue, X. Tian, H. Wang, and others. 2015. Implications for Research. Geneva: WHO. “Survey Using Incognito Standardized Patients Shows Poor WHO, Association for Trauma Surgery and Intensive Quality Care in China’s Rural Clinics.” Health Policy and Care, and International Society of Surgery. 2009. Planning 30 (3): 322–33. Guidelines for Trauma Quality Improvement Programmes. Tafreshi, M. Z., M. Pazargadi, and Z. Abed Saeedi. 2007. Geneva: WHO. “Nurses’ Perspectives on Quality of Nursing Care: A WHO and World Bank. 2014. Monitoring Progress towards Qualitative Study in Iran.” International Journal of Health Universal Health Coverage at Country and Global Levels: Care Quality Assurance 20 (4): 320–28. Framework, Measures, Targets. Geneva: WHO; Washington, Thaddeus, S., and D. Maine. 1994. “Too Far to Walk: Maternal DC: World Bank. Mortality in Context.” Social Science and Medicine 38 (8): Wisniewski, M., and H. Wisniewski. 2005. “Measuring Service 1091–110. Quality in a Hospital Colposcopy Clinic.” International Thorsen, V. C., T. Meguid, J. Sundby, and A. Malata. 2014. Journal of Health Care Quality Assurance 18 (2–3): 217–29. “Components of Maternal Healthcare Delivery System Woolf, S. H. 2000. “Changing Physician Practice Behavior.” Contributing to Maternal Deaths in Malawi: A Descriptive Journal of Family Practice 49 (2): 126–29. 212 Disease Control Priorities: Improving Health and Reducing Poverty World Bank. 2003. World Development Report 2004: Making China: A Condition-Based Outcome Approach.” PLoS One Services Work for Poor People. Washington, DC: World Bank. 10 (10): e0138948. ———. 2014. Health Results Innovation Trust Fund Annual Zaslavsky, A. M. 2001. “Statistical Issues in Reporting Report 2014. Washington, DC: World Bank. Quality Data: Small Samples and Casemix Variation.” Xu, Y., Y. Liu, T. Shu, W. Yang, and M. Liang. 2015. “Variations International Journal for Quality in Health Care 13 (6): in the Quality of Care at Large Public Hospitals in Beijing, 481–88. Quality of Care 213 Chapter 11 High-Quality Diagnosis: An Essential Pathology Package Kenneth A. Fleming, Mahendra Naidoo, Michael Wilson, John Flanigan, Susan Horton, Modupe Kuti, Lai Meng Looi, Christopher P. Price, Kun Ru, Abdul Ghafur, Jinaxiang Wang, and Nestor Lago INTRODUCTION fungal infection, Epstein-Barr virus infection (infectious A young child living in Sub-Saharan Africa presents to a mononucleosis), and malignant lymphoma—and that rural health care clinic with a one-week history of fevers, accurate diagnosis requires pathology investigations, night sweats, chills, and malaise. The child’s mother does including both microbiology and anatomic pathology. not know if the child has lost weight in the recent past; The family leaves the clinic, and the patient is lost to when weighed, the child is significantly below the follow-up. expected weight for her age. No other family members, This scenario is played out daily in many countries including other young siblings, report similar symp- across the world and illustrates one aspect of the crucial toms. Physical examination reveals a fever, mild increase role that pathology has in ensuring effective health care, in heart and respiratory rates, and enlarged lymph nodes namely, diagnosis. Despite recent progress in controlling along both sides of her neck. The clinic does not have communicable disease, the need for pathology is growing access to imaging studies, and the only available pathol- as the burden of noncommunicable diseases increases. ogy laboratory tests show that the patient does not have There were approximately 14 million new cases of serologic evidence of human immunodeficiency virus/ cancer and 8.2 million cancer-related deaths in 2012 acquired immune deficiency syndrome (HIV/AIDS) (Stewart and Wild 2014), but treating these cases accu- infection or malaria. She is mildly anemic as measured rately is impossible unless the pathological diagnosis is by a manual spun hematocrit. The physician wants to known. Cancer is predicted to increase by 70 percent by refer the patient to a hospital in a nearby city, but the 2032, with more than 60 percent of these new cases in family does not have sufficient resources. Asia, Central and South America, and Sub-Saharan The physician offers to collect blood for pathology Africa. Similarly, diagnosing and treating patients with testing and send it to that hospital for testing, but because diabetes mellitus—another developing epidemic in low- the hospital requires advance payment for pathology and middle-income countries (LMICs)—is impossible tests, the family again does not have the resources. without the ability to measure the levels of glucose in the The physician completes the notes, indicating that the blood. The diagnosis and risk stratification of cardiovas- differential diagnosis is broad—including tuberculosis, cular disease requires pathology, for example, to check nontuberculous mycobacterial infection, disseminated levels of serum lipids such as cholesterol. Corresponding author: Kenneth A. Fleming, Center for Global Health, National Cancer Institute, Washington, DC, United States; Green Templeton College, University of Oxford, United Kingdom; kenneth.fleming@medsci.ox.ac.uk. 215 This chapter specifies an essential minimal package of other fluids and involves, for instance, clinical bio- services that should be available in LMICs to provide chemistry, microbiology, and hematology access to pathology services that are of acceptable qual- • Anatomic pathology, which is concerned with cell ity, affordable, and timely to a majority of the popula- and tissue analysis and involves cytology, histology, tion, especially outside of major cities. and autopsy. In high-income countries (HICs), pathology services RANGE OF PATHOLOGY SERVICES typically are provided in one of three ways: The term pathology means the study of disease. The knowledge gained from this study has led to development • Central laboratories that deliver most of their ser- of the many diagnostic tests used in clinical practice. vices in hospital settings. Central laboratories have These tests are performed on body fluids, including a common infrastructure that supports their various blood, urine, sweat, saliva, and sputum; on tissue biopsies; components, including specimen collection services, and on cells obtained from needle-aspirated specimens. transport and reception, and a mechanism for trans- The diagnostic role is a key aspect of what pathology mitting the results of tests and accompanying reports laboratories do and is fundamental to the effective work- to the ordering clinicians and patients. They have ing of any health care system. An interview-based study laboratory information systems (LIS) that are ideally of cardiologists and oncologists in Germany and the connected to electronic patient records. United States indicated that 66 percent of clinical deci- • Smaller laboratories in more rural environments sions are based on results of in vitro diagnostic tests that offer a more limited repertoire of tests, as well as (Rohr and others 2016). point-of-care testing (POCT) in community settings. Pathology also supports clinical care by assessing • A small number of laboratories, often in conjunction disease severity and prognosis, for example, determin- with university departments, that provide the most ing the staging and grading of a cancer by histopathol- specialized tests. These laboratories also undertake ogy; this information is fundamental to deciding and research, both in the field of pathology itself and with managing treatment plans for patients. Equally impor- other disciplines as part of multidisciplinary teams. tant is the role of the pathology laboratory in monitor- They also organize and provide education and train- ing clinical response to treatment, for example, analyzing ing in pathology and related disciplines. blood levels of markers of renal function in patients with renal failure. Although the core of laboratory activities may be Pathology plays a number of other key roles. One is considered the performance of tests and the analysis of quality assurance within the health care system. In 2013, the results (the analytical phase), it is important to rec- autopsies showed an estimated 20 percent major dis- ognize that the pre- and postanalytical phases are crepancy between the pre-mortem clinical diagnosis and equally important for generating accurate laboratory the autopsy diagnosis (Kuijpers and others 2014). test results (box 11.1). These phases range from the Similarly, through the examination of surgical speci- selection of the most appropriate tests or investigations mens, surgeons can learn whether they are fully excising to the interpretation of their results and the provision of tumors; through the use of microbiological culturing, clinical advice across the spectrum of medical special- physicians can correctly identify the cause of a fever. ties. In practice, this involvement may require a review Pathology contributes to disease surveillance by helping of medical records and discussions with ordering identify new and emerging diseases such as the Zika clinicians. An example is the multidisciplinary meeting virus; pathology facilitates the maintenance of disease (tumor boards in the United States), in which patholo- registries that help inform national health policy and gists, surgeons, and chemotherapy and radiation oncol- allocation of resources. Finally, forensic pathology is ogists, radiologists, nurses, and others involved in cancer integral to legal systems around the world. care of a patient meet to review all relevant information In all of these roles, pathology services encompass a and decide on the best approach for treatment and number of disciplines and subspecialties; table 11.1 management. describes the main ones. In the United States and most Pathologists may also provide leadership for hospital- other regions, these pathology disciplines are divided wide quality assurance efforts. Increasingly, pathologists into two main groups: are assuming additional clinical roles in many health systems, for example, serving as infectious disease doc- • Clinical pathology, also called laboratory medicine, tors, managing patients with metabolic disorders, and which is largely concerned with analysis of blood and providing specialized oncology services. 216 Disease Control Priorities: Improving Health and Reducing Poverty Table 11.1 Major Pathology Disciplines and Roles Clinical biochemistry Study of the biochemical basis of disease Cytopathology Study of disease in individual cells Forensic pathology Determination of cause and manner of death for legal purposes Hematology Study of blood disorders a Histopathology Study of disease in human tissue Immunopathology Study of the immunologic basis of disease Medical microbiology Study of infection Molecular pathology and genetics Study of the molecular and genetic basis of diseases and heritable conditions Pediatric and perinatal pathology Study of the diseases of pregnancy, childbirth, and children Transfusion medicine Study of the collection, preparation, storage, and clinical use of blood products Note: A selection of the major disciplines was derived, in part, from https://www.rcpath.org. a. Histopathology includes a number of subdisciplines, such as dermatopathology, neuropathology, and others that focus on diseases of a single organ or organ system. Box 11.1 Three Phases of Laboratory Testing • Preanalytical phase. Selecting the appropriate In HICs, the largest proportion of errors in test, obtaining the specimen, labeling it with the pathology occurs in the pre- and postanalytical patient’s name, providing timely transport to the phases (Plebani 2009). In the preanalytical phase, laboratory, registering receipt in the laboratory, these errors include failing to ensure that the spec- and processing before testing. imen is collected from the right patient, that the • Analytical phase. Performing the test and inter- correct specimen type is collected, and that the preting the result. specimen is collected at the right time. In the post- • Postanalytical phase. Preparing a report detailing analytical phase, errors include reporting the the result and its interpretation, authorizing the wrong result and failing to read the report, making report, and transmitting the report to the clinician the wrong or no decision, or taking the wrong or so that the clinician can institute appropriate action. no action. Clearly, pathology is not a stand-alone service. Its Ideally, the public sectors of LMICs should have value is as a crucial and integral part of the system of care three tiers of laboratories—with a small additional in which the outcomes for patients and the operational number of national or regional research or refer- and economic benefits for the system depend on all of ence laboratories (WHO AFRO and U.S. CDC 2010). the parts working effectively together. Without accurate The tier 1 laboratories are widely distributed in the diagnosis, everything else is compromised. community and typically perform a small number of simple clinical pathology tests. Tier 2 and tier 3 labo- CHALLENGES TO PATHOLOGY SERVICES IN ratories are progressively fewer in number, provide tests of increasing complexity and capacity, and are LMICs found in progressively larger population centers. The child described in the clinical vignette at the begin- In many countries, however, especially poorer ones, ning of this chapter needed access to microbiology, such structures do not exist. Their absence has several hematology, and immunology services, and she almost causes, the most important of which is lack of human certainly would have needed access to the expertise of a capacity, resulting in far too few trained personnel to histopathologist. Yet access to diagnostic pathology ser- staff the laboratories to provide adequate population vices is not available in many countries and regions. coverage at all levels. High-Quality Diagnosis: An Essential Pathology Package 217 Inadequate Staffing A characteristic of many LMICs is that private Data on staffing are lacking for much of the world, but laboratories—most staffed by pathologists from the the available data illustrate the problem. In Sub-Saharan public sector—are often run in parallel to the public Africa, at least five countries have no anatomic patholo- sector and provide services to the population. The facil- gist. Surveys of the other countries in the region have ities in some of these laboratories can be as good as any shown that the number of anatomic pathologists per internationally, but many are much less satisfactory. In patient population is approximately 1:1,000,000, or about India, where 70 percent of the laboratories are private, one-fiftieth the ratio in the United Kingdom and the only 1 percent are accredited (Singh 2013). In Kampala, United States (Adesina and others 2013; African Strategies Uganda, which had more than 900 laboratories in for Advancing Pathology Group Members 2015). In 2011—96 percent of which were private—only 45 labo- China, there were approximately 10,300 pathologists in ratories achieved the first step of the five-step process for all disciplines in 2015 (unpublished data from Chinese international accreditation (Elbireer and others 2013). Society of Pathology 2015), constituting an estimated The result of these challenges is that much of the shortfall of 60,000–120,000. In 2014, there were only eight population in LMICs does not have access to quality pathologists in a population of 14 million in Cambodia pathology services. As noncommunicable diseases that (Vathana and Stauch 2014); the ratio of pathologists per are particularly reliant on pathology for diagnosis and patient population in Vietnam was estimated to be management become more prevalent, the level of mis- 1:254,000 (Van Dang 2014). In upper-middle-income diagnosis is likely to rise. This increase will result in countries, the situation is somewhat better; for example, unnecessary deaths and avoidable prolonged illness in Malaysia the ratio is 1:75,000 (Looi 2008). and distress, with attendant social disruption and neg- ative impacts on productivity. The deficiencies also mean that data needed for disease surveillance and Variable Standards registries, and other types of population data needed to In addition to staff shortages, widely variable standards guide public policy and resource allocation, are not are an issue. Although the quality of services, particularly available. In addition, good quality pathology is neces- those provided in large cities in middle-income coun- sary for the achievement of 11 of the 13 goals of the tries, can be good, frequently it is seriously inadequate in United Nation’s health-related Sustainable Development both urban and rural areas (Daramola and others 2016; Goals (table 11.2); the deficiencies will impede attain- Orem and others 2012). ment of these goals. Table 11.2 Health-Related Sustainable Development Goals and Pathology Is pathology Sustainable Development Goals relevant? Specific pathology examples 3.1: By 2030, reduce the global maternal mortality ratio to less than Yes Testing for most common causes of maternal mortality, 70 per 100,000 live births for example, infections; also blood transfusion for hemorrhage and autopsy to establish cause of death 3.2: By 2030, end preventable deaths of newborns and children Yes Testing and monitoring for most common causes of under age five years, with all countries aiming to reduce neonatal infant mortality, for example, infections, autopsy mortality to at least as low as 12 per 1,000 live births and under- five mortality to at least as low as 25 per 1,000 live births 3.3: By 2030, end the epidemics of HIV/AIDS, tuberculosis, malaria, Yes Testing for communicable diseases, for example, blood and neglected tropical diseases, and combat hepatitis, water-borne tests for HIV/AIDS and malaria, antiretroviral resistance diseases, and other communicable diseases 3.4: By 2030, reduce by one-third premature mortality from Yes Histo- and cytopathology for cancer diagnosis; noncommunicable diseases through prevention and treatment and hematology and biochemistry for diabetes diagnosis promote mental health and well-being and management; pathology support for surveillance and other data platforms, for example, cancer registries 3.5: Strengthen the prevention and treatment of substance abuse Yes Toxicology tests 3.6: By 2020, halve the number of global deaths and injuries from Yes Autopsy reports, blood banks for transfusion support road traffic accidents table continues next page 218 Disease Control Priorities: Improving Health and Reducing Poverty Table 11.2 Health-Related Sustainable Development Goals and Pathology (continued) Is pathology Sustainable Development Goals relevant? Specific pathology examples 3.7: By 2030, ensure universal access to sexual and reproductive Yes Blood and urine testing for pregnancy and for sexually health care services, including family planning, information, and transmitted diseases education; and the integration of reproductive health into national strategies and programs 3.8: Achieve universal health coverage, including financial risk No — protection; access to quality essential health care services; and access to safe, effective, quality, and affordable essential medicines and vaccines 3.9: By 2030, substantially reduce the number of deaths and Yes Toxicology testing and diagnosis of related diseases illnesses from hazardous chemicals, and air, water, and soil pollution and contamination 3a: Strengthen the implementation of the World Health Yes Testing for smoking cessation in urine Organization Framework Convention on Tobacco Control in all countries, as appropriate 3b: Support the research and development of vaccines and Yes Pathology systems provide data, for example, medicines for the communicable and noncommunicable diseases surveillance, and research platforms that primarily affect LMICs; provide access to affordable essential medicines and vaccines, in accordance with the Doha Declaration on the Trade-Related Aspects of Intellectual Property Rights (TRIPS) Agreement and Public Health 3c: Substantially increase health financing and the recruitment, No — development, training, and retention of the health workforce in LMICs, particularly in LICs and small island LMICs 3d: Strengthen the capacity of all countries, particularly LMICs, for Yes Surveillance for emerging disease and through cancer early warning, risk reduction, and risk management of national and registries global health risks Note: HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome; LICs = low-income countries; LMICs = low- and middle-income countries; — = not applicable. THE ESSENTIAL PATHOLOGY PACKAGE the country and its pathology capacity, there may be a highly specialized laboratory performing complex test- The essential pathology package consists of a minimal ing that can act as a referral center for the country or suite of services that should be available in LMICs to even a region. These last two levels will often have edu- provide access to pathology that is of reasonable quality, cational and research capacity and be part of a university affordable, and timely to a majority of the population, medical school. especially that outside the main cities. The key concept This model is similar to the three-tier model in many is an integrated network of tiered laboratories (box 11.2, LMICs (WHO AFRO and U.S. CDC 2010); the crucial table 11.3), the tiers being similar to that described in aspect is that the model must be an integrated network the previous section. Thus, tier 1 is widely distributed in of laboratories for more efficient and effective referral of the community (both rural and urban). It has limited patients across networks than would be the case with pathology capacity and staffing but can perform some independent laboratories. This approach provides econ- basic tests and can refer patients and specimens to the omies of scale, such as sharing use and costs of staff, next tier. The next tier has many fewer laboratories, equipment, and reagents. Other benefits include better probably located in sizable towns. It has greater capacity, communication, exchange of staff and knowledge, pro- performing most routine tests and when necessary, can vision of education and training, and opportunities for refer more specialized tests to the to the next tier. These research. This integrated approach would result in devel- next-tier laboratories will probably be based in the opment of a critical mass of expertise and the optimal largest towns and are capable of performing all routine use of scarce resources. tests and many specialized ones. Finally, depending on High-Quality Diagnosis: An Essential Pathology Package 219 Box 11.2 Definition of Laboratory Tiers • Tier 1. Primary care and health center laboratories • Tier 4. Laboratories in national or teaching primarily serving outpatients in community set- hospitals that receive specimens from their own tings, performing point-of-care tests and single-use patients and receive referrals from tier 1, 2, and tests, and referring more complex work to tier 2 or 3 facilities. They provide routine tests and highly tier 3. These laboratories are staffed at the techni- specialized tests. In small countries, these facil- cian level. ities may be regional and shared by more than • Tier 2. Laboratories in first-level hospitals that one country. receive specimens from their own patients and receive referrals from tier 1 facilities. Usually, they Each country and region has a different burden have a pathologist and perform a selected number of disease and availability of staff, and some of routine tests. shifting of capacity may occur across the tier • Tier 3. Laboratories in second-level hospitals that boundaries. For example, if a tier 2 pathologist receive specimens from their own patients and makes regular visits, then fine needle aspiration receive referrals from tier 1 and 2 facilities. These cytology could be performed and reported in laboratories have significant numbers of pathol- a tier 1 laboratory. In many countries, shortages ogy staff and cover all routine testing in the major of staff require that one laboratory fulfill the pathology disciplines. functions of both tier 3 and 4. Table 11.3 Pathology Tiers Laboratory Tier 2 (includes tier 1 Tier 3 (includes tier 2 Tier 4 (includes tier 3 features Tier 1 capabilities) capabilities) capabilities) Tests and test POCT and single- Many routine diagnostic and All routine and some specialized Specialized services as categories use tests: malaria, prognostic tests tests appropriate, surveillance, tuberculosis, urinalysis, Clinical biochemistry Clinical chemistry toxicology studies, support for pregnancy, blood glucose, transplantation, rare tumors, hemoglobin/hematocrit, urea and electrolytes, HBA1c for Endocrine tests: thyroid; cardiac nutritional studies, support ESR, blood typing diabetes, liver, renal, bone, and markers, troponin, BNP; dynamic for clinical trials, mutational Slide microscopy: malaria, lipid profiles function tests, GTT; tumor studies (cytogenetics, wet preparation, stool Hematology markers: AFP, Ca-125, blood gases; molecular analysis), gene parasites therapeutic drug monitoring; serum analysis complete blood counts, CD4 and urine electrophoresis Preparation of FNAC and count, simple coagulation tissue specimens to send studies and thalassemia Microbiology to tier 2 facilities tests, support for whole Additional antimicrobial blood transfusion susceptibility testing, Microbiology culture fungal cultures, mycobacterial cultures, viral load blood, urine, cerebrospinal fluid, sputum; simple Hematology antimicrobial susceptibility More advanced blood analysis, testing; serology for hepatitis A, for example, component therapy, B, or C and common infections hemolysis, bone marrow studies, hematological malignancies, immunological studies table continues next page 220 Disease Control Priorities: Improving Health and Reducing Poverty Table 11.3 Pathology Tiers (continued) Laboratory Tier 2 (includes tier 1 Tier 3 (includes tier 2 Tier 4 (includes tier 3 features Tier 1 capabilities) capabilities) capabilities) Anatomic pathology Anatomic pathology FNAC, tissue biopsies and Same as for tier 2, but with surgical excisions—processing, special stains including H&E stain and interpretation immunohistochemistry: ER, Hospital autopsy PR for breast cancer Specialized autopsy Staffing Laboratory technicians General pathologist, laboratory Mono-specialty pathologists, clinical Same as for tier 3 plus supervised by general technicians, laboratory scientists, specialized laboratory clinical trial specialists, data pathologist from distance assistants; one of technicians technicians, laboratory assistants, specialist manages laboratory dedicated laboratory manager, Additional specialist possibly laboratory information educational capacity systems coordinator, quality care manager Facilities and responsibilities for education and training of all levels of medical and nonmedical staff Communication Paper or electronic, mobile Paper or electronic or laboratory Electronic or laboratory information Same as tier 3 but more infrastructure information system system; telepathology (optional) data linkages to trials and registries Equipment Simple microscope Automated blood and Automated tissue processor, Molecular biology and Rapid diagnostic tests biochemistry analyzers; equipment for full autopsy, cytogenetics microbiology analyzers and immunohistochemistry station Immunofluorescence POCT and single-use tests incubators; blood typing including refrigerators; tissue Electron microscopy for renal processor and microtome for disease anatomic pathology Specimen and patient Possible biobanking for identification research FNAC and biopsy fixation Turnaround time Rapid, POCT, and single- An hour to several days Routine: 1 hour to several days Same as tier 3 use tests: 0–3 hours Complex: 7 days Send-outs, several days Autopsy: 30–60 days Networks and Accumulates and forwards Report to emerging disease, Links to emerging disease, AST, Research on disease incidence surveillance incidence data to higher AST, cancer, and other NCD cancer, and other NCD registries trends, including AST and tier registries emerging diseases Note: AFP = alpha-fetoprotein; AST = antimicrobial susceptibility testing; BNP = brain natriuretic peptide; Ca-125 = cancer antigen 125; ER and PR = receptor tests for breast cancer; ESR = erythrocyte sedimentation rate; FNAC = fine needle aspiration cytology; GTT = glucose tolerance test; H&E = hematoxylin and eosin stain (basic histopathology test); HBA1c = glycated hemoglobin test; NCD = noncommunicable disease; POCT = point-of-care tests. Assumptions 1. Tiers may be adjusted as necessary to reflect the local burden of disease or local practice patterns and availability of trained staff. 2. Changes in technologies over time can shift tests and workloads across tiers. 3. Tests are examples (as applied to broad groups of infectious disease, cancer, and other NCD) and are not an exhaustive list. In 2008, such national integrated laboratory sys- model was subsequently endorsed in the Freetown tems were proposed as a key development for pathol- Declaration of 2015 (ASLM and WHO AFRO 2015) ogy services in Sub-Saharan Africa in the Maputo as the cornerstone of effective health care. Although Declaration on Strengthening of Laboratory Systems infectious diseases were the focus of the original (WHO AFRO 2008). Ethiopia was one of the first model, the principles are equally applicable to non- countries to successfully develop such a model; the communicable diseases. High-Quality Diagnosis: An Essential Pathology Package 221 A key component in ensuring the sustainability of This administrative oversight is a key leadership res- such a model is the tier 4 laboratory. These centers would ponsibility required by International Organization offer specialized services as well as develop and provide for Standardization (ISO) 15189:2012, the international research, education, and training, especially to the linked reference document for best laboratory practice (ISO 2012). tier 1 and 2 facilities. Furthermore, these centers are Laboratories produce information that result from most likely to develop innovations appropriate to the their processes, personnel, and equipment. This informa- country’s needs. Without these fostering and supporting tion is also influenced by the clinical settings in which the roles, the long-term sustainability of the lower-tier laboratories operate and from which they receive speci- laboratories will not be feasible. Linking such facilities to mens. Patient-specific, disease-specific, and therapy- other centers of excellence (North-South, South-South) specific factors may influence the information that the to provide access to further expertise and resources is laboratories produce. Those in leadership positions need important for continuing long-term development. to understand the interactions between these factors, The model outlined in box 11.2 is intended to especially as those interactions affect how the informa- represent the minimum that a lower-middle-income tion will be used for patient care. The Joint Commission country would provide. Countries at higher levels of International’s accreditation standards for hospitals state development can build on this model to deliver increased that for the purpose of clinical consultation and render- provision appropriate to their needs. Conversely, the ing of medical opinion, the laboratory should be led model serves as a goal for LICs to achieve as resources by physicians, preferably pathologists (JCI 2014). become available and are invested. Pathologists, as clinicians, have insights into the thought To ensure this network is sustainable, effective, and of processes behind requests for laboratory tests and the good quality, five components are vital: decisions that may be made with the information received. These insights are not only invaluable in determining • Leadership how to most effectively organize and direct laboratory • Education, training, and continuing professional services, but they are also crucial to provision of clinical development advice on the further investigation and management of • Emerging technologies individual patients. Clinical scientists, who have had • Quality management and accreditation training significantly similar to that received by clinical • Reimbursement policies for pathology services. pathologists, may also provide this level of leadership. Reflecting the integral role that pathology plays in the wider heath care system, laboratory leadership also needs Leadership to be involved in the development of national strategic The effective and efficient operation of a pathology labo- plans for laboratories. These plans detail the long-term ratory is a multidisciplinary effort. Pathology services are vision and mission of the nation’s laboratory services. primarily delivered by three groups of professionally To be effective, development of this national blueprint qualified staff—pathologists, clinical scientists, and tech- needs to recognize the local disease burden, available nicians (also referred to as technologists)—supported by clinical skills and services, clinical requirements for diag- assistants, managers and administrators, and technology nosis and monitoring, and technical realities. The pri- specialists. In most places, clinical scientists or techni- mary involvement of clinical laboratory leadership, in cians undertake the role of administrator or manager. conjunction with other clinicians, is to provide guidance Pathologists provide leadership and serve as the interface for the definition of policy that delineates the organiza- between laboratory and clinical services; in some coun- tion, scope, and nature of the laboratory service accord- tries and specialties, pathologists share these roles with ing to the tiers providing health care in the respective clinical scientists. Pathologists and clinical scientists countries (WHO AFRO and U.S. CDC 2010). also oversee quality improvement and service develop- Pathologists provide leadership at the operational level. ment as well as pathology-led research and development. Doing so entails the ability to read about and understand Laboratory technologists are responsible for delivering scientific and technological advances in the field of medi- the technical aspects of the service. cine as well as improvements in laboratory technology. The goal of this joint effort is to provide a service that is Changing clinical demands for patient care, as docu- patient oriented and meets clinical needs. These clinical mented in new and revised versions of locally applicable needs are defined by standards of care, expectations of clinical care guidelines, require a laboratory director’s individual physicians, and patients. Accordingly, laboratory involvement and informed response. Similarly, advances in leadership needs to monitor the activities of staff to the technical capacity of laboratories, including the intro- ensure that clinically relevant services are being provided. duction of new tests and the withdrawal of obsolete ones, 222 Disease Control Priorities: Improving Health and Reducing Poverty need to be assessed in relation to their ability to improve These training courses are largely experiential in nature, the clinical effectiveness of the laboratory, as well as the with considerable hands-on involvement in pathology clinical effectiveness and cost-effectiveness of the whole service delivery supplemented by small group teaching care pathway. To effectively lead the response to such and formal lectures. changes, pathologists need the authority to alter aspects of the operations to ensure that laboratories remain true to Clinical Scientists and Technologists their goal of enhancing the quality of patient care. In some countries, clinical scientists perform functions similar to those of pathologists. They follow a similar pathway of education and training to achieve the Education, Training, and Continuing Professional required competence, for example, in clinical biochem- Development istry, immunology, microbiology, or virology. Clinical Educating and training larger numbers of qualified per- scientists may also be responsible for the performance of sonnel is clearly of paramount importance in developing specialized services, such as molecular genetics, toxicol- a sustainable pathology network. There are three major ogy investigations, and electron microscopy. These indi- categories of staff: pathologists, clinical scientists and viduals generally have degrees in chemistry, biological technologists, and technicians. Their education consists science, or biomedical science, usually followed by a largely of a combination of formal courses for degrees master’s or doctoral degree in such areas as microbiology and diplomas and hands-on training and experience or clinical biochemistry. The training period is four to under the supervision of qualified individuals. eight years. There may be subsequent subspecialization in such fields as virology. Pathologists Technologists are also sometimes referred to as medi- Historically, pathologists in LMICs were educated in cal laboratory scientists or biomedical scientists. Their Australia, Europe, and North America; the individuals education and training in some places involves the often resided in the HICs for the duration of their train- acquisition of a university degree, while in others it is ing programs. Although those funded by governments similar to that of technicians. or charities were expected or required to return home when the training was completed, large numbers stayed Technicians in HICs. In contrast, clinical scientists and technicians Technical staff are usually educated and trained through predominantly received their education locally. college courses, often part-time over several years. The Pathologists are medically qualified practitioners who education may encompass all of the specialities of have undergone postgraduate education and training in pathology or it may be restricted to one of the major pathology. There are three main models of training; the specialities, such as anatomic pathology or microbiol- first two are common in LMICs: ogy; such specialization is a feature of more developed laboratory services. In some countries, technical staff do • In the first training model, pathologists are trained not have formal qualifications and only receive hands-on as generalists dealing with all aspects of pathology, training in the laboratory. both clinical and anatomic; this is also called general In most countries, in addition to the professional pathology. This postgraduate training period is usu- qualification or appropriate university degree, individu- ally two to four years. In some countries, the course als need to be registered with the national registration entails a university degree. body as an indication of required competence before • In the second model, pathologists are trained only being allowed to practice. as either clinical or anatomic pathologists. The post- LMICs have increasingly developed their own pathol- graduate training period is two to three years. ogist postgraduate educational and training systems. In • In the third model, pathologists are trained as mono- Sub-Saharan Africa, 21 countries have developed train- specialists, for example, as hematologists, microbiolo- ing programs in the past 25 years. In the 14 countries for gists, or clinical biochemists. Such individuals tend to be which comparative data are available, the number of employed in academic centers. This model reflects coun- pathologists increased from 70 in 1990 to 370 in 2015 tries with more-developed health care systems, such as (Nelson and others 2016). Similarly, in Malaysia, the South Africa. The postgraduate training period is usually number of pathologists increased from approximately a minimum of four years. In much of South America, 50 in the 1980s (Jegathesan and de Witt 1982) to more pathologists are only trained as mono-specialty ana- than 500 in 2016 (Looi 2008). tomic pathologists; the other disciplines of pathology are However, in many countries, especially low-income staffed by clinical scientists, such as clinical biochemists. countries (LICs), the shortage is such that training High-Quality Diagnosis: An Essential Pathology Package 223 enough pathologists to fully staff all relevant sections of encounter. These considerations need to be balanced health care systems is not possible, even in the medium against the generally higher cost of providing POCT, term. Accordingly, the expansion of the training of scien- albeit resulting in savings elsewhere in the care pathway, tists and technicians and the exploration of task-shifting and the technical challenges of generating accurate test and task-sharing are needed, with parallel development results at that level. of shorter training programs focused on specific tasks, A tiered system of laboratory testing that focuses on such as cytology screening. the type of care provided within each tier, as well as the A program of continuing professional development number of tests performed within each tier, can be used (CPD) is necessary to maintain the standards and long- to design approaches to testing. For example, tier 1 facil- term sustainability of the pathology network. Many ities would most benefit from POCT; tier 3 facilities individuals and institutions provide CPD events, often would benefit most from centralized laboratory testing. delivered by visiting individuals and organizations, on Test devices used for disease surveillance purposes can an informal basis; systematic institutional and national be designed for centralized use only. programs are rare in LMICs. One of the most common Device manufacturers and public-private partner- support requests from pathologists in LMICs is for pro- ships have developed new technologies for laboratory vision of and access to CPD. Without such programs, testing to provide both POCT and centralized testing the knowledge and skills of individuals can become out within a tiered system of health care delivery, increase of date, especially as the pace of advances accelerates. and improve access to laboratory testing in general, and bring new diagnostic tests to the public. Key challenges for the development and use of emerging tests are shown Emerging Technologies in box 11.3. In particular, simplicity of specimen collec- Diagnostics tion, device use, and interpretation and communication In all health care systems, the need for medical tests at of test results are critically important because new any point in the care pathway requires that specimens be devices will be used in many LMICs by persons with collected and sent to laboratories for analysis and inter- widely varying languages, backgrounds, training, and pretation. Laboratory testing can be centralized, pro- expertise. vided at the point of care, or more typically a combination Many of today’s laboratory analyzers require a reli- of the two. The selection of which approach to take is able external power supply, and because electricity partly driven by the availability of a given test at the supply can be intermittent in many LMICs, even with point of care, the level of test volumes, and the need back-up facilities such as diesel generators, there is to have test results available at the time of the patient increasing focus on developing devices that require no Box 11.3 Effectiveness Criteria for Emerging Tests • Any new tests should provide results for a speci- • Manufacturers’ claims regarding test perfor- fied clinical problem to guide clinical decisions, for mance characteristics should be independently monitoring disease status or response to therapy, verified. or for data collection for disease surveillance. • Test platforms should be usable and stable in • Results of tests designed to be used in clinical care locations of intended use. should be available in a time frame that will guide • Test platforms should meet procurement require- clinical decision making. ments for supply chain, maintenance, availability • Tests should be easy to perform, and results must of quality control standards, durability, and sta- be easy to interpret and communicate. bility in variable climatic conditions. • Target performance characteristics—such as • Test costs should be affordable in locations of sensitivity, specificity, predictive values, precision, intended use. and accuracy—for the intended uses should be specified before test development. Source: Based on Wu and Zaman 2012. 224 Disease Control Priorities: Improving Health and Reducing Poverty power or have built-in power generation (Pollock and and limited antimicrobial susceptibility testing for others 2013; Whitesides and Wilding 2012; Yetisen, tuberculosis has significantly enhanced global efforts in Akram, and Lowe 2013). In addition, because of the diagnosis and treatment (WHO 2015a). challenges of supply chains and storage in many LMICs, However, the use of small specimen volumes causes interest is growing in developing POCT devices that substantial challenges in the design of systems that can require minimal or no reagents other than the devices yield consistent test results (Bond and Richards-Kortum and that can be stored for long periods in hot and humid 2015). As a result, POCT may not produce test results climates with no performance degradation. For larger that agree with those generated by larger laboratory ana- analyzers used in central laboratories, one goal is to lyzers. The results from POCTs need to be harmonized develop test platforms that can support a number of with those from a central laboratory analyzer so that different assays rather than platforms that are unique to health care providers are familiar with any variations in one set of tests. The development of flexible platforms the results. would minimize the number of devices needed, with associated reductions in acquisition and maintenance Data Handling costs; it would also allow for rapid introduction of new Clinical laboratories generate large volumes of data for assays, a particularly important consideration in light of patient care as well as for quality control and other emerging diseases in LMICs. laboratory-management operations. As access to labora- Molecular diagnostic techniques have historically tory services increases in LMICs, paper reporting systems been substantially more expensive and required techni- will not support the high volumes of data. An integrated, cal expertise and laboratory infrastructure unavailable in tier-based laboratory system requires the ability to trans- most LMICs. This field of diagnostics is rapidly evolving mit data to and from multiple testing sites as well as to to the point where some tests are becoming practicable forward results to clinicians and selected test results to for use in LMICs (St. John and Price 2014), and this patients for self-monitoring, to public health authorities, trend is likely to accelerate. Access to these tests is and to disease registries. These data-handling needs will becoming a routine part of health care delivery because only be achieved by the use of LIS (NPP 2014). Although a number of diseases and conditions are only detect- many commercial systems are not affordable in LMICs, able using these methods. For example, many cancers open-source systems are available that may provide are now classified using molecular tests, and the use of opportunities for local use. Development of robust, reli- some drugs requires molecular testing to determine able LIS that can be integrated with other parts of health whether specific biomarkers are present. care data systems needs to be a priority in all regions. Mobile phones may facilitate the process. Point-of-Care Testing Part of the data used in diagnostic testing consists of POCT is usually performed by medical staff, nurses, or images, including for surgical pathology (histopathology) medical assistants using small, mobile testing devices. It and cytopathology, hematology (blood smear examina- can be used anywhere on the care pathway—first-level, tions), microbiology (identification of parasites based on second-level, or third-level care—as well as in patients’ morphologic examination), microscopic examination of homes. This approach differs from centralized laboratory urine specimens, and malaria smears. One approach to testing, which is performed by specialized technicians diagnostic testing, consultation, and quality control is the using large-capacity (high-throughput) analyzers. use of telepathology—the transmission of images via Although POCT technologies are broadly based on Internet connections to and from remote sites. Previously, the same techniques used in centralized laboratory ana- this technology was expensive and required access to lyzers, they have reduced reagent and sample volume bandwidth not available in most of the world. More requirements, rely upon stabilization of reagents, and recently, costs have decreased, and improved Internet typically use single-use cassettes for testing. connectivity is available in many regions. In LMICs, POCT has been used extensively to help guide the treatment of several diseases and conditions. Expanded access to POCT is cost-effective in extending Quality Management and Accreditation life expectancy in patients with HIV/AIDS (Cassim and Although access to quality pathology laboratory testing others 2014; Hyle and others 2014; Wu and Zaman is an essential part of modern medical practice, in some 2012). Access to smear microscopy, rapid malaria diag- settings most laboratories are not accredited and do not nostic testing, or both has played an important role in meet minimal standards for good laboratory practice. decreasing malaria-related morbidity and mortality These laboratories are unlikely to consistently generate (WHO 2015b). Access to rapid detection of infection accurate or reliable test results. The absence of accurate High-Quality Diagnosis: An Essential Pathology Package 225 and reliable results can lead to incorrect diagnoses, inap- Audit practices have extended beyond internal activ- propriate treatment, wasted resources, and even lost ities to assessments by third parties using national and lives. Such situations give credence to the saying that “no international peer-determined standards. The formal test is better than a bad test.” assessment of laboratories by independent external agencies against such standards, known as accredita- Causes of Suboptimal Testing tion, is the norm in HICs, where requirements for Laboratory testing is a complex process with preanalytical, laboratory practices are often mandated by law. Apart analytical, and postanalytical phase variables (box 11.1). from ensuring quality, accreditation status affects Considering analytical influences alone, test methodolo- the profitability and marketability of laboratories; only gies affect the magnitude of false positive and false accredited tests are reimbursed by health insurance. negative results. Sensitivity and specificity profiles influ- Through mutual recognition agreements, such as the ence choices for screening and confirmatory tests. The Asia-Pacific Laboratory Accreditation Cooperation, competence of personnel, regular quality control, state the Inter-American Accreditation Cooperation, and the of equipment and laboratory infrastructure, and access International Laboratory Accreditation Cooperation, to reagents affect the accuracy of test results. A lapse in the tests performed by accredited laboratories are rec- any step in the long chain of processes can result in ognized by signatories across country boundaries. incorrect and potentially harmful test results. Ethics and In LMICs, the culture of interlaboratory comparison, accountability are as important in laboratories as in any audit, and accreditation has yet to become firmly estab- other component of health care. lished. In India, it is estimated that fewer than 1 percent of the approximately 100,000 pathology laboratories are Quality Management accredited (Singh 2013). A 2013 survey reported that To control these variables, it is essential that laboratories more than 90 percent of countries in Sub-Saharan Africa make the commitment to a quality management system had no laboratories accredited to international quality and organization structure that ensures that tests are fit- standards; of the laboratories that were accredited, more for-purpose, standard operating procedures are docu- than 90 percent were in South Africa (Schroeder and mented and followed, personnel are suitably qualified and Amukele 2014). Laboratory accreditation has not been trained, and regular audits are conducted. The practice of established in many LMICs in Southeast Asia, partly interlaboratory comparisons, such as external quality because most LMICs do not have national health insur- assurance (EQA) and proficiency testing (PT) programs, ance plans, and the incentive of reimbursement for tests has evolved to encourage laboratories to meet validated conducted by accredited laboratories does not apply. In performance benchmarks. Many comprehensive EQA addition, most LMICs lack strong regulatory oversight and PT programs are available regionally and globally of laboratory practice. Laboratory tests performed (box 11.4). These programs vary in strength; some are by public laboratories, which are frequently resource educational, while others have a validation focus. constrained, are heavily subsidized by governments, Box 11.4 Examples of External Quality Assessment Programs International Programs National and Local Programs • Royal College of Pathologists of Australasia Quality • Bureau of Laboratory Quality Standards, Thailand Assurance Programs, Australia • External Quality Assessment schemes of Faculty of • National External Quality Assessment Services, Medical Technology, Mahidol University, Thailand United Kingdom • Laboratory Quality Assurance Scheme, Malaysia • College of American Pathologists, United States • National Center for Clinical Laboratories, China • Randox International Quality Assessment Scheme, • Indian Association of Medical Microbiologists, India international • National Health Laboratory Service, South Africa • International Academy of Pathology, international (this program extends to other Sub-Saharan with regional and national divisions. African countries). 226 Disease Control Priorities: Improving Health and Reducing Poverty while private laboratories benefit from out-of-pocket Laboratory Improvement Process Towards Accreditation payments. EQA and PT are not mandatory. The situa- checklist and the Strengthening Laboratory Management tion pits profit against quality, and many LMICs struggle Toward Accreditation training curriculum. These pro- with the mushrooming of corner shop–type private lab- grams were jointly developed with the U.S. Centers for oratories with substandard practices and questionable Disease Control and Prevention, the Clinton Health accountability. Access Initiative, and the American Society for Clinical However, practices in many emerging economies are Pathology to assist laboratories to move toward accred- rapidly changing, and laboratory accreditation is now itation status (Gershy-Damet and others 2010). actively sought. Although most laboratories started by Although much remains to be done, these tools have seeking accreditation from foreign agencies (for exam- transformed the laboratory mindset and practice land- ple, Australia’s National Association of Testing Agencies scape in Sub-Saharan Africa (Alemnji and others 2014; and the College of American Pathologists), this approach Yao and others 2014). has proved unsustainable because of the high expense. The cooperation of the WHO, governments, and Today, government-backed national accreditation agen- national professional bodies has been crucial in the cies adopting international standards, especially ISO global paradigm shift in laboratory testing to quality and 15189 for medical testing laboratories, provide assess- international standardization. However, many challenges ments at a more reasonable cost. Examples of accredita- remain for LMICs; the most important are resource con- tion agencies are listed in box 11.5. straints; establishment of national EQA, PT, and accred- However, legislation-backed regulation of laboratories itation programs; and legislation-backed regulation of in LMICs remains the exception (Looi 2008; Wattanasri, laboratories. Ensuring the long-term, good quality of the Manoroma, and Viriyayudhagorn 2010), and partici- services provided by the essential pathology package pation in EQA or PT programs and accreditation is requires the adoption of an appropriate form of accred- entirely voluntary. For these emerging economies, the itation, within which EQA is embedded. impetus to gain accreditation has been competition and market driven, especially in light of trade agreements such as the ASEAN (Association of Southeast Asian Reimbursement Policies for Pathology Services Nations) Free Trade Area, the World Trade Organization, Pathology tests are almost universally costed according and the imminent Trans-Pacific Partnership Agreement. to the complexity and the volume of tests performed, In Sub-Saharan Africa, because public laboratories often referred to as the cost-per-test or activity-based are the main providers of services, the WHO Regional costing. Who pays for the tests varies and is closely Office for Africa in 2009 introduced the Stepwise related to overall health reimbursement policies. Box 11.5 Examples of Accreditation Bodies • College of American Pathologists, Laboratory • China National Accreditation Service for Accreditation Program, United States Conformity Assessment, China • Joint Commission International, United States • Hong Kong Accreditation Service, Hong Kong • National Association of Testing Authorities, SAR, China Australia • National Accreditation Board for Testing and • South African National Accreditation System, Calibration Laboratories, India South Africa • Bureau of Laboratory Quality Standards, • United Kingdom Accreditation Service, United Thailand Kingdom • Medical Technology Council, Thailand • International Accreditation New Zealand, New • Department of Standards Malaysia, Malaysia Zealand • General Coordination for Accreditation, Brazil • Comité Français d’Accréditation, France • Bureau of Accreditation, Vietnam • Standards Council of Canada, Canada • Komite Akreditasi Nasional, Indonesia. High-Quality Diagnosis: An Essential Pathology Package 227 China has a complex reimbursement system for countries, self-payment is more common. Payment for pathology services. The national health care system testing is made in advance, with patients and families accounts for the majority of medical reimbursement, but purchasing the necessary supplies to perform the tests in individual provinces and cities have their own differing addition to paying the fee required for testing. reimbursement policies. This variation is reflected in the Some LMICs have community-based health insur- big gap in health care benefits between wealthy and poor ance programs that households can join, but the cover- regions in China (Chen, Zhao, and Si 2014; Pan and oth- age provided varies. Ghana’s program covers only ers 2016). In Tianjin, a large city with a population in hospital-based services. In Bangladesh, nongovernmen- excess of 13 million people, the health care policy states tal organizations operate insurance programs and cover that public medical insurance covers approximately services in their own clinics. Whether laboratory tests 70 percent of laboratory testing provided in local hospi- are covered in these programs depends on the details of tals. The remaining laboratory tests are paid on an out- the particular programs (Robyn, Sauerborn, and of-pocket basis. In practice, however, the government Bärnighausen 2013; Soors and others 2010; Wang 2012). usually only reimburses basic laboratory tests; because The key factor that applies to all programs is that complex tests carry high price tags, only 40 percent of the both patients and clinicians worldwide have a tendency actual cost of pathology testing is covered (Lei, Chen, and to prefer to use their limited financial resources for treat- Lu 2014; Mao 2012; Pan and others 2014). In addition, ment rather than diagnosis. If payment is out of pocket, the circumstances under which pathology tests can be the tendency is for fewer, less complex, and lower-quality used are restricted. The result is that most of the burden tests; the opposite is the case when reimbursement is of the costs of laboratory tests falls on patients. In some provided by national or private programs. Invariably, rural areas, especially the more rural regions of western this bias reduces the eventual quality of the outcome. China, coverage of medical costs, including pathology Moreover, it adversely affects the ability of health care services, is even less generous. systems and governments to standardize health care In India—with more than 40,000 hospitals and 100,000 delivery, collect epidemiological data, and assess the diagnostic laboratories—the private sector delivers 70 effectiveness of policies and interventions. percent of health care, including laboratory services. To optimize the benefits of pathology provision, as lit- Public financing for health care is less than 1 percent of tle as possible of the costs should be on an out-of-pocket gross domestic product; only 17 percent of the population basis. Where countries adopt a model of universal health is covered by any kind of health insurance. Accordingly, coverage, we propose that pathology reimbursement be an more than 70 percent of health expenditures, including integral component of the reimbursement system. Clearly, for pathology services, is borne by families as out- it will be important to ensure that in such a model, pathol- of-pocket payments (The Hindu 2014). ogy costs are kept in check, for example, by the institution In Sub-Saharan Africa, the picture is mixed. In South of guidelines on the use of tests. Africa 80 percent of the population has health care, including pathology, paid for by the government. Patients only make a payment if they can afford to. About Economics of Pathology in Different Countries 7 percent have personal insurance, while the remainder This section analyzes the costs of pathology laboratories pay out of pocket. A similar situation exists in Zimbabwe using data from countries with different income levels and Botswana. In East Africa, there is a mixture of and with varied health systems (table 11.4). These analy- government, insurance, and self-payment. In other ses provide some interesting insights, although data are Table 11.4 Approximate Annual Salary of Pathology Staff, by Country Income Category, 2010 U.S. Dollars WHO employee category and corresponding pathology staff Low-income country Lower-middle-income country 2: laboratory assistant (secondary education or diploma) 2,220 4,800 3: laboratory technician (bachelor’s degree) 2,870 6,170 4: scientific officer (master’s degree) 4,550 9,800 Pathologist (physician with additional training) 13,650 29,400 Source: Based on ongoing estimates from Serje 2015. Note: WHO = World Health Organization. The WHO data are from International Labour Organization salary databases. Equivalencies for technicians, and construction of the top category at three times the salary of category 4 by authors, also is based on unpublished data for Tata Memorial Hospital, Mumbai, as a guideline. 228 Disease Control Priorities: Improving Health and Reducing Poverty limited and not always readily comparable. These varia- Some diagnostic areas are more standardized and tions on unit costs of tests help explain why estimating the more automated than others. Data from the United costs of an essential pathology package is challenging. Kingdom (Department of Health, United Kingdom 2008) found that the median direct cost—excluding equipment Pathology’s Share of Health Costs costs, costs of space, and overhead costs—of a specific One study for the United States suggested that labora- routine test in biochemistry across a sample of laborato- tory tests account for 4.5 percent to 10 percent of total ries was £1.00 compared with £2.40 in hematology, £6.90 health expenditures (Avivar 2012), compared with in microbiology, and £48.10 in histopathology (2006/07 5 percent for Spain (Avivar 2012), 3.3 percent for the costs) (the corresponding costs in 2012 U.S. dollars are United Kingdom (Department of Health, United US$1.94, US$9.03, US$13.39, and US$93.31). In some Kingdom 2006), and 3 percent for Australia (CIE 2016). disciplines, it has been possible to use equipment, such as The payment system in the United States, in which doc- large analyzers, to lower the costs per test. In these areas, tors receive payment on a per test basis (and are particu- staff costs are a smaller proportion of the test cost larly conscious of potential litigation) means that the (68 percent to 87 percent for biochemistry tests across United States is likely to be an outlier among HICs. different sites and 74 percent to 89 percent for hematol- In South Africa, the costs of pathology are about 3.5 ogy, with one outlier). In other disciplines in which auto- percent of total health care expenditure (Pillay 2012). We mation is not as extensive, the unit costs are higher, and have no data on the share of pathology costs in overall staff costs are a higher proportion of test costs at health expenditure in other LMICs. 72 percent to 92 percent for microbiology and 93 percent to 97 percent for histopathology (Department of Health, Cost per Laboratory Test United Kingdom 2008). As science and technology Cost per laboratory test undertaken varies considerably. progress, areas such as microbiology may become more Important factors include the type of test (the diagnostic automated and less costly; however, newer and less auto- area), the volume of tests undertaken in the laboratory mated tests will continue to be developed. (the scale), the level of national income and salaries There are strong economies of scale in laboratory test- of technical personnel, whether the test is undertaken ing (for example, Department of Health, United Kingdom in the normal workflow or on an urgent or rapid- 2008; Cunnama and others 2016 for tuberculosis tests in turnaround basis, and a hard-to-measure efficiency fac- South Africa). However, the tradeoff is that increased tor. Since the level of the laboratory (tiers 1 through 4) centralization of tests is also associated with increased affects the mix of tests undertaken, the cost per test also turnaround time and potential loss of patients to varies with the level of the laboratory. follow-up. In table 11.5 the smallest laboratory performs Table 11.5 Estimated Ingredients for General Pathology Laboratories at Different Levels, Lower-Middle-Income Countries Assumptions Tier 1 laboratory Tier 2 laboratory Tier 3 laboratory Facility description 5 health workers; no inpatients 100 beds 200–400 beds 5 surgeries per day 15–20 surgeries per day 500 outpatients per week 1,500 outpatients per week Population served 30,000 50,000–200,000 3 million to 6 million Approximate annual hospital budget US$150,000 US$6 million US$18 million Laboratory staff, excluding 1 laboratory technician 1 general pathologist 4 pathologists administrative support 4 laboratory technicians 2 clinical scientists 2 laboratory assistants 12 laboratory technicians 8 laboratory assistants 1 medical officer Laboratory test volume per week 100 malaria slides plus point-of 850 2,500 -care tests table continues next page High-Quality Diagnosis: An Essential Pathology Package 229 Table 11.5 Estimated Ingredients for General Pathology Laboratories at Different Levels, Lower-Middle-Income Countries (continued) Assumptions Tier 1 laboratory Tier 2 laboratory Tier 3 laboratory Equipment needs US$2,000–US$5,000 US$150,000–US$200,000 Varies according to functions (microscope; small devices) Annual salary cost, (using US$4,800 US$63,680 US$259,440 table 11.4) Overall annual laboratory budget, n.a. US$318,400 US$1.3 million assuming consumables: salaries are (5.3 percent of hospital budget) (7.2 percent of hospital 4:1 in hospitals budget) Sources: Based on economic ratios from table 11.6, salaries from table 11.4, and expert judgment. Published data were for hospitals; insufficient data were available to make complete estimates for a tier 1 facility. Note: n.a. = not applicable. about one test per employee per day, compared with 24 in ratio for the two big hospital laboratories in India and the medium-sized laboratory in India, and 43 billable Thailand, summarized in table 11.6). In Sub-Saharan tests in the largest laboratory in the United States. We Africa, the current ratio is closer to 1:1 (Kuti, personal used 300 days worked per person per year as a rough communication); this ratio is likely not to be optimal given guide for this calculation. No data on staff were available that too few tests are undertaken in Sub-Saharan Africa. for Thailand. These inputs yield estimates of recurrent laboratory The level of national income affects the technology costs as a proportion of hospital budget of slightly more used in conducting tests, and hence the relative shares of than 5 percent for a first-level hospital, and slightly different cost components. In LMICs, salary costs are more than 7 percent for a second-level hospital. Our lower relative to the cost of reagents and test kits, so tests estimates can be compared with data for Ghana, where tend to be less automated; however, staff costs form a the share of laboratories in total hospital costs was 2.3 smaller proportion of overall costs. In HICs, salary costs percent for a first-level hospital with 117 beds and one are higher relative to the cost of consumables, and there doctor, and 4.1 percent for a second-level hospital with is more automation; but salary costs form a higher pro- 100 beds and three doctors (Aboagye, Degboe, and portion of overall costs (see table 11.6; some caution in Obuobi 2010). In India, the corresponding shares were interpretation is needed because the four laboratories in 7.3 percent for a first-level hospital of 400 beds and the table do not serve identical functions). In the United 24 doctors, and 9.2 percent for a second-level hospital States, the ratio of staff to consumables in total costs has of 778 beds and 237 doctors (Chatterjee, Levin, and increased. The ratio was 40:60 in 1980 for one clinical Laxminarayanan 2013). biochemistry laboratory in a university hospital, but We do not have enough data to estimate laboratory rose to 60:40 by 1990 (Benge, Csako, and Parl 1993). It is costs for primary health centers. One study of 12 govern- likely that LMICs will follow a similar trend as salaries ment primary health centers in Ghana (Dalaba and increase and drive increased automation. others 2013) estimated that the costs of laboratory sup- plies amounted to less than 1 percent of the overall cost of the center. This figure excludes the cost of consum- Estimated Costs for the Essential Pathology Package ables for POCT that do not enter the laboratory. Although the variations in the unit costs of tests make Because of too little published data, our confidence estimating laboratory costs challenging, systematic factors that these numbers apply in LICs is low. Professional are involved as well. We first estimate salary costs for tech- salaries in LICs are about half the level of those in nical staff using the WHO-CHOICE data (table 11.4) for lower-middle-income countries (table 11.4). However, it the average LMIC. We then construct stylized laboratories is unlikely that the costs of laboratories would be half as using expert judgment combined with published data well. The volume of tests is likely to be lower, and unit summarized in table 11.6. We combine these stylized data costs are likely to be higher by an unknown amount. The with the salary data and with the estimate that consum- data from Malawi (Gopal, personal communication) ables in the laboratory cost approximately four times as show that salaries of laboratory personnel are closer to much as salaries in Asia (which is slightly lower than the the levels of lower-middle-income countries than the 230 Disease Control Priorities: Improving Health and Reducing Poverty Table 11.6 Structure and Annual Cost of Tier 3 and 4 Laboratories in Four Settings Tata Memorial Hospital, King Chulalongkorn Major teaching hospital, Lilongwe, Malawi India Hemopathology lab Memorial Hospital, Thailand United States Types of test 91% histology, 9% Primarily hematological 85% biochemistry; 15% Full service cytology malignancies hematology Staff 2 pathologists 2 physicians n.a. 7 pathologists 2 laboratory technicians 2 senior residents 7 technical supervisors 1 laboratory assistant 6 scientists (2 PhDs) 19 phlebotomists 2 technical officers (MSc) 4 blood banks 13 technicians (BSc) 18 molecular and 6 assistants microbiology labs Total 31 26 clinical biochemistry and hematology labs 11 processors 25 outpatient laboratory technicians 117 total, excluding administration Approximate population 1 of only 2 such City of 21 million, state of City of 6.3 million City of 650,000 coverage laboratories, country of 15 112 million, diagnostic center State of 5.3 million million for region Annual number of tests 1,680 227,000 2.16 million 1.5 million billable (7 million total) Annual budget US$ (year) 243,000 (2012) 976,270 (2012) 25.3 million in 2002 ($ 2012) 18 million (2015) (2.7% of hospital budget) Budget shares (%) —c • Space, utilities n.a. 2.8 1.9 (equipment + space) • Equipment 22.6 11.2 13.2 • Staff 61.7 13.9 84.9 • Consumables 14.4 71.1 0 a b • Miscellaneous 1.2 1.1 n.a. Sources: Gopal (personal communication) and Gopal and others 2013; Gujral and others 2010 for India; budget shares calculated by chapter author from published data; Charuruks, Chamnanpai, and Seublinvog 2004 for Thailand. Note: n.a. = not available. a. Communications costs, telepathology link with University of North Carolina. b. Quality control, usually additional tests. c. Data (Christopher Price, personal communication) from a hospital trust in the United Kingdom suggest that the split is 72 percent staff, 26 percent equipment rental, 1 percent equipment maintenance, and 1 percent other. WHO data predict, likely because technically qualified more modest, but most of the testing at this level is point staff are sufficiently scarce that if they were paid less, they of care, and we do not have data on the cost of POCT. would not remain in public laboratories in LICs. What is known from HICs is that POCT is generally In summary, our rough estimates (table 11.5) are that more expensive on a cost-per-test basis compared with recurrent laboratory costs for a first-level hospital should centralized testing, primarily because POCT is based on be slightly more than 5 percent of the hospital budget; single-use technology. for a second- or third-level hospital, they should be The cost of setting up a laboratory is estimated to be slightly more than 7 percent of the budget. Of this share, US$2,000–US$5,000 for a tier 1 laboratory; US$150,000– about 16 percent consists of staff costs, and the balance US$200,000 for a tier 2 laboratory at a second-level hos- consists of consumables. Costs for a tier 1 laboratory are pital; and a considerably larger amount at a third-level High-Quality Diagnosis: An Essential Pathology Package 231 hospital, but no estimates were made because of the wide external (accreditation) audits. Reimbursement systems, variety of equipment choices available. In comparison, especially for universal health coverage, need to include the equipment for a specialized (primarily histopathol- pathology to minimize out-of-pocket expenses and dis- ogy) laboratory in Malawi cost US$150,000 to set up; incentives to appropriate use. Finally, ongoing research is about half of this cost is in addition to the cost of training important to obtain more accurate data on the eco- two technicians in other countries (Gopal, personal com- nomic benefits of pathology and on the most cost- munication). The cost of training two technicians in effective solutions. other countries was a further US$74,000. NOTE CONCLUSIONS World Bank Income Classifications as of July 2014 are as fol- The differential diagnosis of the child in the vignette at lows, based on estimates of gross national income (GNI) per capita for 2013: the beginning of this chapter, ranging from tuberculosis to lymph node cancer, was wide, and each diagnosis • Low-income countries (LICs) = US$1,045 or less would have required completely different treatments • Middle-income countries (MICs) are subdivided: and management. Most of the possible diagnoses were (a) lower-middle-income = US$1,046–US$4,125 life threatening; without the appropriate treatment, the (b) upper-middle-income (UMICs) = US$4,126–US$12,745 prognosis was poor. Conversely, with the right diagnosis • High-income countries (HICs) = US$12,746 or more. and resultant treatment, the prognosis would have been good. The widespread availability of and timely access to good quality pathology, as described in the essential REFERENCES pathology package, would have provided that accurate Aboagye A. Q. Q., A. K. B. Degboe, and A. A. D. Obuobi. diagnosis. 2010. “Estimating the Cost of Healthcare Delivery in Three Hospitals in Southern Ghana.” Ghana Medical Journal 44 (3): 83–92. Key Messages Adesina, A., D. Chumba, A. M. Nelson, J. Orem, D. J. Roberts, and Pathology is a cross-cutting discipline upon which the others. 2013. “Improvement of Pathology in Sub-Saharan other health disciplines depend and a crucial component Africa.” The Lancet Oncology 14 (4): e152–57. doi:10.1016 in the care pathway. Pathologists are diagnosticians who, /S1470-2045(12)70598-3. as part of the clinical team, play a key role in linking African Strategies for Advancing Pathology Group Members. clinical services with laboratory services, providing lead- 2015.“Quality Pathology and Laboratory Diagnostic Services ership, and capitalizing on the opportunities arising Are Key to Improving Global Health Outcomes: Improving from rapidly emerging new technologies. Pathology Global Health Outcomes Is Not Possible without Accurate Disease Diagnosis.” American Journal of Clinical Pathology contributes to research in both communicable and non- 143 (3): 325–28. doi:10.1309/AJCP6K0DZCNVCSCI. communicable diseases, and it plays a central role in Alemnji, G. A., C. Zeh, K. Yao, and P. N. Fonjungo. 2014. national policy planning. “Strengthening National Health Laboratories in Sub-Saharan Africa: A Decade of Remarkable Progress.” Tropical Medicine and International Health 19 (4): 450–58. Recommendations ASLM (African Society for Laboratory Medicine) and WHO Implementation of the essential pathology package is AFRO (World Health Organization Regional Office for needed to address the lack of timely, accurate pathology in Africa). 2015. “The Freetown Declaration.” Statement at many LMICs; the rapidly increasing burden of noncom- Regional Global Health Security Consultation on Laboratory municable diseases makes such implementation a prior- Strengthening, Freetown, Sierra Leone, October 15–16. ity. Our economic analysis shows that provision of the http://new.aslm.org/programmes/global-health-security essential pathology package is affordable at approximately /freetown-declaration. 6 percent of a hospital’s budget. An integrated network is Avivar, C. 2012. “Strategies for the Successful Implementation of Viral Laboratory Automation.” Open Virology Journal 6 crucial to achieving the benefits of shared knowledge, (S1): 151–21. expertise, communication, and economies of scale. Benge, H., G. Csako, and F. F. Parl. 1993. “A 10-Year Analysis The sustainability and quality of the essential pathol- of Revenues, Costs, Staffing and Workload in an Academic ogy package depends on investment in education and Medical Centre Clinical Chemistry Laboratory.” Clinical training and in appropriate emerging technologies, Chemistry 39: 1780–87. including LIS. Standards of practice need to be assessed Bond, M. M., and R. R. Richards-Kortum. 2015. “Drop-to-Drop across the network by an ongoing system of internal and Variation in the Cellular Components of Fingerprick Blood: 232 Disease Control Priorities: Improving Health and Reducing Poverty Implications for Point-of-Care Diagnostic Development.” Cancer Diagnoses.” PLoS One 8 (8): e70361. doi:10.1371 American Journal of Clinical Pathology 144 (6): 885–94. /journal.pone.0070361. doi:10.1309/AJCP1L7DKMPCHPEH. Gujral, S., K. Dongre, S. Bhindare, P. G. Subramanian, Cassim, N., L. M. Coetzee, K. Schnippel, and D. K. Glencross. H. K. V. Narayan, and others. 2010. “Activity-Based Costing 2014. “Estimating Implementation and Operational Costs Methodology as Tool for Costing in Haematopathology of an Integrated Tiered CD4 Service Including Laboratory Laboratory.” Indian Journal of Pathology and Microbiology and Point of Care Testing in a Remote Health District in 53 (1): 68–74. South Africa.” PLoS One 9 (12): e115420. The Hindu. 2014. “Only 17% have Health Insurance Cover.” Charuruks, N., S. Chamnanpai, and T. Seublinvog. 2004. December 22. http://www.thehindu.com/news/national/only “Cost Analysis of Laboratory Tests: A Study of the Central -17-have-health-insurance-cover/article6713952.ece. Laboratory of King Chulalongkorn Memorial Hospital.” Hyle, E. P. L., K. Meye, K. Kelly, S. Christensen, K. Daskilewicz, Journal of the Medical Association of Thailand 87: 955–63. and others. 2014. “The Clinical and Economic Impact of Chatterjee, S., C. Levin, and R. Laxminarayanan. 2013. “Unit Point-of-Care CD4 Testing in Mozambique and Other Cost of Medical Services at Different Hospitals in India.” Resource-Limited Settings: A Cost-Effectiveness Analysis.” PLoS One 8 (7): e69728. doi:10.1371/journalpone 0069728. PLoS Medicine 11 (9): e1001725. Chen, M., Y. Zhao, and L. Si. 2014. “Who Pays for Health Care ISO (International Organization for Standardization). in China? The Case of Heilongjiang Province.” PLoS One 2012. Medical Laboratories: Requirements for Quality and 9 (10): e108867. doi:10.1371/journal.pone.0108867. Competence. ISO 15189:2012. Geneva: ISO. CIE (Centre for International Economics). 2016. The Economic JCI (Joint Commission International). 2014. Accreditation Value of Pathology: Achieving Better Health, and a Better Standards for Hospitals, fifth edition. Oak Brook, IL: JCI. http:// Use of Health Resources. Canberra: CIE. http://www.thecie www.jointcommissioninternational.org/jci-accreditation .com.au/wp-content/uploads/2016/04/Economic-value-of -standards-for-hospitals-5th-edition/. -pathology_-Final-Report-April-2016.pdf. Jegathesan, M., and G. F. de Witt. 1982. “Organisation of Cunnama, L., E. Sinanovic, L. Ramma, N. Foster, L. Berrie, and Laboratory Services in Malaysia.” Malaysian Journal of others. 2016. “Using Top-Down and Bottom-Up Costing Pathology 5: 1–5. Approaches in LMICS: The Case for Using Both to Assess Kuijpers, C. C., J. Fronczek, F. R. van de Goot, H. W. Niessen, the Incremental Costs of New Technologies at Scale.” Health P. J. van Diest, and others. 2014. “The Value of Autopsies Economics 25: 53–66. in the Era of High-Tech Medicine: Discrepant Findings Dalaba, M. A., P. Akweongo, G. Savadogo, H. Saronga, J. Williams, Persist.” Journal of Clinical Pathology 67 (6): 512–19. and others. 2013. “Cost of Maternal Health Services in doi:10.1136/jclinpath-2013-202122. Selected Primary Health Centres in Ghana: A Step-down Lei, H. D., J. B. Chen, and X. S. Lu. 2014. “Analysis of the Allocation Approach.” BMC Health Services Research 13 (1): Personal Burden of Inpatients in a Hospital in Tianjin.” 287–95. Foreign Medical Sciences 31 (1): 38–40. Daramola, A. O., A. A. Banjo, A. Bennett, F. Abdulkareem, and Looi, L. M. 2008. “The Pathology Act Explained.” Malaysian A. M. Shaaban. 2016. “Breast Cancer Reporting in Lagos, Journal of Pathology 30 (1): 1–10. Nigeria: Implications for Training and Education in Africa.” Mao, A. Y. 2012. “Study on the Impact of Public Health Insurance Journal of Global Oncology 2 (2). doi:10.1200/JGO.2015.003079. System Reform over 20 Years in China, How Patients’ Actual Department of Health, United Kingdom. 2006. Report of the Reimbursement Rate Changed.” Chinese Health Economics Review of NHS Pathology Services in England. London. 31 (10): 19–22. http://collection.europarchive.org/tna/20070706124823 Nelson, A. M., D. A. Milner, T. R. Rebbeck, Y. Iliyasu, and /http://dh.gov.uk/en/Publicationsandstatistics/Publications K. A. Fleming. 2016. “Africa: Pathology in the Sub-Sahara.” /PublicationsPolicyAndGuidance/DH_4137606. In The State of Oncology in Africa 2015, edited by P. Boyle, ———. 2008. Report of the Second Phase of the Review of NHS T. Ngoma, R. Sullivan, N. Ndlovu, P. Autier, and others. Pathology Services in England. London. http://webarchive Lyon, France: iPRI Scientific Publication 4, iPRI. .nationalarchives.gov.uk/20130107105354/http:/www.dh NPP (National Pathology Programme). 2014. Digital First: Clinical .gov.uk/prod_consum_dh/groups/dh_digitalassets/@dh Transformation through Pathology Innovation. London: NPP. /@en/documents/digitalasset/dh_091984.pdf. https://www.england.nhs.uk/wp-content/uploads/2014/02 Elbireer, A. M., J. B. Jackson, H. Sendagire, A. Opio, D. Bagenda, /pathol-dig-first.pdf. and others. 2013. “The Good, the Bad, and the Unknown: Orem, J., S. Sandin, C. E. Weibull, M. Odida, H. Wabinga, and Quality of Clinical Laboratories in Kampala, Uganda.” PLoS others. 2012. “Agreement between Diagnoses of Childhood One 8 (5): e64661. doi:10.1371/journal.pone.0064661. Lymphoma Assigned in Uganda and by an International Gershy-Damet, G. M., P. Rotz, D. Cross, F. Cham, J. B. Ndihokubwayo, Reference Laboratory.” Clinical Epidemiology 4: 339–47. and others. 2010. “The World Health Organization African doi:10.2147/CLEP.S35671. Region Laboratory Accreditation Process.” American Journal of Pan, L., P. He, H. F. Wu, Y. Den, W. B. Zhang, and Clinical Pathology 134 (3): 393–400. others. 2014. “Investigation and Analysis of City Public Gopal, S., R. Krysiak, N. G. Liomba, M.-J. Horner, C. G. Shores, Hospital Outpatient and Inpatient Medical Expenses of and others. 2013. “Early Experience after Developing Inpatients and Payment.” Chinese General Practice 17 a Pathology Laboratory in Malawi, with Emphasis on (34): 4127–32. High-Quality Diagnosis: An Essential Pathology Package 233 Pan, J., S. Tian, Q. Zhou, and W. Han. 2016. “Benefit Distribution Vathana, C. S., and G. Stauch. 2014. “Pathology and of Social Health Insurance: Evidence from China’s Urban Telepathology in Cambodia.” Pathology 46 (Suppl 2): S8. Resident Basic Medical Insurance.” Health Policy and Planning Wang, L. 2012. “Study on the Healthcare Benefit in Sub- 1–7. doi:10.1093/heapol/czv141. Saharan African Countries.” Foreign Medical Sciences (Health Pillay, T. S. 2012. “Containing Costs in the Era of National Economics) 29 (4). Health Insurance—The Need for and Importance of Wattanasri, N., W. Manoroma, and S. Viriyayudhagorn. Demand Management in Laboratory Medicine.” South 2010. “Laboratory Accreditation in Thailand: A Systemic African Medical Journal 103 (1): 24. doi: 10.7196/samj.6383. Approach.” American Journal of Clinical Pathology 134 (4): Plebani, M. 2009. “Exploring the Iceberg of Errors in Laboratory 534–40. Medicine.” Clinica Chimica Acta 404: 16–23. Whitesides, G. M., and P. Wilding. 2012. “Laboratory on a Pollock, N. R., S. McGray, D. J. Colby, F. Noubary, H. Nguyen, Stamp: Paper-Based Diagnostic Tools.” Interview by Molly and others. 2013. “Field Evaluation of a Paper-Based Point- Webster and Vikram Sheel Kumar. Clinical Chemistry 58: of-Care Fingerstick Transaminase Test.” PLoS One 8 (9): 956–58. e75616. doi:10.1371/journal.pone.0075616. eCollection 2013. WHO (World Health Organization) 2008. Consultation on Robyn, P. J., R. Sauerborn, and T. Bärnighausen. 2013. “Provider Technical and Operational Recommendations for Clinical Payment in Community-Based Health Insurance Schemes Laboratory Testing Harmonization and Standardization. WHO: in Developing Countries: A Systematic Review.” Health Geneva. http://www.who.int/healthsystems/round9_9.pdf. Policy and Planning 28 (2): 111–22. ———. 2015a. Global Tuberculosis Report. 20th edition. Rohr U.-P., C. Binder, T. Dieterle, F. Giusti, C. G. M. Messina, Geneva: WHO. and others. 2016. “The Value of In-Vitro Diagnostic ———. 2015b. World Malaria Report 2015. Geneva: WHO. Testing in Medical Practice: A Status Report.” PLoS One 11: WHO AFRO (World Health Organization Regional Office e0149856. doi:10.1371.journal.pone.0149856. for Africa). 2008. The Maputo Declaration on Strengthening Schroeder, L. F., and T. Amukele. 2014. “Medical Laboratories of Laboratory Systems. Statement issued at Consensus in Sub-Saharan Africa that Meet International Quality Meeting on Clinical Laboratory Testing Harmonization Standards.” American Journal of Clinical Pathology 141 (6): and Standardization, Maputo, Mozambique, January 791–95. 22–24. http://www.who.int/diagnostics_laboratory/Maputo Serje, J. 2015. “Estimates of Health Sector Salaries across Four -Declaration_2008.pdf. Occupational Levels for UN Member States.” Unpublished, ———, and US CDC (United States Centers for Disease WHO, Geneva. Control and Prevention). 2010. Guidance for Development Singh, T. D. 2013. “Accreditation of Diagnostic Laboratories.” of National Laboratory Strategic Plans. Atlanta: CDC. The Sangai Express, September 28. http://www.who.int/hiv/amds/amds_guide_dev_nat_lab Soors, W., N. Devadasan, V. Durairaj, and Criel B. 2010. _strat.pdf. “Community Health Insurance and Universal Coverage: Wu, G., and M. H. Zaman. 2012.“Low-Cost Tools for Diagnosing Multiple Paths, Many Rivers to Cross.” World Health Report and Monitoring HIV Infection in Low-Resource Settings.” Background Paper 48, WHO, Geneva. Bulletin of the World Health Organization 90 (12): 914–20. Stewart, B., and C. P. Wild, eds. 2014. World Cancer Report. Yao, K., M. Talkmore, E. T. Luman, and J. N. Nkengasong. 2014. Lyon: International Agency for Research on Cancer. “The SLMTA Programme: Transforming the Laboratory St. John, A., and C. P. Price. 2014. “Existing and Emerging Landscape in Developing Countries.” African Journal of Technologies for Point-of-Care Testing.” Clinical Biochemistry Laboratory Medicine 3 (2). doi:10.4102/ajlm.v3i2.194. Review 35: 155–67. Yetisen, A. K., M. S. Akram, and C. R. Lowe. 2013. “Paper-Based Van Dang, D. 2104. “Pathology and Telepathology in Vietnam.” Microfluidic Point-of-Care Diagnostic Devices.” Laboratory Pathology 46 (Suppl 2): S8. on a Chip 13: 2210–51. 234 Disease Control Priorities: Improving Health and Reducing Poverty Chapter 12 Palliative Care and Pain Control Eric L. Krakauer, Xiaoxiao Kwete, Stéphane Verguet, Hector Arreola-Ornelas, Afsan Bhadelia, Oscar Mendez, Natalia M. Rodriguez, Zipporah Ali, Silvia Allende, James F. Cleary, Stephen Connor, Kristen Danforth, Liliana de Lima, Liz Gwyther, Ednin Hamzah, Dean T. Jamison, Quach Thanh Khanh, Suresh Kumar, Emmanuel Luyirika, Anne Merriman, Egide Mpanumusingo, Diana Nevzorova, Christian Ntizimira, Hibah Osman, Pedro Perez-Cruz, M. R. Rajagopal, Lukas Radbruch, Dingle Spence, Mark Stoltenberg, Neo Tapela, David A. Watkins, and Felicia Knaul INTRODUCTION and others 2013; Knaul and others 2017; Summers 2016). For these reasons, it is recognized globally as an Palliative care has been shown to provide significant ethical responsibility of all health care systems and a and diverse benefits for patients with serious, complex, necessary component of universal health coverage or life-limiting health problems. Such benefits include (World Health Assembly 2014). Yet palliative care is the following: rarely accessible in low- and middle-income countries (LMICs). This chapter describes an essential package • Reduced physical, psychological, and spiritual suf- (EP) of palliative care services and treatments that could fering (Abernethy and others 2003; Gwyther and and should be accessible to everyone everywhere, as well Krakauer 2011; Higginson and others 2014; Krakauer as the sites or platforms where those services and treat- 2008; Singer and others 2016; Temel and others 2010; ments could be offered. Thus, it was necessary to make WHO 2008; Zimmerman and others 2014) a preliminary estimate of the burden of health-related • Improved quality of life (Singer and others 2016; suffering requiring palliative care. Zimmerman and others 2014) To roughly estimate the need for palliative care, we • Prolonged survival in some situations (Connor and identified the serious, complex, or life-limiting condi- others 2007; Temel and others 2010). tions listed in the International Classification of Diseases (ICD)-10 that most commonly result in physical, psy- Palliative care also can lower costs to health care chological, social, or spiritual suffering (WHO 2015a). systems (Chalkidou and others 2014; DesRosiers and We then estimated the types, prevalence, and duration others 2014; Gomez-Batiste and others 2012; Jamison of suffering resulting from each condition. On the basis Felicia Knaul is senior author. Corresponding author: Eric L. Krakauer, Harvard Medical School, Boston, Massachusetts, United States; eric_krakauer@hms.harvard.edu. 235 of this characterization of the burden of suffering, we almost all from HICs, revealed that 75 percent of patients propose an EP of palliative care and pain control receiving anti-cancer treatment or with advanced, meta- designed to do the following: static disease had pain, most of which was moderate or severe (Doyle and others 2017; Van den Beuken-van • Prevent or relieve the most common and severe suf- Everdingen, Hochstenbach, and Joosten 2016). Dyspnea— fering related to illness or injury. shortness of breath—is especially common among people • Be affordable, even in LMICs. who die of COPD and heart failure and only slightly less • Provide financial risk protection for patients and common among those who die of malignant neoplasms families by providing a realistic alternative to expen- and HIV/AIDS (Moens, Higginson, and Harding 2014). sive, low-value treatment. Depressed mood and anxiety are quite com- mon among patients with a variety of advanced We costed the EP in one low-income country life-threatening illnesses. Data on prevalence of social (Rwanda), one lower-middle-income country (Vietnam), and spiritual distress among these patients are scant. and one upper-middle-income country (Mexico) and A study in the United States found that 44 percent of projected these costs for LMICs in general (Knaul and patients with advanced cancer experienced spiritual pain. others 2017). At the conclusion of this chapter, we provide In an impoverished rural district in Malawi, 76 percent guidance on how to integrate the EP into health systems of patients receiving palliative care needed social sup- as an essential element of universal health coverage ports. In Germany, approximately 50 percent of patients (UHC) in LMICs. We also discuss how to augment the EP receiving palliative care needed such supports (Herce and as soon as is feasible to further prevent and relieve others 2014; Ostgathe and others 2011). suffering. The Lancet Commission on Global Access to Palliative This chapter draws directly on the work of the Lancet Care and Pain Relief (Knaul and others 2017) identified Commission on Global Access to Palliative Care and Pain (a) the 20 ICD-10 conditions that most commonly Control (the Lancet Commission) (Knaul and others result in a need for palliative care and (b) the specific 2017). categories of suffering typically caused by each condi- tion (table 12.1). Almost all of the identified conditions can cause any of the four categories of suffering. THE NEED FOR PALLIATIVE CARE In addition, psychological and social distress can be a cause of at least some of the ICD-10 conditions (Farmer In 2015, there were 56 million deaths, including nearly and others 2006). To determine the number of deaths 9 million from malignant neoplasms, more than per year from each condition, and hence gain insight 1 million from human immunodeficiency virus/aquired into the need for palliative care, the Commission used immune deficiency syndrome (HIV/AIDS), more than mortality data from the WHO Global Health Estimates 17 million from cardiovascular diseases, and more than (GHE) for 2015 (Mathers and others 2018, chapter 4 of 3 million from chronic obstructive pulmonary disease this volume) and aligned these data with the ICD-10 (COPD) (WHO 2016). These and other serious, com- conditions using a conversion document from the plex, or life-limiting health problems generate multiple WHO (2017). The Commission then estimated the per- kinds of suffering, typically categorized in the palliative centage of people who die from each condition (“dece- care literature as follows (WHO 2002): dents”) who have health-related suffering that requires palliative care. • Pain and other physical distress The Commission also identified the conditions that • Psychological distress often lead to physical, psychological, social, or spiritual • Social distress suffering, even among nondecedents, defined as people • Spiritual distress. who do not die in a given year. These conditions include some that may be curable (drug-resistant tuberculosis Existing data, mostly from high-income countries and some malignancies), others that may be well con- (HICs), indicate that well over 50 percent of patients trolled for long periods (HIV/AIDS and musculoskeletal who die of malignant neoplasms and HIV/AIDS experi- disorders), and others from which patients may recover ence pain (Foley and others 2006). Pain is also common (serious injuries). It also was necessary to identify the among those who die of heart disease, COPD, renal fail- specific types of suffering within each category (for ure, neurologic disease, and dementia (Moens, Higginson, example, pain, dyspnea, and nausea are types of physical and Harding 2014; Solano, Gomes, and Higginson suffering) and to estimate the prevalence and duration 2006). A recent meta-analysis of pain prevalence studies, of each type. 236 Disease Control Priorities: Improving Health and Reducing Poverty Table 12.1 Conditions Responsible for the Need for Palliative Care Decedents Patients Symptoms (%) Total Physical Psychological All symptoms All symptoms All symptoms Condition/disease Rank % Total days At least days Total days At least days Total days At least days a. LMICs Malignant neoplasms 1 26 47 45 50 46 36 36 CVD 2 17 11 12 12 12 7 9 Lung disease 3 11 9 11 8 11 12 12 Injuries 4 6 0 1 0 1 1 1 TB 5 6 6 6 4 4 10 9 Premature birth and trauma 6 5 0 0 0 0 0 0 HIV 7 5 12 8 11 8 12 12 Liver disease 8 5 3 3 3 3 2 1 NI heart disease 9 4 3 3 3 3 4 4 Dementia 10 3 4 4 3 3 10 10 All other 11 5 8 5 8 6 6 All (millions) 20.6 9,143 2,473 7,191 2,378 1,952 1,054 b. Global Malignant neoplasms 1 30 51 49 54 51 39 39 CVD 2 16 10 10 11 11 6 8 Lung disease 3 11 8 10 7 10 11 11 Injuries 4 6 0 1 0 1 1 1 TB 5 5 4 5 3 3 8 7 Dementia 6 5 6 6 4 4 13 13 Liver disease 7 5 2 3 3 3 2 1 Premature birth and trauma 8 4 0 0 0 0 0 0 HIV 9 4 9 6 9 6 10 9 NI heart disease 10 4 3 3 3 3 3 4 All other 11 6 8 6 9 7 7 All (millions) 25.6 11,900 3,231 9,347 3,105 2,553 1,376 Source: Adapted with permission from Knaul and others 2017. Note: CVD = cardiovascular disease; HIV = human immunodeficiency virus; LMICs = low- and middle-income countries; NI = non-ischemic; TB = tuberculosis. a. The other illness conditions that commonly result in a need for palliative care are hemorrhagic fever, leukemia, dementia, inflammatory disease of the central nervous system, degeneration of the central nervous system; chronic ischemic heart disease, renal failure, congenital malformations, atherosclerosis, chronic musculoskeletal disorders, and malnutrition. The Commission identified 20 conditions that account deaths occurred in LMICs in 2015; of these, about 20 million for 81 percent of global deaths and 80 percent of deaths or 45 percent experienced health-related suffering that in LMICs. Based on mortality figures for 2015 and our required palliative care. Patients in LMICs account estimates, at least 50.5 million people each year with these for 17 billion days per year of need for palliative conditions in LMICs, including decedents and nondece- care—80 percent of the annual global total. Among dece- dents, require palliative care. Approximately 60 percent of dents in LMICs, 10 conditions account for approximately these patients are nondecedents. More than 46 million 90 percent of patients and 95 percent of total days of Palliative Care and Pain Control 237 health-related suffering. The other 10 conditions each typically lack access to relief of pain and other types of account for less than 3 percent of decedents and days with suffering that result from common health problems that health-related suffering. may be cured (drug-resistant tuberculosis and some malignancies) or controlled for a long period (HIV/AIDS and musculoskeletal disorders) or from which patients THE GLOBAL SUFFERING DIVIDE: DISPARATE are likely to recover (serious injuries). The need for pal- ACCESS TO PALLIATIVE CARE AND PAIN liative care in low-resource settings is often determined by the magnitude of suffering, the inadequacy of existing CONTROL capacity to respond, and the resultant need for relief. Despite compelling evidence of a huge burden of reme- Therefore, the EP of palliative care and pain control that diable health-related suffering and of the efficacy of we propose should be as follows: palliative care and pain treatment, these essential health services are rarely accessible in LMICs. Data from the • Accessible at all levels of health care systems and in International Narcotics Control Board show that patients’ homes. 91 percent of the morphine consumed worldwide in • Adapted to local cultures, as well as clinical and social 2013 was consumed in HICs, which have only 19 percent situations. For example, in resource-poor settings, of the world’s population; people in LMICs, which the social circumstances of the patient and family account for 81 percent of the world’s population, only members may be a major source of the patient’s suf- consumed 9 percent (Pain and Policy Studies Group fering and may need to be a focus for palliative care 2017). Given that morphine is essential to relieve mod- (Gwyther and Krakauer 2011). erate and severe pain (WHO 1996, 2012) and that mor- • Integrated with disease prevention and treatment pro- phine consumption is the most common—although grams, although not considered a substitute for these, imperfect—measure of palliative care accessibility, the and assist patients in accessing and adhering to optimum data reveal an enormous disparity between rich and disease treatment—if they desire such treatment and poor in meeting the need for palliative care. if it may be more beneficial than harmful according Available data indicate that 74 percent of countries— to patients’ values, balanced with scientific evidence. virtually all of them LMICs—had at best isolated pallia- Further, palliative care workers have a responsibility tive care provision as of 2013 (Connor and Sepulveda to advocate for access to comprehensive health care Bermedo 2014). Among the 9 percent of countries where including, but not restricted to, disease-modifying palliative care is “at a stage of advanced integration into treatments, such as cancer chemotherapy, antiretro- mainstream service provision,” only Romania and viral treatment, or effective medicines for multidrug Uganda are LMICs, and most people in need lack access resistant tuberculosis (Gwyther and Krakauer 2011; to palliative care even in these two countries (Connor Shulman and others 2014). and Sepulveda Bermedo 2014, 39–40). The global suffer- • Applied not only to persons who are dying but also ing divide appears to be one of the world’s largest health to those living with long-term physical, psychologi- care inequities. The EP of palliative care that we propose cal, social, or spiritual sequelae of serious, complex, is designed specifically to be the minimum acceptable or life-limiting illnesses or of their treatment. The package for the lowest income settings. Accordingly, EP should be applied to relieve acute pain and other although necessary for all countries, the EP is not acute symptoms when medically indicated. exhaustive; palliative care can be improved by expanding • With adequate levels of palliative care training and the package to include additional medicines, equipment, skill, applied by health care workers of various kinds, and human resources. including primary care providers, generalists, and specialists in many disciplines and from basic to intermediate to specialist. AN ESSENTIAL PACKAGE OF PALLIATIVE CARE AND PAIN CONTROL Design of the Essential Package Patients with life-threatening illnesses are the sole focus The EP that we propose is a key component of health of palliative care according to the current WHO defini- systems and is designed to relieve the most common tion, and there are calls for it to be revised and expanded and severe suffering related to illness or injury, to be (Gwyther and Krakauer 2011; WHO 2002). There is low cost and feasible to deliver in LMICs, and to pro- large-scale, unrelieved health-related suffering among tect patients and their families from catastrophic other groups as well. In particular, patients in LMICs health expenditures (table 12.2). It consists of a list of 238 Disease Control Priorities: Improving Health and Reducing Poverty Table 12.2 Delivery Platforms for the Essential Palliative Care Interventions Delivery platform First-, second-, and third-level Intervention Intersectoral Mobile outreach or home care Health center (PHC) hospitals Control of chronic pain related to serious, • Routine social assessment • Surveillance and emotional • Oral immediate-release morphine — complex, or life-limiting health problems • Income and in-kind support a support by community health and other essential medicines and workers as often as needed simple equipment for prevention (sometimes daily) and relief of chronic pain • Visits by PHC nurse or doctor as needed Control of other types of physical and • Routine social assessment • Emotional support and suffering • Essential medicines and simple — psychological sufferingb related to serious, • Income and in-kind supporta surveillance by community equipment for prevention and complex, or life-limiting health problems health workers as often as relief of other types of physical needed (sometimes daily) and psychological suffering • Visits by PHC nurse or doctor • Psychological counseling as needed Control of refractory suffering (chronic pain, • Routine social assessment • Oral immediate-release other types of physical and psychological • Income and in-kind support a morphine and injectable sufferingb that have not or cannot be controlled morphine and other essential at lower level) medicines and simple equipment for prevention and relief of chronic pain and other types of physical and psychological sufferingc • Psychological counselingc Acute pain related to surgery or serious injury — — — • Essential medicines and simple equipment for prevention and relief of acute painc Note: PHC = public health care. — = this type of care not provided in this setting. Palliative Care and Pain Control a. Support provided only for patients living in extreme poverty and for one caregiver per patient. b. Physical suffering includes breathlessness, fatigue, weakness, nausea, vomiting, diarrhea, constipation, pruritus, bleeding, and wounds. Psychological suffering includes anxiety or worry, depressed mood, confusion or delirium, and dementia. c. Care devolves to lower level once effective treatment is established. 239 medicines, based on the WHO Model List of Essential essential uses in palliative care and are safe and easy to Medicines for Palliative Care (WHO 2015b, 2015c). prescribe. For example, haloperidol is the first-line Centrally important in this list is immediate-release medicine in many cases for relief of nausea, vomiting, morphine. The EP includes equipment (lock boxes) agitation, delirium, and anxiety. An SSRI, such as and procedures to assure against misuse of opioids. fluoxetine, is the first-line pharmacotherapy for The package also includes some small and inexpensive depressed mood or persistent anxiety, both of which equipment. In addition, the package specifies several are common among patients with serious, complex, or types of palliative care interventions and the platform life-limiting health problems. All physicians at any or health care system level at which each intervention level of the health care system and who care for and each item in the package should be available. patients with these symptoms should be trained and Finally, the package includes intersectoral inputs in the permitted to prescribe these medicines—not solely form of income and in-kind support required by any psychiatrists or neurologists. patient or family caregiver living in extreme poverty. (See annex table 12A.1 for an exhaustive list of medi- Equipment cines and other inputs required for the EP.) The EP includes equipment that often is needed for palliative care yet may not be available in all health cen- Medicines ters and hospitals in LMICs. Such equipment includes Morphine, in oral immediate-release and injectable pressure-reducing mattresses, adult diapers, opioid lock preparations, is the most clinically important of the boxes nasogastric tubes, and urinary catheters (annex essential palliative care medicines (WHO 2011). It must table 12A.1). For the sake of efficiency, the EP does not be accessible in the proper form and dose by any patient include materials needed for palliative care that should with terminal dyspnea or with moderate or severe pain be standard equipment for any health center or hospital, that is either acute, chronic and associated with malig- such as gauze and tape for dressing wounds, nonsterile nancy, or chronic in a patient with a terminal prognosis. examination gloves, syringes, and angiocatheters. We do not recommend the use of opioids for chronic pain outside of cancer, palliative, and end-of-life care, Psychological and Spiritual Counseling except under special circumstances and with strict Interventions to relieve psychological distress may be monitoring (Dowell, Haegerich, and Chou 2016). provided not only by psychologists but also by ade- All physicians who ever care for patients with moderate quately trained and supervised physicians, nurses, or or severe pain of the types described, or for patients social workers at any level of the health care system with terminal dyspnea, should be able to prescribe oral (Belkin and others 2011; Patel 2014; Rahman and others and injectable morphine for inpatients and outpatients 2016). For patients or family members with complicated in any dose necessary to provide adequate relief as psychological problems, such as suicidality, psychotic determined by the patients. Physicians should be able to disorders, or bipolar disorder, referral should be made to prescribe an adequate supply of morphine so that psychiatrists, if possible. In addition, hospital-based staff obtaining refills is feasible for patients or families with- members should routinely ask patients with serious, out requiring unreasonably frequent, expensive, or complex, or life-limiting health problems if they desire arduous travel. spiritual counseling, and hospitals should allow local Although ensuring access to morphine for anyone in volunteer spiritual counselors to visit inpatients upon need is imperative, it also is necessary to take reasonable request by the patient or family. precautions to prevent diversion and nonmedical use. Model guidelines for this purpose are available (Joranson, Social Supports Maurer, and Mwangi-Powell 2010). Oral immediate Social supports should be accessible both for any patient release and injectable morphine should be accessible at in need of palliative care and for their main caregiver in all third-, second, and first-level hospitals. Personnel at instances of extreme poverty. Given that extreme poverty health centers also should be trained in opioid analgesia is both a cause and an effect of serious, complex, or and safe storage so that morphine may be safely dis- life-limiting health problems, it is crucial that meaningful pensed by prescription in these settings as well. social supports are accessible (Bamberger 2016). Such Among the other essential palliative medicines are social supports include transportation vouchers, cash oral and injectable haloperidol and oral fluoxetine or payments, food packages, and other types of in-kind sup- another selective serotonin reuptake inhibitor (SSRI). port (annex table 12A.1) (Carrillo and others 2011; Syed, Although these medicines are considered psychiatric Gerber, and Sharp 2013). In most cases, funding for these or psychotropic medicines, they have multiple social supports should come not from health care 240 Disease Control Priorities: Improving Health and Reducing Poverty budgets but from antipoverty or social welfare programs. COSTS OF THE ESSENTIAL PACKAGE OF Thus, to be able to implement all aspects of the full EP, PALLIATIVE CARE AND PAIN CONTROL there must be intersectoral coordination. In most LMICs, the cost of caring for patients with seri- Human Resources ous, complex, or life-limiting health problems is borne The EP should include adequate time for trained per- primarily not by governments but out-of-pocket by sonnel at each level of the health care system to provide patients and their families. Serious, complex, or life- palliative care consisting of the interventions, medicines, limiting health problems put patients’ families at risk of equipment, counseling, and social supports described financial ruin and caregivers at risk of exhaustion and earlier. These personnel include doctors, nurses, coun- health problems of their own (Emanuel and others 2008; selors such as social workers or psychologists, pharma- Emanuel and others 2010). Data on the obvious and cists, community health workers, and family caregivers hidden costs of palliative care and any cost savings are (annex table 12A.1). Community health workers require important to inform governmental decisions about a minimum of several hours of training to prepare them including palliative care among public health care ser- to recognize and report any uncontrolled suffering to a vices and about covering palliative care with government supervisor. Capable family caregivers should be trained, health insurance. equipped, and encouraged by staff at health centers to provide basic nursing care such as wound and mouth care and medicine administration. Nurses and doctors Data on Costs and Cost Savings at health centers who provide palliative care or who Multiple studies from HICs indicate that palliative care instruct family caregivers need basic training and all can reduce costs for patients and families, as well as for doctors who care for patients with serious, complex, or health systems (Chalkidou and others 2014; DesRosiers life-limiting health problems at hospitals require inter- and others 2014; Gomez-Batiste and others 2012; mediate training. Ideally, all countries should have palli- Jamison and others 2013; Summers 2016). Not only can ative care specialist physicians to lead training and palliative care improve patient outcomes, it also can service implementation and to advise governments on reduce health care costs by reducing length of stay in the palliative care policy (World Health Assembly 2014). hospital, hospital admissions, and demand for expensive disease-modifying treatments of dubious benefit near the end of life. Patients who receive palliative care, espe- Augmenting the Essential Package cially early in their disease course, incur lower health care After the EP of palliative care and pain control is univer- costs (Albanese and others 2013; May 2016; Morrison sally accessible, additional palliative medicines, equip- and others 2008), have shorter hospitalization (Morrison ment, and services should be made accessible by all and others 2008; Postier and others 2014), enjoy equal or countries to further prevent and relieve health-related higher quality of life (Zimmerman and others 2008), and suffering as soon as resources permit (Lutz, Jones, and live equally long or longer (Elsayem and others 2004) Chow 2014; Miner 2005; Shulman and others 2015). than patients who do not receive palliative care. Palliative This augmentation would consist of the following: care also has been shown to increase satisfaction of fam- ily caregivers (Zimmerman and others 2008). Thus, evi- • Generic slow-release oral morphine or generic trans- dence indicates that palliative care can generate positive dermal fentanyl patches externalities and can lower indirect costs to society, but • Palliative surgery data on costs and cost savings of palliative care in LMICs • Palliative radiotherapy are limited (Emanuel and others 2010; Hongoro and • Palliative cancer chemotherapy Dinat 2011; Mosoiu, Dumitrescu, and Connor 2014). • Canes and wheelchairs. Method for Costing an Essential Package of Palliative In many LMICs, rehabilitation and long-term care Care and Pain Control in LMICs services are either inadequate or inaccessible by the To estimate the cost of delivering high-quality palliative poor. As a result, community-based palliative care teams care and pain control in LMICs, we used a method often assume responsibility for these tasks (Ratcliff and developed by the Lancet Commission (Knaul and oth- others 2017). However, all countries should develop ers 2017). We obtained input from palliative care clini- policies and allocate funding specifically for the imple- cians and global health experts with extensive experience mentation of these much-needed services (World Health in LMICs to devise a method of costing the EP of palli- Assembly 2016). ative care and pain control described in this chapter. Palliative Care and Pain Control 241 After creating the package of interventions, inexpensive in need of palliative care in one low-income country, one essential medicines, simple equipment, human lower-middle-income country, and one upper-middle- resources, and intersectoral social supports, as well as income country, we applied our method in Rwanda, the sites or platforms where each part of the package Vietnam, and Mexico (table 12.3) using the prices should be accessible (table 12.2 and annex table 12A.1), reported in each. we then estimated amounts of each item that would be Not surprisingly, the cost of achieving universal required by patients with each ICD-10 health condition access to the EP would require a much higher share of that generates a need for palliative care or pain control. total government expenditure on health in Rwanda We also estimated the staffing needs in full-time equiv- (between 7.0 and 10.0 percent) than in Mexico (less alents (FTEs) to apply the package at each site. Using than 1.0 percent) or Vietnam (between 1.0 and WHO GHE mortality data (WHO 2016), we were able 1.7 percent). As a proportion of gross domestic product, to estimate the total amount of medicines, equipment, there would be an almost tenfold difference in cost of and personnel FTEs, as well as the intersectoral social the EP between Rwanda (0.25 percent) and Mexico supports needed to provide palliative care and pain (0.03 percent) or Vietnam (0.04 percent). control to all patients in need in any country. We also produced preliminary estimates of the costs To determine the cost of delivering the EP in a spe- of the intersectoral social supports previously mentioned, cific country, we then identified the reported unit price considering only patients living in extreme poverty of all medicines, equipment, social supports, and (daily income below US$1.9) and a patient’s one main monthly FTE salaries of the palliative care providers in caregiver living in extreme poverty (World Bank 2017). that country (De Lima and others 2014; McCoy and (These illustrative estimates assume that a stringent others 2008). The total cost of the EP was the total cost means test (screening process) can be implemented to of all components. To cost human resources, we used identify those living in extreme poverty. However, expe- monthly total pretax (including mandatory benefits), rience with means tests in many places suggests that they public sector, FTE-reported salaries. We also considered may be costly to administer and subject to abuse.) Our the most basic operational inputs required to support assumption is that intersectoral social supports are both the provision of the EP at every level of care and added, financed and provided by sectors of government work- on average, 8 percent to our overall figures. ing on poverty alleviation, and not by ministries of health. On the basis of data on subsidies provided to families by Application of the Method in Specific Countries existing anti-poverty programs in Mexico, and given the To provide examples for policy makers of the expected small proportion of families below the poverty line (3 cost of the health care components of the EP per patient percent), the social supports would represent a very Table 12.3 Per Patient Cost of the Health Care Components of the Essential Package of Palliative Care and Pain Control in Mexico, Rwanda, and Vietnam, US$, 2015 Current Value Rwandad Vietname Mexico Medicines 52 27 122 Morphine (oral or injectable) 20 14 90 Equipment 31 5 31 Palliative care team (HR) 121 78 584 Operational Costs (8% of total) 16 9 59 Total 219 119 796 % GDPa 0.25 0.04 0.03 % health expenditureb 3.35 0.56 0.50 c % public health expenditure 8.79 1.04 0.97 Note: GDP = gross domestic product; HR = human resources. a. GDP, World Development Indicators, World Bank, http://data.worldbank.org/indicator/NY.GDP.MKTP.CD. b. Health expenditure, total (% of GDP), World Development Indicators, World Bank, http://data.worldbank.org/indicator/SH.XPD.TOTL.ZS. c. Health expenditure, public (% of total health expenditure), World Development Indicators, World Bank, http://data.worldbank.org/indicator/SH.XPD.PUBL. d. For costing in Rwanda, the following substitutions were made: Fluoxetine was substituted with SSRI and reusable cloth diapers instead of disposable. e. Costing in Vietnam does not include Parenteral Fluconazole as pricing for this medicine was unavailable in the country. 242 Disease Control Priorities: Improving Health and Reducing Poverty small additional cost, about 1 percent, compared to the ANNEX health components of the EP. For Rwanda, as would be the case for other low-income countries, the total cost The annex to this chapter is available at http://www would be quite high, largely because more than .dcp-3.org/DCP. 60 percent of families live in extreme poverty. Thus, the EP would have an anti-poverty function for the most • Annex 12A. The Essential Package of Palliative financially vulnerable patients with palliative care needs Care: Interventions, Medicines, Equipment, Human and caregivers. Resources, and Intersectoral Supports Limitations of the Method NOTES This costing method has several limitations. First, it does not include the costs of initial palliative care This chapter was adapted from Knaul, F. M., P. E. Farmer, E. L. Krakauer, L. de Lima, A. Bhadelia, and others. 2017. capacity building, including secure supply-chain “Alleviating the Access Abyss in Palliative Care and Pain Relief: building for controlled substances, human resource An Imperative of Universal Health Coverage. Report of the training and policy changes to officially integrate Lancet Commission on Global Access to Palliative Care and Pain palliative care into the health care system and to Control.” The Lancet. doi:10.1016/S0140-6736(17)32513-8. ensure essential medicine accessibility. Second, our World Bank Income Classifications as of July 2015 are calculations are based on a particular model of palli- as follows, based on estimates of gross national income ative care delivery. For example, our model assumes (GNI) per capita for 2014: that inpatient care is available at health centers for one patient at a time whose family is unable to pro- • Low-income countries (LICs) = US$1,045 or less vide adequate care in the home but who does not • Middle-income countries (MICs) are subdivided: require higher-level care. Where inpatient care at (a) lower-middle-income = US$1,046 to US$4,125 health centers is not available, costs may differ. (b) upper-middle-income (UMICs) = US$4,126 to US$12,735 Our model includes estimated FTEs of all palliative • High-income countries (HICs) = US$12,736 or more. care team members. Where larger or smaller FTEs are devoted to palliative care, costs will differ. REFERENCES Abernethy, A. P., D. C. Currow, P. Frith, B. S. Fazekas, CONCLUSIONS A. McHugh, and others. 2003. “Randomised, Double Blind, Placebo Controlled Crossover Trial of Sustained Release A universally accessible EP of palliative care and pain Morphine for the Management of Refractory Dyspnea.” control can prevent and relieve suffering for chronically BMJ 327: 523–29. or terminally ill patients. It is indispensable for achieving Albanese, T. H., S. M. Radwany, H. Mason, C. Gayomali, universal health coverage and for realizing Sustainable and K. Dieter. 2013. “Assessing the Financial Impact of Development Goal 3: “ensure healthy lives and promote an Inpatient Acute Palliative Care Unit in a Tertiary Care Teaching Hospital.” Journal of Palliative Medicine 16 (3): well-being for all at all ages” (United Nations General 289–94. Assembly 2015, 16). It therefore is a medical and moral Bamberger, J. 2016. “Reducing Homelessness by Embracing imperative to include such a package in publicly financed Housing as a Medicaid Benefit.” JAMA Internal Medicine universal health coverage. In addition, the EP of pallia- 176 (8): 1051–52. tive care and pain control that we propose may reduce Belkin, G. S., J. Unutzer, R. C. Kessler, H. Verdell, G. J. Raviola, costs for health care systems and national economies and and others. 2011. “Scaling Up for the ‘Bottom Billion’; ‘5×5’ provide financial risk protection for patients and their Implementation of Community Mental Health Care in Low- families. Income Regions.” Psychiatric Services 62 (12): 1494–1502. To ensure that the EP of palliative care and pain con- Carrillo, J. E., V. A. Carrillo, H. R. Perez, D. Salas-Lopez, trol is universally accessible, governments should enact A. Natale-Pereira, and others. 2011. “Defining and Targeting appropriate policies and ensure that health care provid- Health Care Access Barriers.” Journal of Health Care for the Poor and Underserved 22 (2): 562–75. ers have the necessary competencies by including train- Cassell, E. J. 1982. “The Nature of Suffering and the Goals ing in palliative care and pain control in standard of Medicine.” New England Journal of Medicine 306 (11): undergraduate and postgraduate curricula in medicine, 639–45. nursing, and other clinical fields (Stjernswärd, Foley, and Chalkidou, K., P. Marquez, P. K. Dhillon, Y. Teerawattananon, Ferris 2007; World Health Assembly 2014). T. Anothaisintawee, and others. 2014. “Evidence-Informed Palliative Care and Pain Control 243 Frameworks for Cost-Effective Cancer Care and Prevention Gomez-Batiste, X., C. Caja, J. Espinosa, I. Bullich, M. Martınez- in Low, Middle, and High-Income Countries.” The Lancet Muñoz, and others. 2012. “The Catalonia World Health Oncology 15 (3): e119–31. http://dx.doi.org/10.1016/S1470 Organization Demonstration Project for Palliative Care -2045(13)70547-3. Implementation: Quantitative and Qualitative Results at 20 Connor, S., B. Pyenson, K. Fitch, C. Spence, and K. Iwasaki. 2007. Years.” Journal of Pain Symptom Management 43 (4): 783–94. “Comparing Hospice and Non-Hospice Patient Survival Gwyther, L., and E. L. Krakauer. 2011. “WPCA Policy Statement among Patients Who Die within a Three-Year Window.” on Defining Palliative Care.” Worldwide Palliative Care Journal of Pain and Symptom Management 33 (3): 238–46. Alliance, London. http://www.thewhpca.org/resources Connor, S., and M. C. Sepulveda Bermedo, eds. 2014. Global /item/definging-palliative-care. Atlas of Palliative Care at the End of Life. London: Worldwide Herce, M. E., S. N. Elmore, N. Kalanga, J. W. Keck, E. B. Wroe, Palliative Care Alliance. http://www.who.int/nmh/Global and others. 2014. “Assessing and Responding to Palliative _Atlas_of_Palliative_Care.pdf. Care Needs in Rural Sub-Saharan Africa: Results from a De Lima, L., T. Pastrana, L. Radbruch, and R. Wenk. 2014. “Cross- Model Intervention and Situation Analysis in Malawi.” Sectional Pilot Study to Monitor the Availability, Dispensed PLoS One 9 (10): e110457. http://www.plosone.org/article Prices, and Affordability of Opioids around the Globe.” /info%3Adoi%2F10.1371%2Fjournal.pone.0110457. Journal of Pain and Symptom Management 48 (4): 649–59. Higginson, J., C. Bausewein, C. C. Reilly, W. Gao, M. Gysels, and DesRosiers, T., C. Cupido, E. Pitout, L. van Niekerk, M. Badri, others. 2014. “An Integrated Palliative and Respiratory Care and others. 2014. “A Hospital-Based Palliative Care Service Service for Patients with Advanced Disease and Refractory for Patients with Advanced Organ Failure in Sub-Saharan Breathlessness: A Randomised Controlled Trial.” The Lancet Africa Reduces Admissions and Increases Home Death Respiratory Medicine 2 (12): 979–87. Rates.” Journal of Pain and Symptom Management 47 (4): Hongoro, C., and N. Dinat. 2011. “A Cost Analysis of a Hospital- 786–92. Based Palliative Care Outreach Program: Implications for Dowell, D., T. M. Haegerich, and R. Chou. 2016. “CDC Expanding Public Sector Palliative Care in South Africa.” Guideline for Prescribing Opioids for Chronic Pain— Journal of Pain and Symptom Management 41 (6): 1015–24. United States.” MMWR 65 (1): 1–48. Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, S. Berkley, Doyle, K. E., S. K. El Nakib, M. R. Rajagopal, S. Babu, G. Joshi, and others. 2013. “Global Health 2035: A World Converging and others. 2017. “Predictors and Prevalence of Pain and within a Generation.” The Lancet 382 (9908): 1898–955. Its Management in Four Regional Cancer Hospitals in Joranson, D., M. Maurer, and F. Mwangi-Powell, eds. 2010. India.” Journal of Global Oncology. http://ascopubs.org/doi “Guidelines for Ensuring Patient Access to, and Safe /pdf/10.1200/JGO.2016.006783. Management of, Controlled Medicines.” Kampala: African Elsayem, A., K. Swint, M. J. Fisch, J. L. Palmer, S. Reddy, Palliative Care Association. http://integratepc.org/wp and others. 2004. “Palliative Care Inpatient Service in a -content/uploads/2013/05/patient_access1.pdf. Comprehensive Cancer Center: Clinical and Financial Knaul, F. M., P. E. Farmer, E. L. Krakauer, L. de Lima, A. Bhadelia, Outcomes.” Journal of Clinical Oncology 22 (10): 2008–14. and others. 2017. “Alleviating the Access Abyss in Palliative Emanuel, N., M. A. Simon, M. Burt, A. Joseph, N. Sreekumar, Care and Pain Relief: An Imperative of Universal Health and others. 2010. “Economic Impact of Terminal Illness Coverage. Report of the Lancet Commission on Global and the Willingness to Change It.” Journal of Palliative Access to Palliative Care and Pain Control.” The Lancet. Medicine 13 (8): 941–44. doi:10.1016/S0140-6736(17)32513-8. Emanuel, R. H., G. A. Emanuel, E. B. Reitschuler, A. J. Lee, Krakauer, E. L. 2008. “Just Palliative Care: Responding E. Kikule, and others. 2008. “Challenges Faced by Informal Responsibly to the Suffering of the Poor.” Journal of Pain Caregivers of Hospice Patients in Uganda.” Journal of and Symptom Management 36 (5): 505–12. Palliative Medicine 11 (5): 746–53. Lutz, S. T., J. Jones, and E. Chow. 2014. “Role of Radiation Farmer, P. E., B. Nizeye, S. Stulac, and S. Keshavjee. 2006. Therapy in Palliative Care of the Patient with Cancer.” “Structural Violence and Clinical Medicine.” PLoS 3 Journal of Clinical Oncology 32 (26): 2913–19. (10): e449. http://journals.plos.org/plosmedicine/article Mathers, C., G. Stevens, D. Hogan, A. Mahanani, and J. Ho. ?id=10.1371/journal.pmed.0030449. 2018. “Global and Regional Causes of Death: Patterns Foley, K. M., J. L. Wagner, D. E. Joranson, and H. Gelband. 2006. and Trends, 2000–15.” In Disease Control Priorities (third “Pain Control for People with Cancer and AIDS.” In Disease edition): Volume 9, Improving Health and Reducing Poverty, Control Priorities in Developing Countries (second edition), edited by D. T. Jamison, H. Gelband, S. Horton, P. Jha, edited by D. T. Jamison, J. G. Breman, A. R. Measham, R. Laxminarayan, C. N. Mock, and R. Nugent. Washington, G. Alleyne, M. Claeson, D. B. Evans, P. Jha, A. Mills, and DC: World Bank. P. Musgrove. Washington, DC: World Bank. May, P., M. M. Garrido, J. B. Cassel, A. S. Kelley, D. E. Meier, Gaskin, D. J., and P. Richard. 2011. “Appendix C: The Economic and others. 2016. “Palliative Care Teams’ Cost-Saving Effect Costs of Pain in the United States.” In Institute of Medicine. Is Larger for Cancer Patients with Higher Numbers of Relieving Pain in America: A Blueprint for Transforming Comorbidities.” Health Affairs 35 (1): 44–53. Prevention, Care, Education, and Research, 301–37. McCoy, D., S. Bennett, S. Witter, B. Pond, B. Baker, and others. Washington DC: National Academies Press. https://www 2008. “Salaries and Incomes of Health Workers in Sub- .ncbi.nlm.nih.gov/books/NBK92521/. Saharan Africa.” The Lancet 371 (9613): 675–81. 244 Disease Control Priorities: Improving Health and Reducing Poverty Miner, T. J. 2005. “Palliative Surgery for Advanced Cancer: Stjernswärd, J., K. M. Foley, and F. D. Ferris. 2007. “The Public Lessons Learned in Patient Selection and Outcome Health Strategy for Palliative Care.” Journal of Pain and Assessment.” American Journal of Clinical Oncology 28 (4): Symptom Management 33 (5): 486–93. 411–14. Summers, L. H. 2016. “No Free Lunches but Plenty of Cheap Moens, K., I. J. Higginson, and R. Harding. 2014. “Are There Ones.” Larry Summers’s Blog, February 7. http://larrysummers Differences in the Prevalence of Palliative Care-Related .com/2016/02/08/no-free-lunches-but-plenty-of-cheap-ones/. Problems in People Living with Advanced Cancer and Syed, S. T., B. S. Gerber, and L. K. Sharp. 2013. “Traveling Eight Non-Cancer Conditions? A Systematic Review.” towards Disease: Transportation Barriers to Health Care Journal of Pain and Symptom Management 48 (4): Access.” Journal of Community Health 38 (5): 976–93. 660–77. Temel, J. S., J. A. Greer, A. Muzikansky, E. R. Gallagher, Morrison, R. S., J. D. Penrod, J. B. Cassel, M. Caust-Ellenbogen, S. Admane, and others. 2010. “Early Palliative Care for A. Litke, and others. 2008. “Cost Savings Associated with Patients with Metastatic Non–Small-Cell Lung Cancer.” U.S. Hospital Palliative Care Consultation Programs.” New England Journal of Medicine 363 (8): 733–742. Archives of Internal Medicine 168 (16): 1783–90. United Nations General Assembly. 2015. “Transforming Our Mosoiu, D., M. Dumitrescu, and S. R. Connor. 2014. World: The 2030 Agenda for Sustainable Development.” “Developing a Costing Framework for Palliative Care UNGA A/RES/70/1. http://www.un.org/ga/search/view Services.” Journal of Pain and Symptom Management 48 _doc.asp?symbol=A/RES/70/1&Lang=E. (4): 719–29. Van den Beuken-van Everdingen, M. H. J., L. J. Hochstenbach, Ostgathe, C., B. Alt-Epping, H. Golla, J. Gaertner, G. Lindena, and E. A. J. Joosten. 2016. “Update on Prevalence of Pain and others. 2011. “Non-Cancer Patients in Specialized in Patients with Cancer: Systematic Review and Meta- Palliative Care in Germany: What Are the Problems?” Analysis.” Journal of Pain and Symptom Management 51 Palliative Medicine 25 (2): 148–52. (6): 1070–90. Pain and Policy Studies Group. 2017. Opioid Consumption WHO (World Health Organization). 1996. Cancer Pain Data. Madison, University of Wisconsin. http://www Relief, with a Guide to Opioid Availability (second edi- .painpolicy.wisc.edu/opioid-consumption-data. tion). Geneva: WHO. http://apps.who.int/iris/bitstream Patel, V. 2014. Where There Is No Psychiatrist. London: Gaskell. /10665/37896/1/9241544821.pdf. Postier, A., J. Chrastek, S. Nugent, K. Osenga, and ———. 2002. “Definition of Palliative Care.” WHO, Geneva. S. J. Friedrichsdorf. 2014. “Exposure to Home-Based http://www.who.int/cancer/palliative/definition/en/. Pediatric Palliative and Hospice Care and Its Impact ———. 2006. “Amended Constitution of the World Health on Hospital and Emergency Care Charges at a Single Organization.” http://www.who.int/governance/eb/who Institution.” Journal of Palliative Medicine 17 (2): 183–88. _constitution_en.pdf. Rahman, A., S. U. Hamdani, N. R. Awan, R. A. Bryant, K. S. ———. 2011. Ensuring Balance in National Policies on Dawson, and others. 2016. “Effect of a Multicomponent Controlled Substances: Guidance for Availability and Behavioral Intervention in Adults Impaired by Accessibility of Controlled Medicines. Geneva: WHO. http:// Psychological Distress in a Conflict-Affected Area of www.who.int/medicines/areas/quality_safety/GLs_Ens Pakistan: A Randomized Clinical Trial.” Journal of the _Balance_NOCP_Col_EN_sanend.pdf. American Medical Association 316 (24): 2609–17. ———. 2012. WHO Guidelines on the Pharmacological Ratcliff, C., A. Thyle, S. Duomai, and M. Manak. 2017. “Poverty Treatment of Persisting Pain in Children with Medical Reduction in India through Palliative Care: A Pilot Project.” Illnesses. Geneva: WHO. http://apps.who.int/iris/bitstream Indian Journal of Palliative Care 23 (1): 41–45. /10665/44540/1/9789241548120_Guidelines.pdf. Shulman, L. N., T. Mpunga, N. Tapela, C. M. Wagner, T. Fadelu, and ———. 2015a. WHO International Classification of others. 2014. “Bringing Cancer Care to the Poor: Experiences Diseases (ICD)-10. Geneva: WHO. http://apps.who.int from Rwanda.” Nature Reviews Cancer 14 (12): 815–21. /classifications/icd10/browse/2015/en. Shulman, L. N., C. M. Wagner, R. Barr, G. Lopes, G. Longo, ———. 2015b. “WHO Model List of Essential Medicines, and others. 2015. “Proposing Essential Medicines to Treat 19th list.” Geneva, WHO. http://www.who.int/medicines Cancer: Methodologies, Processes, and Outcomes.” Journal /publications/essentialmedicines/en/index.html. of Clinical Oncology 34 (1): 69–75. ———. 2015c. “WHO Model List of Essential Medicines Singer, A. E., J. R. Goebel, Y. S. Kim, S. M. Dy, S. C. Ahluwalia, for Children, 5th list.” WHO, Geneva. http://www.who and others. 2016. “Populations and Interventions for .int/medicines/publications/essentialmedicines/en/index Palliative and End-of-Life Care: A Systematic Review.” .html. Journal of Palliative Medicine 19 (9): 995–1008. doi:10.1089/ ———. 2016. “Cause-Specific Mortality.” WHO, Geneva. jpm.2015.0367. http://www.who.int/healthinfo/global_burden_disease Solano, J. P., B. Gomes, and I. J. Higginson. 2006. “A Comparison /estimates/en/index1.html. of Symptom Prevalence in Far Advanced Cancer, AIDS, ———. 2017. “WHO Methods and Data Sources for Country- Heart Disease, Chronic Obstructive Pulmonary Disease and Level Causes of Death 2000–2015.” Geneva: WHO. Renal Disease.” Journal of Pain and Symptom Management http://www.who.int/healthinfo/global_burden_disease 31 (1): 58–69. /GlobalCOD_method_2000_2015.pdf. Palliative Care and Pain Control 245 World Bank. 2017. “Poverty Headcount Ratio at $1.90 on Ageing and Health.” WHA A69/17. http://apps.who.int a Day (2011 PPP) (% of Population).” World Bank, /gb/ebwha/pdf_files/WHA69/A69_17-en.pdf. Washington, DC. http://data.worldbank.org/indicator/ Zimmerman, C., R. Riechelmann, M. Krzyzanowska, G. Rodin, SI.POV.DDAY. and I. Tannock. 2008. “Effectiveness of Specialized Palliative World Health Assembly. 2014. “Strengthening of Care: A Systematic Review.” Journal of the American Medical Palliative Care as a Component of Comprehensive Association 299 (14): 1698–709. Care throughout the Life Course.” WHA 67.19. Zimmerman, C., N. Swami, M. Krzyzanowska, B, Hannon, http://apps.who.int/gb/ebwha / pdf_files/WHA67/ N. Leighl, and others. 2014. “Early Palliative Care for A67_R19-en.pdf. Patients with Advanced Cancer: A Cluster-Randomised ———. 2016. “Multisectoral Action for a Life Course Approach Controlled Trial.” The Lancet 383 (9930): 1721–30. http:// to Healthy Ageing: Draft Global Strategy and Plan of Action dx.doi.org/10.1016/S0140-6736(13)62416-2. 246 Disease Control Priorities: Improving Health and Reducing Poverty Chapter 13 Strengthening Health Systems to Provide Emergency Care Teri A. Reynolds, Hendry Sawe, Andrés M. Rubiano, Sang Do Shin, Lee Wallis, and Charles N. Mock INTRODUCTION many ways, depending on available resources, and each is essential to the delivery of effective emergency care. All around the world, acutely ill and injured people of all Each of the previous eight volumes of this edition of ages seek care every day. They will call neighbors, the Disease Control Priorities (third edition) (DCP3) police, or universal emergency numbers for help. They presents a package of essential services and highlights will be assisted by family members, community mem- urgent services for conditions likely to result in morbid- bers with first-aid training, or professional prehospital ity or mortality if not addressed rapidly. An ECS is an providers. They may travel to a health care facility by integrated mechanism to address these time-sensitive foot, motorcycle, taxi, or ambulance. On arrival, they conditions, and this chapter integrates the urgent inter- may or may not find a designated emergency area and ventions from all the Disease Control Priorities packages providers capable of delivering the care they need. with the WHO ECS framework to derive a package of Emergency care systems (ECSs) address a wide range essential emergency care services, including key policy of acute conditions, including injuries, communicable strategies for system development. This effort is intended and noncommunicable diseases, and complications of to identify ways in which national health care systems pregnancy. Especially when there are barriers to health globally can be strengthened to provide emergency care care access, people may seek care only when acutely ill or more effectively. injured. Emergency care is an essential component of universal health coverage—a critical mechanism for ensuring accessible, affordable, high-quality care—and WHAT IS EMERGENCY CARE? for many people around the world, it is the primary point of access to the health system. Emergency care has been defined by various attributes, The World Health Organization (WHO) has defined such as time-to-care provision and acuity of the condition a series of essential functions for an ECS that span from addressed. Common definitions include care delivered prehospital care and transport through facility-based within minutes or hours (Kobusingye and others 2006) emergency unit care to early operative and critical care and care for conditions that require rapid intervention to (figure 13.1). Each of these functions can be achieved in avoid death or disability (Hirshon and others 2013). Corresponding author: Teri Reynolds, Department for Management of Noncommunicable Diseases, Disability, Violence, and Injury Prevention, World Health Organization, Geneva, Switzerland; reynoldst@who.int. 247 Figure 13.1 WHO Emergency Care System Infographic Human resources Functions Vehicles, equipment, supplies, information technologies Emergency Care System Framework CLINICAL OR OPERATIONAL All around the world, acutely ill and injured people seek care every day. Frontline PROTOCOLS Inpatient providers manage children and adults with injuries and infections, heart attacks CHART and strokes, asthma and acute complications of pregnancy. An integrated ap- proach to early recognition and management saves lives. This visual summary TRANSPORT CARE KIT illustrates the essential functions of a responsive emergency care system and COMMUNICATION the key human resources, equipment, and information technologies needed to TECHNOLOGIES execute them. The reverse side adresses elements of governance and oversight. • Early critical care • Early operative care Ad mis sion sfer an Tr Emergency unit Disposition Disc harge home • Positioning Fie • Intervention Provider Driver ld t • Monitoring • Assessment o • Resuscitation facility com m A B C • Intervention D • Monitoring tion ctiva Allied ma u nic a ti o n ste mber health Sy s nu cces worker A Dispatcher via ons cti tru Clerical ns Provider staff I S B A R Handover Triage Screening Registration Provider Bystander Reception of patients Scene Transport Facility • Bystander response • Patient transport • Reception • Dispatch • Transport care • Emergency unit care • Provider response • Disposition www.who.int/emergencycare · emergencycare@who.int • Early inpatient care Source: WHO, http://www.who.int/emergencycare/emergencycare_infographic/en/. Note: H = hospital; WHO = World Health Organization. Definitions of emergency care that focus on the acuity of However, users of the health care system may not them- the condition itself have the advantage of being indepen- selves be able to judge whether a condition is life- dent of the rapidity or level of care that can be achieved threatening; the belief that an emergency condition by the system and, instead, encompass all rapidly exists requires at least urgent preliminary assessment by progressive conditions. This approach is preferable to health care professionals. definitions grounded in a specific period for care deliv- People in need of care may access the system at many ery, since much emergency care would fall outside of a points, including by activating the prehospital system, time-bound definition in regions where long transport by visiting a primary health center, or by presenting times are the norm and referrals may take days. directly to a hospital-based emergency unit (figure 13.2); To facilitate consistent understanding across systems providers at every level of the health system deliver at varying levels of development, emergency care is con- emergency care, whether or not they have the dedicated sidered here to encompass health services for conditions training and resources to do so effectively. Frontline that require rapid intervention to avert death and dis- emergency care may involve early recognition and initial ability (such as shock or respiratory failure) or for which resuscitation for dangerous conditions followed by delays of hours can worsen prognosis or render care less transfer for definitive care (for example, chest drain effective (such as treatment of infections, management placement, volume resuscitation, and transfusion per- of asthma exacerbations, or suturing of wounds). formed before transfer for surgery) or may encompass 248 Disease Control Priorities: Improving Health and Reducing Poverty Figure 13.2 Access to Emergency Care People may access emergency care at multiple levels of the system... ICU OT Clinic Basic triage, Triage, recognition, recognition, resuscitation, resuscitation, Emergency unit and referral and advanced care Scene PHC : Level 1 Hospital : Level 2/3 Transport and Transfer Patients may receive definitive care, ending the acute episode, at multiple levels of the system or may require transfer for additional care. Source: World Health Organization, http://www.who.int/emergencycare. Note: H = hospital; ICU = intensive care unit; OT = operating theatre; PHC = primary health clinic. definitive therapy (such as administration of antibiotics Further details on emergency care specific to the prehos- for pneumonia, wound repair, or nonoperative fracture pital setting are covered in chapter 14 of volume 1 of management). DCP3 (Thind and others 2015). In keeping with the WHO ECS framework, the use of the term emergency care in this chapter encompasses WHY FOCUS ON EMERGENCY CARE? care that occurs both before and beyond the emergency unit itself (figure 13.1), including prehospital care and Expanding the availability of disease-specific treatments the early operative care and critical care that may occur and procedures is essential. The effectiveness of these in an operating room or an inpatient intensive care unit interventions is compromised, however, without the (ICU). Although the focus of this chapter’s package is initial emergency care interface that links undifferenti- on facility-based emergency care, many of these services ated patient presentations to appropriate definitive care. can be mapped onto prehospital systems at increasing For the most part, people seeking care for acute illness or levels of development. In general, depending on the injury do not know if they have a condition requiring level of development of a prehospital system, the ser- oxygen, antibiotics, pericardiocentesis, or surgery. They vices may be very basic, similar to those available at a generally present complaining of fever, pain, or difficulty community-based health center, or may include sophis- breathing rather than pneumonia, appendicitis, or tam- ticated critical care approaching that available in an ICU. ponade. They do not necessarily know when they are Strengthening Health Systems to Provide Emergency Care 249 critically ill and cannot go directly to ICUs or operating a dedicated area and standards for hospital-based emer- rooms. In most parts of the world, initial emergency care gency care, and a core of nonrotating providers trained in is delivered by frontline providers (often cadres other the care of emergencies and assigned to the emergency than doctors) acting with limited diagnostic resources. unit. These gaps are reflected in wide global discrepancies Emergency care includes both the early assessment that in outcomes across the range of emergency conditions: helps narrow a chief complaint toward a diagnosis, as well as the initial management that allows survival until • Overall mortality rates from diabetic ketoacidosis a diagnosis-oriented therapy can be identified and are less than 1 percent in high-income countries accessed. The failure (a) to designate and staff emer- (HICs) (Nyenwe and Kitabchi 2011) but are as high gency care areas, (b) to train frontline providers in rec- as 30 percent in LMICs (Mbugua and others 2005). ognition of and resuscitation for dangerous conditions, • The estimated lifetime risk of maternal mortality in and (c) to create organized ECSs to match people rapidly HICs is 1 in 3,300, compared to 1 in 41 in LMICs with the care they need, will cost lives, even where (Alkema and others 2016). life-saving resources are already available somewhere in • Although available data are limited and range widely, the system (Dare and others, 2015; Grimes and others mortality from sepsis in LMICs is likely to be more than 2011; Hsiao and others 2013; Irfan, Irfan, and Spiegel twice that in HICs (Silva and others 2004; Stevenson 2012). and others 2014; Tanriover and others 2006). A systematic approach to emergency care—centered • Even within a single country, the discrepancy in out- on acuity-based triage, early recognition and resuscita- comes associated with limited access to emergency tion, and simple initial management and referral—has care can be dramatic: in one Indian study, being poor been shown to decrease the mortality associated with a was associated with reduced access to timely emer- range of medical and surgical conditions. Implementation gency treatments for acute myocardial infarction of a systematic emergency unit approach to early recog- and with a 50 percent relative increase in mortality nition and treatment has been shown to reduce signifi- (Xavier and others 2008). cantly mortality from both pneumonia and sepsis • Finally, modeling studies estimate that between 20 (Gaieski and others 2010; Hortmann and others 2014; and 38 percent of the global injury burden (between Rivers 2001). Better-organized trauma systems have 1 million and 2 million fatalities each year and been shown to decrease preventable deaths among the around 52 million disability-adjusted life years, or severely injured by 50 percent and to improve func- DALYs) could be averted if severe injury outcomes tional outcomes among survivors (Siman-Tov, in LMICs were similar to those in HICs (Higashi and Radomislensky, and Peleg 2013; Tallon and others others 2015; Mock and others 2012). 2012). Recognition and emergency treatment for myo- cardial infarction delivered within 60 minutes rather Overall, the global burden of disease that potentially than hours has been shown to reduce mortality twofold can be addressed by prehospital and facility-based emer- (Terkelsen and others 2010); early noninvasive positive gency care is estimated at a staggering 54 percent of the pressure ventilation reduces in-hospital mortality (RR annual deaths in LMICs (Thind and others 2015) [95% CI]: 0.66 [0.48, 0.89]) in patients with heart fail- (figure 13.3). ure (Vital, Ladeira, and Atallah 2013). Early treatment Although severe global discrepancies exist in out- with aspirin (within 48 hours) for ischemic stroke has comes from emergency conditions, both these modeling been shown to reduce both morbidity and mortality estimates and direct evidence suggest that emergency (Sandercock and others 2014), and early intensive care has the potential to narrow this gap dramatically. blood-pressure lowering (within six hours) has been Powerful examples of feasible life-saving emergency care shown to improve functional outcomes in hemorrhagic interventions in LMICs include the following: stroke (Anderson and others 2013). Three obstetric emergencies—hemorrhage, hypertensive disorders, and • Organizing low-cost prehospital systems was asso- sepsis—are responsible for more than half of the mater- ciated with a dramatic decrease in all-condition nal deaths worldwide (Say and others 2014) and are mortality in Cambodia and Iraq (Husum and others highly treatable with simple emergency care interven- 2003), in road-traffic mortality in Iraq (Murad tions (Holmer and others 2015). and others 2012), and in snakebite mortality in Despite the substantial positive impact emergency Nepal (Sharma and others 2013). A recent review care can have, however, many low- and middle-income and meta-analysis estimated that simple prehospital countries (LMICs) lack the fundamentals of organized systems can reduce injured patients’ risk of death by emergency care: basic prehospital care and transport, 25 percent (Henry and Reingold 2012). 250 Disease Control Priorities: Improving Health and Reducing Poverty Figure 13.3 Burden of Disease That Can Potentially Be Addressed by Prehospital and Emergency Care in Low- and Middle-Income Countries 200 180 160 140 120 Millions 100 80 60 40 20 0 es ns s s ria s se se e ma es l l on iti ion na na as as tio ea ea et ala ing th iti ise tio tio dit ab ise dis dis ec As nd M en en en td on Di nf ld co rt lar M int Int ar yi lc ea ea od cu he Un or na ch rh as ho at er ive ar mi pir ov ild at Di ns he br M es Ch rte re Isc rr Ce pe we Hy Lo Communicable and maternal conditions Chronic conditions Injuries Total addressable deaths = 24.3 million Total addressable DALYs lost = 1,023 million Total addressable YLL = 932 million Total addressable YLD = 91.4 million Source: Thind and others 2015 (data from WHO 2013). Note: DALYs = disability-adjusted life years; LMICs = low- and middle-income countries; YLD = years lived with disability; YLL = years of life lost. • Designating an area for emergency care of all critical • The introduction of standardized resuscitation pro- patients at a third-level hospital in Romania trans- tocols in Colombia reduced hospital length of stay formed care and halved mortality.1 and all-cause mortality among injured patients by a • In Malawi, restructuring a hospital intake area to quarter (Kesinger, Puyana, and Rubiano 2014). create a dedicated emergency care area and initiating • Short course trainings in trauma management were formal triage were associated with halved inpatient associated with reduced mortality in injured patients mortality and a reduction in the proportion of deaths from 19.9 to 15.1 percent in China (Wang and others occurring within 24 hours from 36 to 12.6 percent 2010) and from 8.8 to 6.3 percent in Rwanda with no (Molyneux, Ahmad, and Robertson 2006). significant increase in resource usage (Petroze and • Timely simple interventions (fluids, antibiotics, and others 2015). clinical monitoring) within the first six hours of hos- • Finally, one modeling study, although dependent on pitalization in Ugandan adults with serious infection the assumption of available oxygen, predicted that the reduced mortality from 46 to 33 percent (Jacob and use of pulse oximetry, combined with current WHO others 2012). guidelines for recognition of severe illness, has the • In rural Mali, improved access to emergency obstetric potential to avert up to 148,000 deaths per year in the care halved the risk of maternal mortality and reduced 15 countries across Africa and Asia with the highest the risk nearly threefold among women with hemor- global burden of childhood pneumonia (Floyd and rhage (Fournier and others 2009). Growing evidence others 2015). indicates that a range of simple nonsurgical interven- tions for complications of childbirth can dramati- Evidence from around the world shows that emer- cally improve maternal mortality in LMICs (Kausar gency care is an effective means of saving lives, and and others 2012; Miller, Lester, and Hensleigh 2004; evidence from LMICs suggests that feasible and simple Paxton and others 2005). steps to improve emergency care could rapidly improve Strengthening Health Systems to Provide Emergency Care 251 outcomes and reduce global disparities in outcomes. THE WHO EMERGENCY CARE SYSTEM More broadly, the recently adopted United Nations FRAMEWORK Sustainable Development Goals (SDGs)2 and their asso- ciated targets provide guidance for coordinated action to To facilitate systematic assessment and targeted develop- end poverty, protect the planet, and promote health on a ment of integrated ECSs, the WHO ECS framework global level. ECSs directly address nearly all the health- (annex 13A) was designed with input from more than related SDG targets, as well as those on disasters and 30 LMICs. This consensus-based document defines violence (table 13.1), and the SDG targets are unlikely to essential emergency care functions at the scene of injury be met without strengthening ECSs globally. or illness, during transport, and through emergency unit Table 13.1 Sustainable Development Goals Directly Addressed by Emergency Care Sustainable Development Goal targets Emergency care interventions 3.1: By 2030, reduce the global maternal mortality ratio to less than 70 per 100,000 Interventions include treatment for obstetric emergencies such as live births hemorrhage, hypertensive disorders, and sepsis. 3.2: By 2030, end preventable deaths of newborns and children under age five years, Interventions include treatment of acute pediatric diarrhea, pneumonia, with all countries aiming to reduce neonatal mortality to at least as low as 12 per and sepsis. 1,000 live births and under-five mortality to at least as low as 25 per 1,000 live births 3.3: By 2030, end the epidemics of HIV/AIDS, tuberculosis, malaria, and neglected Interventions include recognition and treatment of acute infections. tropical diseases, and combat hepatitis, water-borne diseases, and other communicable diseases 3.4: By 2030, reduce by one-third premature mortality from noncommunicable Interventions include treatment of acute exacerbations of diseases through prevention and treatment and promote mental health and noncommunicable diseases such as heart attack, stroke, and asthma. well-being 3.5: Strengthen the prevention and treatment of substance abuse Interventions include treatment of overdose and emergency-unit harm- reduction interventions. 3.6: By 2020, halve the number of global deaths and injuries from road-traffic Interventions include postcrash emergency care for injury. accidents 3.7: By 2030, ensure universal access to sexual and reproductive health care Interventions include time-sensitive postexposure treatments. services, including family planning, information, and education; and the integration of reproductive health into national strategies and programs 3.8: Achieve universal health coverage, including financial risk protection; access Interventions include continuous access to timely essential services to quality essential health care services; and access to safe, effective, quality, and for acute illness and injury. Emergency care is the primary point of affordable essential medicines and vaccines access to the health system for many, especially among vulnerable populations. 3.9: By 2030, substantially reduce the number of deaths and illnesses from Interventions include management of acute exposures. hazardous chemicals and air, water, and soil pollution and contamination 3d: Strengthen the capacity of all countries, particularly LMICs, for early warning, The ECS is a critical site for syndromic surveillance and for risk reduction, and risk management of national and global health risks preparedness to mitigate the risk of health system collapse in the face of mass events. 11.5: By 2030, significantly reduce the number of deaths and the number of people The ECS is an essential substrate for emergency response and health affected and substantially decrease the direct economic losses relative to global system resilience. gross domestic product caused by disasters, including water-related disasters, with a focus on protecting the poor and people in vulnerable situations 16.1 Significantly reduce all forms of violence and related death rates everywhere Interventions include treatment for victims of violence and early recognition of vulnerable individuals. Source: Sustainable Development Goal targets, http://www.un.org/sustainabledevelopment/sustainable-development-goals/. Note: ECS = emergency care system; HIV/AIDS = human immunodeficiency virus/acquired immune deficiency syndrome. 252 Disease Control Priorities: Improving Health and Reducing Poverty and early inpatient care. The functions are mapped across assessment differs, shared challenges have been identi- the WHO Health System Building Blocks (WHO 2010), fied across many low- and low-middle-income coun- and each function is associated with general categories of tries, including the following: human and material resources as well as information and governance elements. The framework—intended to facil- • The need for better coordination of prehospital and itate system planning and development activities— facility-based care identifies the components of each essential function to • Limited or no coverage of prehospital systems, espe- allow policy makers and planners to coordinate system cially in rural areas development activities and identify and use existing pro- • Critical emergency care service gaps at first-level hos- cesses and resources more effectively. pitals (some countries report gaps in both equipment Different systems may achieve each function in differ- and skills, whereas several middle-income countries ent ways, based on available resources. For example, report limited emergency care due to first-level hos- system activation may occur in a high-resource setting pital provider knowledge gaps, even when equipment with a universal access number linked to a central, com- is available) puterized dispatch and global positioning system. In • Lack of nonrotating staff assigned to the emergency other settings, system activation may involve the use of unit, which limits coordinated action to improve care simple mobile phone–based protocols that guide dis- and processes patchers to provide advice on first aid and use landmark • Limited data on emergency care delivery and poor maps to identify patient location. links for existing data to system planning and quality- At the same time, the framework is designed to improvement efforts account for all the basic functions of emergency care. • Lack of standard clinical management and documen- Each function corresponds to specific human, material, tation in prehospital and facility settings and governance requirements. In the case of patient trans- • Gaps in dedicated emergency care training across the fer, for example, it is impossible for one person to drive system, especially regarding integration into formal and care for a patient simultaneously, so essential human curricula and ongoing certification requirements resources include both the driver and provider. The • Insufficient funding and lack of dedicated funding authority responsible for medical equipment is not likely streams to be the same as that responsible for vehicle mainte- • Lack of security for prehospital and facility-based nance, and distinct governance components are required. emergency care staff. The framework identifies minimum resource categories and ensures that all essential functions are addressed. Areas targeted for priority action by multiple coun- Since few countries will have the available resources tries include the following: to implement all components of a fully developed ECS at once, the WHO ECS framework is designed to allow • Designating or strengthening the authority of a gov- policy makers to identify gaps in care delivery and to ernment agency to ensure better coordination create context-relevant priority action plans for system • Creating policies to improve access to emergency development.The framework is linked with the WHO care, including legislation mandating access without Emergency Care System Assessment (ECSA) tool, a requirement for prior payment and explicit integra- survey-based tool designed to help policy makers and tion of emergency care services into national insur- planners assess a national or regional ECS and set prior- ance plans ities for system development (WHO, n.d.). The ECSA • Coordinating development of dedicated emergency allows users to rate the level of development of compo- units with fixed staff at first-level hospitals nents of an ECS on a progressive scale. By providing • Establishing dedicated emergency care training pro- specific descriptions of each progressive stage, the tool grams for diverse cadres, including (depending on the provides a road map, allowing users to generate action system) lay people; undergraduate health professions priorities rapidly from identified gaps (figure 13.4). For students; and a range of providers, such as clinical example, for a given component rated at the lowest level officers, nurses, and generalist and specialist doctors (level one), the next most appropriate and feasible tar- • Implementing standardized clinical charts based gets would likely be the elements described in levels two on WHO data sets to facilitate systematic clinical and three. approaches, as well as standardized data collection to WHO ECSAs and associated priority development inform quality improvement and system planning meetings have been conducted in more than 25 coun- • Developing and disseminating formal triage and tries across multiple regions. Although each country’s condition-specific management protocols. Strengthening Health Systems to Provide Emergency Care 253 Figure 13.4 Example of Progressive Ratings in the WHO Emergency Care System Assessment Tool * 8.3 Emergency unit staffing in facilities: An emergency unit is any dedicated intake area for acutely ill and injured patients. This may be referred to as the emergency department/room/ward, accident and emergency, casualty, etc. First-level hospitals are the lowest level of hospital that receives referrals. In many countries these are district hospitals. Third-level hospitals are the highest level of facility. Note that in some countries there may be other facility levels in between first-level and third-level that are not addressed here. First-level hospitals Third-level hospitals [1] There are no dedicated emergency units or no providers with specific responsibility for emergency unit patients until they are admitted. [2] There are staff that register and direct patients in the emergency unit to inpatient areas (the unit has a sorting function, but minimal care is provided). [3] Providers from inpatient services have on-call responsibility to cover emergency unit patients, but are not assigned to be in the emergency unit. [4] Providers from inpatient services are assigned to be in the emergency unit, rotating through for limited intervals (for example, one-month blocks). [5] There are a core of nonrotating providers that permanently staff the emergency unit. I don’t know. Cannot answer for another reason (explain): Source: WHO, http://who.int/emergencycare/activities/en. In addition to guiding in-country system develop- package of basic emergency care services, they identify a ment efforts, these shared priorities and country- range of services that an effective ECS must be able to identified needs also serve to guide WHO technical provide. As such, they serve as a foundation for the resource development and program agendas. package described here. Each essential package defines a set of services, includ- ing the capacity to recognize or manage specific condi- ESSENTIAL PACKAGE OF EMERGENCY CARE tions and to perform specific procedures or other Each volume of this edition of Disease Control Priorities interventions. Although many of the urgent elements highlights a set of urgent time-dependent elements specify a diagnosis (pneumonia or meningitis) or from among its essential package. Although these ele- diagnosis-specific intervention (appendectomy), most ments do not in themselves form a comprehensive emergency care is by its nature syndrome-based 254 Disease Control Priorities: Improving Health and Reducing Poverty (addressing shortness of breath, shock, or altered mental The emergency care package includes nearly all the status). Even in a fully resourced system, the entire arc of urgent elements identified in other packages from this emergency unit assessment and management may occur edition, except where these do not fall in the scope of before establishing a diagnosis. This scenario is especially emergency care (for example, electroconvulsive ther- true where diagnostic resources are limited. In this apy for depression or hepatitis B vaccination). In addi- chapter, the essential urgent services identified in other tion, the critical presenting syndromes in emergency packages from DCP3 are integrated with the components care—difficulty breathing, shock, altered mental necessary to the practice of frontline care for the undif- status—and their commonly associated diagnoses are ferentiated acutely ill patient, creating a comprehensive used to identify additional elements (table 13.3). As package of basic emergency care services (table 13.2). with the other packages in this edition, the essential Table 13.2 Essential Package of Emergency Care Protocols with Training and Capacity to Perform Referral and Primary health center First-level hospital specialized hospitals Crosscutting policy interventions Recognition of danger signs in Acuity-based triage of children • Ensure that the National Ministry of Health has a children and adults and adults directorate dedicated to emergency care (not limited to disaster response). Vital signs measurement • Conduct a standardized national assessment of the emergency care system (using the WHO ECSA or a similar tool) to identify gaps and inform system development. BLS ALS • Ensure that emergency care is explicitly incorporated into the National Health Plan. Neonatal resuscitation (including Full supportive care for preterm • Establish national legislation ensuring access to kangaroo care and thermal care newborns emergency care without regard to ability to pay. for preterm newborns) Basic approach to difficulty in Advanced approach to difficulty Advanced condition- • Ensure that hospitals at all levels include dedicated breathing, shock, altered mental in breathing, shock, altered specific algorithms for life- emergency units—areas dedicated to the provision status, trauma mental status, trauma threatening conditions of emergency care and staffed with at least a core of nonrotating personnel who are specifically trained in the care of emergency conditions. • Disseminate dedicated training for emergency care across cadres, including training in basic emergency care for all prehospital providers, basic emergency care training for all cadres of facility-based providers who treat patients with emergency conditions, dedicated emergency care training integrated into undergraduate medical and nursing curricula, and residency or specialist training programs in emergency medicine. Detection of sepsis Emergency management of • Establish acuity-based triage systems at all facilities sepsis that regularly receive acutely ill and injured patients. • Establish prehospital care systems based on WHO or other international standards, including a dedicated certification pathway for prehospital care providers and a toll-free, universal access number for emergency care. table continues next page Strengthening Health Systems to Provide Emergency Care 255 Table 13.2 Essential Package of Emergency Care (continued) Protocols with Training and Capacity to Perform Referral and Primary health center First-level hospital specialized hospitals Crosscutting policy interventions • Develop critical process and clinical protocols as identified in the WHO ECS framework (including transport and referral protocols, prehospital and facility-based clinical treatment protocols, disaster and mass casualty). Detect and initiate treatment of Advanced treatment of severe • Implement standardized clinical charts and registries severe malnutrition malnutrition incorporating essential data points, such as those based on WHO standards, to facilitate quality improvement efforts. Post exposure prevention of STI/ HIV, emergency contraception, counseling Basic case-based syndromic surveillance and reporting Basic communicable disease Advanced communicable isolation disease isolation Disaster and mass casualty Advanced regional protocols response protocols for disaster and mass casualty Emergency Unit Procedures Referral and Primary health center First-level hospital specialized hospitals Cervical spine immobilization Endotracheal intubation Oral and nasal airway placement Surgical airway Bedside swallow evaluation BVM ventillation Mechanical Ventilation Noninvasive positive pressure ventilation Oxygen administration Needle decompression for tension pneumothorax Placement of chest drain IV fluid infusion (peripheral) for IV infusion (central) neonates, children, adults Pericardiocentesis Defibrillation Pacing Cardioversion (including synchronized) Safe physical restraint NGT placement table continues next page 256 Disease Control Priorities: Improving Health and Reducing Poverty Table 13.2 Essential Package of Emergency Care (continued) Emergency Unit Procedures Referral and Primary health center First-level hospital specialized hospitals Lumbar puncture Passive rewarming techniques Active invasive rewarming techniques Drainage of superficial abscess Basic wound care Suturing laceration Escharotomy/fasciotomy Splinting for extremity injury Nonoperative fracture management (closed reduction and casting) Reduction of simple dislocated joint Placement of external fixator; use of traction Relief of urinary obstruction: catheterization or suprapubic cystostomy Management of labor and (See operative services below) delivery in low risk women (BEMNOC) Procedural sedation Regional block Operating Theatre Services Referral and Primary health center First-level hospital specialized hospitals Spinal anesthesia General anesthesia Repair of perforations: for Surgical intervention for example, perforated peptic gastrointestinal bleeding ulcer, typhoid ileal perforation Appendectomy Colostomy Gallbladder removal Hernia, including incarceration Trauma laparotomy Open reduction and internal fixation for fractures Irrigation and debridement of open fractures Emergency surgery for obstruction Trauma-related amputations table continues next page Strengthening Health Systems to Provide Emergency Care 257 Table 13.2 Essential Package of Emergency Care (continued) Operating Theatre Services Referral and Primary health center First-level hospital specialized hospitals Burr hole Drainage of septic arthritis Surgery for ectopic pregnancy Cesarean section Hysterectomy for uterine rupture or intractable postpartum hemorrhage Dilation and curretage Radiology Services Referral and Primary health center First-level hospital specialized hospitals Comprehensive X-ray services CT services Radiology performed ultrasound Point of care ultrasound Laboratory Services Referral and Primary health center First-level hospital specialized hospitals Point of care testing: glucose, Point of care HIV testing. Comprehensive malaria, urinalysis and urine Laboratory complete blood laboratory services for pregnancy test, hemoglobin. counts, simple coagulation emergency diagnoses, studies, urea, and electrolytes. including troponin and Slide microscopy for cell counts, cardiac markers, blood malaria, and wet preparation. gas, thyroid studies, STI testing. Capcity to collect therapeutic drug levels blood culture in emergency unit prior to antibiotic administration. Medications ABCDEs Referral and Primary health center First-level hospital specialized hospitals IV paralytic (depolarizing and nondepolarizing agent) Oral steroids IV steroids (for airway, CNS, and antenatal) Inhaled bronchodilator Nebulized bronchodilator IM adrenaline IV adrenaline Oral rehydration solution IV fluids for rehydration Transfusion (whole blood, FFP, packed red blood cells) Oral aspirin Systemic anticoagulation table continues next page 258 Disease Control Priorities: Improving Health and Reducing Poverty Table 13.2 Essential Package of Emergency Care (continued) Medications ABCDEs Referral and Primary health center First-level hospital specialized hospitals Thrombolytic (streptokinase for STEMI) Insulin Oral (buccal) glucose IV glucose Antidotes Referral and Primary health center First-level hospital specialized hospitals Activated charcoal Naloxone Antithyroid agents Bicarbonate infusion Atropine Antivenina Pyridoxine Oral Vitamin K IV Vitamin K Cardiac Referral and Primary health center First-level hospital specialized hospitals Oral diuretics IV diuretics Adenosine Advanced vasopressor support Amiodarone Nitroglycerin SL IV Nitroglycerin IV antihypertensive agent IV Betablockers or CCB IV Calcium Oral potassium IV Potassium CNS Referral and Primary health center First-level hospital specialized hospitals Oral antipsychotic IM & IV antipsychotic Oral and rectal benzodiazepine IM & IV benzodiazepine Oral and IM analgesia IV analgesia Local anesthesia for injection table continues next page Strengthening Health Systems to Provide Emergency Care 259 Table 13.2 Essential Package of Emergency Care (continued) Obstetrics and Gynecology Referral and Primary health center First-level hospital specialized hospitals Initiate antenatal steroids IM magnesium sulphate (loading Magnesium sulphate infusiona dose) Oxytocin for IV infusion Second-line agent for PPH Anti-D immunoglobulin Oral agents for management of ectopic pregnancy, emergency contraception Vaccines Referral and Primary health center First-level hospital specialized hospitals Tetanus vaccine Rabies vaccine Antirabies immunoglobulina Antimicrobials Referral and Primary health center First-level hospital specialized hospitals Oral antibiotics for lung, skin, IV antibiotics (for lung, skin, GI, GI, or GU source (including GU, or CNS source; PPROM) syndromic STI treatment); PPROM) Topical antifungals Oral antifungal IV antifungal Oral antimalarials IV antimalarials Oral antihelminthics Oral antiviral (acyclovir or IV antiviral (acyclovir or equivalent) equivalent) Opthalmic topical antibacterial Topical antidermatoparasitic agent Other Referral and Primary health center First-level hospital specialized hospitals Oral paracetamol Oral antiemetic IV antiemetic Oral zinc Topical agents for burn dressing table continues next page 260 Disease Control Priorities: Improving Health and Reducing Poverty Table 13.2 Essential Package of Emergency Care (continued) Other Referral and Primary health center First-level hospital specialized hospitals Topical steroid Mannitol Agents for acute glaucoma (IV acetazolamide, opthalmic topical steroid, opthalmic topical beta-blocker) Surface and skin disinfectants Note: All resources mapped to lower levels are expected to be available at higher levels. ABCDEs = airway, breathing, circulation, disability, exposure; ALS = advanced life support; BEMNOC = basic emergency newborn and obstetric care; BLS = basic life support; BVM = bag valve mask; CCB = calcium channel blocker; CNS = central nervous system; CT = computed tomography scan; ECS = emergency care system; ECSA = emergency care system assessment; FFP = fresh frozen plasma; GI = gastrointestinal; GU = genitourinary infection; HIV = human immunodeficiency virus; IM = intramuscular; IV = intravenous; NGT = nasogastric tube; PPH = postpartum hemorrhage; PPROM = preterm premature rupture of the membranes; SL = sublingual; STEMI = ST-elevated myocardial infarction; STI = sexually transmitted infections; WHO = World Health Organization. a. In many regions, antivenin will be kept centrally by public health authorities. Ensure timely availability to first-level hospitals. Table 13.3 Key Diagnoses Associated with Critical Syndromes Difficulty breathing Shock Altered mental status Airway injury and inflammation Sepsis Coma Foreign body Gastroenteritis and diarrhea Delirium Pneumohemothorax Bradycardia Hypo- and hyperglycemia Pneumonia Hemorrhage Hypoxia Pleural effusion Cardiac valvular disease Hypo- and hyperthermia Asthma Abnormal cardiac rhythm or cardiac failure Electrolyte or thyroid abnormality Chronic obstructive pulmonary disease Gastrointestinal bleeding Liver disease Anemia Tension pneumothorax Kidney disease Myocardial ischemia Anaphylaxis Poisoning and envenomation Cardiac failure Spinal cord injury Psychosis Pericardial effusion Seizure Pulmonary embolism Stroke Drug overdose Tumor Chest wall injury Traumatic brain injury Paralysis Central nervous system infections, including HIV-related Note: HIV = human immunodeficiency virus. package for emergency care highlights interventions intravenous therapies are only included at the first-level that should be considered part of universal health cov- hospital and above, and therapies dependent on a diag- erage (Jamison and others 2013). nosis of cardiac rhythm are included only at the second- The following general assumptions were used as a or third-level hospitals. Practice conditions will vary guide in assigning components to levels of the system. It among countries and regions, and so this constitutes a is not assumed that primary health centers have the minimum package. Countries and regions with greater capacity to deliver intravenous infusions or that capacity at lower levels of the health system may want to emergency units in first-level hospitals have electrocar- map package components from higher levels to lower- diogram and cardiac monitoring available. Hence, level facilities. Strengthening Health Systems to Provide Emergency Care 261 COST-EFFECTIVENESS OF EMERGENCY CARE averted primarily through the institution of emer- SERVICES IN LMICS gency obstetric services (McCord and Chowdhury 2003). Many examples of individual emergency care services are highly cost-effective in LMICs, including the following:3 PRIORITIES FOR ACTION • Dedicated emergency unit with formal triage. The The following key priorities for policy makers and creation of the dedicated emergency unit in Malawi planners were derived from the WHO ECS frame- described earlier (associated with halved inpatient work (annex 13A) and represent key policy compo- mortality and a reduction in the proportion of nents to support delivery of the essential package early deaths from 36 to 12.6 percent) had a cost of care: of US$1.95 per patient (Molyneux, Ahmad, and Robertson 2006). • Ensure that the national ministry of health has a • Oxygen for pneumonia. In Papua New Guinea, intro- directorate dedicated to emergency care (not limited duction of an improved oxygen system (oxygen con- to disaster response). centrators and pulse oximeters) decreased mortality • Conduct a standardized national assessment of the risk for children with pneumonia by 35 percent. The ECS (using the WHO ECSA or a similar tool) to iden- estimated cost of this system was US$118 per patient tify gaps and inform system development. treated, US$3,868 per life saved, and US$116 per • Ensure that emergency care is explicitly incorporated DALY averted (Duke and others 2008). into the national health plan. • Pulse oximetry for childhood pneumonia. The mod- • Establish national legislation ensuring access to emer- eling study that described the impact of imple- gency care without regard to ability to pay. menting pulse oximetry combined with WHO • Ensure that hospitals at all levels include dedicated guidelines showed that the intervention was emergency units—areas dedicated to the provision extremely cost-effective, with estimates ranging of emergency care and staffed with at least a core of from US$3.26 to US$72.01 per DALY averted, in nonrotating personnel who are specifically trained in the 15 countries across Africa and Asia with the the care of emergency conditions. highest global burden of childhood pneumonia • Disseminate dedicated training for emergency care (Floyd and others 2015). across cadres, including training in basic emergency • Treatment of acute myocardial infarction. In India, the care for all prehospital providers, basic emergency incremental cost of treating, with either aspirin to a care training for all cadres of facility-based provid- 95 percent coverage level or aspirin plus streptokinase ers who treat patients with emergency conditions, to an 80 percent coverage level, treatment-eligible dedicated emergency care training integrated into patients with acute myocardial infarction who were undergraduate medical and nursing curricula, and not yet being treated was US$0.56 and US$701 per residency or specialist training programs in emer- DALY averted, respectively (Megiddo and others gency medicine. 2014). Early electrocardiogram diagnosis to facilitate • Establish acuity-based triage systems at all facilities triage and referral was shown to be cost-effective in that regularly receive acutely ill and injured patients. India at US$17 per quality-adjusted life year (QALY) • Establish prehospital care systems based on WHO or (Gaziano and others 2017). other international standards, including a dedicated • Emergency obstetric care. In rural India, an emergency certification pathway for prehospital care providers obstetric hospital provided services at an estimated and a toll-free, universal access number for emer- cost of US$0.43 per capita per year for the commu- gency care. nity (or US$1.50 per woman of childbearing age) • Develop critical process and clinical protocols as (McCord and others 2001). identified in the WHO ECS framework (including • Trauma surgery. In a Cambodian hospital dealing transport and referral protocols, prehospital and almost exclusively with injury (about 90 percent of facility-based clinical treatment protocols, and disas- cases), surgical interventions (though not all were ter and mass casualty protocols. emergency surgeries) cost approximately US$133 per • Implement standardized clinical charts and reg- DALY averted (Gosselin and Heitto 2008). istries incorporating essential data points, such as • Emergency obstetric services. A small hospital in those based on WHO standards, to facilitate quality rural Bangladesh demonstrated substantial DALYs improvement efforts. 262 Disease Control Priorities: Improving Health and Reducing Poverty CONCLUSIONS 2. For more information on the SDGs, see http://www.un.org /sustainabledevelopment/sustainable-development-goals/. Conditions that can be addressed by emergency care 3. All costs are adjusted to 2012 US$. (such as injuries, infections, obstetrical complica- tions, stroke, myocardial infarction, and respiratory failure) account for a substantial health burden. REFERENCES The interventions needed to address these conditions Alkema, L., D. Chou, D. Hogan, S. Zhang, A. B. Moller, and are very cost-effective, even in limited-resource set- others. 2016. “Global, Regional, and National Levels and tings, but critical gaps remain in emergency care gov- Trends in Maternal Mortality between 1990 and 2015, with ernance and delivery in LMICs. Improving the Scenario-Based Projections to 2030: A Systematic Analysis organization of and planning for emergency care by the UN Maternal Mortality Estimation Inter-Agency services substantially improves the outcomes of Group.” The Lancet 387 (10017): 462–74. patients with emergency conditions. Most of the evi- Anderson, C. S., E. Heeley, Y. Huang, J. Wang, C. Stapf, and oth- dence for such improvements comes from HICs, but ers. 2013. “Rapid Blood-Pressure Lowering in Patients with there is growing evidence that such improvements Acute Intracerebral Hemorrhage.” New England Journal of can also be made—affordably, sustainably, and with Medicine 368 (25): 2355–65. Dare, A. J., J. S. Ng-Kamstra, J. Patra, S.H. Fu, P. S. Rodriguez, dramatic impact—in LMICs. The WHO ECS frame- and others. 2015. “Deaths from Acute Abdominal work and the complementary essential package of Conditions and Geographical Access to Surgical Care in emergency care services represent a mechanism by India: A Nationally Representative Spatial Analysis.” The which emergency care can be scaled up globally, accel- Lancet Global Health 3 (10): e646–e653. erating progress toward universal health coverage and Duke, T., F. Wandi, M. Jonathan, S. Matai, M. Kaupa, and a range of other SDG targets. others. 2008. “Improved Oxygen Systems for Childhood Pneumonia: A Multihospital Effectiveness Study in Papua New Guinea.” The Lancet 372 (9646): 1328–33. ACKNOWLEDGMENT Floyd, J., L. Wu, D. Hay Burgess, R. Izadnegahdar, D. Mukanga, and others. 2015. “Evaluating the Impact of Pulse Oximetry The authors acknowledge with thanks the contributions on Childhood Pneumonia Mortality in Resource-Poor of Stas Salerno Amato, Morgan Broccoli, and Jennifer Settings.” Nature 528 (7580): S53–S59. Nash to the tables and figures in the chapter. Fournier, P., A. Dumont, C. Tourigny, G. Dunkley, and S. Dramé. 2009. “Improved Access to Comprehensive Emergency Obstetric Care and Its Effect on Institutional ANNEX Maternal Mortality in Rural Mali.” Bulletin of the World Health Organization 87 (1): 30–38. The annex to this chapter is as follows. It is available at Gaieski, D. F., M. E. Mikkelsen, R. A. Band, J. M. Pines, http://www.dcp-3.org/DCP. R. Massone, and others. 2010. “Impact of Time to Antibiotics on Survival in Patients with Severe Sepsis or • Annex 13A: WHO Emergency Care Systems Septic Shock in Whom Early Goal-Directed Therapy Was Framework. Initiated in the Emergency Department.” Critical Care Medicine 38 (4): 1045–53. Gaziano, T., M. Suhrcke, E. Brouwer, C. Levin, I. Nikolic, and NOTES others. 2017. “Costs and Cost-Effectiveness of Interventions and Policies to Prevent and Treat Cardiovascular and World Bank Income Classifications as of July 2014 are as fol- Respiratory Diseases.” In Disease Control Priorities (third lows, based on estimates of gross national income (GNI) per edition): Volume 5, Cardiovascular, Respiratory, and Related capita for 2013: Disorders, edited by D. Prabhakaran, S. Anand, T. Gaziano, J.-C. Mbanya, Y. Wu, and R. Nugent. Washington, DC: • Low-income countries (LICs) = US$1,045 or less World Bank. • Middle-income countries (MICs) are subdivided: Gosselin, R. A., and M. Heitto. 2008. “Cost-Effectiveness of (a) lower-middle-income = US$1,046 to US$4,125 a District Trauma Hospital in Battambang, Cambodia.” (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 World Journal of Surgery 32 (11): 2450–53. • High-income countries (HICs) = US$12,746 or more. Grimes, C. E., K. G. Bowman, C. M. Dodgion, and C. B. Lavy. 2011. “Systematic Review of Barriers to Surgical Care 1. R. Arafat, Ministry of the Interior, Department for in Low-Income and Middle-Income Countries.” World Emergency Situations, Government of Romania, personal Journal of Surgery 35 (5): 941–50. communication with the author based on internal facility Henry, J., and A. Reingold. 2012. “Prehospital Trauma Systems data, March 2016. Reduce Mortality in Developing Countries: A Systematic Strengthening Health Systems to Provide Emergency Care 263 Review and Meta-Analysis.” Journal of Trauma Acute Care Mbugua, P. K., C. F. Otieno, J. K. Kayima, A. A. Amayo, and Surgery 73 (1): 261–68. S. O. McLigeyo. 2005. “Diabetic Ketoacidosis: Clinical Higashi, H., J. J. Barendregt, N. J. Kassebaum, T. G. Weiser, Presentation and Precipitating Factors at Kenyatta National S. W. Bickler, and others. 2015. “Burden of Injuries Avertable Hospital, Nairobi.” East African Medical Journal 82 (Suppl by a Basic Surgical Package in Low- and Middle-Income 12): S191–S196. Regions: A Systematic Analysis from the Global Burden of McCord, C., and Q. Chowdhury. 2003. “A Cost-Effective Small Disease 2010 Study.” World Journal of Surgery 39 (1): 1–9. Hospital in Bangladesh: What It Can Mean for Emergency Hirshon, J., N. Risko, E. J. Calvello, S. Stewart de Ramirez, Obstetric Care.” International Journal of Gynecology and M. Narayan, and others. 2013. “Health Systems and Obstetrics 81 (1): 83–92. Services: The Role of Acute Care.” Bulletin of the World McCord, C., R. Premkumar, S. Arole, and R. Arole. 2001. Health Organization 91 (5): 386–88. “Efficient and Effective Emergency Obstetric Care in Rural Holmer, H., K. Oyerinde, J. G. Meara, R. Gillies, J. Liljestrand, Indian Community Where Most Deliveries Are at Home.” and others. 2015. “The Global Met Need for Emergency International Journal of Gynaecology and Obstetrics 75 (3): Obstetric Care: A Systematic Review.” British Journal of 297–307. Obstetrics and Gynaecology 122 (2): 183–89. Megiddo, I., S. Chatterjee, A. Nandi, and R. Laxminarayan. Hortmann, M., H. J. Heppner, S. Popp, T. Lad, and M. Christ. 2014. “Cost-Effectiveness of Treatment and Secondary 2014. “Reduction of Mortality in Community-Acquired Prevention of Acute Myocardial Infarction in India: Pneumonia after Implementing Standardized Care Bundles A Modeling Study.” Global Heart 9 (4): 391–98. in the Emergency Department.” European Journal of Miller, S., F. Lester, and P. Hensleigh. 2004. “Prevention and Emergency Medicine 21 (6): 429–35. Treatment of Postpartum Hemorrhage: New Advances for Hsiao, M., S. K. Morris, A. Malhotra, W. Suraweera, and P. Jha. Low-Resource Settings.” Journal of Midwifery and Women’s 2013. “Time-Critical Mortality Conditions in Low-Income and Health 49 (4): 283–92. Middle-Income Countries.” The Lancet 381 (9871): 998–94. Mock, C., M. Joshipura, C. Arreola-Risa, and R. Quansah. 2012. Husum, H., M. Gilbert, T. Wisborg, Y. Van Heng, and M. Murad. “An Estimate of the Number of Lives That Could Be Saved 2003. “Rural Prehospital Trauma Systems Improve Trauma through Improvements in Trauma Care Globally.” World Outcome in Low-Income Countries: A Prospective Study Journal of Surgery 36 (5): 959–63. from North Iraq and Cambodia.” Journal of Trauma 54 (6): Molyneux, E., S. Ahmad, and A. Robertson. 2006. “Improved 1188–96. Triage and Emergency Care for Children Reduces Inpatient Irfan, F. B., B. B. Irfan, and D. A. Spiegel. 2012. “Barriers to Mortality in a Resource-Constrained Setting.” Bulletin of Accessing Surgical Care in Pakistan: Healthcare Barrier the World Health Organization 84 (4): 314–19. Model and Quantitative Systematic Review.” Journal of Murad, M. K., D. B. Issa, F. M. Mustafa, H. O. Hassan, and Surgical Research 176 (1): 84–94. H. Husum. 2012. “Prehospital Trauma System Reduces Jacob, S. T., P. Banura, J. M. Baeten, C. C. Moore, D. Meya, and Mortality in Severe Trauma: A Controlled Study of Road others. 2012. “The Impact of Early Monitored Management Traffic Casualties in Iraq.” Prehospital Disaster Medicine 27 on Survival in Hospitalized Adult Ugandan Patients with (1): 36–41. Severe Sepsis: A Prospective Intervention Study.” Critical Nyenwe, E. A., and A. E. Kitabchi. 2011. “Evidence-Based Care Medicine 40 (7): 2050–58. Management of Hyperglycemic Emergencies in Diabetes Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, Mellitus.” Diabetes Research and Clinical Practice 94 (3): S. Berkley, and others. 2013. “Global Health 2035: A World 340–51. Converging within a Generation.” The Lancet 382 (9908): Paxton, A., D. Maine, L. Freedman, D. Fry, and S. Lobis. 2005. 1898–955. “The Evidence for Emergency Obstetric Care.” International Kausar, F., J. L. Morris, M. Fathalla, O. Ojengbede, A. Fabamwo, Journal of Gynecology and Obstetrics 88 (2): 181–93. and others. 2012. “Nurses in Low Resource Settings Save Petroze, R. T., J. C. Byiringiro, G. Ntakiyiruta, S. M. Briggs, Mothers’ Lives with Non-Pneumatic Anti-Shock Garment.” D. L. Deckelbaum, and others. 2015. “Can Focused Trauma Maternal Child Nursing: The American Journal of Maternal/ Education Initiatives Reduce Mortality or Improve Child Nursing 37 (5): 308–16. Resource Utilization in a Low-Resource Setting?” World Kesinger, M. R., J. C. Puyana, and A. M. Rubiano. 2014. Journal of Surgery 39 (4): 926–33. “Improving Trauma Care in Low- and Middle-Income Rivers, E., B. Nguyen, S. Havstad, J. Ressler, A. Muzzin, Countries by Implementing a Standardized Trauma and others. 2001. “Early Goal-Directed Therapy in the Protocol.” World Journal of Surgery 38 (8): 1869–74. Treatment of Severe Sepsis and Septic Shock,” New England Kobusingye, O., A. A. Hyder, D. Bishai, M. Joshipura, E. R. Hicks, Journal of Medicine 8, no. 345 (19): 1368–77. and others. 2006. “Emergency Medical Services.” In Disease Sandercock, P. A., C. Counsell, M. C. Tseng, and E. Cecconi. Control Priorities in Developing Countries, second edition, 2014. “Oral Antiplatelet Therapy for Acute Ischaemic edited by D. T. Jamison, J. G. Breman, A. R. Measham, Stroke.” Cochrane Database of Systematic Reviews 26 (3): G. Alleyne, M. Claeson, D. Evans, P. Jha, A. Mills, and CD000029.pub3. P. Musgrove, 1261–80. Washington, DC: World Bank and Say, L., D. Chou, A. Gemmill, Ö. Tunçalp, A. B. Moller, and Oxford University Press. others. 2014. “Global Causes of Maternal Death: A WHO 264 Disease Control Priorities: Improving Health and Reducing Poverty Systematic Analysis.” The Lancet Global Health 2 (6): Primary Percutaneous Coronary Intervention.” Journal of e323–e333. the American Medical Association 304 (7): 763–71. Sharma, S. K., P. Bovier, N. Jha, E. Alirol, L. Loutan, and others. Thind, A., R. Hsia, J. Mabweijano, E. Romero Hicks, A. Zakariah, 2013. “Effectiveness of Rapid Transport of Victims and and others. 2015. “Prehospital and Emergency Care.” In Community Health Education on Snake Bite Fatalities in Disease Control Priorities (third edition): Volume 1, Essential Rural Nepal.” American Journal of Tropical Medicine and Surgery, edited by H. T. Debas, P. Donkor, A. Gawande, D. Hygiene 89 (1): 145–50. T. Jamison, M. Kruk, and C. N. Mock. Washington, DC: Silva, E, P. M. de Almeida, A. C. B. Sogayar, T. Mohovic, C. L. Silva, World Bank. and others. 2004. “Brazilian Sepsis Epidemiological Study Vital, F. M. R., M. T. Ladeira, and A. N. Atallah. 2013. “Non- (BASES Study).” Critical Care Medicine 8 (4): R251–R261. Invasive Positive Pressure Ventilation (CPAP or Bilevel Siman-Tov, M., I. Radomislensky, and K. Peleg. 2013. NPPV) for Cardiogenic Pulmonary Oedema.” Cochrane “Reduction in Trauma Mortality in Israel during the Database of Systematic Reviews 5: CD005351. Last Decade (2000–2010): The Impact of Changes in the Wang, P., N. P. Li, Y. F. Gu, X. B. Lu, J. N. Cong, and others. 2010. Trauma System.” Injury 44 (11): 1448–52. “Comparison of Severe Trauma Care Effect before and after Stevenson, E. K., A. R. Rubenstein, G. T. Radin, R. S. Wiener, Advanced Trauma Life Support Training.” Chinese Journal and A. J. Walkey. 2014. “Two Decades of Mortality Trends of Traumatology 13 (6): 341–44. among Patients with Severe Sepsis: A Comparative Meta- WHO (World Health Organization). 2010. Monitoring the Analysis.” Critical Care Medicine 42 (3): 625–31. Building Blocks of Health Systems: A Handbook of Indicators Tallon, J. M., D. B. Fell, S. A. Karim, S. Ackroyd-Stolarz, and and Their Measurement Strategies. WHO: Geneva. http:// D. Petrie. 2012. “Influence of a Province-Wide Trauma www.who.int/healthinfo/systems/WHO_MBHSS_2010 System on Motor Vehicle Collision Process of Trauma Care _full_web.pdf. and Mortality: A 10-Year Follow-Up Evaluation.” Canadian ———. 2013. “Global Health Estimates for Deaths by Cause, Journal of Surgery 55 (1): 8. Age, and Sex for Years 2000–2011.” WHO, Geneva. http:// Tanriover, M. D., G. S. Guven, D. Sen, S. Unal, and O. Uzun. www.who.int/healthinfo/global_burden_disease/en/. 2006. “Epidemiology and Outcome of Sepsis in a Tertiary- ———. n.d. “Emergency and Trauma Care,” WHO, Geneva. Care Hospital in a Developing Country.” Epidemiology and http://www.who.int/emergencycare/activities/en. Infection 134 (2): 315–22. Xavier, D., P. Pais, P. J. Devereaux, C. Xie, D. Prabhakaran, and Terkelsen, C. J., J. T. Sørensen, M. Maeng, L. O. Jensen, others. 2008. “Treatment and Outcomes of Acute Coronary H. H. Tilsted, and others. 2010. “System Delay and Syndromes in India (CREATE): A Prospective Analysis of Mortality among Patients with STEMI Treated with Registry Data.” The Lancet 371 (9622): 1435–42. Strengthening Health Systems to Provide Emergency Care 265 Chapter 14 Community Platforms for Public Health Interventions Melissa Sherry, Abdul Ghaffar, and David Bishai INTRODUCTION Then the cycle repeats. A community that has the ability Community health platforms are the partnerships to engage successfully in the cycle shown in figure 14.1 formed to assess and ensure public health. They provide has a platform that can support all types of community the context in which outside interventions should be health initiatives. implemented and sustained, and they offer a way to The provision of legal authority for community develop and maintain community-centered solutions. health platforms can be traced to England’s first health Although local boards of health and health departments law, the Public Health Act of 1848, which gave cities the are the official bodies with the mandate to sustain strong option to create local health boards (Rosen 1958; community health platforms, they do not always achieve Szreter 1988). In the mid-nineteenth century, func- their full potential (Bellagio District Public Health tional health departments were established throughout Workshop Participants 2016). In the absence of an effec- Canada, Europe, and the United States before the tive government presence, nongovernmental organiza- development of effective medical care and drove the tions (NGOs) can build community health platforms. dramatic decline in mortality in the twentieth century Well-functioning community health platforms can (McKeown, Record, and Turner 1975). However, west- serve as vehicles for health information and advocacy ern governments had largely omitted the creation of and can convene local resources to support successful functioning local health departments when they public health interventions. Well-designed and well- formed colonies in the Americas, Africa, and Asia; implemented community health platforms can function countries that gained independence in the mid-1900s as the engine in the public health cycle of convening faced an urgent need to catch up. By the late 20th cen- communities to monitor, review, and act (figure 14.1). tury, the growing recognition that public health and These are functional tasks that are best conducted in a primary care were lagging became the topic of interna- partnership among public health professionals, politi- tional concern. In 1978, an International Conference cians, and community members. Effective partnerships on Primary Health Care in Alma-Ata, USSR, attended among these parties ensure that health data are collected by nearly all member nations of the World Health to answer questions posed by the community, that local Organization and the United Nations Children’s Fund, health data are shared with the community to guide demonstrated the degree of concern about access to actions, and that actions marshal all of a community’s primary health care (Lawn and others 2008). It resulted human and capital resources as well as public revenue. in the Declaration of Alma-Ata. Corresponding author: Melissa Sherry, Johns Hopkins Bloomberg School of Public Health, Baltimore, Maryland, United States; msherry4@jhu.edu. 267 Figure 14.1 Public Health Cycle: Monitor, Review, and Act Alma-Ata (Lawn and others 2008; Rohde and others 2008). The Cold War fostered a culture of development planning that emphasized interventions that were rapidly deployed and easily measured. Health commodities, such as vaccines, oral rehydration solutions, micronutrients, Monitor contraceptives, and antibiotics, became the focus of health Act population care systems (Lawn and others 2008; Perry 2013). The health emphasis of global health donors on results and short Community + project cycles made the focus on commodities rather than Multisector systems more expedient. stakeholders The urgency of saving lives in the moment and the truth that the commodities really did save lives perpetu- ated a stronger emphasis on delivery of medical services and health care goods and a lighter emphasis on communities’ development of Alma-Ata—style plat- Review data, prioritize actions forms. The term vertical was used to define projects focused on getting a selected health commodity or ser- vice to households in the most expedient way, typically using a stand-alone organization of staff, vehicles, and Chapter Overview capital. The term horizontal was used to define initiatives The Declaration of Alma-Ata asserted that health is a to build more comprehensive institutions of primary fundamental human right and that community consul- care services and for population-level public health. A tation and participation in health care are essential short-term focus on vertical programs delivering good elements of successful programs (Lawn and others 2008; health at low cost crowded out attention to building Rohde and others 2008). Following the declaration, long-term horizontal platforms. The World Development global health indicators improved despite inadequate Report 1993: Investing in Health (World Bank 1993) adherence to the principles laid out in the declaration. offered an excellent listing of population-level public The recent transition from the Millennium Development health interventions that could be implemented, but it Goals to the Sustainable Development Goals of the neglected any discussion of how to make them happen, United Nations has renewed attention to strategies that other than by raising money. This report was novel in build on local capacity to strengthen community health that it demonstrated for the first time that international platforms (Open Working Group of the General health investments could be justified on the basis of hav- Assembly 2014). ing measureable outcomes and effects. This chapter presents a brief review of how the public Volume 1 of the first edition of Disease Control health cycle supports the sustained success of any of the Priorities in Developing Countries (DCP1) also offered a interventions discussed in the Disease Control Priorities comprehensive list of public health policies, with recom- volumes. It offers a typology of the stages of develop- mendations for developing and financing state capacity in ment of community health platforms, as well as a frame- data collection and data analysis (Mosley, Bobadilla, and work for assessing their success. We illustrate four stages Jamison 1993). The authors shared aspirations for better of development of community health platforms with policy environments that would be conducive to struc- four case studies that range from a most developed case tural approaches to public health. Volume 2 of the second in Indonesia to a primitive case of near-paralysis of the edition of Disease Control Priorities in Developing state’s efforts in public health. The chapter closes with Countries (DCP2) explicitly recognized the need for a discussion of investment opportunities for policy community-driven global health efforts to strengthen makers who are interested in strengthening community health systems and infrastructure and suggested the health platforms. need to strengthen platforms that would allow communi- ties to hold health systems accountable for improved quality and access to services (Mills, Rasheed, and Tollman Background and Historical Context 2006). DCP2 also emphasized that a lack of intersectoral The lack of a clear roadmap to implement community action through cross-sector partnerships and the failure involvement, combined with changes in the global econ- of health systems to address community-level barriers to omy, slowed the progress of low- and middle-income accessing the health system were key constraints for health countries in achieving the primary health care goals set by system strengthening (Mills, Rasheed, and Tollman 2006). 268 Disease Control Priorities: Improving Health and Reducing Poverty However, DCP1, DCP2, and the World Development (Bishai and others 2016). For example, the Pan American Report 1993 did not offer specific recommendations Health Organization’s (PAHO) list of Essential Public about how to create conducive policy environments that Health Functions (EPHFs) is as follows (PAHO 2001): could enable and sustain public health interventions, cross-sectoral partnerships, and community engagement 1. Monitor health status to identify community health with local health departments (Macinko, Starfield, and problems. Erinosho 2009; Mosley, Bobadilla, and Jamison 1993; 2. Diagnose and investigate health problems and Rohde and others 2008). health hazards in the community. The lack of a roadmap for creating community health 3. Inform, educate, and empower people about health platforms and cross-sectoral action made room for ver- issues. tical programming to dominate the policy landscape 4. Mobilize community partnerships to identify and (Lawn and others 2008; Macinko, Starfield, and Erinosho solve health problems. 2009; Rohde and others 2008). These vertical programs 5. Develop policies and plans that support individual saved lives, but they left populations vulnerable by failing and community health efforts. to create resilient systems in situ that would marshal 6. Enforce laws and regulations that protect health and local political will and local resources to address the root ensure safety. causes of poor population health. 7. Link people to needed personal health services and Actions that improve public health are often met assure the provision of health care when otherwise with resistance about who will pay for them, because unavailable. results are often less tangible and urgent than medical 8. Assure a competent public and personal health care interventions. Further, public health actions often workforce. threaten the livelihoods of industries and occupations 9. Evaluate effectiveness, accessibility, and quality of whose harmful aspects are regulated. Resistance is to personal and population-based health services. be expected. Examples of public health actions range 10. Research for new insights and innovative solutions from the need to pay for sewers and waterworks to the to health problems. need to enact and enforce restrictions on tobacco, food 11. Engage in disaster preparedness to reduce the impact labeling, and road safety. Solving these problems is of emergencies and disasters on health. fundamental to public health. Solutions are often political, and vertical approaches are only partial PAHO’s 11 items fall into the same basic cycle of monitor, responses. review, and act shown in figure 14.1. EPHFs 1 and 2 are The inability to sustain a local consensus and to for monitoring; EPHFs 3–5 are for reviewing, typically mobilize community buy-in regarding the health risks through participatory multistakeholder community leads to difficulty in imposing the measures needed to engagement; and EPHFs 6–11 are for acting. The best control health threats. Poorly performing public health community health platforms successfully make their pop- departments are part of the reason that HIV/AIDS ulations healthy by understanding what constitutes health (human immunodeficiency virus/acquired immune threats and by sharing this information with community deficiency syndrome) and the Ebola virus arose and members from multiple sectors. Community health plat- overwhelmed many health systems. forms mobilize parts of a coherent solution using the strengths and resources present in the community. Essential Public Health Functions To improve public health functioning, between 1989 Health Care and Health Facilities and 1994, groups at the Centers for Disease Control and The care of the sick and the delivery of health commodities Prevention and the U.S. Public Health Service developed are integral parts of public health practice and are parts of a list of 10 essential public health functions to bench- the work plan of community health platforms. Community mark the quality of practice in public health agencies health workers can play multiple roles in generating health (Dyal 1995). The consensus was that country health data (PAHO 2001, EPHFs 1–2), informing and mobilizing ministries and regional offices needed to define national- communities (PAHO 2001, EPHFs 3–5), and helping to level lists of functions and items deemed essential and provide primary care services (PAHO 2001, EPHFs 7–9). that the lists should be country specific (Bettcher, Many of the interventions discussed in Volume 4 of the Sapirie, and Goon 1998). Countries and regions have third edition of Disease Control Priorities (DCP3) rely on adapted their own priority lists of essential public facilities and community health workers (Patel and others health functions on the basis of local stakeholder input 2015). When community health platforms fulfill their Community Platforms for Public Health Interventions 269 mandate to provide essential public health functions like concept can be applied. Types of platforms described in those mentioned earlier, interventions based in facilities published and gray literature generally fall into the fol- and involving community health workers become inte- lowing categories: grated and sustained by local support and action. • Health committees • Community health worker interventions MEASURING SUCCESS IN COMMUNITY • Community-based participatory research and health HEALTH PLATFORMS scorecards The literature shows that community health platforms • NGOs or academic community partnerships for specific that enable participation and engagement lead to community interventions (Beracochea 2015; Draper, improved health outcomes (Edmunds and Albritton Hewitt, and Rifkin 2010; George and others 2015; Kenny 2015; George and others 2015; Kenny and others 2013; and others 2013; Marmot and others 2008; Meier, Pardue, McCoy, Hall, and Ridge 2012; O’Mara-Eves and others and London 2012; Rifkin 1996; Tiwari, Lommerse, and 2015; Rifkin 1996, 2014). Measuring health outcomes Smith 2014; UK Aid and DFID/HDRC 2011). associated with community participation can be diffi- cult, but community participation in public health The literature also covers concepts of community generally leads to improvements in health knowledge, engagement, participation, and mobilization as they service quality, and health-related outcomes (Kenny and relate to multiple types of community platforms (Cyril others 2013; Russell and others 2008). and others 2015; Draper, Hewitt, and Rifkin 2010; The degree to which a community health platform is Frumence and others 2014; Meier, Pardue, and London high functioning lies along a continuum. At one end is 2012; Rifkin 1996, 2014; Rosato and others 2008; Russell development that extends from mere delivery of services. and others 2008; UK Aid and DFID/HDRC 2011). At the other end is facilitation of an active community The likelihood that community engagement will result through an engagement platform whereby communities in improved health outcomes depends on many factors. are informed and enabled to take shared responsibility for Cyril and others (2015) identified the following compo- addressing their changing health risks and concerns nents of success: engaging in real power sharing, building (Beracochea 2015; Cyril and others 2015; Dooris and collaborative partnerships, providing bidirectional learn- Heritage 2013; Draper, Hewitt, and Rifkin 2010; George ing, incorporating the voice and agency of beneficiary and others 2015; McCoy, Hall, and Ridge 2012; Raeburn communities in research protocol, and using multicultural and others 2006; Rosato and others 2008; Russell and health care workers for intervention delivery. Draper, others 2008). Hewitt, and Rifkin (2010) suggested a continuum of The breadth of the literature on community health process measures for use in evaluating community platforms demonstrates the range of ways that the participation in a health system context (table 14.1). Table 14.1 Example of Process Indicators for Participation Continuum of community participation Indicators of participation Values for mobilization Values for collaboration Values for empowering Leadership: Professionals Health professionals assume Collaborative decision making occurs between Program is led by community members introducing interventions, leadership. health professionals and community leaders. who are selected through a representative or by community of Local leadership does not Local leadership tries to present the interests process. intended beneficiaries necessarily try to widen the of different groups. Health professionals give leadership decision-making base in the training, if necessary. community. Planning and Health professionals tell Health professionals initiate collaboration. Partnerships between communities and management: The way the community how it may Communities are invited to participate within health professionals are created and partnerships between participate. They decide the a predetermined remit. institutionalized. leadership and the program’s focus, goals, and Professionals facilitate, and the community community are forged activities and provide the Activities reflect community priorities and involve local people and existing community defines priorities and manages the program. necessary resources. organizations. Local people learn skills they need for management and evaluation. table continues next page 270 Disease Control Priorities: Improving Health and Reducing Poverty Table 14.1 Example of Process Indicators for Participation (continued) Continuum of community participation Indicators of participation Values for mobilization Values for collaboration Values for empowering Women’s involvement The inclusion of women is not Women actively participate in some aspects The active participation of women specifically sought outside their of the program, but they have minor decision- in positions of decision making and traditional roles. making roles. responsibility is a program objective. External support for Funding comes from outside the The majority of funding comes from outside Community members work to find ways program development: community and is controlled by the community, but local people are asked to of mobilizing resources, including through In terms of finance and health professionals. contribute time, money, and materials. external funding and their own resources program design Program components Health professionals allocate resources, (for example, microfinancing). are designed by health although they may consult community The program is designed by community professionals. members. members with technical advice from health The program is designed by health professionals on request. The design is professionals in discussion with community flexible and incorporates wide community representatives. participation, including that of women and minority groups. Each role in the program, including those for women and minority groups, is negotiated. Monitoring and Health professionals design Health professionals design mixed method Communities do a participatory evaluation evaluation: monitoring and evaluation monitoring and evaluation protocols and that produces locally meaningful findings. The way intended protocols, choose outcomes, and perform analyses, but community members are A variety of data collection methods is used, beneficiaries are analyze data in ways to suit their involved in data collection. A broad definition and the community chooses the indicators involved in these information needs. of “success” is used. for success. activities The approach is mainly one of Responses to monitoring findings are jointly Health professionals assist at the request of hypothesis testing and statistical decided, and community feedback is both the community. analyses of health-related sought and given. outcomes. Communities are actively involved in participatory monitoring and decide how to Communities might not be made respond to monitoring findings. aware of the findings. Communities contribute to wider external evaluations. Score given 1–2 3–4 5 Source: Draper, Hewitt, and Rifkin 2010. Note: Scores range from a low of 1 (lowest level of community participation) to 5 (highest level of community participation). Figure 14.2 From Passive to Active Community Participation through in their health systems as they improve in their ability to empower communities to take on Increasing empowerment health challenges. Using themes that emerged from Information-sharing Collaboration Full responsibility the literature, we identified broad domains of func- consultation tion in the development of community health Source: Rosato and others 2008. platforms: Figure 14.2 summarizes a process of increasing empower- • Level of community engagement: To what extent ment in the development of community participation. was the community empowered to engage with the health care system? • Health-system context and role of the government: INTERVENTIONS, POLICIES, AND Was the health system decentralized? Did local health EFFECTIVENESS departments have power to innovate and to work with communities? Was the government a support or Community Health Platform Case Studies a hindrance to community health platforms? We describe the continuum of developmental stages • Breadth of intersectoral partnerships: Was the com- that low- and middle-income countries move munity able to work with NGOs, community-based Community Platforms for Public Health Interventions 271 organizations, local governments, and other sectors offer informative examples of community health plat- in addition to the health sector? Did this ability forms that have been sustainable and high achieving predict the comprehensiveness of improvements? (Altobelli 2008; Blas, Sommerfeld, and Kurup 2011; Was the community able to influence action across Kowitt and others 2015; Rasanathan and others sectors? 2012; Siswanto 2009; Tanvatanakul and others 2007; • Sustainability: Was the community health platform’s Tiwari, Lommerse, and Smith 2014; Westphal and ability to be both scalable and sustainable a key factor others 2011). Table 14.2 shows a staged typology of in its success and longevity? This category includes community health platforms as countries move from the financing strategies and the ability to create lasting low-functioning platforms with little accountability change while reducing inefficiencies across the system. (level 1) to high-functioning platforms that promote Is the community health platform legally recognized? intersectoral action (level 4). • Leadership and platform structure that promotes integration across all partners: Who initiated the community’s involvement with the health system? Factors That Support Successful Community Did the platform create opportunities for shared Health Platforms vision, shared leadership and decision making, and Supportive factors that emerged from the case study shared financing across sectors? review and that contribute to sustainability include government participation, advocacy, cross-sectoral part- nerships, and community-owned vetting mechanisms. Identifying Case Studies Demonstrating Community Successful community health platforms were devel- Health Platform Development oped to fit in the political and cultural context of the Among the countries with recent rapid reductions local area they served, but they were strengthened by in mortality under age 5 years, Indonesia and Peru advocacy from NGOs or universities, which also Table 14.2 Continuum of Functioning, from High to Low, across Functional Domains of Community Health Platforms Level 4 Frontier of intersectoral collaboration where all sectors Level 3-> and community are involved in Level 2-> Sectorwide partnerships, creating health aspects in all Level 1-> Contractor and donor working to address burden of policies, intersectoral action, Poor functioning, not driven, uncoordinated disease, but unsuccessful in existence of a global budget, and Features accountable across sectors improving health outcomes successful health outcomes Community No platform exists for Limited community Community is engaged and able Community works closely with engagement community engagement or engagement is through to voice needs to government government sectors, NGOs, and other priority setting. select organizations or and other sectors. community organizations to ensure contractors working with needs are met. community for specific purposes. Role of Government is centralized. Contractors and donors guide Government participates in Government is decentralized, focuses government Health system is government decisions. cross-sectoral partnerships. on partnerships with community and fragmented and lacks Government does not work other sectors, has high accountability resources to support to integrate sectors or and transparency, has sufficient funding intersectoral action address community needs. to support the public health and for health. medical system, has legislation that enables public health and community No accountability exists. integration, and uses global budgeting. Partnerships No substantial Partnerships exist between Multiple partnerships exist Action across sectors is fully realized. across partnerships exist across sectors, but they are limited across sectors; integrating entity community sectors. to a few partners working brings together government together at a time, not sectors, community, NGOs, sectorwide. and others. table continues next page 272 Disease Control Priorities: Improving Health and Reducing Poverty Table 14.2 Continuum of Functioning, from High to Low, across Functional Domains of Community Health Platforms (continued) Level 4 Frontier of intersectoral collaboration where all sectors Level 3-> and community are involved in Level 2-> Sectorwide partnerships, creating health aspects in all Level 1-> Contractor and donor working to address burden of policies, intersectoral action, Poor functioning, not driven, uncoordinated disease, but unsuccessful in existence of a global budget, and Features accountable across sectors improving health outcomes successful health outcomes Sustainability Spending is wasteful, Sustainability is low, and Sustainability is moderate: Partnerships across sectors are used to and duplication of efforts platform is reliant on outside intersectoral action improves fill gaps in funding across government. occurs; platform is not assistance to maintain efficiency and reduces Improved social determinants result sustainable without health system. duplication of efforts. However, in improved health and less medical continuous donor funding. continued reliance on outside spending, with minimal reliance on funding remains necessary outside funding sources. Health outcomes Health MDGs are not met. Few MDGs are met. Improvements are achieved in Majority of MDGs are met, and social reaching MDGs, but substantial determinants are being addressed. improvements are still needed. Type of There is no integrator. Contractor or NGO integrates Government or community board Government brings together all sectors integrator with one or two other integrates with multiple sectors. in partnership to improve health. sectors at a time. Source: Authors. Note: MDGs = Millennium Development Goals; NGOs = nongovernmental organizations. provided technical support for emerging platforms. that involved communities in primary health care. In the Support from the government was essential for 1980s and 1990s, these posts offered limited services, and longer-term sustainability, but strong internal and quality and performance varied (Blas, Sommerfeld, and external advocates from nongovernment sectors Kurup 2011; Siswanto 2009). After Indonesia decentral- helped communities engage with governments and ized in 2001, district governments were empowered to health systems, which led to more formal structures. run district-level health systems. Successful community health platforms relied on The experience of Lumajang district in East Java is coordination across sectors to meet health goals, notable as an example of a health-in-all-policies which resulted in reduced duplication of efforts and approach driven by public health and community par- more efficient use of government funding. Successful ticipation, as well as for its ability to adapt and sustain platforms also provided a mechanism to vet new pro- itself despite political and environmental changes over jects or accept funding from external donors or NGOs time. The district health office originally created enriched based on the priorities of communities. The ability of health posts with three key functions: community edu- platforms to set their own health agenda further cation, community empowerment, and community ser- reduced duplication of efforts and empowered com- vices. The enriched health post hosted activities such as munities to establish control over their own health clinical maternal and child health, family planning, priorities. nutrition, immunization, diarrhea control, under-five growth stimulation, and early childhood education. Other sectors outside of health care, such as education, Case Study: Gerbangmas Movement as a Community became involved (Blas, Sommerfeld, and Kurup 2011; Health Platform, Lumajang District, Indonesia, Level 4 Siswanto 2009). Among lower-middle-income countries, Indonesia has Starting in 2005, with encouragement from the achieved the highest reduction in the rate of mortality governor, the district health office led subsequent under age 5 years in recent decades (Rohde and others efforts to create the Gerbangmas movement, a plat- 2008; Siswanto 2009). One component in this success form for communities, the public health sector, and was a network of community health posts (posyandus) other government sectors to work collaboratively Community Platforms for Public Health Interventions 273 to achieve 21 indicators of concern (Blas, Sommerfeld, to compete for community dollars in their respective and Kurup 2011; Siswanto 2009). The more specific programs, while preventing duplication of efforts and objectives were to achieve 14 indicators for human competition across sectors (Blas, Sommerfeld, and development, 1 indicator for the economy, and 6 indi- Kurup 2011; Siswanto 2009). cators for the household environment that together The district governor mandated that all community represented the priorities of the government and the empowerment programs use the Gerbangmas move- community, as well as the religious, education, indus- ment as an entry point, thereby reducing competition try and trade, health, family planning, agriculture, and and keeping outside interests (such as those from NGOs) public works sectors (Blas, Sommerfeld, and Kurup from affecting the success of the partnerships across 2011; Siswanto 2009). The sectors worked together sectors (Blas, Sommerfeld, and Kurup 2011; Siswanto with support from the district governor, leadership 2009). With the community at the center of the partner- from a local NGO to address family welfare issues, and ship structure, a hierarchy that placed all sectors on a funding stream that allowed all sectors to contribute equal footing, and a common set of indicators to work to progress on the chosen indicators. toward, the Gerbangmas movement helped sectors work This movement for community development resulted together effectively. in improvements in all indicators (Marmot and others 2008). The multisectoral Gerbangmas movement was Leadership and Integration sustainable and successful, even in the context of a The district health office was the initial champion for the changing economic and government landscape. Gerbangmas movement, which eventually assumed the Although the Gerbangmas movement has experienced role of the integrated health platform. During the initial numerous changes over time, its central tenants of build- scale-up from health posts to enriched health posts, the ing a community health platform to lead cross-sector district health office garnered support from local gov- partnerships has remained relevant in Indonesia for the ernment and encouraged involvement of other sectors past 15 years. Lessons learned from this case study illus- (such as education) while demonstrating the importance trate important roles for local government, cross-sector of involving other sectors in achieving common health partnerships, and leadership. goals. As the health posts evolved into enriched posts, or Gerbangmas health posts (figure 14.3), the district Heath Systems and Role of Local Government health office took a step back to participate as a member The development of the Gerbangmas movement of a team engaging other sectors; an NGO took on a stemmed from decentralization of the Indonesian health more significant role as an integrator and coordinator of system, allowing peripheral innovation. The local gov- the movement. Part of the significance of an NGO’s ernment offered support and leadership for the initia- heading an integrated platform is that such an organiza- tive, as well as a mechanism for funding. Once the tion can be sector neutral, allowing each sector equal movement was planned and funded, the district health weight in achieving agreed-upon goals. Notably, the office created a single vehicle through which the com- community itself held power over the management of munities, the health system, and other sectors could the programs and the priorities of the Gerbangmas collaborate around common goals without competing movement. for volunteers or resources. The district health office did not dominate the partnership; it included itself as a Role of Communities stakeholder, with leadership provided by a neutral entity. Community volunteers conducted needs surveys in their respective villages, and maps of community needs were Partnerships across Sectors developed on the basis of the data gathered. The com- The partnership structure provided clear roles for each munity problems in each village were discussed in open sector to develop programs to help achieve the shared forums where members created action plans. Final pro- indicators. The district PKK (a family welfare semi- posals were drawn up that became the community governmental NGO consisting of spouses of government action plans. Community members had input on the officials and community members) helped coordinate allocation of funds. Financing came from government and support the partnership across organizations. The funding allocation and from financial contributions funding structure created a common pool of funds from from the community. Community volunteers also par- which communities were able to draw for investment in ticipated in the monitoring and evaluation of activities interventions in multiple sectors. Some sectors also con- that had been carried out each year. tributed funds to achieve action plans. Essentially, the District Gerbangmas teams trained subdistrict partnership structure of this movement allowed sectors teams. A training-of-trainers approach helped educate 274 Disease Control Priorities: Improving Health and Reducing Poverty Figure 14.3 Evolution of Conventional Health Posts to Gerbangmas Health Posts in Lumajang District Pre-2001 2001–04 2005 and after Conventional health posts Transitional phase Gerbangmas health posts Activities: Five health services Activities: Five health services Activities: Five health services (for example, maternal and plus under-5 growth plus family endurance, clean child health, family planning, stimulation, early childhood and healthy behavior, nutrition, immunization, and education, and health post for education for children under diarrhea control) the elderly age 5 years and illiterate people, mental and spiritual building, Population targets: Population targets: and productive economy Mothers and children under Mothers, children under age 5 years age 5 years, and the elderly Population targets: Mothers, children under Place of activities: House Sectors involved: age 5 years, the elderly, and yard of community leaders Health, family planning, and all communities education Sectors involved: Health Place of activities: and family planning Number of cadres: Five Posyandu Hall, household persons and community groups Number of cadres: Five persons Sectors involved: Multiple sectors Role of district health officer Championed conversion of conventional Drafted concept of community health posts to enriched health posts. empowerment. Secured funding and helped expand targets. Encouraged role of NGO as neutral integrator. Advocated for multisector collaboration, resulting in the Created vehicle by which all sectors had an equal voice, addition of the education sector. with the district health officer as member. Formed comprehensive plan for more Promoted community assessment and advanced health posts. evaluation cycles. Source: Adapted from Siswanto 2009. Note: NGO = nongovernmental organization. many community volunteers and village staff mem- partnership structure did not depend on the success of bers on the way to assess community health, facilitate any single organization or leader. The largest hurdle to community dialogue about the findings to lead to sustainability was the turnover of government offi- community involvement in proposing and imple- cials. Sustainability relied on the new district gover- menting action plans, and evaluate the results of those nor’s approval of the Gerbangmas movement in the plans (Blas, Sommerfeld, and Kurup 2011; Siswanto subsequent five-year plan. In response, the district 2009). government created an official book on the Gerbangmas movement, including write-ups of the success of Sustainability the movement in the governor’s accountability report. Sustainability was supported by an overall structure The report covered a summary of the governor’s that included resources, funding, and training from achievements during his term and included the move- partnering organizations and did not rely on grant ment as policy in the district regulation, which was rat- funds or external donor dollars (Blas, Sommerfeld, ified by the district legislative body (Blas, Sommerfeld, and Kurup 2011; Siswanto 2009). In addition, the and Kurup 2011; Siswanto 2009). Community Platforms for Public Health Interventions 275 Case Study: Local Health Administration areas, guiding interventions, and choosing where Committees, Peru, Level 3 resources should be allocated. The CLAS structure also The Peruvian government has legalized, regulated, and allowed communities to control the quality of care and institutionalized community participation as a means of distribution of services. Unlike a community advisory ensuring its role in primary health care (Altobelli 2008; board in which participation is often based on board Beracochea 2015; Blas, Sommerfeld, and Kurup 2011; members’ advising those with the power to make deci- Iwami and Petchey 2002). Local health administration sions and allocate funding, each CLAS had the power communities (comunidades locales de administración and resources to act as the local health department for en salud, or CLAS) are private, nonprofit civil associa- its respective community. tions that have agreements with the government to The CLAS’s financing came from direct government receive and administer public funding for the purpose of transfers from general revenue, reimbursements from implementing primary health care services responsive to the government health insurance program for the poor, community needs. and in-kind stocks of medicines and supplies from the regional health directorates. Control over allocation of Evolution of Local Health Administration these funds resided in the hands of the CLAS (Altobelli Committees 2008; Beracochea 2015; Blas, Sommerfeld, and Kurup The path to development of the CLAS movement was 2011; Iwami and Petchey 2002). The CLAS assemblies a complicated one. The CLAS movement emerged in conducted community assessments for health needs and 1994, following the collapse of the health sector in helped identify unmet health needs to determine how Peru. Terrorism and hyperinflation were major national best to tailor primary health care services to local con- challenges, and decentralization was beginning texts (Iwami and Petchey 2002). (Altobelli 2008; Beracochea 2015; Blas, Sommerfeld, and Kurup 2011; Iwami and Petchey 2002). Rural areas Sustainability had a strong mistrust of the government; initial efforts The CLAS movement began as a pilot with 250 health to expand primary health care in these areas resulted in facilities incorporated into the program. Early evalua- further mistrust, because community members often tions showed improved equity, quality, and coverage of felt mistreated by physicians (Altobelli 2008; health services in CLAS facilities, compared to non- Beracochea 2015; Blas, Sommerfeld, and Kurup 2011; CLAS facilities (Beracochea 2015). Advocates helped Iwami and Petchey 2002). When Jaime Freundt became demonstrate the positive effects of the model, and in the minister of health in the mid-1990s, he sought 2007, the Peruvian Congress approved a statute for citi- reform through a process that involved convening zen participation in primary health care at local levels. technical experts and community members. As a The passage of this law ensured the sustainability of the result, a new form of CLAS was proposed (Altobelli CLAS movement and confirmed Peru’s commitment to 2008; Beracochea 2015; Blas, Sommerfeld, and Kurup empowering communities to have some control over 2011; Iwami and Petchey 2002). their own health care (Altobelli 2008). Role of Communities CLAS Achievements In the new CLAS, community members were part of a The CLAS movement increased the availability of phy- civil association under the authority of the Peruvian sicians in rural areas; improved access to care for the Civil Code. Community members had a formal relation- poor; improved usage rates, especially for children; ship with the government by electing community repre- improved quality in health facilities; and improved con- sentatives to a general assembly that worked with the nections among people in Peruvian communities regional health directorates (Altobelli 2008; Beracochea (Altobelli 2008; Beracochea 2015; Blas, Sommerfeld, 2015; Blas, Sommerfeld, and Kurup 2011; Iwami and and Kurup 2011; Iwami and Petchey 2002). These Petchey 2002). achievements were the result of the communities’ ability The elected assembly provided a way to demand to allocate budgets to attract higher numbers of physi- accountability from health personnel (Altobelli 2008; cians to areas where they were needed and to provide Beracochea 2015; Blas, Sommerfeld, and Kurup 2011; full or partial fee exemptions based on financial need. In Iwami and Petchey 2002). The CLAS became a platform addition, the number of women members of the CLAS through which community representatives and volun- general assembly grew substantially (Altobelli 2008; teers could perform public health roles of community Beracochea 2015; Blas, Sommerfeld, and Kurup 2011; assessment, identifying health priorities across local Iwami and Petchey 2002). 276 Disease Control Priorities: Improving Health and Reducing Poverty Heath Systems and Role of Government Poverty Action Lab 2015; Björkman and Svensson One interesting lesson learned from the CLAS move- 2009, 2010). The intervention sought to create a com- ment is that public mistrust of the government can be munity-led process of monitoring to ensure that counteracted through structures for communities to health care workers were performing their assigned take ownership and oversight of public programs tasks (Abdul Latif Jameel Poverty Action Lab 2015; (Altobelli 2008; Beracochea 2015; Blas, Sommerfeld, and Björkman and Svensson 2009, 2010). The results of Kurup 2011; Iwami and Petchey 2002). The CLAS move- the study indicated that, compared to control com- ment was a key driver in creating transparency, partici- munities, community-based monitoring improved pation, and social control over the health system that the quality and quantity of primary care delivered, built community trust and improved relations between reduced the number of deaths among children under communities and the government (Altobelli 2008). The age 5 years, improved outpatient service use, and Ministry of Health, with internal and external champi- improved quality measures such as wait time in pri- ons, was instrumental in helping the CLAS expansion to mary care (Abdul Latif Jameel Poverty Action Lab continue and become law (Altobelli 2008). 2015; Björkman and Svensson 2009, 2010). Partnerships across Sectors Analysis of Uganda District Scorecards In addition to primary health care needs, CLAS began The example of the district scorecard study in Uganda to focus on the development needs of communities represented a limited intervention that was driven by through community work plans that used discretionary outside agencies for the purposes of involving the com- funds and partnerships with local municipalities to allo- munity in health service improvement. Despite positive cate dollars to community-identified development proj- outcomes, ongoing success was reliant on ongoing col- ects (Beracochea 2015). CLAS appears to be well on its lection of scores from scorecards by third-party entities way to transitioning from level 3 to level 4 in the typol- (Abdul Latif Jameel Poverty Action Lab 2015; Björkman ogy of table 14.2; CLAS is already a community plat- and Svensson 2009, 2010). form for addressing health needs and is broadening its intersectoral reach to partner with additional sectors. Health Systems and Role of Government The CLAS movement has been spreading through the In Uganda’s decentralized system, local health unit man- SEED-SCALE model of sustainability (Taylor and agement committees monitored the day-to-day health Taylor 2002). Successful models in each region served as service activities of the public dispensaries. The govern- training centers and hubs for lateral diffusion of ment was not the driver of the interventions and did not innovations. have a large role in the improvements to community health, other than through its role in running the com- mittees (Abdul Latif Jameel Poverty Action Lab 2015; Case Study: Community Scorecards in Björkman and Svensson 2009, 2010). Nine Districts, Uganda, Level 2 Examples of contractor- and donor-driven platforms Partnerships across Sectors (level 2 in table 14.2) are fairly common in practice, and Partnerships across sectors were limited in this example. extensive literature documents this approach. We pres- NGOs and community organizations participated in ent a district scorecard program conducted in Uganda in community meetings, but there were few other partner- 2004 to promote community oversight of health services ships across sectors or across government agencies at the primary care level. (Abdul Latif Jameel Poverty Action Lab 2015; Björkman The goal of the intervention was to strengthen pro- and Svensson 2009, 2010). vider accountability through a process that used com- munity organizations as facilitators of village-level Leadership and Integration meetings to inform communities about the status of The community health platform was originally devel- health service delivery in their area relative to the oped by researchers at the University of Stockholm and standards held in surrounding areas (Abdul Latif the World Bank, and the researchers generated the report Jameel Poverty Action Lab 2015; Björkman and cards that served as the basis for the program. Local Svensson 2009, 2010). Facilitators encouraged com- NGOs facilitated program meetings and served as com- munity members to identify areas for improvement in munity leaders for the intervention. There was no means health service provision and to develop action plans for integration across sectors (Björkman and Svensson that could lead to improvements (Abdul Latif Jameel 2009, 2010). Community Platforms for Public Health Interventions 277 Role of Communities that did exist was destroyed, and the vulnerability of the The role of communities was to attend meetings where state and subsequent reliance on NGOs, faith-based health care provider performance and quality were organizations, and formal providers for care was fur- examined, discuss health care delivery problems that ther exposed (Hill and others 2014). could be improved, and develop action plans for needed Given the diversity of NGOs working throughout changes (Abdul Latif Jameel Poverty Action Lab 2015; Haiti, health care delivery was largely inconsistent in Björkman and Svensson 2009, 2010). Although the com- quality, quantity, and coordination across the country munities’ ability to hold health care providers account- (Hill and others 2014). The role of the Ministry of Public able was limited, they were able to participate in the Health and Population was marginal, and external improvement process and were given a voice for address- resources were often allocated according to the priorities ing their concerns. of NGOs or donors (Hill and others 2014; Zanotti 2010). Ultimately, many of these NGOs did not have local ori- Sustainability gins, did not understand local context, and did not focus Because the scorecards—determined to be a crucial on creating sustainable, responsive platforms where piece of this intervention—were not developed by communities could be empowered to address their own communities or the government, this intervention was health needs (Zanotti 2010). scalable and sustainable only as long as researchers continued to provide data, or until a cheaper and Analysis of Haiti’s Challenges with Development of more direct way of creating the scorecards was estab- Successful Community Health Platforms lished (Abdul Latif Jameel Poverty Action Lab 2015; Unreliable health services and access to those services Björkman and Svensson 2009, 2010). Without further promoted health inequities and created a reliance on government and community buy-in to allocate external entities that created difficulties for communities resources to these activities, the district scorecard to voice their own needs (Hill and others 2014). Lack of intervention faced many challenges in scalability and service integration and coordination led to further frag- sustainability. mentation and duplication of efforts, and Haitians often relied on traditional medicine that was widely available (Hill and others 2014). Case Study: Weak Government Platforms for Despite the challenges, Haiti’s structure also provides Community Empowerment, Haiti, Level 1 the opportunity for NGOs to develop community health Challenges to Development of Community Health platforms that are responsive and engage local commu- Platforms nities. Several NGOs engaged the needs of communities Haiti faces many challenges in developing local govern- and helped build community capacity in the areas of ment engagement of community health platforms. It development, health, and education. Successful NGOs provides a case study where important lessons can be had several factors in common: learned about the role of NGOs and donor agencies in helping promote or hinder development of community • They had local origins in Haiti. health platforms. • They had a diverse international network of donors Haiti has long suffered from natural disasters, disease and were not accountable to a single funder or gov- outbreaks, poverty and social divisions, political insta- ernment agency. bility, and other social and political inequalities that • They focused on addressing local needs and the needs have led to instability (Fatton 2006; James 2010). of the poorest individuals. Numerous NGOs arrived with varying agendas; before • They shared a vision that tied economy, politics, and the 2010 earthquake, an estimated 8,000–9,000 were human rights (Zanotti 2010). working in the country (Batley and McLoughlin 2010; Zanotti 2010). Nearly all of the interventions in the edu- Health Systems and Role of Government cation, health, and development sectors were led by The weakness of the state and the reliance on NGOs NGOs, which provided 70 percent of health care services created an environment in which external entities often and 85 percent of education support (Vaux and Visman influenced resource allocation and priority setting. The 2005; Zanotti 2010). The flow of funds through NGOs lack of a focus on Haitian governance and the subsequent rather than the government weakened the elected gov- lack of health system structure and community input ernment, created instability, and further undermined created difficulties for the community to engage mean- the accountability and sustainability of the state (Zanotti ingfully in the public health process and hampered the 2010). After the earthquake, the negligible state capacity creation of sustainable and responsive health care systems. 278 Disease Control Priorities: Improving Health and Reducing Poverty The ability of communities to hold the government arise whether the priority is (a) implementing or scaling accountable for health service access and quality was up delivery of commodities, services, and programs or nearly absent. (b) building the capacity of communities to identify and address long-standing and emerging public health Partnership across Sectors problems. Coordination among health and other sectors has been The benefits of stronger platforms arise because the slow owing to lack of government leadership. However, more health platforms develop along the continuum in successful NGOs acknowledged the importance of other table 14.2, the better they can carry out the essential sectors in improving health outcomes and worked on public health functions and the cycle of monitoring, issues of sanitation, economic development, and educa- reviewing, and acting to achieve solutions. Strength tion, in addition to health (Zanotti 2010). NGOs served means the capability of health data collection through as providers of services, as well as social advocates pur- local surveillance and outbreak investigation. Strength suing reforms to address poverty and social injustice means that public health personnel can find ways to (Zanotti 2010). share the data with their communities and to engage communities in developing local solutions that mobilize Leadership and Integration external resources as well as untapped resources in com- One of the key difficulties that Haiti faces in creating munities. Strength also means that local public health community health platforms is that the country’s lead- personnel can facilitate implementation of existing pro- ers are highly influenced by external funding sources. grams and develop modifications in response to emerg- The ability of an NGO to make decisions on the basis ing issues. of community needs would be much greater if it did Because only some communities have community not depend on external agencies with specific agendas. health platforms that can effectively carry out essential Addressing community needs requires flexibility in public health functions, outsiders often develop action setting agendas that not all NGOs possess. plans that can succeed in the absence of these platforms. The unintended consequence of neglecting core strength Role of Communities in community health platforms is the continued build- Successful NGOs were those that were able to engage ing of partial substitutes for what community health communities, to set priorities for community input, and platforms ought to be doing. The partial substitutes to include communities in identifying problems and crowd out the necessary business of building indigenous developing and delivering solutions. These included, for strength. example, community health workers and health care providers (Zanotti 2010). Factors That Strengthen Community Health Platforms Sustainability Our review found the following identifiable factors that One of Haiti’s most significant challenges is creating strengthen community health platforms: sustainable solutions in the presence of NGOs that pro- vide the majority of the health-related services in the • Access to data about health problems and health country. NGOs that can create a platform through which threats communities can carry out basic public health functions • The means and will to share data and control with and partner with other sectors to address the social community members determinants of health represent a way forward. NGOs • Achievement of a balance between delivering clinical that can empower communities and provide them with services and preventing disease in whole populations the necessary skills are setting the stage for the sustain- • Advocacy to maintain community engagement ability and effectiveness of a future health system. against pressure to consolidate control. In some cases, these factors were present fortuitously. STRENGTHENING COMMUNITY HEALTH However, evidence suggests that the success factors can PLATFORMS be present as the result of intention and effort. A com- mitment to engage community stakeholders cannot be Benefits of Strengthening Community Health Platforms maintained for long simply because of circumstances. The reviewed literature and the focal case studies high- However, a widespread political movement toward light the benefits of and provide a framework for openness and grassroots engagement can make main- strengthening community health platforms. The benefits taining a community orientation easier. Community Platforms for Public Health Interventions 279 Priorities for Investment in Strengthening platforms can marshal new resources to the service of Community Health Platforms public health. Effective strategies must come from taking stock of the current position of a community on the develop- ment continuum shown in table 14.2. Tools to mea- Valuing Community Health Platforms sure a community’s performance of essential public Given the common misinterpretation that cost- health functions have been used extensively in the effectiveness (as dollars per disability-adjusted life Americas (Corso and others 2000; PAHO 2001; year averted) is the key to understanding an interven- Upshaw 2000). Measurement of current strength in tion’s value, one might be lulled into thinking that any public health care services through a performance investment that cannot show its disability-adjusted and quality improvement tool that targets the essen- life years averted is wasteful—perhaps even unethical, tial public health functions can help identify areas of given that people are dying of preventable causes emphasis within a district if the measures are pro- every day. vided to the public health staff to help create a per- Without initiatives to help community health plat- formance improvement plan (Bishai and others forms flourish around the world, the health gains 2016). promised by interventions will cost more and deliver A strategy to develop community health platforms less. Communities will miss opportunities to activate requires a modest investment in a central unit devoted partners and resources that can shift health determi- to the quality of public health practice. Quality units nants in schools and workplaces and the commerce, are a growing feature in public health departments transport, and culture sectors. Political will to make (Gunzenhauser and others 2010). The best practice for changes in public health law enforcement and regulation a quality unit is to use measurement of practice as a and to hold governments accountable is a precious conversation starter rather than a disciplinary blud- resource that community health platforms can nurture geon. A public health practice quality unit for a central and maintain. With the availability of local data, local or regional health ministry requires a small invest- forums for sharing data, and local multisectoral stake- ment. The budget should allow a team of district holder engagement, the solutions will work better and supervisors to make quarterly supervisory visits to deliver more. This human infrastructure has been specified districts and remain in regular electronic neglected for far too long. communication. Checklists and protocols for supervi- sory visits have been developed and are available from several sources. (The library of these resources A Way Forward for Health Systems can be found at http://www.ianphi.org/documents With the Sustainable Development Goals and calls for /pdfs/evaluationtool and https://sites.google.com/site health system resilience, we are entering a new era in /ephfjhu/.) which this neglect of community engagement and capacity is ending (Bellagio District Public Health Workshop Participants 2016). Community health plat- CONCLUSIONS forms require a respectful trust that people being pre- Communities vary in their level of sophistication in sented with data about their health problems and conducting a cycle of monitoring, reviewing, and act- evidence about what works to solve the problems will ing on the basis of local data and local multisector choose wisely. Community health platforms require a community-engaged partnerships. Helping communi- recognition that health is too big for the health care sec- ties do this well is a concept that goes back to the tor alone; we need a decision-making forum that includes foundations of the field of public health. Because the education sector, commercial interests, transporta- good health can exist at low cost with vertical pro- tion, law enforcement, and media. These partnerships grams that rescue people regardless of their communi- are essential if we are to address upstream social ty’s functional level, making the case for investing in determinants. community resilience can be challenging. The situa- Our model of community health platforms is explic- tion does not need to be an “either-or” option; the way itly drawn at the local level. The national and global forward ought to be a “both-and” option. Rescuing policy makers have important roles in setting up expec- and building resilience are complementary. Especially tations and tools to support local communities. where budgets are finite, strong community health Fundamentally, human bodies are small objects; most 280 Disease Control Priorities: Improving Health and Reducing Poverty of the time, what makes a body sick (or worse) is a NOTE microbe from across the street or a cigarette from the World Bank Income Classifications as of July 2014 are as local store or a speeding car with a drunk driver behind follows, based on estimates of gross national income (GNI) per the wheel. Protecting a body requires a protector that is capita for 2013: close to that body. The emerging burden of noncom- municable diseases caused by health behavior choices, • Low-income countries (LICs) = US$1,045 or less lifestyles, mental health trauma, and injuries under- • Middle-income countries (MICs) are subdivided: scores the need for local approaches. High-income (a) lower-middle-income = US$1,046 to US$4,125 country data show that noncommunicable disease (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 burdens differ intensely at the scale of a census tract. • High-income countries (HICs) = US$12,746 or more. Modern cities are seeing life-expectancy differentials of 20 years across neighborhoods. The other advantage of local communities is their sheer number. For a failed state, efforts to work at the REFERENCES national level can remain frustrating for decades. At Abdul Latif Jameel Poverty Action Lab. 2015. “The Power the local level, one can find failed communities, but of Information in Community Monitoring.” J-PAL Policy one can also find successful communities. One can Briefcase, July. https://www.povertyactionlab.org/sites/default even find successful communities inside failed states /files/publications/Community%20Monitoring_2.pdf. Altobelli, L. 2008. “Case Study of CLAS in Peru: Opportunity and accomplish at subnational levels what cannot be and Empowerment for Health Equity.” Prepared for Case done when a central government is not prioritizing Studies of Programmes Addressing Social Determinants health. of Health and Equity, Priority Public Health Conditions– The model of community health platforms asks Knowledge Network, World Health Organization (WHO) local government health officials to play a prominent Commission on Social Determinants of Health, with role as conveners and integrators. Government pres- support from the Alliance for Health Policy and Systems ence does not suggest that government workers per- Research, WHO, Geneva. form all of the roles in the public health cycle. The Batley, R., and C. McLoughlin. 2010. “Engagement with Non- decisions about who does what emerge from the com- State Service Providers in Fragile States: Reconciling State munity, on the basis of its own stock of possible actors Building and Service Delivery.” Development Policy Review and doers. Community health platforms can mobilize 28 (10): 131–54. Bellagio District Public Health Workshop Participants. 2016. resources through volunteers and voluntary activities “Public Health Performance Strengthening at Districts: independent of the budgets of governments and Rationale and Blueprint for Action.” White paper prepared donors. after the Bellagio Conference, New York City, November A Chinese proverb says that the best time to plant 21–25. www.who.int/alliance-hpsr/bellagiowhitepaper.pdf. a tree was 20 years ago, and the second-best time is Beracochea, E. 2015. Improving Aid in Global Health. New York: today. High-functioning community health plat- Springer. forms are the trees that we wish our ancestors had Bettcher, D. W., S. Sapirie, and E. H. Goon. 1998. “Essential planted in every community many years ago. Future Public Health Functions: Results of the International generations cannot afford to have us spend the next Delphi Study.” World Health Statistics Quarterly 51 (1): 20 years attending to local epidemics and global pan- 44–54. demics that could have been snuffed out and quickly Bishai, D., M. K. Sherry, C. Pereira, S. Chicumbe, F. Mbofana, and others. 2016. “Development and Usefulness of a controlled if all local communities had been per- District Health Systems Tool for Performance Improvement forming all of the essential public health functions in Essential Public Health Functions in Botswana and and engaging their communities in building a culture Mozambique.” Journal of Public Health Management and of health. Practice 22 (6): 586–96. Björkman, M., and J. Svensson. 2009. “Power to the People: Evidence from a Randomized Field Experiment on ACKNOWLEDGMENTS Community-Based Monitoring in Uganda.” Quarterly Journal of Economics 124 (2): 735–69. The authors gratefully acknowledge helpful comments ———. 2010. “When Is Community-Based Monitoring from Henry Mosley, MD, Johns Hopkins Bloomberg Effective? Evidence from a Randomized Experiment in School of Public Health, Johns Hopkins University, Primary Health in Uganda.” Journal of European Economic Baltimore, MD, United States. Association 8 (2–3): 571–81. Community Platforms for Public Health Interventions 281 Blas, E., J. Sommerfeld, and A. S. Kurup. 2011. Social Health: A Scoping Review.” BMC Health Services Research Determinants Approaches to Public Health: From Concept to 13: 64–72. Practice. Geneva: World Health Organization. Kowitt, S. D., D. Emmerling, E. B. Fisher, and C. Tanasugarn. Corso, L. C., P. J. Wiesner, P. K. Halverson, and C. K. Brown. 2015. “Community Health Workers as Agents of 2000. “Using the Essential Services as a Foundation for Health Promotion: Analyzing Thailand’s Village Health Performance Measurement and Assessment of Local Public Volunteer Program.” Journal of Community Health 40 (4): Health Systems.” Journal of Public Health Management and 780–88. Practice 6 (5): 1–18. Lawn, J. E., J. Rohde, S. Rifkin, M. Were, V. K. Paul, and Cyril, S., B. J. Smith, A. Possamai-Inesedy, and A. M. Renzaho. M. Chopra. 2008. “Alma-Ata 30 Years On: Revolutionary, 2015. “Exploring the Role of Community Engagement in Relevant, and Time to Revitalise.” The Lancet 372 (9642): Improving the Health of Disadvantaged Populations: A 917–27. Systematic Review.” Global Health Action 8: 29842. Macinko, J. S., B. Starfield, and T. Erinosho. 2009. “The Impact Dooris, M., and Z. Heritage. 2013. “Healthy Cities: Facilitating of Primary Healthcare on Population Health in Low- and the Active Participation and Empowerment of Local Middle-Income Countries.” Journal of Ambulatory Care People.” Journal of Urban Health 90 (Suppl 1): 74–91. Management 32 (2): 150–71. Draper, A. K., G. Hewitt, and S. Rifkin. 2010. “Chasing the Marmot, M., S. Friel, R. Bell, T. A. J. Houweling, and S. Taylor, Dragon: Developing Indicators for the Assessment of for the Commission on Social Determinants of Health. Community Participation in Health Programmes.” Social 2008. “Closing the Gap in a Generation: Health Equity Science and Medicine 71 (6): 1102–09. through Action on the Social Determinants of Health.” The Dyal, W. W. 1995. “Ten Organizational Practices of Public Lancet 372 (9650): 1661–69. Health: A Historical Perspective.” American Journal of McCoy, D. C., J. A. Hall, and M. Ridge. 2012. “A Systematic Preventive Medicine 11 (Suppl 6): 6–8. Review of the Literature for Evidence on Health Facility Edmunds, M., and E. Albritton. 2015. “Global Public Health Committees in Low- and Middle-Income Countries.” Systems Innovations: A Scan of Promising Practices.” Health Policy Plan 27 (6): 449–66. AcademyHealth, Washington, DC, and Robert Wood McKeown, T., R. G. Record, and R. D. Turner. 1975. “An Johnson Foundation, Princeton, NJ. Interpretation of the Decline of Mortality in England and Fatton, R. 2006. “Haiti: The Saturnalia of Emancipation and Wales during the Twentieth Century.” Population Studies the Vicissitudes of Predatory Rule.” Third World Quarterly 29: 391–422. 27 (1): 115–33. Meier, B. M., C. Pardue, and L. London. 2012. “Implementing Frumence, G., T. Nyamhanga, M. Mwangu, and A. K. Hurtig. Community Participation through Legislative Reform: 2014. “Participation in Health Planning in a Decentralised A Study of the Policy Framework for Community Health System: Experiences from Facility Governing Participation in Western Cape Province of South Africa.” Committees in the Kongwa District of Tanzania.” Global BMC International Health and Human Rights 12: 15–29. Public Health 9 (10): 1125–38. Mills, A., F. Rasheed, and S. Tollman. 2006. “Strengthening George, A., K. Scott, S. Garimella, S. Mondal, R. Ved, and K. Health Systems.” In Disease Control Priorities in Developing Sheikh. 2015. “Anchoring Contextual Analysis in Health Countries, second edition, edited by D. T. Jamison, J. G. Policy and Systems Research: A Narrative Review of Bremen, A. R. Measham, G. Alleyne, M. Claeson, D. B. Contextual Factors Influencing Health Committees in Low Evans, P. Jha, A. Mills, and P. Musgrove. Washington, DC: and Middle Income Countries.” Social Science and Medicine World Bank and Oxford University Press. 133: 159–67. Mosley, W. H., J. L. Bobadilla, and D. T. Jamison. 1993. Gunzenhauser, J., Z. Eggena, J. Fielding, K. Smith, D. Jacobson, “The Health Transition: Implications for Health Policy and N. Bazini-Barakat. 2010. “The Quality Improvement in Developing Countries.” In Disease Control Priorities in Experience in a High-Performing Local Health Department: Developing Countries, edited by D. T. Jamison, W. H. Mosley, Los Angeles County.” Journal of Public Health Management A. R. Measham, and J. L. Bobadilla. Washington, DC: World and Practice 16 (1): 39–48. Bank; New York: Oxford University Press. Hill, P. S., E. Pavignani, M. Michael, M. Murru, and M. Beesley. O’Mara-Eves, A., G. Brunton, S. Oliver, J. Kavanagh, F. Jamal, 2014. “The ‘Empty Void’ Is a Crowded Space: Health Service and J. Thomas. 2015. “The Effectiveness of Community Provision at the Margins of Fragile and Conflict Affected Engagement in Public Health Interventions for States.” Conflict and Health 8 (20). Disadvantaged Groups: A Meta-Analysis.” BMC Public Iwami, M., and R. Petchey. 2002. “A CLAS Act? Community- Health 15: 129. Based Organizations, Health Service Decentralization and Open Working Group of the General Assembly. 2014. “Open Primary Care Development in Peru.” Journal of Public Working Group Proposal for Sustainable Development Health Medicine 26 (4): 246–51. Goals.” A/68/970, United Nations, New York. http://undocs. James, E. C. 2010. “Ruptures, Rights, and Repair: The Political org/A/68/970. Economy of Trauma in Haiti.” Social Science and Medicine PAHO (Pan American Health Organization). 2001. Public 70 (1): 106–13. Health in the Americas. Washington, DC: PAHO. Kenny, A., N. Hyett, J. Sawtell, V. Dickson-Swift, J. Farmer, Patel, V., D. Chisholm, T. Dua, R. Laxminarayan, and and P. O’Meara. 2013. “Community Participation in Rural M. E. Medina-Mora, eds. 2015. Disease Control Priorities 282 Disease Control Priorities: Improving Health and Reducing Poverty (third edition): Volume 4, Mental, Neurological, and Lumajang District.” Jurnal Managemen Pelayanan Substance Use Disorders. Washington, DC: World Bank. Keshehatan 12 (1). Perry, H. B. 2013. “Primary Health Care: A Redefinition, Szreter, S. 1988. “The Importance of Social Intervention in History, Trends, Controversies and Challenges.” White Britain’s Mortality Decline c. 1850–1914: A Reinterpretation paper, Unpublished. of the Role of Public Health.” Social History of Medicine Raeburn, J., M. Akerman, K. Chuengsatiansup, F. Mejia, and O. 1 (1): 1–38. Oladepo. 2006. “Community Capacity Building and Health Tanvatanakul, V., C. Vicente, J. Amado, and S. Saowakontha. Promotion in a Globalized World.” Health Promotion 2007. “Strengthening Health Development at the International 21 (Suppl 1): 84–90. Community Level in Thailand: What Events Should Be Rasanathan, K., T. Posayanonda, M. Birmingham, and V. Managed?” World Health and Population 9 (1): 65–73. Tangcharoensathien. 2012. “Innovation and Participation Taylor, D. C., and C. E. Taylor. 2002. A Just and Lasting Change: for Healthy Public Policy: The First National Health When Communities Own Their Own Futures. Baltimore, Assembly in Thailand.” Health Expectations 15 (1): 87–96. MD: Johns Hopkins University Press. Rifkin, S. B. 1996. “Paradigms Lost: Toward a New Tiwari, R., M. Lommerse, and D. Smith. 2014. M2 Models Understanding of Community Participation in Health and Methodologies for Community Engagement. Singapore: Programmes.” Acta Tropica 61 (2): 79–92. Springer. ———. 2014. “Examining the Links between Community UK Aid and DFID/HDRC (Department for International Participation and Health Outcomes: A Review of the Development, United Kingdom/Human Development Literature.” Health Policy and Planning 29 (Suppl 2): ii98–106. Resource Centre). 2011. “Helpdesk Report: Community Rohde, J., S. Cousens, M. Chopra, V. Tangcharoensathien, Engagement in Health Service Delivery.” November 16, R. Black, and others. 2008. “30 Years after Alma-Ata: Has DFID, London. Primary Health Care Worked in Countries?” The Lancet Upshaw, V. 2000. “The National Public Health Performance 372 (9642): 950–61. Standards Program: Will It Strengthen Governance of Local Rosato, M., G. Laverack, L. H. Grabman, P. Tripathy, N. Nair, Public Health?” Journal of Public Health Management and and others. 2008. “Community Participation: Lessons for Practice 6 (5): 88–92. Maternal, Newborn, and Child Health.” The Lancet 372 Vaux, T., and E. Visman. 2005. Service Delivery in Countries (9642): 962–71. Emerging from Conflict. Report, Centre for International Rosen, G. 1958. A History of Public Health. Baltimore, MD: Co-operation and Security, University of Bradford, Johns Hopkins University Press. Bradford, U.K. Russell, N., S. Ingras, N. Johri, H. Kuoh, M. Pavin, and J. Westphal, M. F., F. Zioni, M. F. Almeida, and P. R. Nascimento. Wickstrom. 2008. “The Active Community Engagement 2011. “Monitoring Millennium Development Goals in Continuum.” Project Working Paper, July, ACQUIRE Brazilian Municipalities: Challenges to Be Met in Facing Up to Project, New York, and United States Agency for Iniquities.” Cadernos de saude publica 27 (Suppl 2): S155–63. International Development, Washington, DC. http://www World Bank. World Development Report 1993: Investing in .acquireproject.org/fileadmin/user_upload/ACQUIRE Health. Washington, DC: World Bank. /Publications/ACE-Working-Paper-final.pdf. Zanotti, L. 2010. “Cacophonies of Aid, Failed State Building and Siswanto, E. S. 2009. “Community Empowerment through NGOs in Haiti: Setting the Stage for Disaster, Envisioning Intersectoral Action: A Case Study of Gerbangmas in the Future.” Third World Quarterly 31 (5): 755–71. Community Platforms for Public Health Interventions 283 Chapter 15 Rehabilitation: Essential along the Continuum of Care Jody-Anne Mills, Elanie Marks, Teri Reynolds, and Alarcos Cieza INTRODUCTION surgical services, preventing complications, and ensuring The World Health Organization (WHO) has defined that the optimal functional outcome is achieved. rehabilitation as “a set of interventions designed to opti- Rehabilitation in the community similarly aims to opti- mize functioning and reduce disability in individuals mize functioning in those who are not in the hospital with health conditions, in interaction with their system, to identify needs, and to provide services in a environment” (Nas and others 2015, 1). Rehabilitation person’s typical environment. Community rehabilitation interventions optimize well-being by addressing impair- services frequently are accessed by those with chronic ments, limitations, and restrictions in many areas (areas health conditions or sensory impairment, as well as by as disparate as mobility, vision, and cognition), as well as children with developmental conditions. by considering personal and environmental factors (Nas The demand for community- and hospital-based and others 2015). rehabilitation services will continue to grow as the result Individuals with health conditions or injuries may of several factors. First is the significant epidemiological require rehabilitation across the course of their lifespan. transition and demographic shift underway globally The timing and type of intervention that a rehabilitation (Dalal and others 2015; Dias and others 2013; GBD provider selects depend greatly on several factors. These 2015 Disease and Injury Incidence and Prevalence include the etiology and severity of the person’s health Collaborators 2016). Second, as access to care expands to condition, the prognosis, the way in which the person’s universal health coverage, rehabilitation is essential for condition affects the person’s ability to function in maximizing the effectiveness of a range of medical and the environment, as well as the individual’s identified surgical interventions. Finally, injuries (which remain an personal goals. escalating public health concern in some countries) also Rehabilitation services may be delivered in any setting contribute substantially to the demand for rehabilitation (including in hospitals and in communities), depending services (WHO 2014). These factors suggest that the on individuals’ needs and situation. In hospitals, acute positive health, social, and economic effects of rehabili- rehabilitation is particularly important in facilitating tation will have a more profound influence on popula- recovery, maximizing the effect of emergency and tion health in coming years (WHO 2016a). Corresponding author: Alarcos Cieza, Department for Management of Noncommunicable Diseases, Disability, Violence, and Injury Prevention, World Health Organization, Geneva, Switzerland; ciezaa@who.int. 285 GROWING DEMAND FOR admissions (Beswick and others 2008; Gillespie and REHABILITATION SERVICES others 2012). Ensuring that the disabilities associated with aging are minimized is a major priority for policy Health and Population Trends development (UN 2015); health systems need to take The prevalence of noncommunicable diseases has concerted action to ensure that they can provide older increased by 13.7 percent in the past 10 years (GBD populations with the requisite services. 2015 Disease and Injury Incidence and Prevalence The potential benefits of rehabilitation services are Collaborators 2016). Noncommunicable diseases and not restricted to aging and adult populations. Children associated health complications can have a profound constitute a significant and important portion of users effect across functioning domains, such as mobility, of rehabilitation services. Although fertility rates are respiration, vision, cognition, and communication. slowly declining in many low- and lower-middle-income Studies have shown that rehabilitation can effectively countries, populations continue to expand. For example, assist in prevention of and recovery from various health 48 percent of the population of Chad and 42 percent conditions. Stroke and cardiac rehabilitation have been of the population of Timor-Leste are between ages shown to be effective in increasing independence, reduc- 0–14 years (World Bank 2016). Furthermore, while child ing mortality, and reducing hospital readmissions mortality rates are declining, not all who survive actually (Jolliffe and others 2000; Stroke Unit Trialists’ thrive (Grantham-McGregor and others 2007; WHO Collaboration 2013; Taylor and others 2010). Similarly, 2016b). Early interventions that optimize developmental rehabilitation following amputation improves physical outcomes for children with various health conditions functioning and improves the likelihood of home dis- (including neurological, congenital, and intellectual charge from the hospital (Fleury, Salih, and Peel 2013; impairments), as well as injuries, can positively affect Kurichi and others 2009). participation rates in education, community activities, Demand for rehabilitation services directly corre- and future capacity to work. sponds to the incidence of injuries (such as those caused by road traffic crashes, burns, near drownings, falls, and poisonings). For every one of the more than Expanded Access to Health Care 5 million people who die as the result of injuries every As access to more advanced emergency, trauma, and year, 10 to 15 more people are estimated to survive, medical care expands, rehabilitation becomes propor- many with ensuing impairment. A significant portion tionally more important. It constitutes an essential of injuries are caused by road traffic injuries (WHO aspect of care for many of those who experience, or are 2014), which are predicted to become the seventh lead- at risk of experiencing, short-term or long-term residual ing cause of death by 2030. The number of road traffic impairment and functioning limitations following inju- injuries is anticipated to increase, especially in low- ries or illnesses. These include the following: income countries as economies develop and more peo- ple use vehicles (Gosselin and others 2009). Along with • Individuals with injuries or medical conditions surgical and medical interventions, rehabilitation helps requiring lower-limb amputation. Amputations to mitigate the profound socioeconomic impact of non- may effectively save lives, but mobility will decline fatal injuries. substantially without proper postoperative stump The consequences of the demographic shift currently management, strengthening, and training in the use underway globally are substantial; the number of indi- of a mobility device such as a prosthesis (Fleury, viduals older than age 60 years is projected to increase Salih, and Peel 2013; Godlwana, Stewart, and 56 percent globally by 2030 (UN 2015). Aging is associ- Musenge 2015). ated with natural decrements in intrinsic capacity (such • Children with spastic cerebral palsy. Antispasmodic as declines in musculoskeletal strength and cognitive medication may be effective, but children’s inde- function) that increase vulnerability to health conditions pendence may be largely unchanged without and injuries (WHO 2015). Widespread availability of adequate supported seating, splinting, and func- rehabilitation services is essential for health systems to tional retraining (Aisen and others 2011; Novak and be able to respond effectively to the needs of older others 2013). populations. Numerous studies have concluded that • People with burn injuries. Such individuals may ben- community-based rehabilitation increases the safety and efit from skin grafting, but rehabilitation is required independence of older people, reduces the risk of falls, from the acute to long-term phase of recovery to pre- and decreases the need for hospital and nursing home vent or minimize skin contractures, to regain strength 286 Disease Control Priorities: Improving Health and Reducing Poverty and dexterity, and to maximize functional outcomes the most economically disadvantaged. The lack of robust (Proctor 2010). impact studies notwithstanding, substantial evidence on • Those with spinal cord injuries (particularly complete the effectiveness of rehabilitation on health, economic, and high-level injuries) who have received optimal and quality of life outcomes provides ample impetus to care in the acute phase. Without access to appropriate adopt a systematic approach to building and strengthen- rehabilitation and long-term care, such individuals ing rehabilitation services. Several examples from may experience potentially fatal complications, such upper-middle-income countries demonstrate the feasi- as pressure sores and urinary tract infections (Nas bility of implementing rehabilitation interventions in and others 2015). health systems with limited resources for health and a diversity of approaches to doing so. Integrating rehabilitation into health care systems and providing early access to services can benefit both indi- viduals and health systems. Such integration can help to Expanding the Availability of Rehabilitation in Mexico ensure optimal outcomes from medical and surgical Mexico responded to its population’s growing rehabilita- interventions, and it can mitigate the risks of ongoing tion needs by developing 46 first-level rehabilitation complications that may burden the health system. units that provide evaluation, therapy, and referral; these Furthermore, the benefits of rehabilitation are realized units are staffed by physiatrists, physiotherapists, social beyond the health system. By restoring functioning, reha- workers, and nurses. The development of these units has bilitation can enable people to take up or resume family increased Mexico’s rehabilitation services capacity by and work roles and can enable them to participate in 60 percent. In addition to these services, Mexico also has education and community life, with potentially substan- 1,444 community-based basic rehabilitation units dis- tial economic and social implications (WHO 2017). tributed across the country, and rehabilitation services are integrated in general and specialized hospitals and institutions (Guzman and Salazar 2014). UNMET REHABILITATION NEEDS AND PROMISING PROGRAMS IN MIDDLE- Speeding Access to Acute Rehabilitation in Brazil INCOME COUNTRIES The Orthopaedic and Traumatology Institute at a hos- In many parts of the world, the capacity to provide reha- pital in São Paulo, Brazil, has created a simplified reha- bilitation is limited or nonexistent, and the needs of the bilitation program to address the rehabilitation needs population remain largely unmet (Anchique Santos and of those in its care. Before the program’s development, others 2014; Atijosan and others 2009). Analysis sug- people who sustained spinal cord injuries, amputations gests that 92 percent of the burden of disease in the following limb injuries, and severe musculoskeletal world is related to an etiology for which rehabilitation injuries had to wait to receive therapy for up to one year may be required; it further suggests that a strong nega- following their injuries. For many people, this delay tive relationship exists between countries with the resulted in devastating secondary complications that highest rehabilitation need and the availability of reha- easily could have been prevented, such as pressure bilitation professionals (Gupta, Castillo-Laborde, and sores, joint deformities, and chronic pain. The program Landry 2011). has had a profound effect on the prevention of compli- The true effect of this unmet need is difficult to cations and resulting functional outcomes, and it dem- capture, partly because the benefits of rehabilitation are onstrates how facilities with limited resources can realized longitudinally and in outcomes that are more benefit from basic rehabilitation strategies (Mock and challenging to measure (such as participation in work others 2006). and education). Moreover, few studies have assessed the long-term and comprehensive effects of rehabilitation; ECONOMIC CASE FOR INVESTMENT these effects may be made manifest in the ability to return to or engage in meaningful occupation and gain- The diversity in the scope of rehabilitation interventions ful employment, to participate in education, and to and the settings in which they are provided create a chal- achieve a degree of independence with self-care tasks. lenge for cost-effectiveness assessments. This limitation Deductive inference suggests, however, that the health notwithstanding, several examples of the application of and social impacts of failing to receive necessary rehabil- rehabilitation in the context of specific conditions itation services will fall most heavily on those who are demonstrate cost savings. These tend to capture cost Rehabilitation: Essential along the Continuum of Care 287 benefits in the acute phase of care for health systems and and a 40 percent reduction in health care costs (for indi- not the economic advantages for service users, which viduals with short-term disabilities) (Beal 2007). Another may be more profound. study found that that for every dollar invested in return- Cost savings associated with rehabilitation are not to-work rehabilitation, $2.35 is returned to society always fully accrued by the health sector. They may be (Na 2016). The magnitude of return on investment to realized through reduction in ongoing care costs pro- taxpayers is dependent on the disability scheme in the vided by social services, the persons themselves, or their country; regardless, without return-to-work programs, family members. A multicenter cohort analysis from 62 employees affected by injury or illness may face substan- rehabilitation services in third-level hospitals in the tial reductions in standard of living. Depending on the United Kingdom (Turner-Stokes and others 2016) found availability of resources, such programs could be adjusted specialized rehabilitation for complex neurological con- for most settings. ditions to be highly cost-efficient. The weekly care costs for a person with a spinal cord injury who was highly dependent were reduced by £847; approximately ESSENTIAL PACKAGE OF REHABILITATION 22.7 months were needed to offset the cost of the reha- bilitation episode. INTERVENTIONS Rehabilitation also has been found to be cost-effective The essential package of interventions presented in in the context of preoperative and postoperative care. table 15.1 is an initial attempt to compile rehabilitation Provision of rehabilitation before and after lumbar spine interventions in a minimum essential set of services. The fusion surgery in a hospital in Denmark resulted in interventions are based on the International Classification lower costs for both the hospital and patients. In addi- of Functioning, Disability, and Health (WHO 2001) and tion to the benefit of reduced hospital length of stay, the International Classification of Health Interventions costs were 1,625€ lower per patient once direct (hospital (WHO 2016c). As such, the interventions are not fees) and indirect fees (financial burden for patients mapped to specific diagnoses and may be performed in before returning to work) were considered (Nielsen and the context of many health conditions. The rehabilita- others 2008). tion interventions included in the essential package are Whereas large, high-quality methodical studies of targeted at resource-constrained settings, such as a rehabilitation cost-effectiveness originate predomi- district hospital in Sub-Saharan Africa. However, coun- nately from high-income countries, studies from low- tries are not restricted to this level; when the package is and middle-income countries (LMICs) suggest that the applied in settings with greater resource availability, same is true in these settings. Cardiac rehabilitation in countries are encouraged to expand the scope, quality, LMICs, for example, has been found to save costs, com- and availability of interventions. pared with routine management based on provider Certain important adjuncts to rehabilitation have judgment. In Brazil, cardiac rehabilitation leads to mean not been included in this package of interventions. monthly savings per patient of US$190. In Colombia, Prescription of medication (for example, analgesia to the economic benefit was calculated as significantly assist with pain management or antispasmodic medi- higher; the cost-effectiveness of a typical cardiac reha- cation to assist with tone or spasticity) also may be bilitation program for patients with heart failure is considered if it is in the scope of practice of the pro- estimated to be US$998 per quality-adjusted life year, vider. Use of medication during selected interven- compared with usual care with five-year follow-up tions, or as an intervention in its own right, can assist (Oldridge, Pakosh, and Thomas 2016). with patient comfort and ability to participate in Although the literature is limited to high-income functional activities. Psychological interventions also countries, promising evidence of the cost-effectiveness are an important component of rehabilitation, not of rehabilitation programs for reintegration into the only in the context of mental health, but also for peo- workplace exists (European Agency for Safety and Health ple experiencing different impairments (such as phys- at Work 2016; Franche and others 2005). Studies suggest ical or sensory). Mental health interventions for adults that although there is an initial investment in return-to- and children are exclusively covered in the third edi- work programs (typically incurred by the employer), tion of Disease Control Priorities (DCP3), volume 4, there can be a substantial return for society. Cost savings Mental, Neurological, and Substance Use Disorders are almost entirely due to foregone benefit payments (chapters 4 [Hyman and others 2015] and 8 [Scott and (Bardos, Burak, and Ben-Shalom 2015). One study others 2015]). found that return-to-work rehabilitation programs The rehabilitation workforce is potentially the resulted in a 25 to 30 percent reduction in lost workdays most important mechanism for delivering the package 288 Disease Control Priorities: Improving Health and Reducing Poverty Table 15.1 Essential Package of Rehabilitation Interventions Platform for delivery Intervention area Communitya Primary health center Hospitalb Musculoskeletal system Transfer training Mobility training (including gait training) Prescriptionc of mobility techniques customized to the condition and individual Acute mobilization—inpatients and outpatients Basic lower limb, upper limb, and Simple lower limb, upper Prescriptionc of lower limb, upper trunk/spine exercise and symptom limb, and trunk/spine exercise limb, and trunk/spine exercise management programs according and symptom management and symptom management to standard protocols based on programs based on diagnosis programs customized to the presentation (condition specific) condition and individual • Joint mobilization • Stretches/range of movement • Strengthening Postamputation management • Stump care • Limb positioning Ponseti clubfoot treatment Body repositioning for • Pressure area care • Supportive seating, in wheelchairs Upper limb functional retraining Prescriptionc of upper limb • Functional gross and fine motor functional retraining techniques movement patterns customized to the condition and individual • Compensatory strategies Prescriptionc of scar and contracture management techniques to optimize range of movement Cardiorespiratory system Cardiac rehabilitation (such as Prescriptionc of a cardiac recommendations for physical activity, rehabilitation program customized nutrition, and risk factor management) to the condition and individual Breathing exercises to improve Chest function interventions, respiratory function, including sputum including sputum clearance clearance techniques techniques table continues next page Rehabilitation: Essential along the Continuum of Care 289 Table 15.1 Essential Package of Rehabilitation Interventions (continued) Platform for delivery Intervention area Communitya Primary health center Hospitalb Neurological systems and Basic swallow retraining/interventions Prescriptionc of swallow communication retraining techniques customized to the condition and individual Acute swallow management for inpatients Speech and communication Prescriptionc of speech and interventions communication techniques • Interventions for aphasia and ataxia customized to the condition and individual • Sign language • Other alternative mechanisms of communication Cognitive interventions Prescriptionc of cognitive • Training in basic-level cognitive interventions customized to the functions condition and individual • Cognitive compensatory strategies (techniques and provision of assistive products) • Early stimulation for children Mechanical stabilization Prosthesis review and referral Fabrication, fitting, and training in and assistive products to hospital if indicated the use of a prosthesisd Splinting and orthosis review Splinting and orthosise for upper and referral to hospital if limb, lower limb, and spine indicated immobilization and stability Postoperative splinting and orthosise Upper limb positioning • Slings • Casting Compression therapyf for postamputation management, burns, and vascular and lymphatic conditions Provision and training in the use of Provision and training in the use assistive products, assistive technology, of assistive products, assistive and compensatory strategies for technology, and compensatory • Mobility, activities of daily living, strategies for and skin care • Hearing aids and hearing • Vision loss (such as white canes, loopsd braille displays, magnification, and other aids) • Communication devices table continues next page 290 Disease Control Priorities: Improving Health and Reducing Poverty Table 15.1 Essential Package of Rehabilitation Interventions (continued) Platform for delivery Intervention area Communitya Primary health center Hospitalb Cross-cutting areas Self-care training Prescriptionc of self-care techniques customized to the condition and individual Early childhood development rehabilitation interventions (such as motor, sensory, and language stimulation) Environmental modifications (such as a grab rail or ramp installation) Note: This table identifies a package of essential rehabilitation interventions that an effective rehabilitation system must be able to provide. The interventions selected are based on expert opinion from key stakeholders representing a broad range of rehabilitation disciplines. • Interventions in red are considered acute and urgent. • All interventions assigned to a given level also should be available at higher levels. • Medications (such as pain medication to assist with pain management, and antispasmodic medication to assist with tone or spasticity) are not included here, but they may be essential adjuncts to these interventions. • Interventions have been broadly categorized into intervention areas for the purposes of readability; however, substantial overlap exists in interventions between categories. For example, a person may require mobility training for musculoskeletal, cardiorespiratory, and neurological conditions; however, within the package it has been categorized under the musculoskeletal system intervention area. A glossary of intervention terms is available in annex 15A. a. The rehabilitation interventions in the community may need to be delivered by a specialized rehabilitation provider, whereas others may be delivered by generalist community- health workers or other care providers. The level of skill required of the provider depends on the complexity of a person’s needs. Where warranted, interventions should be done under the prescription or supervision of a specialized rehabilitation provider based in the community or in the hospital setting. b. Hospital-based rehabilitation interventions, in first-level and third-level hospitals, are highly variable across countries. Thus, first-level and third-level hospitals are considered as a single delivery setting for the purposes of this package. c. A rehabilitation prescription is the provision of interventions customized for an individual’s condition or specific needs, for ongoing self-management, or to be carried out by another provider. Education is provided to the individual and others involved in the individual’s care to enable them to carry out the prescribed interventions safely and effectively. Such education may include instruction on correct technique, precautions, and specifications of the regime. Prescription and education usually require the input of a specialized rehabilitation provider. d. This intervention also can occur in outpatient settings, although it usually takes place in hospitals. e. This intervention requires access to immobilizing materials (such as thermoplastic, casting, or locally sourced materials) and knowledge of fabrication and application principles, techniques, and precautions. f. This intervention can be done only if the providers are adequately trained in compression bandaging or garment fitting and provision and only if they are aware of precautions and contraindications. It is usually provided in a specialist outpatient service setting (such as a burn unit, plastic surgery facility, or vascular clinic). of interventions. Specialized rehabilitation providers The package does not indicate specific rehabilitation include but are not limited to physiatrists, physiothera- disciplines that will be held responsible for providing the pists, occupational therapists, and speech and language interventions, so as to be applicable to a range of settings pathologists, who together have the capability to provide and levels of rehabilitation workforce capability. interventions across the full scope of needs existing in However, an underlying assumption exists that provid- populations. However, the availability of such a work- ers at the primary health center level will be general- force is rare in countries where rehabilitation is young ists with minimal rehabilitation training, whereas and underdeveloped. In such cases, the skills required to hospital-based providers will have specialized training. conduct basic-level rehabilitation interventions (those Unlike other areas, rehabilitation interventions in the that do not require complex clinical reasoning and community may need to be delivered by a specialized are compatible with foundational health knowledge, rehabilitation provider, whereas others may be delivered skills, and competencies) may be distributed across the by generalist community-health workers or other existing health workforce by using transdisciplinary providers. In the Essential Package of Interventions, a approaches and by developing or strengthening a mid- broad spectrum of skills is required to deliver many of level rehabilitation workforce. Where possible, models of the interventions, largely dependent on the complexity service delivery in which supervision or oversight by a of the needs of the person (such as the presence of rehabilitation professional is provided to less qualified comorbidities, the severity of the health condition, and providers can expand access to services while reducing other personal and environmental factors). The effec- the risk of inappropriate interventions. tiveness of the interventions depends heavily on a Rehabilitation: Essential along the Continuum of Care 291 provider’s skills, experience, and clinical reasoning. At a comprehensive package of services possible at the most minimum, interventions need to be delivered on the accessible level of the delivery. basis of the person’s underlying health condition; apply- Substantial evidence supports the provision of reha- ing interventions irrespective of etiology can be both bilitation at the earliest possible stages and across dangerous and ineffective. the continuum of care: acute, subacute, and long-term Although interventions ideally should be customized care (Choi and others 2008; Parker, Sricharoenchai, and to specific conditions and individual needs and goals Needham 2013; Scivoletto, Morganti, and Molinari (referred to as “prescriptions” in the package), the provi- 2005; Stucki and others 2005). Depending on the etiol- sion of rehabilitation should not be dependent on such ogy of their condition, people may need to access reha- an approach. Prescribing customized interventions bilitation at any level of the health system and likely will requires a level of clinical reasoning that may be avail- continue to require services as they move in and between able by providers at the hospital level, but beyond the levels. Community-based services are necessary to capabilities of a mid-level rehabilitation provider or ensure that those people requiring rehabilitation who generalist health worker in primary health centers or the are not in hospital systems (such as children with sen- community. In such instances, interventions can sory and developmental conditions) are identified and be delivered according to standardized protocols on the receive early intervention. Provision of rehabilitation in basis of presentation and condition. In the package, it is hospitals (including acute wards) is similarly imperative assumed that interventions delivered in the community to prevent complications, to speed recovery, and to link may be delivered by providers capable of following people to follow-up care beyond discharge (Stucki and preexisting standardized protocols for different presen- others 2005). tations that, although not customized, can have great effect. Providers at the primary health center level, where a diagnosis may be more readily available, may be capa- AVAILABLE TOOLS TO INFORM ble of delivering condition-specific interventions, but REHABILITATION SYSTEM PLANNING may not be able to customize them according to individ- The WHO has developed tools to assist countries in ual or complex needs. strengthening rehabilitation in their health systems, Given the variability in training and level of special- including the following: ization of the rehabilitation workforce in LMICs, for the sake of both quality and safety, countries and services must consider the competencies of their workforce (as • The WHO Rehabilitation System Assessment Tool1 well as other resource and contextual factors) when plan- The Rehabilitation System Assessment Tool comprises ning to implement the package. Annex 15A is a glossary (1) a survey-based tool on system-wide rehabilitation that provides a brief description of the interventions that capacity and (2) a field component that assesses the can be used to further guide decision making. rehabilitation system performance. A clear under- The interventions are organized across three service standing of the various elements of the rehabilitation delivery platforms: community, primary health centers, system that are available and how the system is work- and inpatient hospitals. Because of the substantial ing is essential to inform which interventions should global variability in the rehabilitation capacity of first- be offered and how best to deliver them. level and referral hospitals, no differentiation is made • Rehabilitation in Health Systems between these settings. The service delivery platforms The publication Rehabilitation in Health Systems do not correspond with the providers’ level of expertise; outlines nine fundamental recommendations for some community and primary health center–based strengthening rehabilitation in health systems (WHO interventions (such as recommendations for specific 2017). The recommendations highlight the strong environmental modifications and cognitive interven- need for rehabilitation to be integrated across all levels tions) should be delivered by specialist providers. The of the health system, as well as the need for finan- package reflects the necessity of providing rehabilitation cial allocation to ensure sustainable, quality service in both the community and hospital settings. Delivery delivery. of the intervention is not restricted to the service deliv- ery platform under which it is positioned; this position- Further information on both resources, as well as ing reflects only the intervention’s typical point of others under development (such as a toolkit for delivery. In particular, the package has been targeted to rehabilitation development), is available at the WHO low-resource health systems; systems with greater rehabilitation website: http://www.who.int/disabilities resource availability should aim to provide the most /care/en/. 292 Disease Control Priorities: Improving Health and Reducing Poverty PRIORITIES FOR ACTION of the rehabilitation workforce, and the financing and monitoring of rehabilitation delivery. Whereas rehabili- The following actions are key for policy makers seeking tation plays a critical role in optimizing health outcomes, to strengthen and extend quality rehabilitation services: advances in the field have lagged those in other areas with comparable effects. Recognizing rehabilitation’s contri- • Establish education and certification pathways for bution to improving functioning and the quality of life dedicated rehabilitation providers. and its importance to the effectiveness of other health • Ensure the availability of appropriately skilled reha- interventions is fundamental to correcting this disparity. bilitation providers in specialized inpatient settings. • Include rehabilitation in national health plans and financing schemes. ANNEX • Ensure that health insurance (where it exists or is to The annex to this chapter is as follows. It is available at be implemented) covers rehabilitation interventions. http://www.dcp-3.org/DCP. • Integrate rehabilitation into both community- and hospital-based health services. • Annex 15A. Glossary of Rehabilitation Intervention • Implement financial and procurement policies to Terminology ensure that high-quality assistive products (as well as training in their proper use) are available to all who need them. ACKNOWLEDGMENTS The following people contributed to the development Research Priorities of the Essential Package of Rehabilitation Interventions through research and peer review: Li-Rong Cheng Critical gaps exist in the evidence base for rehabilitation. (International Association of Logopedics and Phoniatrics), A substantial increase in research is urgently needed to Christoph Gutenbunner and Boya Nugraha (International guide priority setting for system planning and to increase Society of Physical and Rehabilitation Medicine), Kaloyan the availability and effectiveness of rehabilitation services. Kamenov (WHO and Instituto de Salud Carlos III, Centro Several of the research priorities included in the WHO’s de Investigación Biomédica en Red, CIBERSAM), Pauline Rehabilitation in Health Systems (WHO 2017) are partic- Kleinitz (Ludwig-Maximillians University), Ritchard ularly pertinent to rehabilitation policy: Ledgerd (World Federation of Occupational Therapists), Chiara Retis (Handicap International), and Catherine • Research to ascertain the cost benefit of rehabilitation Sykes (World Confederation of Physical Therapy). • Research to identify facilitators and barriers to access- ing rehabilitation • Research to enable a standardized measure of reha- NOTES bilitation effect. World Bank Income Classifications as of July 2014 are as follows, based on estimates of gross national income (GNI) per capita for 2013: CONCLUSIONS Given the increasing demand for rehabilitation around • Low-income countries (LICs) = US$1,045 or less the world, the need to extend the availability of essential • Middle-income countries (MICs) are subdivided: rehabilitation interventions is urgent. Commendable (a) lower-middle-income = US$1,046 to US$4,125 efforts in several LMICs demonstrate the feasibility of (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 • High-income countries (HICs) = US$12,746 or more. improving rehabilitation capacity and performance in resource-poor settings. The DCP3 package of essential 1. The WHO Rehabilitation System Assessment tool is not rehabilitation interventions is designed to help scale up publicly available but is provided by the WHO on request rehabilitation services to reach those who need them when appropriate. Contact details are located on the home most. The package is informed by expert consensus and page: http://www.who.int/disabilities/care/en. the limited evidence base available. As further evidence emerges, future iterations may reflect changes to the package of interventions. REFERENCES To have the greatest effect on population health, care- Aisen, M. L., D. Kerkovich, J. Mast, S. Mulroy, T. A. Wren, and ful attention needs to be given to the systems that deliver others. 2011. “Cerebral Palsy: Clinical Care and Neurological rehabilitation services, the training and skills Rehabilitation.” The Lancet Neurology 10 (9): 844–52. Rehabilitation: Essential along the Continuum of Care 293 Anchique Santos, C. V., F. Lopez-Jimenez, B. Benaim, Lower Limb Amputation from an Under Resourced G. Burdiat, R. Fernandez Coronado, and others. 2014. Community: A Randomized Controlled Trial.” Physiotherapy “Cardiac Rehabilitation in Latin America.” Progression in 101 (1): e458. Cardiovascular Diseases 57 (3): 268–78. Gosselin, R., D. A. Spiel, R. Coughlin, and L. G. Zirkle. 2009. Atijosan, O., V. Simms, H. Kuper, D. Rischewski, and C. Lavy. “Injuries: The Neglected Burden in Developing Countries.” 2009. “The Orthopaedic Needs of Children in Rwanda: Bulletin of the World Health Organization 87: 246. Results from a National Survey and Orthopaedic Service Grantham-McGregor, S., Y. B. Cheung, S. Cueto, P. Glewwe, Implications.” Journal of Pediatric Orthopedics 29 (8): 948–51. L. Richter, and others. 2007. “Developmental Potential in Bardos, M., H. Burak, and Y. Ben-Shalom. 2015. Assessing the the First 5 Years for Children in Developing Countries.” Cost and Benefits of Return-to-Work Programs. Consultant The Lancet 369 (9555): 60–70. report by Mathematica Policy Research for the U.S. Gupta, N., C. Castillo-Laborde, and M. D. Landry. 2011. Department of Labor, Washington, DC. https://www.dol “Health-Related Rehabilitation Services: Assessing the .gov/odep/topics/pdf/RTW_Costs-Benefits_2015-03.pdf. Global Supply of and Need for Human Resources.” BMC Beal, R. W. 2007. “Survey of Rehabilitation and Return-to- Health Services Research 11 (1): 1. Work Practices among U.S. Disability Carriers.” Report pre- Guzman, J. M., and E. G. Salazar. 2014. “Disability and pared by Milliman Consultants and Actuaries for Cornell Rehabilitation in Mexico.” American Journal of Physical University School of Industrial and Labor Relations, Ithaca, Medicine & Rehabilitation 29 (1 Suppl 1): 36–38. NY. http://digitalcommons.ilr.cornell.edu/cgi/viewcontent Hyman, S., R. Parikh, P. Y. Collins, and V. Patel. 2015. “Adult .cgi?article=1445&context=gladnetcollect. Mental Disorders.” In Disease Control Priorities (third Beswick, A. D., K. Rees, P. Dieppe, S. Ayis, R. Gooberman-Hill, edition): Volume 4, Mental, Neurological, and Substance and others. 2008. “Complex Interventions to Improve Use Disorders, edited by V. Patel, D. Chisholm, T. Dua, Physical Function and Maintain Independent Living in R. Laxminarayan, and M. E. Medina-Mora. Washington, Elderly People: A Systematic Review and Meta-Analysis.” DC: World Bank. The Lancet 371 (9614): 725–35. Jolliffe, J. A., K. Rees, R. S. Taylor, D. Thompson, N. Oldridge, Choi, J. H., M. Jakob, C. Stapf, R. S. Marshall, A. Hartmann, and others. 2000. “Exercise-Based Rehabilitation for and H. Mast. 2008. “Multimodal Early Rehabilitation and Coronary Heart Disease.” Cochrane Database of Systematic Predictors of Outcome in Survivors of Severe Traumatic Reviews (4): Cd001800. Brain Injury.” Journal of Trauma 65 (5): 1028–35. Kurichi, J. E., D. S. Small, B. E. Bates, J. A. Prvu-Bettger, P. L. Dalal, H. M., P. Doherty, S. Rod, and R. S. Taylor. 2015. “Cardiac Kwong, and others. 2009. “Possible Incremental Benefits Rehabilitation.” BMJ 351: h5000. of Specialized Rehabilitation Bed Units among Veterans Dias, J. M., B. F. Mazuquin, F. Q. Mostagi, T. B. Lima, M. A. after Lower Extremity Amputation.” Medical Care 47 (4): Silva, and others. 2013. “The Effectiveness of Postoperative 457–65. Physical Therapy Treatment in Patients Who Have Mock, C. N., S. Nguyen, R. Quansah, C. Arreola-Risa, R. Viradia, Undergone Arthroscopic Partial Meniscectomy: Systematic and M. Joshipura. 2006. “Evaluation of Trauma Care Review with Meta-Analysis.” Journal of Orthopaedic & Capabilities in Four Countries Using the WHO-IATSIC Sports Physical Therapy 43 (8): 560–76. Guidelines for Essential Trauma Care.” World Journal of European Agency for Safety and Health at Work. 2016. Research Surgery 30: 946-56. doi:10.1007/s00268-005-0768-4, PMID Review on Rehabilitation and Return to Work. Luxembourg: :16736320. European Union. Na, W. 2016. “Cost-Benefit Analysis of Vocational Rehabilitation Fleury, A. M., S. A. Salih, and N. M. Peel. 2013. “Rehabilitation Program for Persons with Significant Disabilities in Korea.” of the Older Vascular Amputee: A Review of the Literature.” Niepełnosprawność—Zagadnienia, Problemy, Rozwiązania Geriatrics & Gerontology International 13 (2): 264–73. 2 (19): 7–15. Franche, R. L., K. Cullen, J. Clarke, E. Irvin, S. Sinclair, and others. Nas, K., L. Yazmalar, V. Şah, A. Aydin, and K. Öneş. 2015. 2005. “Workplace-Based Return-to-Work Interventions: A “Rehabilitation of Spinal Cord Injuries.” World Journal of Systematic Review of the Quantitative Literature.” Journal Orthopedics 6 (1): 8. of Occupational Rehabilitation 15 (4): 607–31. Nielsen, P. R., J. Andreasen, M. Asmussen, and H. Tonnesen. GBD 2015 Disease and Injury Incidence and Prevalence 2008. “Costs and Quality of Life for Prehabilitation and Collaborators. 2016. “Global, Regional, and National Early Rehabilitation after Surgery of the Lumbar Spine.” Incidence, Prevalence, and Years Lived with Disability BMC Health Services Research 8: 209. for 310 Diseases and Injuries, 1990–2015: A Systematic Novak, I., S. Mcintyre, C. Morgan, L. Campbell, L. Dark, Analysis for the Global Burden of Disease Study 2015.” and others. 2013. “A Systematic Review of Interventions The Lancet 388: 1545–602. for Children with Cerebral Palsy: State of the Evidence.” Gillespie, L. D., M. C. Robertson, W. J. Gillespie, C. Sherrington, Developmental Medicine & Child Neurology 55 (10): S. Gates, and others. 2012. “Interventions for Preventing 885–910. Falls in Older People Living in the Community.” Cochrane Oldridge, N. B., M. T. Pakosh, and R. J. Thomas. 2016. “Cardiac Database of Systematic Reviews (9): CD007146. Rehabilitation in Low- and Middle-Income Countries: Godlwana, L., A. Stewart, and E. Musenge. 2015. “Mobility A Review on Cost and Cost-Effectiveness.” International during the Intermediate Stage of Rehabilitation after Health 8 (2): 77–82. 294 Disease Control Priorities: Improving Health and Reducing Poverty Parker, A., T. Sricharoenchai, and D. M. Needham. 2013. Turner-Stokes, L., H. Williams, A. Bill, P. Bassett, and “Early Rehabilitation in the Intensive Care Unit: Preventing K. Sephton. 2016. “Cost-Efficiency of Specialist Inpatient Impairment of Physical and Mental Health.” Current Rehabilitation for Working-Aged Adults with Complex Physical Medicine and Rehabilitation Reports 14: 307–14. Neurological Disabilities: A Multicentre Cohort Analysis of Proctor, F. 2010. “Rehabilitation of the Burn Patient.” Indian a National Clinical Dataset.” BMJ Open 6 (2). Journal of Plastic Surgery 43 (Suppl): S101–13. UN (United Nations). 2015. World Population Ageing 2015. Scivoletto, G., B. Morganti, and M. Molinari. 2005. “Early New York: Department of Economic and Social Affairs, UN. versus Delayed Inpatient Spinal Cord Injury Rehabilitation: WHO (World Health Organization). 2001. International An Italian Study.” Archives of Physical and Rehabilitation Classification of Functioning, Disability and Health: ICF. Medicine 86 (3): 512–16. Geneva: WHO. Scott, J. G., C. Mihalopoulos, H. E. Erkskine, J. Roberts, and ———. 2014. “Injuries and Violence: The Facts 2014.” WHO, A. Rahman. 2015. “Childhood Mental and Developmental Geneva. Disorders.” In Disease Control Priorities (third edition): ———. 2015. World Report on Ageing and Health. Geneva: WHO. Volume 4, Mental, Neurological, and Substance Use Disorders, ———. 2016a. World Health Statistics 2016: Monitoring Health edited by V. Patel, D. Chisholm, T. Dua, R. Laxminarayan, for the SDGs. Geneva: WHO. and M. E. Medina-Mora. Washington, DC: World Bank. ———. 2016b. “Global Health Observatory (GHO) Data.” Stroke Unit Trialists’ Collaboration. 2013. “Organised Inpatient http://www.who.int/gho/child_health/mortality/mortality (Stroke Unit) Care for Stroke.” Cochrane Database of _under_five_text/en/. Systematic Reviews 9: Cd000197. ———. 2016c. International Classification of Health Stucki, G., M. Stier-Jarmer, E. Grill, and J. Melvin. 2005. Interventions. Geneva: WHO. http://www.who.int “Rationale and Principles of Early Rehabilitation Care after /classifications/ichi/en/. an Acute Injury or Illness.” Disability and Rehabilitation ———. 2017. Rehabilitation in Health Systems. Geneva: WHO. 27 (7–8): 353–59. WHO (World Health Organization) and World Bank. 2011. Taylor, R. S., H. Dalal, K. Jolly, T. Moxham, and A. Zawada. 2010. World Report on Disability. Geneva: WHO. “Home-Based versus Centre-Based Cardiac Rehabilitation.” World Bank. 2016. “Population Ages 0–14 (% of Total).” http:// Cochrane Database of Systematic Reviews (1): Cd007130. data.worldbank.org/indicator/SP.POP.0014.TO.ZS. Rehabilitation: Essential along the Continuum of Care 295 Part 5 Intersectoral and International Topics Chapter 16 Development Assistance for Health Eran Bendavid, Trygve Ottersen, Liu Peilong, Rachel Nugent, Nancy Padian, John-Arne Rottingen, and Marco Schäferhoff INTRODUCTION patterns of sources, channels, flows, and targets of donor Development assistance holds promise for alleviating the resources—is available from other sources, which we death and suffering of impoverished children, women, and reference throughout this chapter. Instead, we address men from readily preventable and treatable conditions key questions that challenge our understanding of the and to support global economic development, demo- present and planning for the future of international graphic sustainability, and political stability. Although cooperation on health. the desirability of these goals is widely shared, there is little The first section addresses the measurement of health agreement on who should shoulder the financial responsi- aid, including an overview of common definitions and bility or how best to use development assistance to achieve measurements of how health aid flows, from whom, to these goals. whom, and to what intended ends. The section also sum- How much financing should be provided and in what marizes recent efforts to reconsider the scope of health form, who is eligible, and what health areas and interven- aid, including aid originating in non–Organisation for tions should be prioritized? How should institutions Economic Co-operation and Development (OECD) coun- balance the financing for current interventions and for tries and support for R&D and other global public goods. future priorities? Should funding for research and devel- The second section addresses the normative land- opment (R&D) be a health aid priority? And what exactly scape of health aid: What are the goals for the provision counts as health aid? Does a favorable loan to build a of health aid and the criteria guiding its allocation? hospital in rural China count? How about in rural Mali? We illustrate the role of the implicit and explicit goals of How much health aid flows through recognized chan- health aid, including the alleviation of death and suffer- nels, and how much falls outside well-documented ing, human development, national relationships, global channels? What criteria should be used to allocate scarce health equity, and international security. We also address health aid resources? Which countries and populations how implicit and explicit goals guide the provision of have the strongest claims to assistance or favorable health aid across regions and countries and across dis- financing? This chapter provides frameworks for address- ease and intervention areas. ing these questions and understanding the crossroads for The third section provides two case studies that illus- foreign aid to the health sector. This chapter does not trate patterns of health aid sources and the breadth of provide a systematic review of current patterns of health aid health aid efforts. The fourth draws lessons learned from allocation. The descriptive epidemiology of health aid—the the experience with health aid and identifies guiding 299 principles for organizing and implementing health aid producing different results (IHME 2016a; PMNCH 2014; resources. We end with a summary and recommendations Victora and others 2015). In addition, information on the for future health aid investments. Investing these resources purpose of a grant or loan in the CRS can be vague or wisely will play an important role in achieving a grand short, often no longer than a few words, making it diffi- convergence in global health and a decent life for all cult to link the resources with their intended priorities. (Jamison and others 2013). The second data source contains development assis- tance for health, a term introduced by the Institute for Health Metrics and Evaluation (IHME) in 2009 that TRACKING HEALTH AID quickly became commonly used in the global health Health aid can be broadly defined as the transfer of community. In contrast to ODA, development assistance resources from multilateral organizations, foundations, or for health includes financing from private sources and governments to the health sector of a country or a financial transfers that target the private sector such population. Although much health aid is in the form of as advance market commitments. Unlike the CRS, grants and in-kind gifts, some is in the form of concession- which contains project-level information on more than ary agreements, loans, and preferable trade agreements. 3 million projects, IHME’s data contain global health Beneath the broad definitions, however, lie several major financing data aggregated by source, channel, recipient, challenges to the definitions and measurement of health aid. and disease areas (IHME 2016a). Largely lacking from both the CRS and IHME data- bases is information on investments in global public Definitions goods. Such goods include R&D for diseases that dispro- An important challenge to any discussion on prioritiza- portionately affect people living in poor countries or for tion of health aid is the lack of agreement on what priorities with global benefits such as epidemic outbreak exactly counts as health aid. This section describes two preparedness. The extent to which investments in global data sources that track health aid and highlights the public goods should count as health aid is an area of differences between them. active debate. The concept of health ODA plus attempts The most detailed source of data on health aid comes to provide a more complete picture of donor flows to from the OECD’s Development Assistance Committee global health by including flows to global functions (DAC). The DAC is charged with tracking and measur- (Schäferhoff and others 2015). Specifically, health ODA ing all forms of donor financing, including official devel- plus includes (1) health aid reported by donors to the opment assistance (ODA).1 ODA includes mostly grants OECD and (2) public funding for R&D for neglected and loans that are concessional in character and contain diseases, including funding channeled by donors to orga- a grant element of at least 25 percent. The OECD main- nizations working on R&D without a specific focus on tains the Creditor Reporting System (CRS), a database of low- and middle-income countries (LMICs). For exam- ODA coded into 36 broad sectors, including two sectors ple, ODA plus includes financing by the Swedish govern- that are noted as “health” and “population and repro- ment for research on antimicrobial resistance through ductive health.” The database contains information on the Karolinska Institute, as well as financing of research grants and loans starting as early as 1973, but health aid on a vaccine to prevent human immunodeficiency data are sparse before 2000. Thirty donor nations virus/acquired immunodeficiency syndrome (HIV/ (mostly members of the DAC); several multilateral AIDS) through federal institutions such as the National organizations, such as the World Health Organization Institutes of Health or private pharmaceutical and bio- (WHO) and the African Development Bank; as well as technology companies. Because these resources do not the Bill & Melinda Gates Foundation provide specific flow to LMICs and support research priorities with information about the purpose, amount, and intended global importance, this funding is not included in the recipients of grants and loans qualifying as development CRS and IHME databases. The concept of health ODA assistance. plus posits that funding for health priorities that affect Because the CRS database has become an important lower-income countries is a valuable component of public source of health aid data, its limitations deserve health aid even if it does not flow directly to LMICs (and further mention. First, information about the purpose of adds to the complexity of health aid measurements). an aid item may be too general for many health purposes. What other types of assistance are not measured or For example, the CRS lacks a code for reproductive, tracked reliably? Aid to priority areas such as neglected maternal, newborn, and child health. As a result, recent tropical diseases may be substantially underrepresented initiatives that aim to monitor the international flow of in the CRS and other data sets. Aid for noncommunicable resources for such priorities use different measurements, diseases (NCDs) is not officially represented at all. 300 Disease Control Priorities: Improving Health and Reducing Poverty Health aid from non-OECD countries such as Brazil, needs in the recipient country. The dependence of health China, India, and South Africa is substantial, but these aid on nonhealth priorities and dynamics that may be countries do not report to the DAC. Data on South- entirely exogenous to events in the recipient country South cooperation are hard to track and may include makes health aid particularly vulnerable to swings. The items not considered aid by other definitions, even impact of volatility may be particularly detrimental to though it makes up an important supplement to more funding streams that finance health services with few established forms of assistance (described in case study 2 substitutes and long-term commitments, such as antiret- in this chapter). roviral therapy. In addition, substantial amounts of aid, including As shown in figure 16.1, substantial variation was health aid, flow between Arab nations and territories. experienced in the global increase in health aid. Between Data on the magnitude or nature of this aid are very 2002 and 2014, some countries have given increasing limited. Kuwait and the United Arab Emirates report to amounts of health aid (for example, the United States), the CRS, and Qatar publishes aggregate information on while others have given stable or declining amounts aid provided mostly to other Arab nations (Kharas (Norway). 2015). The United Arab Emirates has been increasing its Global political and economic cycles also shape ODA contributions since 2010, including a 608 percent donor priorities, with recessions leading governments increase in real terms in 2013 (mainly support to the to rearrange their spending priorities, often in ways that Arab Republic of Egypt), of which a little less than do not favor foreign aid. Following the 2007–08 reces- 10 percent is designated for health. Support for health sion, the 2010 health aid budget of OECD countries multilaterals is also growing. For example, Oman, Qatar, became more volatile and that of the United States Saudi Arabia, and the United Arab Emirates all provide stagnated. In 2005, several OECD countries committed funding to Gavi, the Vaccine Alliance. Changing the CRS to tethering foreign aid—including health aid—to a database to include reporting from non-DAC donor portion (0.7 percent) of their gross domestic product. countries could add to our understanding of interna- If the recommendation to peg foreign aid to gross tional cooperation and unmet needs. There is precedent domestic product is followed, health aid will grow for expanding the CRS database: several non-DAC during economic booms and shrink during economic donors already report ODA funding to the CRS, includ- downturns, making future levels of commitment highly ing Lithuania, Saudi Arabia, and Thailand. Expanding uncertain. Following the 2016 elections in the United the global accounting of ODA—including health aid— States, the new administration expressed decreasing would relieve areas of great uncertainty in deciding on commitment to foreign aid programs, including explicit resource allocations to countries and regions with lim- large reductions in health aid. ited information. For example, the resource flows to Private sources and foundations have been playing an conflict regions such as Syria from non-OECD sources increasingly important role in the health aid landscape. are largely unknown and likely substantial. Overall, private sources made up more than 25 percent of health aid between 2010 and 2015, a relatively small com- ponent of direct contributions. The Bill & Melinda Gates Sources and Flows of Health Aid Foundation was the third-largest overall contributor of Growth in health aid slowed considerably between 2010 health aid between 2010 and 2015, above most European and 2015. During the “golden age of global health” countries (IHME 2016a). However, the influence of pri- (2000–10), health aid grew 11.4 percent a year on average. vate aid transcends its direct financial contribution. Since then, average growth has dropped to 1.2 percent. Unencumbered by public interests, organizations such as In 2015, health aid (as measured by the IHME) totaled the Bill & Melinda Gates Foundation and the Hewlett US$36.4 billion, below the 2013 peak level of US$38.0 Foundation are at relative liberty to take strategic risks and billion (see IHME 2016a). Much donor support for set new agendas. The Bill & Melinda Gates Foundation, health originates in national budgets supported by taxes for example, provided critical early financing to Gavi, the and other sources of national income in wealthier coun- Vaccine Alliance, and its emphasis on financing novel tries. Unlike domestic health spending that originates technological solutions for global health problems has from governments that are, in many cases, accountable to generated funding for more than 1,000 exploratory high- the populations they serve, health aid is unstable. Because risk, high-reward projects such as farming grasshoppers a donor government does not have the same fiduciary as a source of protein or developing odorants to block the relationship to the population of another country as it ability of malaria-causing Anopheles mosquitoes to does to its own population, health aid amounts and pri- detect humans. The Hewlett Foundation has been a orities may shift for reasons that have little to do with leader in advancing rigorous program evaluations to Development Assistance for Health 301 Figure 16.1 Health Aid as a Percentage of Total Aid for Major Country Donors, 2002–14 0.35 0.30 US, 0.288 Overall gross ODA to health, percent 0.25 0.20 Canada, 0.204 UK, 0.178 Ireland, 0.174 0.15 Norway, 0.142 Australia, 0.120 0.10 Netherlands, 0.105 Sweden, 0.090 France, 0.106 Germany, 0.059 0.05 Japan, 0.054 0 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Source: Organisation for Economic Co-operation and Development Creditor Reporting System, gross disbursements in constant prices (sector codes 120 and 130) and imputed multilateral contributions to the health sector. Development Assistance Committee Secretariat estimates (as of January 2016). Note: ODA = overseas development assistance. understand what works, such as through its support for of several global initiatives, such as the Group of Eight’s the International Initiative for Impact Evaluation and Muskoka Initiative on Maternal, Newborn and Child leadership of the Effective Philanthropy Group. Health.) Health areas not targeted by the MDGs have More than 300 foundations were registered with the received even less attention. These overlooked condi- U.S. Agency for International Development (USAID) tions include NCDs such as cardiovascular disease and in 2014, and most of them operate independently cancers as well as neglected tropical diseases, even (USAID 2015). Their portfolios can be wide, including though the burden of these diseases is large in many infectious diseases, reproductive health, and complemen- aid-recipient countries. tary areas such as education, health systems, and gover- Of the total amount of health aid in 2015, 30 percent nance. One implication of the relatively small size of each and 28 percent were allocated to HIV/AIDS and to foundation and their independent operation is that maternal, child, and newborn health, respectively, while foundations commonly identify their own (often nar- 6 percent was targeted to malaria control, and only row) strategic focus rather than align their investments 1 percent to NCDs, even though NCDs are responsible for within a streamlined, global strategy. more deaths than any other major category in every Health financing is distributed unevenly across health region except Sub-Saharan Africa (Dieleman and others areas. Since 2000, the launch year of the Millennium 2015). Box 16.1 discusses this issue in greater detail. Development Goals (MDGs), the largest growth in Financing has shifted slightly with the launch of global health aid funding has been related to the control of initiatives focusing on child health, maternal health, and infectious diseases, particularly HIV/AIDS and malaria. nutrition (Darmstadt and others 2014; Kirton, Kulik, Child health and especially maternal and reproductive and Bracht 2014). Mirroring these shifts, health aid for health have received more modest attention. (This trend HIV/AIDS and tuberculosis has declined from peak levels has changed somewhat since 2010 following the launch in 2013 (IHME 2016a). 302 Disease Control Priorities: Improving Health and Reducing Poverty Box 16.1 Funding for Noncommunicable Diseases Unlike in many other areas of health, households bear If health aid declines or stagnates in the coming years, much of the burden of noncommunicable diseases (NCDs). domestic governments will have to provide the bulk of Governments in low- and middle-income countries have new funding. The following actions could help align NCD allocated very little to NCD prevention and care. More than funding with needs: 50 percent of current spending for cardiovascular diseases in low-income countries is out of pocket from patients and • Aim for a closer alignment of funders’ health aid with their households, 33 percent is from domestic governments, health burden in poor countries and 13 percent is from donors; in high-income countries, • Link funders’ priorities with NCD prevention and out-of-pocket spending on NCDs is a far lower share of the treatment programs—for example, integrate NCD total (WHO n.d.). Government financing for NCDs also prevention, such as blood pressure management, into varies substantially across countries. primary care settings Figure B16.1.1 provides estimates of development assistance • Link investments in health system strengthening with for NCDs and all health aid from 2000 to 2014. investments in NCD prevention. Figure B16.1.1 Development Assistance for Noncommunicable Diseases and All Health, 2000–14, 2011 US$ 700 40,000 600 35,000 All health funding (US$, millions) 30,000 NCD funding (US$, millions) 500 25,000 400 20,000 300 15,000 200 10,000 100 5,000 0 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 NCD funding 160 181 219 179 217 251 295 334 442 458 474 528 515 608 611 Total health funding 11,601 12,026 13,821 15,859 18,057 19,965 21,886 25,194 29,236 30,120 33,935 34,912 33,129 36,456 35,890 NCD funding as % of total health funding 1.38 1.51 1.58 1.13 1.20 1.26 1.35 1.33 1.51 1.52 1.40 1.51 1.55 1.67 1.70 Source: IHME 2016a. Note: NCD = noncommunicable disease. Development Assistance for Health 303 GOALS AND CRITERIA FOR ALLOCATING Criteria for Allocation across Geographic Areas HEALTH AID Guiding the allocation of health aid across countries or geographic areas is often of importance to donors. Donors’ and recipients’ normative views and goals Recent and ongoing economic transitions, however, have inherently shape decisions about whether to provide aid, made decisions about country allocation more difficult how much, in what form, to whom, toward what, and for donors seeking to direct health aid toward individu- how (Centre on Global Health Security Working Group als or communities with large needs relative to their on Health Financing 2014). These views and goals capacity (rather than to countries that may have large underpin the variation in health aid across countries and relatively well-off populations). Economic growth rates partly explain, for example, why health aid per capita have been impressive in many countries, including many ranges from US$0.7 to US$32 in LMICs (IHME 2016a). formerly low-income countries (LICs), over the past two This section examines stated and unstated goals under- decades, and many countries have moved from low- lying the allocation of health aid and discusses criteria income to middle-income status, including populous for guiding the allocation of health aid resources across countries such as China, India, and Nigeria. At the same geographic and health areas. time, many of these countries have pronounced inequal- ities in income and health. One consequence is that most of the world’s poor and the world’s disease burden are no Goals of Health Aid longer located in LICs, but in middle-income countries Averting preventable deaths and suffering, especially in (MICs) (IHME 2016b; Sumner 2012). countries with limited domestic capacity to address Questions arise about the role of MICs with regard to health needs, is a shared goal of health aid providers and health aid and, more generally, the central role currently recipients. For example, the mission of the Global given to mean national income, such as gross national Fund is to invest the world’s money to defeat HIV/AIDS, income (GNI) per capita, in allocation decisions. Agreement tuberculosis, and malaria, and the global health mission is growing that GNI per capita is an inadequate basis for of USAID is to support partner countries in preventing deciding which countries are eligible for health aid and and managing major health challenges of poor, under- how much each country should receive. Therefore, with served, and vulnerable people (Global Fund 2016a; respect to cross-country allocation of health aid, a central USAID 2012). Between 2000 and 2015, many donors task for many donors in the coming years will be to con- also explicitly sought to help countries reach the sider resituating GNI per capita as one tool among several MDGs on child mortality (MDG 4); maternal health in the overall decision-making process. (MDG 5); and HIV/AIDS, tuberculosis, and other major The larger debate about how health aid can better diseases (MDG 6) (Ravishankar and others 2009). These target the communities and individuals in greatest need priorities and funding streams remained dominant even revolves around three broad approaches. One is deter- after the deadline for the MDG 2 at the end of 2015. mining whether GNI per capita thresholds should be At the same time, many donors cite broader goals for used at all to determine eligibility for health aid. Many health aid, including goals related to poverty alleviation, have called for donors to raise their thresholds, in effect economic growth, educational outcomes, and security. reducing the role GNI per capita plays in determining Starting in 2016 with the Sustainable Development eligibility. What other criteria should be used if GNI per Goals, health-related aims could be further integrated capita does not provide an eligibility benchmark remains with broader development objectives. an open issue. Second, others argue for maintaining Donors may also have goals that have less to do with GNI per capita as a criterion, but supplementing it with recipient need and more to do with donor interests. criteria directly linked to health needs in the country. For These goals can occasionally be gleaned from revealed example, the Global Fund hosted the Equitable Access donor preferences without being made explicit. For Initiative in 2015, which concluded that countries’ health example, some donors provide health aid to protect their needs and fiscal capacity are important factors for own populations, such as targeting rapidly spreading donors to consider when allocating funds (Global Fund infectious diseases, like Ebola virus disease; or to promote 2016b). Again, the specific metrics to use and how to their political and economic interests (Berthélemy 2006; integrate them remain open issues. Finally, some suggest Hoeffler and Outram 2011). Irrespective of whether that donors ought to go beyond countries and average explicit or implicit goals are pursued, the Paris measures such as GNI per capita and focus more on the Declaration on Aid Effectiveness calls for donors to align subnational allocation of health aid. Options for linking their support, whenever possible, with recipient-country eligibility and other allocation criteria directly to subna- government priorities. tional units need more study. 304 Disease Control Priorities: Improving Health and Reducing Poverty Criteria for Allocation across Health Areas Epidemiological and other transitions are creating Allocation across disease and health priorities requires new challenges for allocating health aid across disease additional consideration. Health aid resources cannot areas. NCDs now account for almost 60 percent of the fully subsidize the health sector of even the poorest global burden of disease (Murray and others 2015), and countries, and decisions for prioritizing disease areas 80 percent of NCD deaths occur in LMICs. Donors need and programs are unavoidable. In 2014, US$10.8 billion to carefully balance their responses to NCDs with their and US$10.1 billion were allocated to HIV/AIDS and responses to maternal, neonatal, and child health prob- maternal and child health, respectively, while US$2.4 lems and with the unfinished agenda of infectious dis- billion was allocated to malaria and only US$475 million eases. Weighing these choices may involve further inquiry to NCDs (IHME 2016a). What accounts for such varia- into how criteria related to cost-effectiveness, disease tion? What principles appear to guide—and ought to burden, and the worse off can be specified and traded guide—the distribution of health aid? off. At the same time, transnational health threats, Although some donors clearly state their general pri- including pandemics and antimicrobial resistance, are orities, few provide the explicit criteria used to allocate increasingly being viewed as within the purview of health aid across disease areas. Perhaps the most straight- health aid. Chapter 18 of Major Infectious Diseases forward way to prioritize financing decisions would be to (volume 6 of this series) on antimicrobial infections allocate resources in proportion to the burden of disease provides additional arguments supporting the role of such that if the death and disability from disease A is twice health aid in curbing antimicrobial resistance (Miller- that from disease B, then twice the resources should go Petrie, Pant, and Laxminarayan 2017). What share of toward controlling disease A (Sridhar and Batniji 2008). health aid should be allocated to these kinds of threats While the equitability of this resource-allocation heuristic will be a key question. The interpretation and generation is appealing, its principal shortcoming is that, without of cost-effectiveness estimates for interventions in these considering the cost of reducing disease burden, alloca- areas will also be important because such estimates are tion proportional to burden may not reduce as much currently lacking or are highly uncertain. disease burden as prioritizing diseases for which the most cost-effective interventions exist. Disease burden esti- mates can be useful for identifying the conditions causing CASE STUDIES the most mortality and morbidity, but they do not show This section presents two case studies illustrating the where health aid resources could yield the greatest bene- historical trajectory of health aid and the changing land- fits (Bendavid and others 2015). For example, stroke is a scape of donor-recipient relationships. The first describes leading cause of death and disability in China, but financ- the role played by the U.S. President’s Emergency Plan ing stroke treatment in China may yield relatively few for AIDS Relief (PEPFAR) and illustrates the tensions in benefits in comparison with treating and preventing setting priorities and strategies with ambiguous goals tuberculosis (Coyle and others 2013; Prabhakaran, Ruff, and motivations. While the PEPFAR case study delves and Bernstein 2015). To identify the investment priorities into the challenges of archetypal health aid institutions, that provide the greatest benefits with the available the second case study—describing China’s approach health aid resources, information is needed on the cost- to development cooperation on health (South-South effectiveness of potential interventions. One of the cooperation)—represents a complementary approach to principal objectives of the third edition of Disease Control health aid. Priorities is to provide this information. A third proposed criterion for choosing disease prior- ities for health aid (in addition to disease burden and Case Study 1: PEPFAR cost-effectiveness of interventions) would be to provide The spread of HIV/AIDS in Sub-Saharan Africa and the resources to the diseases the afflict the most ill, globally or United States in the 1980s and 1990s preceded—and nationally (Ottersen and others 2014). For example, arguably caused—the expansion of health aid in the priority could be assigned to interventions benefiting 1990s and 2000s. Health aid for HIV/AIDS increased persons with lower healthy life expectancy. Although this from effectively zero in the mid-1990s to the largest sin- criterion might yield different allocation guidance than gle disease priority a decade later. The rapid global a cost-effectiveness criterion, many interventions will response was related to the spread of HIV/AIDS in score high on both—for example, cheap and highly effec- Europe and the United States, where it became the lead- tive interventions targeting potentially life-threatening ing cause of death among young men and created a conditions, such as diarrhea, malaria, and pneumonia, groundswell of activism and growing recognition of the in children living in poverty. security and economic threats of infectious diseases in Development Assistance for Health 305 an increasingly globalized world. The United Nations However, many of PEPFAR’s U.S.-based partners resisted General Assembly Declaration of Commitment on HIV/ the withdrawal of support, resulting in a gradual and AIDS, endorsed in 2001, singled out HIV/AIDS as an (as of 2016) still-incomplete transition of implementa- exceptional priority. tion to local organizations. That exceptionalism was backed by substantial increases Another example of an effort to bridge short-term in commitments and new disbursements toward global and long-term goals is PEPFAR’s support for medical control of HIV/AIDS. The largest of those commitments, education in partner countries. Through a large grant announced in early 2003, became PEPFAR. In this section, program, PEPFAR supported the creation of a dozen we draw on published materials and an interview with a medical training programs in Sub-Saharan African part- former director of the Office of the U.S. Global AIDS ner countries (Fogarty International Center 2015; Kim Coordinator, the agency tasked with implementing and Evans 2014). While this program reflects a commit- PEPFAR, to understand historical and future trajectories ment to creating long-term, in-country capacity, it also of health aid and the challenges of identifying and stand- represents a rethinking of PEPFAR’s original priorities. ing by clear goals and criteria in aid allocation. PEPFAR receives little credit for its attempts to balance PEPFAR changed what was considered possible in short-term targets and long-range vision. These tensions health aid, directing billions of U.S. dollars annually were an integral part of PEPFAR’s implementation. toward a single issue in a small group of high-priority In part because of the need for an epidemic control strat- countries. The model adopted by PEPFAR involved egy that is responsive to a changing epidemic and in part rapid and concentrated deployment of resources as a because of changing leadership, PEPFAR has altered its response to a global public health emergency. The trade- strategy from responding to emergencies to increasing offs of this approach included occasionally downplaying country ownership and integration, and, more recently, to long-term considerations, such as international parity in achieving global public health goals that extend beyond resource allocation, that are more characteristic of mul- HIV/AIDS control (Fauci and Folkers 2012). tilateral organizations like the World Bank or United The challenges facing PEPFAR’s strategic decisions Nations agencies and that may lead to these thinly spread possibly reflect its attempts to balance short-term and organizations’ relatively slow operations. long-term strategic goals. For example, the U.S. Global The program funded implementers with established AIDS Coordinator at the end of the George W. Bush track records, including multilateral U.S.-based organi- administration was replaced swiftly after President zations such as Columbia University, the Elizabeth Obama took office, and the future of PEPFAR was, for a Glaser Pediatric AIDS Foundation, the Harvard School while, highly uncertain (McNeil 2010). By 2017, PEPFAR of Public Health, and Catholic Relief Services. Driven by had matured into an established health aid program expediency, the first phase of implementation included with wide-ranging support and a broad mandate. From capacity building and service provision that largely cir- this position, it could adopt a long-term, stable set of cumvented the public sector in partner countries and guiding principles that could help relieve some of the created a tension that is still evident today: success from pressures to shift strategies in response to leadership and PEPFAR’s perspective meant creating a parallel system of funding changes. health care delivery. This allowed for short-term reduc- Many see an opportunity for PEPFAR to leverage the tion in mortality, but created longer-term challenges. It infrastructure it created to focus on multiple diseases, took several years before PEPFAR prioritized capacity including NCDs, and, in the process, to integrate with building in its partner countries and began moving other health sectors (Fogarty International Center U.S.-based partners to a technical assistance role. That 2016). Although this may be an intuitive direction for model, in which in-country partners were supported to improving the care of HIV/AIDS patients treated in provide health services and the role of U.S.-based part- PEPFAR-supported programs, it also signals a broaden- ners was more advisory, was viewed as more sustainable. ing of PEPFAR’s mandate at the same time that PEPFAR This tension between short-term goals and long-term is poised to deepen its commitment to the highly ambi- vision is evident in many of PEPFAR’s decisions. As tious goals of achieving both “90-90-90” (90 percent recently as 2016, efforts to shift contracts to in-country of persons with HIV/AIDS aware of their status, organizations were met with resistance from the original 90 percent in regular treatment, and 90 percent of those U.S.-based implementers. Shifting to in-country organi- in treatment virally suppressed) and an “AIDS-free zations was thought to enable further scale-up of services generation.” If “90-90-90” is achieved, 55 million indi- (by eliminating the payment of overhead to U.S.-based viduals are estimated to need treatment by 2030, more organizations) and to foster local capacity, sustainabil- than 3.5 times the number of people on treatment ity, and competence (Vermund and others 2012). at the end of 2016 (Hoos, El-Sadr, and Dehne 2016). 306 Disease Control Priorities: Improving Health and Reducing Poverty Successfully broadening and deepening its mandates, for Disease Control and regional medical research cen- possibly with flat or declining resources, is likely to be ters, assisted African countries to improve disease sur- among PEPFAR’s principal challenges. veillance systems, and funded 100 maternal and child health projects for LMICs. China also contributes to the Global Fund; Gavi, the Vaccine Alliance; the WHO; Case Study 2: China’s Contributions to Global Health the African Union; the World Food Programme; and Health aid is an integral part of China’s foreign aid, which the United Nations’ health programs. China’s normative it has been providing for more than 60 years, mostly as approaches to health aid have also evolved, with more South-South partnerships (Zhou, Zhang, and Zhang emphasis on mutually beneficial goals and shared devel- 2015). Beginning in 1950 with aid to socialist neighbor- opment, while emphasizing noninterference in internal ing countries and extending in the mid-1950s to LMICs affairs and avoiding political conditions for aid. in other regions, notably Africa, China has provided a Official data on the financial flows of China’s health large quantity of goods and materials in support of devel- aid are not available. According to Liu and others opment projects.2 After the political and economic reform (2014), between 2007 and 2011, Chinese medical teams in 1978 and the subsequent rise in national income, in Africa were equivalent to about US$60 million in aid China continued to expand the level of foreign aid and annually, donated facilities were about the same, and the diversity of aid forms. As of 2009, China’s total foreign total health aid to Africa averaged about US$150 million aid equaled US$37.6 billion after increasing nearly annually. However, these data include only central gov- 30 percent annually from 2003 to 2009 (China State ernment health aid. They do not include basic salaries Council 2011; Zhou, Zhang, and Zhang 2015). From of medical team members, which are covered by their 2010 to 2012, China contributed an additional US$14.4 home hospitals; support provided by provincial govern- billion in foreign aid (China State Council 2014). During ments to the medical teams they dispatch; scholarships this period, China focused more on LICs; basic infra- for students from LMICs to study medicine in China, structure projects such as roads, ports, and water supply; which are funded by the Ministry of Education; or social projects linked to personal welfare; and technical R&D on neglected tropical diseases, which is funded by training (Zhou, Zhang, and Zhang 2015). other sources. China’s health aid, although a small portion of overall China’s role in health and development is not limited Chinese foreign aid, increased over time, especially to to the direct provision of health aid through bilateral Africa, with the launch of the Forum on China-Africa channels. Since the outbreak of severe acute respiratory Cooperation. Unlike most OECD donors, China does not syndrome, China has participated in global action on offer direct transfers to the health sector. It uses a project health security. China has also engaged in global health approach and provides health aid through grants. China’s policy debates and worked with global health institu- in-kind health aid focuses more on specific aspects of the tions. Although not counted as health aid by most his- health system, such as the delivery of health care services; torical yardsticks, these activities support shared global provision of essential medical products, procedures, and functions with benefits to LMICs. traditional Chinese medicine technologies; improvement of health infrastructure; development of a health work- force; and, more recently, malaria control and emergency GUIDING PRINCIPLES FOR THE response to the Ebola epidemic. The main focus is Africa, NEXT DECADE where almost 90 percent of the dispatched medical teams and 80 percent of donated health facilities—the domi- Health Aid Effectiveness nant forms of China’s regular health aid—are targeted. A growing body of evidence suggests that the surge in China’s variable aid components emerged gradually. health aid, especially since 2000, has helped reduce the In 1963, China first dispatched medical teams with morbidity and mortality from many infectious diseases donated drugs and medical equipment. Since 1970, and the burden of child and maternal mortality in China has constructed health facilities, and in 2000, it many LMICs, occasionally to levels approaching those in launched the Human Resources Development Fund for wealthier regions (Bendavid 2014b; Bendavid and Africa. Since 2006, China has been involved in malaria Bhattacharya 2014). The declines in child mortality dur- control, and in 2014, it provided four rounds of emer- ing the past 30 years coincided with the increase in gency aid, totaling US$120 million, to control the Ebola health aid targeting the causes of child mortality such as outbreak in West Africa. Recently, to support the 2030 vaccine-preventable illnesses. While this supports the Agenda for Sustainable Development, a series of new role of health aid in the decline of child mortality, direct initiatives has helped establish an African Union Center attribution is difficult because child mortality has Development Assistance for Health 307 declined for many reasons. The proliferation of effective countries can finance some if not most health care deliv- organizations committed to expanding the provision of ery for their own populations. In the past 20 years, health highly efficacious, low-cost child health goods, such as aid grew, in part, because many countries did not ade- insecticide-treated bednets and vaccinations, suggests quately finance priorities that donors perceived as urgent that health aid has played an important role, in addition (for example, HIV/AIDS) or exceptionally high value to factors such as economic growth, improved education (for example, vaccinations). However, as countries con- and nutrition, and the diffusion of knowledge such as tinue to develop economically, including many in Latin the benefits of breastfeeding (Levine 2004). Health aid is America, South and South-East Asia, and Sub-Saharan associated with the convergence of mortality rates not Africa, the domestic resources dedicated to supporting only among different countries, but also within countries. health care could grow with, or even faster than, general The geographic and wealth distribution of child mortal- economic growth (Moon and Omole 2013; Resch, ity has been narrowing within most aid-recipient Ryckman, and Hecht 2015). Additional domestic resources countries, most precipitously after 2000, coinciding with could finance goods and services, including child health, the largest rise in health aid (Bendavid 2014a). maternal health, reproductive health, and the prevention and treatment of some infectious diseases such as soil-transmitted helminths and malaria. Changing Aid Commitments Economic development of aid recipients, changing dis- tribution of disease burden, and growing recognition of Which Health Aid Investments Work? the importance of global functions are creating new Health aid would have more impact if resources were conditions and new opportunities that would intuitively guided by evidence of effectiveness and cost-effectiveness. lead to shifts in the allocation and emphasis of health A proliferation of randomized field trials during the past aid. As countries are increasingly able to finance the two decades has added a new layer of specificity to the delivery of health goods, and mortality from causes evidence on what works for health improvements in financed by health aid continue to decline (for example, LMICs. However, similar to the role of randomized clin- vaccine-preventable illnesses or malaria), the allocation ical trials in clinical medicine, the interventions exam- of health aid resources may be better used to address ined in each trial are specific, and the study populations new priorities. may not be broadly representative. This limited general- Outside of a spike in funding earmarked for Ebola izability notwithstanding, the widespread popularity of response, health aid funding remained largely flat randomized controlled trials could point to other ways in between 2010 and 2016. Unless new resources become which evidence could improve health aid. available, any increases in financing of some priorities Randomized evaluations could be incorporated into will require trade-offs and the deprioritization of existing the design of major programs. Currently, most random- high priorities. This is a challenging endeavor for some ized evaluations are organized by academic institu- streams of health aid, where resources are tied up in long- tions and result in attempts to infer generalizable insights term commitments. A striking example of this limited about the process of successful development from flexibility is the financing of antiretroviral therapy (ART) high-quality evidence in specific instances. Despite the for millions of persons living with HIV/AIDS. ART is proliferation of randomized evaluations, however, con- costly, life saving, and lifelong, and efforts to move ART cerns about generalizability of trial insights have only programs from donor to domestic funding have been increased over time (Deaton 2009; Pritchett 2004). met with vociferous resistance (McNeil 2010). A shift in focus would greatly improve their utility: ran- Liberating aid committed from long-term programs domized evaluations could replace traditional monitor- would bring flexibility in responding to new challenges ing and evaluation. Trials provide credible estimates of and opportunities, but the transition will be gradual and the effectiveness of specific interventions and the mech- may not be feasible in the near term. In the meantime, anisms of action. They are less biased than traditional resources could be diverted from low-value priorities monitoring and evaluation and could be streamlined so lacking long-term commitments with relatively low that routine field evaluations could be carried out. Using opportunity cost. It could be expedient to start examin- rigorous evidence to guide the allocation of health aid ing such priorities before tackling entitlements and long- would lend credibility, improve resource allocation, and term commitments. ultimately improve health. Increasing the domestic ownership of health invest- Randomized trials are not the only approach to ments is one way to shift the allocation of health aid discovering “what works.” They are part of a broader commitments. National governments in aid-recipient context of scientific understanding and discovery. 308 Disease Control Priorities: Improving Health and Reducing Poverty For many issues in global health, randomized trials aid recipient believes that the sum total of health aid and may not be feasible for practical or ethical reasons. For domestic resources flowing to the same area exceeds the example, studying the effect of good governance on social optimum. The evidence for health aid displace- health is not readily amenable to randomized assign- ment is consistent with this process (Lu and others ment (Kudamatsu 2012). For such questions, observa- 2010). To prevent or reduce the likelihood of displace- tional analyses are the only way to discover meaningful ment, donors might fund interventions for diseases that insights. The accumulation of evidence is a gradual are relatively underfunded. process, but lessons learned through cumulating evi- Using health aid to fund cost-effective interventions dence have been important in guiding interventions that for underfunded high-burden diseases could yield high save many lives (Glassman and Levine 2016). returns. Local context will determine the appeal of a particular intervention, given that the burden, cost (cost-effectiveness), domestic prioritization, and effec- Identifying Investment Opportunities tiveness of an intervention are locally determined. Future The burden (or projected burden) of disease is a predom- work comparing the appeal of interventions based on inant consideration in choosing new health aid invest- local conditions could have important implications for ments, with high-burden conditions arguably deserving health aid decisions. more attention than low-burden conditions. However, efficient distribution of resources is also needed. To allo- cate resources efficiently, the cost-effectiveness of available Investments in Global Functions interventions must be taken into account. For example, The Lancet Commission on Investing in Health made the coronary bypass surgery may be an efficacious option for case that, as LMICs undergo economic growth, the value a high-burden condition, but it is not cost-effective of health aid investments in “global functions”—that relative to preventing coronary artery disease (Basu, is, the provision of global public goods and protection Bendavid, and Sood 2015). against global cross-border health threats (Jamison and Interventions that are similarly cost-effective may others 2013)—might become more appealing in com- have different effectiveness (and different costs). Decision parison with country-specific investments. This concept makers may have to choose among options that provide has been echoed in several high-impact policy analyses greater benefits to fewer people and similarly cost- (Blanchet and others 2014; Centre on Global Health effective options that provide fewer benefits to more Security Working Group on Health Financing 2014; people. A stylized example is a trade-off between two Frenk and Moon 2013; Ottersen and others 2014). interventions with similar cost-effectiveness. Intervention Based on work by the Lancet Commission on A averts 1.0 disability-adjusted life year per person, while Investing in Health, one study estimated how much intervention B averts only 0.1 disability-adjusted life year donors spend on global functions versus how much they per person; intervention A also costs 10 times more than spend on country-specific support (Schäferhoff and intervention B to treat one person. With a fixed budget, others 2015). Global functions were characterized by choosing intervention A yields the same population-level their ability to address transnational issues and were benefits at the same cost as intervention B, and while only divided into those providing global public goods one-tenth of the people can be treated, people success- (conducting R&D of new health tools, generating and fully treated with intervention A will realize greater gains sharing knowledge), those managing cross-border exter- (on average) than those treated with intervention B nalities (preparing for outbreaks, tackling antimicrobial (Rose 2001). An efficiency (cost-effectiveness) framework resistance), and those fostering leadership and steward- cannot distinguish between the two interventions. The ship (convening leaders to build consensus). Country- greater number of beneficiaries could advantage inter- specific support, in contrast, tackles current health vention B under an equity framework, but the greater priorities that justify international collective action. The effectiveness of intervention A may reduce the uncer- study found that about one-fifth of health ODA plus was tainty about impact, which may be an important consid- for key global health functions, with the rest channeled eration in some circumstances. to country-specific support. Strengthening donor sup- Effectiveness and cost-efficiency are important crite- port for global functions could have several benefits that ria for health aid (Denny and Emanuel 2008), but aid are not immediately obvious. displacement is also a consideration. Health aid flowing First, every country benefits from investments in to disease areas from which domestic resources could global health, and the costs of inaction are potentially easily be diverted is likely to lead to displacement, possi- very high—for example, a severe influenza pandemic bly outside the health sector. This is especially true if the could result in as much as US$3 trillion in global losses Development Assistance for Health 309 (Gostin and Friedman 2015). The returns on investing in how to design and implement taxation polices to increase R&D are potentially among the largest of all investments domestic financing, and how to engage in cross-sectoral in global health, but actual investments in R&D for work, including human rights and education. neglected and poverty-related diseases are limited. For example, a 70 percent efficacious vaccine would reduce new HIV/AIDS infections by 44 percent (Harmon and CONCLUSIONS others 2016), leading to large reductions in incidence and potential epidemic control. The WHO has therefore Health aid is a relatively large component of all health called for a doubling of current R&D expenditures expenditures in LICs and one of the key tools for reduc- for poverty-related and neglected diseases—from ing preventable death and suffering among the world’s US$3 billion to US$6 billion a year, approximately poorest. Several key challenges and opportunities exist 3 percent of total health R&D (Consultative Expert for the future of health aid: Working Group on Research and Development: Financing and Coordination 2012). Market-shaping • Health aid has an opportunity to continue driving activities such as advanced market commitments also health improvements among the poorest. Although have led to important gains, especially in the fields of more deliberate and nuanced allocation is needed, immunization and diagnostics. However, only a small especially across countries, populations, and disease fraction of current health aid has market-shaping effects. areas, opportunities exist for high-impact investments Second, enhanced capacity for global disease surveil- in programs that address high-burden disease, finance lance and detection and improved international coordi- cost-effective interventions, and address domestically nation are important for responding to emerging health underfunded priorities. threats, such as the Ebola outbreak in West Africa. • Donors should clarify and explicitly state their goals and Donors invested less than US$1 billion in 2013 for their criteria for health aid allocation. There are many management of cross-border externalities (including legitimate goals for providing health aid, including outbreak preparedness but also environmental chal- reducing global inequalities, averting preventable lenges and other global threats). In the years leading up human suffering, engaging in self-protection from to the Ebola outbreak, the WHO’s budget for outbreak border-crossing threats, and promoting peaceful and crisis response was cut from US$469 million in national bonds. However, these goals are often only 2012–13 to US$241 million in 2014–15. A pandemic of implicit. Clear standards are needed to align strategy larger proportions could be extraordinarily costly, esti- with goals. In their absence, organizational priorities mated at about US$500 billion per year in losses remain vague, and short-term pressures may move (Fan, Jamison, and Summers 2016). On the other hand, organizations away from their core priorities. Poor implementing a framework to improve preparedness for alignment with core priorities may jeopardize suc- such an event is estimated to cost about US$4.5 billion a cess, which in the case of health aid has important year and could lead to large savings (Sands, Mundaca- human costs because it reduces the potential benefits Shah, and Dzau 2016). to the poorest. Third, investments in global functions would help • As domestic resources rise in LMICs, a growing portion address the “middle-income country dilemma”: although of health care should be financed by domestic resources, most of the poor now live in pockets of poverty in MICs and a declining portion should be financed by health and face high mortality rates, these countries are consid- aid. In other words, many LMICs should require ered to be sufficiently wealthy to finance health care for less health aid as their own domestic resources grow. their entire populations and are therefore commonly not However, such transitions will need to occur carefully eligible for health aid. Poor individuals in MICs would because abrupt shifts may disrupt aid-dependent benefit from donor support for global functions, such as health programs and jeopardize health gains (Isenman R&D, knowledge sharing, market shaping, and better 2015; Katz, Bassett, and Wright 2013). systems for controlling and managing outbreaks. China • Health aid should gradually target global functions. and India, for example, would substantially benefit from Enormous benefits could be gained from the dis- collective purchasing of commodities, market shaping to covery of new vaccines and therapeutics or the reduce drug prices, and international efforts to control design of effective pandemic surveillance systems. multidrug-resistant tuberculosis. These countries As more countries make the transition from health would also benefit from greater global leadership and aid, donor funding could be directed to global dialogue on topics such as how to fight the double functions. This shift would help support poor burden of infectious and noncommunicable diseases, populations in all countries. However, the value of 310 Disease Control Priorities: Improving Health and Reducing Poverty these investments is incompletely understood and Behave the Same?” Review of Development Economics should be a research priority. 10 (2): 179–94. • As the composition of donors, channels, and forms of Blanchet, N., M. Thomas, R. Atun, D. Jamison, F. Knaul, and health aid changes, data systems need to capture a fuller others. 2014. “Global Collective Action in Health: The WDR+ 20 Landscape of Core and Supportive Functions.” breadth of health aid. Newer donors like China engage WIDER Working Paper, United Nations University World in global health in ways that are poorly captured in Institute for Development Economics Research, Helsinki. the current data systems, and changing this situation Centre on Global Health Security Working Group on Health would have large benefits. Financing. 2014. Shared Responsibilities for Health: A Coherent Global Framework for Health Financing. London: Chatham House. NOTES China State Council. 2011. “China’s Foreign Aid.” Information Office of the State Council, Beijing. http://www.gov.cn World Bank Income Classifications as of July 2014 are as /english/official/2011-04/21/content_1849913.htm. follows, based on estimates of gross national income (GNI) ———. 2014. “China’s Foreign Aid.” Information Office of the per capita for 2013: State Council, Beijing. http://news.xinhuanet.com/english /china/2014-07/10/c_133474011.htm. • Low-income countries (LICs) = US$1,045 or less Consultative Expert Working Group on Research and • Middle-income countries (MICs) are subdivided: Development: Financing and Coordination. 2012. Research (a) lower-middle-income = US$1,046 to US$4,125 and Development to Meet Health Needs in Developing (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 Countries: Strengthening Global Financing and Coordination. • High-income countries (HICs) = US$12,746 or more. Geneva: WHO. Coyle, D., K. Coyle, C. Cameron, K. Lee, S. Kelly, and others. 1. Two important exclusions from ODA are other official 2013. “Cost-Effectiveness of New Oral Anticoagulants flows (broadly, financial transfers that are not clearly Compared with Warfarin in Preventing Stroke and Other intended to promote development of the recipient country) Cardiovascular Events in Patients with Atrial Fibrillation.” and grants from private sources. More details on the exact Value in Health 16 (4): 498–506. definition of these concepts are available at http://www Darmstadt, G. L., M. V. Kinney, M. Chopra, S. Cousens, L. Kak, .oecd.org/dac/stats/dac-glossary.htm#ODA. and others. 2014. “Who Has Been Caring for the Baby?” 2. Complete projects refer to construction or civil projects The Lancet 384 (9938): 174–88. completed in recipient countries supported by Chinese Deaton, A. S. 2009. “Instruments of Development: grants or interest-free loans. The Chinese side is responsible Randomization in the Tropics, and the Search for the Elusive for all or part of the construction process. After a project is Keys to Economic Development.” Working Paper 14690, completed, China hands it over to the recipient country. National Bureau of Economic Research, Cambridge, MA. Denny, C. C., and E. J. Emanuel. 2008. “U.S. Health Aid beyond PEPFAR.” Journal of the American Medical Association 300 (17): 2048–51. REFERENCES Dieleman, J. L., C. Graves, E. Johnson, T. Templin, M. Birger, and Basu, S., E. Bendavid, and N. Sood. 2015. “Health and others. 2015. “Sources and Focus of Health Development Economic Implications of National Treatment Coverage Assistance, 1990–2014.” Journal of the American Medical for Cardiovascular Disease in India: Cost-Effectiveness Association 313 (23): 2359–68. Analysis.” Circulation: Cardiovascular Quality and Outcomes Fan, V. Y., D. J. Jamison, and L. H. Summers. 2016. “The 8 (6): 541–51. Inclusive Loss of Pandemic Influenza Risk.” NBER Working Bendavid, E. 2014a. “Changes in Child Mortality over Time Paper No. 22137, National Bureau of Economic Research, across the Wealth Gradient in Less-Developed Countries.” Cambridge, MA. Pediatrics 134 (6): e1551–59. Fauci, A. S., and G. K. Folkers. 2012. “The World Must Build ———. 2014b. “Is Health Aid Reaching the Poor? Analysis of on Three Decades of Scientific Advances to Enable a Household Data from Aid-Recipient Countries.” PLoS One New Generation to Live Free of HIV/AIDS.” Health Affairs 9 (1): e84025. 31 (7): 1529–36. Bendavid, E., and J. Bhattacharya. 2014. “The Relationship of Fogarty International Center. 2015. Global Health Matters Health Aid to Population Health Improvements.” JAMA Newsletter (September-October). https://www.fic.nih.gov Internal Medicine 174 (6): 881–87. /News/GlobalHealthMatters/september-october-2015 Bendavid, E., A. Duong, C. Sagan, and G. Raikes. 2015. “Health /Pages/default.aspx. Aid Is Allocated Efficiently, but Not Optimally: Insights ———. 2016. “Research to Guide Practice: Enhancing HIV/ from a Review of Cost-Effectiveness Studies.” Health Affairs AIDS Platforms to Address NCDs in Low-Resource Settings.” 34 (7): 1188–95. Fogarty International Center, Washington, DC. https://www Berthélemy, J. C. 2006. “Bilateral Donors’ Interest vs. Recipients’ .fic.nih.gov/About/Staff/Policy-Planning-Evaluation/Pages Development Motives in Aid Allocation: Do All Donors /pepfar-ncd-project.aspx. Development Assistance for Health 311 Frenk, J., and S. Moon. 2013. “Governance Challenges in Global Data.” Journal of the European Economic Association 10 (6): Health.” New England Journal of Medicine 368 (10): 936–42. 1294–317. Glassman, A., and R. Levine. 2016. Millions Saved: New Levine, R. 2004. Millions Saved: Proven Successes in Global Cases of Proven Success in Global Health. Washington, DC: Health. Washington, DC: Peterson Institute. Brookings Institution Press. Liu, P., Y. Guo, X. Qian, S. Tang, Z. Li, and others. 2014. “China’s Global Fund. 2016a. “History of the Global Fund.” Global Distinctive Engagement in Global Health.” The Lancet Fund, Geneva. http://www.theglobalfund.org/en/history/. 384 (9945): 793–804. ———. 2016b. “Overview: Equitable Access Initiative.” Lu, C., M. T. Schneider, P. Gubbins, K. Leach-Kemon, Global Fund, Geneva. http://www.theglobalfund.org/en D. Jamison, and others. 2010. “Public Financing of Health /equitableaccessinitiative/. in Developing Countries: A Cross-National Systematic Gostin, L. O., and E. A. Friedman. 2015. “A Retrospective Analysis.” The Lancet 375 (9723): 1375–87. and Prospective Analysis of the West African Ebola Virus McNeil, D., Jr. 2010. “At Front Lines, AIDS War Is Falling Disease Epidemic: Robust National Health Systems at the Apart.” New York Times, May 9. http://www.nytimes.com Foundation and an Empowered WHO at the Apex.” The /2009/01/31/health/31aids.html. Lancet 385 (9980): 1902–9. Miller-Petrie, M., S. Pant, and R. Laxminarayan. 2017. Harmon, T. M., K. A. Fisher, M. G. McGlynn, J. Stover, M. J. “Drug Resistant Infections.” In Disease Control Priorities Warren, and others. 2016. “Exploring the Potential Health (third edition), Volume 6, Major Infectious Diseases. Edited Impact and Cost-Effectiveness of AIDS Vaccine within a by K. K. Holmes, S. Bertozzi, B. R. Bloom, and P. Jha. Comprehensive HIV/AIDS Response in Low- and Middle- Washington, DC: World Bank. Income Countries.” PLoS One 11 (1): e0146387. Moon, S., and O. Omole. 2013. “Development Assistance for Hoeffler, A., and V. Outram. 2011. “Need, Merit, or Self- Health: Critiques and Proposals for Change.” Chatham Interest—What Determines the Allocation of Aid?” Review House, London. of Development Economics 15 (2): 237–50. Murray, C. J., R. M. Barber, K. J. Foreman, A. A. Ozgoren, F. Abd- Hoos, D., W. El-Sadr, and K. L. Dehne. 2016. “Getting the Allah, and others. 2015. “Global, Regional, and National Balance Right: Scaling up Treatment and Prevention.” Disability-Adjusted Life Years (DALYs) for 306 Diseases Global Public Health (April): 1–15. and Injuries and Healthy Life Expectancy (HALE) for 188 IHME (Institute for Health Metrics and Evaluation). 2016a. Countries, 1990–2013: Quantifying the Epidemiological “Financing Global Health 2015: Development Assistance Transition.” The Lancet 386 (10009): 2145–91. Steady on the Path to New Global Goals.” IHME, Seattle, WA. Ottersen, T., A. Kamath, S. Moon, and J.-A. Røttingen. 2014. ———. 2016b. “Global Burden of Disease Study 2013 “Development Assistance for Health: Quantitative Allocation (GBD 2013) Results by Location, Cause, and Risk Factor.” Criteria and Contribution Norms.” Centre on Global Health IHME, Seattle, WA. Security Working Group Paper, Chatham House, London. Isenman, P. 2015. “Defining Development Drastically PMNCH (Partnership for Maternal, Newborn, and Child Downward: Premature Graduation from IDA and ODA.” Health). 2014. The PMNCH 2014 Accountability Report: Brookings Institution blog, Washington, DC, March 5. http:// Tracking Financial Commitments to the Global Strategy www.brookings.edu/blogs/future-development/posts for Women’s and Children’s Health. Geneva: World Health /2015/03/05-ida-eligibility-isenman. Organization, PMNCH. Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, Prabhakaran, S., I. Ruff, and R. A. Bernstein. 2015. “Acute S. Berkley, and others. 2013. “Global Health 2035: A World Stroke Intervention: A Systematic Review.” Journal of the Converging within a Generation.” The Lancet 382 (9908): American Medical Association 313 (14): 1451–62. 1898–955. Pritchett, L. 2004. “An Homage to the Randomistas on the Katz, I. T., I. V. Bassett, and A. A. Wright. 2013. “PEPFAR in Occasion of the J-PAL 10th Anniversary: Development as a Transition: Implications for HIV Care in South Africa.” Faith-Based Activity.” Center for Global Development blog, New England Journal of Medicine 369 (15): 1385–87. Washington, DC, March 10. http://www.cgdev.org/blog Kharas, H. 2015. “Trends and Issues in Qatari Foreign Aid.” /homage-randomistas-occasion-j-pal-10th-anniversary Working Paper 15-11, Silatech, Doha. -development-faith-based-activity. Kim, J. Y., and T. G. Evans. 2014. “Redefining the Measure Ravishankar, N., P. Gubbins, R. J. Cooley, K. Leach-Kemon, of Medical Education: Harnessing the Transformative C. M. Michaud, and others. 2009. “Financing of Global Potential of MEPI.” Academic Medicine 89 (8): S29–31. Health: Tracking Development Assistance for Health from Kirton, J., J. Kulik, and C. Bracht. 2014. “The Political Process 1990 to 2007.” The Lancet 373 (9681): 2113–24. in Global Health and Nutrition Governance: The G-8’s Resch, S., T. Ryckman, and R. Hecht. 2015. “Funding AIDS 2010 Muskoka Initiative on Maternal, Child, and Newborn Programmes in the Era of Shared Responsibility: An Analysis Health.” Annals of the New York Academy of Sciences 1331 (1): of Domestic Spending in 12 Low-Income and Middle- 186–200. Income Countries.” The Lancet Global Health 3 (1): e52–61. Kudamatsu, M. 2012. “Has Democratization Reduced Infant Rose, G. 2001. “Sick Individuals and Sick Populations.” Mortality in Sub-Saharan Africa? Evidence from Micro International Journal of Epidemiology 30 (3): 427–32. 312 Disease Control Priorities: Improving Health and Reducing Poverty Sands, P., C. Mundaca-Shah, and V. J. Dzau. 2016. “The ———. 2015. 2015 VolAg Report: The Report of Voluntary Neglected Dimension of Global Security: A Framework for Agencies Engaged in Overseas Relief and Development. Countering Infectious-Disease Crises.” New England Journal Washington, DC: USAID. of Medicine 374 (13): 1281–87. Vermund, S. H., M. Sidat, L. F. Weil, J. A. Tique, T. D. Moon, Schäferhoff, M., S. Fewer, J. Kraus, E. Richter, L. H. Summers, and others. 2012. “Transitioning HIV Care and Treatment and others. 2015. “How Much Donor Financing for Health Programs in Southern Africa to Full Local Management.” Is Channelled to Global versus Country-Specific Aid AIDS 26 (10): 1303–10. Functions?” The Lancet 386 (10011): 2436–41. Victora, C. G., J. H. Requejo, A. J. Barros, P. Berman, Z. Bhutta, Sridhar, D., and R. Batniji. 2008. “Misfinancing Global Health: and others. 2015. “Countdown to 2015: A Decade of A Case for Transparency in Disbursements and Decision Tracking Progress for Maternal, Newborn, and Child Making.” The Lancet 372 (9644): 1185–91. Survival.” The Lancet 387 (10032): 2049–59. Sumner, A. 2012. “Where Do the Poor Live?” World Development WHO (World Health Organization). n.d. “Health Accounts.” 40 (5): 865–77. WHO, Geneva. http://www.who.int/health-accounts/en. USAID (U.S. Agency for International Development). 2012. Zhou, H., J. Zhang, and M. Zhang. 2015. Foreign Aid in China. USAID’s Global Health Strategic Framework: Better Health Berlin: Springer Verlag and Social Sciences Academic Press for Development. Washington, DC: USAID. (China). Development Assistance for Health 313 Chapter 17 Pandemics: Risks, Impacts, and Mitigation Nita Madhav, Ben Oppenheim, Mark Gallivan, Prime Mulembakani, Edward Rubin, and Nathan Wolfe during the 2009 influenza pandemic (Katz 2009). INTRODUCTION International donors also have begun to invest in Pandemics are large-scale outbreaks of infectious dis- improving preparedness through refined standards ease that can greatly increase morbidity and mortality and funding for building health capacity (Wolicki and over a wide geographic area and cause significant eco- others 2016). nomic, social, and political disruption. Evidence sug- Despite these improvements, significant gaps and gests that the likelihood of pandemics has increased challenges exist in global pandemic preparedness. over the past century because of increased global travel Progress toward meeting the IHR has been uneven, and and integration, urbanization, changes in land use, and many countries have been unable to meet basic require- greater exploitation of the natural environment (Jones ments for compliance (Fischer and Katz 2013; WHO and others 2008; Morse 1995). These trends likely will 2014). Multiple outbreaks, notably the 2014 West Africa continue and will intensify. Significant policy attention Ebola epidemic, have exposed gaps related to the timely has focused on the need to identify and limit emerging detection of disease, availability of basic care, tracing of outbreaks that might lead to pandemics and to expand contacts, quarantine and isolation procedures, and pre- and sustain investment to build preparedness and paredness outside the health sector, including global health capacity (Smolinsky, Hamburg, and Lederberg coordination and response mobilization (Moon and 2003). others 2015; Pathmanathan and others 2014). These The international community has made progress gaps are especially evident in resource-limited settings toward preparing for and mitigating the impacts of and have posed challenges during relatively localized pandemics. The 2003 severe acute respiratory syn- epidemics, with dire implications for what may happen drome (SARS) pandemic and growing concerns about during a full-fledged global pandemic. the threat posed by avian influenza led many coun- For the purposes of this chapter, an epidemic is tries to devise pandemic plans (U.S. Department of defined as “the occurrence in a community or region of Health and Human Services 2005). Delayed reporting cases of an illness . . . clearly in excess of normal expec- of early SARS cases also led the World Health Assembly tancy” (Porta 2014). A pandemic is defined as “an epi- to update the International Health Regulations (IHR) demic occurring over a very wide area, crossing to compel all World Health Organization member international boundaries, and usually affecting a large states to meet specific standards for detecting, report- number of people” (Porta 2014). Pandemics are, there- ing on, and responding to outbreaks (WHO 2005). fore, identified by their geographic scale rather than the The framework put into place by the updated IHR severity of illness. For example, in contrast to annual contributed to a more coordinated global response seasonal influenza epidemics, pandemic influenza is Corresponding Author: Nita Madhav, MSPH, Metabiota, San Francisco, California, United States; nmadhav@metabiota.com. 315 defined as “when a new influenza virus emerges and • Some pandemic mitigation measures can cause sig- spreads around the world, and most people do not have nificant social and economic disruption. immunity” (WHO 2010). • In countries with weak institutions and legacies of This chapter does not consider endemic diseases— political instability, pandemics can increase political those that are constantly present in particular localities stresses and tensions. In these contexts, outbreak or regions. Endemic diseases are far more common than response measures such as quarantines have sparked pandemics and can have significant negative health and violence and tension between states and citizens. economic impacts, especially in low- and middle- income countries (LMICs) with weak health systems. Additionally, given the lack of historical data and extreme Mitigation uncertainty regarding bioterrorism, this chapter does • Pathogens with pandemic potential vary widely in the not specifically consider bioterrorism-related events, resources, capacities, and strategies required for miti- although bioterrorism could hypothetically lead to a gation. However, there are also common prerequisites pandemic. for effective preparedness and response. This chapter covers the following findings concerning • The most cost-effective strategies for increasing pan- the risks, impacts, and mitigation of pandemics as well demic preparedness, especially in resource-constrained as knowledge gaps: settings, consist of investing to strengthen core public health infrastructure, including water and sanitation systems; increasing situational awareness; and rapidly Risks extinguishing sparks that could lead to pandemics. • Pandemics have occurred throughout history and • Once a pandemic has started, a coordinated response appear to be increasing in frequency, particularly should be implemented focusing on maintenance because of the increasing emergence of viral disease of situational awareness, public health messaging, from animals. reduction of transmission, and care for and treatment • Pandemic risk is driven by the combined effects of of the ill. spark risk (where a pandemic is likely to arise) and • Successful contingency planning and response spread risk (how likely it is to diffuse broadly through require surge capacity—the ability to scale up the human populations). delivery of health interventions proportionately for • Some geographic regions with high spark risk, includ- the severity of the event, the pathogen, and the pop- ing Central and West Africa, lag behind the rest of the ulation at risk. globe in pandemic preparedness. • For many poorly prepared countries, surge capacity • Probabilistic modeling and analytical tools such as likely will be delivered by foreign aid providers. This exceedance probability (EP) curves are valuable for is a tenable strategy during localized outbreaks, but assessing pandemic risk and estimating the potential global surge capacity has limits that likely will be burden of pandemics. reached during a full-scale global pandemic as higher- • Influenza is the most likely pathogen to cause a severe capacity states focus on their own populations. pandemic. EP analysis indicates that in any given year, • Risk transfer mechanisms, such as risk pooling and a 1 percent probability exists of an influenza pan- sovereign-level catastrophe insurance, provide a via- demic that causes nearly 6 million pneumonia and ble option for managing pandemic risk. influenza deaths or more globally. Knowledge Gaps Impacts • Spending and costs specifically associated with pan- • Pandemics can cause significant, widespread increases demic preparedness and response efforts are poorly in morbidity and mortality and have disproportion- tracked. ately higher mortality impacts on LMICs. • There is no widely accepted, consistent methodology • Pandemics can cause economic damage through for estimating the economic impacts of pandemics. multiple channels, including short-term fiscal shocks • Most data regarding the impacts of pandemics and and longer-term negative shocks to economic growth. the benefits and costs of mitigation measures come • Individual behavioral changes, such as fear-induced from high-income countries (HICs), leading to aversion to workplaces and other public gathering biases and potential blind spots regarding the risks, places, are a primary cause of negative shocks to eco- consequences, and optimal interventions specific to nomic growth during pandemics. LMICs. 316 Disease Control Priorities: Improving Health and Reducing Poverty PANDEMIC RISKS AND CONSEQUENCES impacts as well as the resources, capacities, and strate- gies required for mitigation. Importance of Pandemics One must distinguish between several broad catego- Pandemics can cause sudden, widespread morbidity ries of pandemic threats. At one extreme are pathogens and mortality as well as social, political, and economic that have high potential to cause truly global, severe disruption. The world has endured several notable pandemics. This group includes pandemic influenza pandemics, including the Black Death, Spanish flu, and viruses. These pathogens transmit efficiently between human immunodeficiency virus/acquired immune humans, have sufficiently long asymptomatic infectious deficiency syndrome (HIV/AIDS) (table 17.1). periods to facilitate the undetected movement of infected Because the definition of pandemic primarily is persons, and have symptomatic profiles that present geographic, it groups together multiple, distinct types challenges for differential diagnosis (particularly in the of events and public health threats, all of which have early periods of infection). A second group of pathogens their own severity, frequency, and other disease char- presents a moderate global threat. These agents (for acteristics. Each type of event requires its own optimal example, Nipah virus and H5N1 and H7N9 influenzas) preparedness and response strategy; however this have not demonstrated sustained human-to-human chapter also discusses common prerequisites for effec- transmission but could become transmitted more effi- tive response. The variety of pandemic threats is ciently as a result of mutations and adaptation. A third driven by the great diversity of pathogens and their group of pathogens (for example, Ebola, Marburg, Lassa) interaction with humans. Pathogens vary across multi- has the potential to cause regional or interregional epi- ple dimensions, including the mechanism and dynam- demics, but the risk of a truly global pandemic is limited ics of disease transmission, severity, and differentiability because of the slow pace of transmission or high proba- of associated morbidities. These and other factors bility of detection and containment. determine whether cases will be identified and con- Among all known pandemic pathogens, influenza tained rapidly or whether an outbreak will spread poses the principal threat because of its potential (Fraser and others 2004). As a result, pathogens with severity and semiregular occurrence since at least the pandemic potential also vary widely in the scale of 16th century (Morens and others 2010). The infamous their potential health, economic, and sociopolitical 1918 influenza pandemic killed an estimated 20 million Table 17.1 Notable Epidemics and Pandemics since the Middle Ages Geographic Estimated direct morbidity Estimated economic, social, Starting year Event extent or mortality or political impact 1347 Bubonic plague (Black Eurasia 30–50 percent mortality of the Likely hastened end of the feudal Death) pandemic European population (DeWitte 2014) system in Europe (Platt 2014) Early 1500s Introduction of Americas More than 50 percent mortality in Destroyed native societies, facilitating smallpox some communities (Jones 2006) the hegemony of European countries (Diamond 2009) 1881 Fifth cholera pandemic Global More than 1.5 million deaths Sparked attacks on Russian tsarist (9.7 per 10,000 persons) government and medical officials (Chisholm 1911) (Frieden 1977) 1918 Spanish flu influenza Global 20 million–100 million deaths GDP loss of 3 percent in Australia, pandemic (111–555 deaths per 10,000 persons) 15 percent in Canada, 17 percent (Johnson and Mueller 2002) in the United Kingdom, 11 percent in the United States (McKibbin and Sidorenko 2006) 1957 Asian flu influenza Global 0.7 million–1.5 million deaths GDP loss of 3 percent in Canada, pandemic (2.4–5.1 deaths per 10,000 persons) Japan, the United Kingdom, and (Viboud and others 2016) the United States (McKibbin and Sidorenko 2006) 1968 Hong Kong flu Global 1 million deaths (2.8 deaths per US$23 billion–US$26 billion direct influenza pandemic 10,000 persons) (Mathews and and indirect costs in the United States others 2009) (Kavet 1977) table continues next page Pandemics: Risks, Impacts, and Mitigation 317 Table 17.1 Notable Epidemics and Pandemics since the Middle Ages (continued) Geographic Estimated direct morbidity Estimated economic, social, Starting year Event extent or mortality or political impact 1981 HIV/AIDS pandemic Global More than 70 million infections, 36.7 2–4 percent annual loss of GDP million deaths (WHO Global Health growth in Africa (Dixon, McDonald, Observatory data, http://www.who and Roberts 2001)a .int/gho/hiv/en/) 2003 SARS pandemic 4 continents, 37 8,098 possible cases, 744 deaths GDP loss of US$4 billion in Hong countries (Wang and Jolly 2004) Kong SAR, China; US$3 billion–US$6 billion in Canada; and US$5 billion in Singapore (Keogh-Brown and Smith 2008) 2009 Swine flu influenza Global 151,700–575,500 deaths (0.2–0.8 GDP loss of US$1 billion in the pandemic per 10,000 persons) (Dawood and Republic of Korea (Kim, Yoon, and Oh others 2012) 2013) 2012 MERS epidemic 22 countries 1,879 symptomatic cases, 659 deaths US$2 billion loss in the Republic of (Arabi and others 2017) Korea, triggering US$14 billion in government stimulus spending (Jun 2015; Park and Kim 2015) 2013b West Africa Ebola 10 countries 28,646 cases, 11,323 deaths US$2 billion loss in Guinea, Liberia, virus disease epidemic (WHO 2016a) and Sierra Leone (World Bank 2014) 2015 Zika virus pandemic 76 countries 2,656 reported cases of microcephaly US$7 billion–US$18 billion loss in or central nervous system Latin America and the Caribbean malformation (WHO 2017) (UNDP 2017) Note: List of events is illustrative rather than exhaustive. All U.S. dollar amounts are rounded to nearest billion. GDP = gross domestic product; HIV/AIDS = human immunodeficiency virus/acquired immunodeficiency syndrome; MERS = Middle East respiratory syndrome; SARS = severe acute respiratory syndrome. a. Studies of the effects of HIV/AIDS on per capita gross national product have found smaller effects. b. The West Africa Ebola virus outbreak occurred from 2013 to 2016, but the peak and international response efforts began in 2014. to 100 million persons globally, with few countries Gowtage-Sequeria 2005), and the next pandemic is spared (Johnson and Mueller 2002). Its severity reflects likely to be a zoonosis as well. Zoonoses enter into in part the limited health technologies of the period, human populations from both domesticated animals when no antibiotics, antivirals, or vaccines were avail- (such as farmed swine or poultry) and wildlife. Many able to reduce transmission or mortality (Murray and historically significant zoonoses were introduced others 2006). through increased human-animal interaction follow- During the 1918 pandemic, populations experienced ing domestication, and potentially high-risk zoonoses significantly higher mortality rates in LMICs than in (including avian influenzas) continue to emerge from HICs, likely as a result of higher levels of malnutrition livestock production systems (Van Boeckel and others and comorbid conditions, insufficient access to sup- 2012; Wolfe, Dunavan, and Diamond 2007). Some portive medical care, and higher rates of disease trans- pathogens (including Ebola) have emerged from wild- mission (Brundage and Shanks 2008; Murray and life reservoirs and entered into human populations others 2006). The mortality disparity between HICs and through the hunting and consumption of wild species LMICs likely would be even greater today for a similarly (such as bushmeat), the wild animal trade, and other severe event, because LMICs have disproportionately contact with wildlife (Pike and others 2010; Wolfe, lower medical capacity, less access to modern medical Dunavan, and Diamond 2007). interventions, and higher interconnectivity between Zoonotic pathogens vary in the extent to which they population centers. can survive within and spread between human hosts. As shown in table 17.2, the degree of zoonotic adapta- tion spans a continuum from transmission only within Origin of Pandemics animal populations (stage 1) to transmission only Most new pandemics have originated through the within human populations (stage 5). Most zoonotic “zoonotic” transmission of pathogens from animals pathogens are not well adapted to humans (stages 2–3), to humans (Murphy 1998; Woolhouse and emerge sporadically through spillover events, and may 318 Disease Control Priorities: Improving Health and Reducing Poverty lead to localized outbreaks, called stuttering chains the United States, and Western Europe. Key drivers for (Pike and others 2010; Wolfe and others 2005). These spark risk from domesticated animals include intensive episodes of “viral chatter” increase pandemic risk by and extensive farming and livestock production sys- providing opportunities for viruses to become better tems and live animal markets, as well as the potential adapted to spreading within a human population. for contact between livestock and wildlife reservoirs Pathogens that are past stage 3 are of the greatest con- (Gilbert and others 2014; Jones and others 2008). cern, because they are sufficiently adapted to humans Wildlife zoonosis risk is distributed far more broadly, to cause long transmission chains between humans with foci in China, India, West and Central Africa, and (directly or indirectly through vectors), and their geo- the Amazon Basin (Jones and others 2008). Risk driv- graphic spread is not constrained by the habitat range ers include behavioral factors (such as bushmeat hunt- of an animal reservoir. ing and use of animal-based traditional medicines), natural resource extraction (such as sylviculture and logging), the extension of roads into wildlife habi- Pandemic Risk Factors tats, and environmental factors (including the degree Pandemic risk, as noted, is driven by the combined and distribution of animal diversity) (Wolfe and effects of spark risk and spread risk. The foci of both others 2005). risk factors often overlap, especially in some LMICs (such as in Central and West Africa and Southeast Spread Risk Asia), making these areas particularly vulnerable to After a spark or importation, the risk that a pathogen pandemics and their negative consequences. will spread within a population is influenced by pathogen- specific factors (including genetic adaptation and mode Spark Risk of transmission) and human population-level factors A zoonotic spark could arise from the introduction of (such as the density of the population and the suscepti- a pathogen from either domesticated animals or bility to infection; patterns of movement driven by wildlife. Zoonoses from domesticated animals are con- travel, trade, and migration; and speed and effectiveness centrated in areas with dense livestock production of public health surveillance and response measures) systems, including areas of China, India, Japan, (Sands and others 2016). Table 17.2 Pathogen Adaptation and Pandemic Risk Stage Transmission to humansa Pathogen example Simplified transmission diagram Stage 1: animal reservoir None H3N8 equine influenza transmission only virus Stage 2: primary infection Only from animals Anthrax Stage 3: limited outbreaks Few human-to-human transmission chains Marburg virus Stage 4: sustained outbreaks Many human-to-human transmission Pandemic A (H1N1) 2009 chains influenza virus Stage 5: predominant human Human-to-human Smallpox virus transmission Source: Adapted from Wolfe, Dunavan, and Diamond 2007. a. Direct or indirect transmission through vector. Pandemics: Risks, Impacts, and Mitigation 319 Dense concentrations of population, especially in • Public health infrastructure capable of identifying, urban centers harboring overcrowded informal settle- tracing, managing, and treating cases ments, can act as foci for disease transmission and • Adequate physical and communications infrastruc- accelerate the spread of pathogens (Neiderud 2015). ture to channel information and resources Moreover, social inequality, poverty, and their environ- • Fundamental bureaucratic and public management mental correlates can increase individual susceptibility to capacities infection significantly (Farmer 1996). Comorbidities, • Capacity to mobilize financial resources to pay for malnutrition, and caloric deficits weaken an individual’s disease response and weather the economic shock of immune system, while environmental factors such as lack the outbreak of clean water and adequate sanitation amplify transmis- • Ability to undertake effective risk communications. sion rates and increase morbidity and mortality (Toole and Waldman 1990). Collectively, all these factors suggest Well-prepared countries have effective public institu- that marginalized populations, including refugees and tions, strong economies, and adequate investment in the people living in urban slums and informal settlements, health sector. They have built specific competencies likely face elevated risks of morbidity and mortality dur- critical to detecting and managing disease outbreaks, ing a pandemic. including surveillance, mass vaccination, and risk com- A country’s expected ability to curtail pandemic munications. Poorly prepared countries may suffer from spread can be expressed using a preparedness index political instability, weak public administration, inade- developed by Oppenheim and others (2017). The index quate resources for public health, and gaps in funda- illustrates global variation in institutional readiness to mental outbreak detection and response systems. detect and respond to a large-scale outbreak of infectious Map 17.1 presents the global distribution of epidemic disease. It draws on the IHR core capacity metrics and preparedness, with countries grouped into quintiles. A other publicly accessible cross-national indicators. geographic analysis of preparedness shows that some However, it diverges from the IHR metrics in its breadth areas of high spark risk also are the least prepared. and focus on measuring underlying and enabling institu- Geographic areas with high spark risk from domesticated tional, infrastructural, and financial capacities such as the animals (including China, North America, and Western following (Oppenheim and others 2017): Europe) have relatively higher levels of preparedness, Map 17.1 Global Distribution of Epidemic Preparedness, 2017 IBRD 43200 | SEPTEMBER 2017 Quintile: 1 2 3 4 5 No score Note: Countries are grouped into quintiles of epidemic preparedness (1 = most prepared, 5 = least prepared). 320 Disease Control Priorities: Improving Health and Reducing Poverty although China lags behind its counterparts. However, from person to person via disease transmission dynam- geographic areas with high spark risk from wildlife spe- ics and from place to place via incorporation of long- cies (including Central and West Africa) have some of the range and short-range population movements. The lowest preparedness scores globally, indicating a poten- models also can incorporate mitigation measures, sea- tially dangerous overlap of spark risk and spread risk. sonality, stochastic processes, and other factors that can Table 17.3 presents the average epidemic preparedness vary during an epidemic. Millions of these simulations quintile across each of the World Bank’s country income can be run with wide variation in the initial conditions groups. National income alone offers an incomplete and and final outcomes. potentially misleading metric of preparedness. Although These millions of simulations can be used to quan- income is correlated with epidemic preparedness, many tify the burden of pandemics through a class of prob- countries are substantially better or worse prepared than abilistic modeling called catastrophe modeling, which expected, given their gross national income per capita. the insurance industry uses to understand risks posed by infrequent natural disasters such as hurricanes and earthquakes (Fullam and Madhav 2015; Kozlowski and Burden of Pandemics Mathewson 1997). When applied to pandemics, this Quantifying the morbidity and mortality burden from approach requires statistically fitting distributions of pandemics poses a significant challenge. Although esti- the parameters. These parameter distributions pro- mates are available from historical events (table 17.1), vide weightings of the likelihood of the different the historical record is sparse and incomplete. To over- events. Through correlated statistical sampling based come these gaps in estimating the frequency and severity on the parameter weights, scenarios are selected for of pandemics, probabilistic modeling techniques can inclusion in an event catalog of simulated pandemic augment the historical record with a large catalog of events. A schematic diagram shows how the catastro- hypothetical, scientifically plausible, simulated pandem- phe modeling process is used to develop the event ics that represent a wide range of possible scenarios. catalog (figure 17.1). Modeling can also better account for changes that have Analysis of the event catalog yields annual EP curves occurred since historical times, such as medical advances, (for example, as shown in figure 17.2), which provide a changing demographics, and shifting travel patterns. metric of the likelihood that an event of a given severity, Scenario modeling of epidemics and pandemics can or worse, begins in any given year. The EP curve is a visu- be achieved through large-scale computer simulations of alization of the event catalog, in which the number of global spread, dynamics, and illness outcomes of disease estimated deaths for each event is ranked in descending (Colizza and others 2007; Tizzoni and others 2012). order. Because the event catalog includes scenarios incor- These models allow for specification of parameters that porating spark probabilities and estimates of disease prop- may drive the likelihood of a spark (for example, loca- agation, the EP curve includes the combined impacts of tion and frequency) and determinants of severity (for both spark risk and spread risk. Although a global curve is example, transmissibility and virulence). The models shown in figure 17.2, EP curves can be estimated for other then simulate at a daily time step the spread of disease geographic resolutions, such as a country or province. Table 17.3 Epidemic Preparedness Score, by Country Income Group, 2017 Mean epidemic Top-performing Bottom-performing country Country income groupa preparedness quintileb country in group in group High-income 1.3 Norway Trinidad and Tobago Upper-middle-income 2.9 Malaysia Equatorial Guinea Lower-middle-income 3.7 Armenia Mauritania Low-income 4.8 Nepal Somalia Source: The epidemic preparedness index draws on indicators from the World Health Organization, World Bank, United Nations agencies, and nongovernmental sources (see Oppenheim and others 2017). a. Income groups follow World Bank income classifications for fiscal 2018, based on estimates of 2016 gross national income per capita and calculated using the World Bank Atlas method: high-income (US$12,236 or more), upper-middle-income (US$3,956–US$12,235), lower-middle-income (US$1,006–US$3,955), and low-income (US$1,005 or less). For further explanation, see https://datahelpdesk.worldbank.org/knowledgebase/articles/906519-world-bank-country-and-lending-groups. b. Countries are grouped into quintiles of epidemic preparedness (1 = most prepared, 5 = least prepared). Pandemics: Risks, Impacts, and Mitigation 321 Figure 17.1 Process for Generating the Event Catalog The EP curve is a powerful tool that yields several key findings regarding the frequency and severity of poten- Transmissibility Frequency Virulence tial pandemics. Applied to influenza pandemics, we find the following: • An influenza pandemic having the global mortality rate observed during the 2009 Swine flu pandemic (0.2–0.8 deaths per 10,000 persons) or worse has about a 3 percent probability of occurring in any Modeled events given year. • In any given year, the probability of an influenza pandemic causing nearly 6 million pneumonia and Event catalog influenza deaths (8 deaths per 10,000 persons) or more globally is 1 percent. • The annual probability of an influenza pandemic’s meeting or exceeding the global mortality rate of the 1918 Spanish flu pandemic (111–555 deaths per Figure 17.2 Estimated Annual Exceedance Probability Curve for 10,000 persons) is less than 0.02 percent. Global Pneumonia and Influenza Deaths Caused by Influenza Pandemics, 2017 • As indicated by the heavy tail of the EP curve, most of the potential burden from influenza pandemics comes from the most severe pandemics. Annual exceedance probability (%) 1.0000 Table 17.4 shows select EPs for influenza pandemics 0.1000 in low-, middle-, and high-income countries, based on further analysis of the event catalog. For example, in any given year, all LICs combined have a 3 percent probability 0.0100 of experiencing at least 140,000 deaths attributable to an influenza pandemic and a 0.1 percent chance of 0.0010 experiencing at least 8.3 million deaths. LICs bear a substantial burden of mortality risk from influenza 0.0001 pandemics. Strikingly, LICs contain only about 9 percent 0 50 100 150 200 250 300 350 400 450 of the global population, yet they would contribute Deaths (millions) nearly 25 percent of deaths during an influenza pandemic. Source: Metabiota simulations. Note: Annual exceedance probability is the likelihood that an event of a given severity, or worse, Based on the event catalog, the average estimated begins in any given year. Dashed lines indicate the 5th and 95th percentile bands. global mortality from pneumonia and influenza during Table 17.4 Select Annual Exceedance Probabilities for Pneumonia and Influenza Deaths Caused by Influenza Pandemics, by Country Income Level, 2017 Deaths (millions) Annual exceedance probability (%) Low income Middle income High income Total 3.0 0.1 0.4 0.05 0.6 2.0 0.6 1.5 0.1 2.2 1.0 1.5 4.0 0.4 5.9 0.5 2.7 7.6 0.9 11.2 0.2 5.5 14.8 1.7 22.0 0.1 8.3 22.5 2.5 33.3 Source: Metabiota simulations. Note: Annual exceedance probability is the likelihood that an event of a given severity, or worse, begins in any given year. Rows may not sum to total value due to rounding. 322 Disease Control Priorities: Improving Health and Reducing Poverty an influenza pandemic is more than 7.3 million deaths. lower immunity than older people, which significantly However, because influenza pandemics occur on aver- increases the years of life lost (Viboud and others 2010). age once every 25–30 years, the average annual pneumo- Furthermore, many infectious diseases can have chronic nia and influenza mortality from influenza pandemics is effects, which can become more common or widespread a little more than 230,000 deaths. This is comparable to in the case of a pandemic. For example, Zika-associated seasonal influenza, which worldwide causes at least microcephaly has lifelong impacts on health and 250,000 deaths annually (WHO 2016b). Although both well-being. numbers reflect an annual average, they differ in the The indirect health impacts of pandemics can combination of frequency and severity. Seasonal influ- increase morbidity and mortality further. Drivers of enza deaths occur every year, but pandemic influenza indirect health impacts include diversion or depletion deaths occur much less frequently, are concentrated in of resources to provide routine care and decreased larger spikes, and affect a younger demographic. access to routine care resulting from an inability to When pandemics cause large morbidity and mor- travel, fear, or other factors. Additionally, fear can lead tality spikes, they are much more likely to overwhelm to an upsurge of the “worried well” seeking unnecessary health systems. Overwhelmed health systems and care, further burdening the health care system (Falcone other indirect effects may contribute to a 2.3-fold and Detty 2015). increase in all-cause mortality during pandemics, During the 2014 West Africa Ebola epidemic, lack of although attribution of the causative agent is difficult routine care for malaria, HIV/AIDS, and tuberculosis (Simonsen and others 2013). If indirect deaths are led to an estimated 10,600 additional deaths in Guinea, taken into account, the average annual global deaths Liberia, and Sierra Leone (Parpia and others 2016). from influenza pandemics could be greater than This indirect death toll nearly equaled the 11,300 520,000, although there is a significant uncertainty in deaths directly caused by Ebola in those countries the estimate. (WHO 2016a). Additionally, diversion of funds, medi- Pandemics caused by pathogens other than influenza cal resources, and personnel led to a 30 percent decrease also must be considered. Novel coronaviruses (such as in routine childhood immunization rates in affected SARS-CoV), filoviruses (such as Ebola virus), and flavi- countries (UNDP 2014). During the 2009 influenza viruses (such as Zika virus) have caused large epidemics pandemic, a greater surge in hospital admissions for and pandemics. These viruses, like influenza, are ribonu- influenza and pneumonia was associated with statisti- cleic acid viruses that have high mutation rates. cally significant increases in deaths attributable to acute Noninfluenza viruses typically cause more frequent, myocardial infarction and stroke (Rubinson and others smaller epidemics but also an overall lower burden of 2013). However, during a pandemic, distinguishing morbidity and mortality than pandemic influenza. For which deaths are attributable to the pandemic itself and diseases caused by coronaviruses and filoviruses, the which are merely coincidental may be impossible. lower burden stems from the mode of transmission, During the 2014 West Africa Ebola epidemic, facili- which often requires closer and more sustained contact ties closures as a result of understaffing and fear of than influenza does to spread. contracting the disease played a large role in lack of access to or avoidance of routine health care. One study of 45 public facilities in Guinea found that the Ebola Consequences of Pandemics outbreak led to a 31 percent decrease in outpatient vis- Health Impacts its for routine maternal and child health services The direct health impacts of pandemics can be (Barden-O’Fallon and others 2015). Among children catastrophic. During the Black Death, an estimated under age five years, hospitals witnessed a 60 percent 30–50 percent of the European population perished decrease in visits for diarrhea and a 58 percent decrease (DeWitte 2014). More recently, the HIV/AIDS pan- in visits for acute respiratory illness (ARI), while health demic has killed more than 35 million persons since centers saw a 25 percent decrease in visits for diarrhea 1981 (WHO Global Health Observatory data, http:// and a 23 percent decrease in visits for ARI. In Sierra www.who.int/gho/hiv/en). Leone, visits to public facilities for reproductive health Pandemics can disproportionately affect younger, care fell by as much as 40 percent during the outbreak more economically active segments of the population (UNDP 2014). (Charu and others 2011). During influenza pandemics The availability of health care workers also decreases (as opposed to seasonal outbreaks of influenza), the during a pandemic because of illness, deaths, and fear- morbidity and mortality age distributions shift to driven absenteeism. Viral hemorrhagic fevers such as younger populations, because younger people have Ebola take an especially severe toll on health care Pandemics: Risks, Impacts, and Mitigation 323 workers, who face significant exposure to infectious including direct budgetary support. However, during a material: severe pandemic where HICs confront the same fiscal stresses and may be unable or unwilling to provide assis- • During the first Ebola outbreak in the Democratic tance, LMICs could face larger budget shortfalls, poten- Republic of Congo in 1976 (then called Zaire), the tially leading to weakened public health response or cuts Yambuku Mission Hospital—at the epicenter of the in other government spending. outbreak—was closed because 11 out of the 17 staff The direct fiscal impacts of pandemics generally are members had died of the disease (WHO 1978). small, however, relative to the indirect damage to eco- • During the Kikwit Ebola outbreak in 1995 in the same nomic activity and growth. Negative economic growth country, 24 percent of cases occurred among known shocks are driven directly by labor force reductions or possible health care workers (Rosello and others caused by sickness and mortality and indirectly by 2015). fear-induced behavioral changes. Fear manifests itself • During the 2014 West Africa Ebola epidemic, through multiple behavioral changes. As an analysis health care workers experienced high mortality of the economic impacts of the 2014 West Africa rates: 8 percent of doctors, nurses, and midwives Ebola epidemic noted, “Fear of association with oth- succumbed to Ebola in Liberia, 7 percent in Sierra ers . . . reduces labor force participation, closes places of Leone, and 1 percent in Guinea (Evans, Goldstein, employment, disrupts transportation, motivates some and Popova 2015). governments to close land borders and restrict entry of citizens from affected countries, and motivates private Even if health care workers do not die, their ability to decision makers to disrupt trade, travel, and commerce provide care may be reduced. At the peak of a severe by canceling scheduled commercial flights and reducing influenza pandemic, up to 40 percent of health care shipping and cargo services” (World Bank 2014). These workers might be unable to report for duty because they effects reduce labor force participation over and above are ill themselves, need to care for ill family members, the pandemic’s direct morbidity and mortality effects need to care for children because of school closures, or and constrict local and regional trade. are afraid (Falcone and Detty 2015; U.S. Homeland The indirect economic impact of pandemics has been Security Council 2006). quantified primarily through computable general equi- librium simulations; the empirical literature is less devel- Economic Impacts oped. World Bank economic simulations indicate that a Pandemics can cause acute, short-term fiscal shocks as severe pandemic could reduce world gross domestic well as longer-term damage to economic growth. Early- product (GDP) by roughly 5 percent (Burns, Van der phase public health efforts to contain or limit outbreaks Mensbrugghe, and Timmer 2006). The reduction in (such as tracing contacts, implementing quarantines, demand caused by aversive behavior (such as the avoid- and isolating infectious cases) entail significant human ance of travel, restaurants, and public spaces, as well as resource and staffing costs (Achonu, Laporte, and prophylactic workplace absenteeism) exceeds the eco- Gardam 2005). As an outbreak grows, new facilities may nomic impact of direct morbidity- and mortality- need to be constructed to manage additional infectious associated absenteeism. cases; this, along with increasing demand for consum- These results align with country-specific estimates: an ables (medical supplies, personal protective equipment, analysis of pandemic influenza’s impact on the United and drugs) can greatly increase health system expendi- Kingdom found that a low-severity pandemic could tures (Herstein and others 2016). reduce GDP by up to 1 percent, whereas a high-severity Diminished tax revenues may exacerbate fiscal stresses event could reduce GDP by 3–4 percent (Smith and oth- caused by increased expenditures, especially in LMICs, ers 2009). The World Bank’s estimates from the 2014 where tax systems are weaker and government fiscal West Africa Ebola epidemic suggest that economic dis- constraints are more severe. This dynamic was visible ruption in low-income countries (LICs) could be even during the 2014 West Africa Ebola epidemic in Liberia: greater. For example, the 2015 economic growth esti- while response costs surged, economic activity slowed, mate for Liberia was 3 percent (against a pre-Ebola esti- and quarantines and curfews reduced government mate of 6.8 percent); for Sierra Leone, it was −2 percent capacity to collect revenue (World Bank 2014). (against a pre-Ebola estimate of nearly 9 percent) During a mild or moderate pandemic, unaffected (Thomas and others 2015). HICs can offset fiscal shocks by providing increased offi- Finally, estimates of fiscal and growth shocks are cial development assistance (ODA) to affected countries, significant but do not include the intrinsic value of 324 Disease Control Priorities: Improving Health and Reducing Poverty lives lost. Fan, Jamison, and Summers (2016) consider incumbent politicians accused of leveraging the crisis this additional dimension of economic loss by estimat- and disease control measures to cement political con- ing the value of excess deaths across varying levels of trol and opposition figures accused of hampering dis- modeled pandemic severity, finding that the bulk of the ease response efforts (ICG 2015). Whereas growing expected annual loss from pandemics is driven by the tensions did not lead to large-scale political violence or direct cost of mortality, particularly in the case of instability, they did complicate public health response low-probability, severe events. efforts. In Sierra Leone, quarantine in opposition- During a severe pandemic, all sectors of the dominated regions was delayed because of concerns economy—agriculture, manufacturing, services—face that it would be seen as politically motivated (ICG disruption, potentially leading to shortages, rapid 2015). In countries with high levels of political polar- price increases for staple goods, and economic stresses ization, recent civil war, or weak institutions, sustained for households, private firms, and governments. A sus- outbreaks could lead to more sustained and challeng- tained, severe pandemic on the scale of the 1918 influ- ing political tensions. enza pandemic could cause significant and lasting Pandemics also can have longer-term impacts on economic damage. state capacity (Price-Smith 2001). The HIV/AIDS pan- demic offers one notable example. The 1990s and early Social and Political Impacts 2000s saw extremely high HIV/AIDS prevalence rates Evidence suggests that epidemics and pandemics can among African militaries, leading to increased absentee- have significant social and political consequences, ism, decreased military capacity, and decreased readiness creating clashes between states and citizens, eroding (Elbe 2002). Similar effects may occur during shorter, state capacity, driving population displacement, and more acute pandemics, reducing state capacity to man- heightening social tension and discrimination (Price- age instability. The weakening of security forces can, in Smith 2009). turn, amplify the risk of civil war and other forms of Severe premodern pandemics have been associated violent conflict (Fearon and Laitin 2003). with significant social and political upheaval, driven by Large-scale outbreaks of infectious disease have large mortality shocks and the resulting demographic direct and consequential social impacts. For example, shifts. Most notably, deaths arising from the introduc- widespread public panic during disease outbreaks can tion of smallpox and other diseases to the Americas led lead to rapid population migration. A 1994 outbreak of directly to the collapse of many indigenous societies plague in Surat, India, caused only a small number of and weakened the indigenous peoples’ institutions and reported cases, but fear led some 500,000 people military capacity to the extent that they became vulner- (roughly 20 percent of the city’s population, including a able to European conquest (Diamond 2009; see table 17.1). disproportionately large number of clinicians) to flee Subsequent pandemics have not had such dramatic their homes (Barrett and Brown 2008). Sudden popula- effects on political and social stability, primarily because tion movements can have destabilizing effects, and the potential mortality shock has been attenuated by migrants face elevated health risks arising from poor improvements in prevention and care. sanitation, poor nutrition, and other stressors (Toole Evidence does suggest that epidemics and pandemics and Waldman 1990). Migration also poses the risk of can amplify existing political tensions and spark unrest, further spreading an outbreak. particularly in fragile states with legacies of violence and Finally, outbreaks of infectious disease can cause weak institutions. During the 2014 West Africa Ebola already vulnerable social groups, such as ethnic minority epidemic, steps taken to mitigate disease transmission, populations, to be stigmatized and blamed for the dis- such as the imposition of quarantines and curfews by ease and its consequences (Person and others 2004). security forces, were viewed with suspicion by segments During the Black Death, Jewish communities in Europe of the public and opposition political leaders. This led faced discrimination, including expulsion and commu- directly to riots and violent clashes with security forces nal violence, because of stigma and blame for disease (McCoy 2014). Latent political tensions from previously outbreaks (Cohn 2007). Modern outbreaks have seen warring factions in Liberia also reemerged early in the more subtle forms of discrimination, such as shunning epidemic and were linked with threats to health care and fear, directed at minority populations linked with workers as well as attacks on public health personnel and disease foci. For example, Africans in Hong Kong SAR, facilities. China, reported experiencing social isolation, anxiety, The Ebola epidemic also greatly amplified political and economic hardship resulting from fears of their tensions in Guinea, Liberia, and Sierra Leone, with association with Ebola (Siu 2015). Pandemics: Risks, Impacts, and Mitigation 325 Trends Affecting Pandemic Risk assumptions (Pritchett, Woolcock, and Andrews 2013). In recent decades, several trends have affected pan- Many of these countries are in areas with high spark risk, demic probability, preparedness, and mitigation particularly in Central and West Africa, and thus may capacity. Various factors—population growth, increas- remain vulnerable and require significant international ing urbanization, greater demand for animal protein, assistance during a pandemic. greater travel and connectivity between population Other environmental and population trends that centers, habitat loss, climate change, and increased could increase the severity of pandemics include the interactions at the human-animal interface—affect the persistence of slums, unresponsive health systems, higher likelihood of pandemic events by increasing either the prevalence of comorbidities, weaker sanitation, and probability of a spark event or the potential spread of aging populations (Arimah 2010; UNDESA 2015). The a pathogen (Tilman and Clark 2014; Tyler 2016; Zell increasing threat posed by antibiotic resistance also 2004). With global population estimated to reach could amplify mortality during pandemics of bacterial 9.7 billion by 2050 and with travel and trade steadily diseases such as tuberculosis and cholera and even viral intensifying, public health systems will have less time diseases (especially for influenza, in which a significant to detect and contain a pandemic before it spreads proportion of deaths is often the result of bacterial pneu- (Tyler 2016). monia coinfections) (Brundage and Shanks 2008; Van As for poverty, the trends are mixed. On the positive Boeckel and others 2014). side, enormous gains in poverty reduction have decreased the number of people living in extreme poverty. This PANDEMIC MITIGATION: PREPAREDNESS may attenuate the mortality shock of a mild pandemic somewhat. On the negative side, extreme poverty is now AND RESPONSE concentrated in a small number of low-growth, high- Pandemic preparedness and response interventions can poverty countries (Chandy, Kato, and Kharas 2015). be classified by their timing with respect to pandemic In such countries, progress in building health system occurrence: the prepandemic period, the spark period, capacity also has been far slower. and the spread period, as shown in box 17.1. Likewise, for a subset of countries with endemically Whereas some interventions clearly fall under the pur- weak institutions, building institutional capacity for view of a single authority, responsibility for implement- complex tasks like pandemic mitigation and response is ing and scaling up many critical aspects of preparedness likely to be a slow process even under the most optimistic and response is spread across multiple authorities, which Box 17.1 Examples of Pandemic Preparedness and Response Activities, by Time Period Prepandemic period (before a pandemic starts) • Contact tracing, quarantine, and isolation • Stockpile building • Situational awarenessa • Continuity planning • Public health workforce training Spread period (after a pandemic starts) • Simulation exercises • Global pandemic declaration • Risk transfer mechanism set-up • Risk communications • Situational awarenessa • Contact tracing, quarantine, and isolation • Social distancing Spark period (as a pandemic starts) • Stockpile deployment • Initial outbreak detection • Vaccine or antiviral administration • Pathogen characterization or laboratory • Care and treatment confirmation • Situational awarenessa • Risk communication and community engagement a. Situational awareness includes passive and active animal and human disease • Animal disease control surveillance and monitoring of public health facilities and resources. 326 Disease Control Priorities: Improving Health and Reducing Poverty play complementary, interlocking, and, in some cases, detecting the most effective methods to reduce transmis- overlapping roles (Brattberg and Rhinard 2011). The sibility, and deciding where to allocate resources. During governance of pandemic preparedness and response is a pandemic, situational awareness allows for monitoring complex, with authority fragmented across international, to understand the course a pandemic is taking and national, and subnational institutions, as well as among whether intervention measures are effective. multiple organizations with functional responsibility The ability to detect the presence of a pandemic for specific tasks (Hooghe and Marks 2003). Pandemic requires the health care workforce to recognize the preparedness requires close coordination across public illness and to have the technical and laboratory capac- and private sector actors: vaccine development requires ity to identify the pathogen (or rule out known patho- close coordination between government and vaccine gens) and respond to surges of clinical specimens in a producers; whereas critical response measures—such as timely manner. Rapid identification reduces risk by managing quarantines—requires engagement between enabling infected persons to be isolated and given nonprofit organizations (hospitals, clinics, and nongov- appropriate clinical care. During the 2003 SARS pan- ernmental organizations), public health authorities, demic, a one-week delay in applying control measures affected communities and civil society groups, and the may have nearly tripled the size of the outbreak and security sector. increased its duration by four weeks (Wallinga and Historical pandemics offer only a partial view to Teunis 2004). guide preparedness and response activities. Many coun- Endemic infectious diseases can affect pandemic tries and organizations have used the historical influenza detection by complicating the differential diagnosis and pandemics in 1918, 1957, and 1968 to estimate the rapid identification of pandemic cases. Overlapping potential morbidity and mortality burden during a symptoms between endemic and emerging pathogens— future pandemic (WHO 2016c). However, using these for instance, between dengue and Zika or between moderate-to-severe events to plan for a mild pandemic malaria and Ebola—have hampered the early identifica- (for example, the 2009 influenza pandemic) can lead to tion of cases (de Wit and others 2016; Waggoner and an overzealous response—such as widespread manda- Pinsky 2016). This difficulty suggests a role for invest- tory school closures—that may create unintended nega- ment in the development and deployment of rapid diag- tive economic consequences (Kelly and others 2011). nostic tests in regions with a high burden of endemic And although the 1918 influenza pandemic is sometimes pathogens and high risk of disease emergence or impor- considered a “worst-case scenario” for planning pur- tation (Yamey and others 2017). Additional constraints poses, possible scenarios today could be far more affecting epidemic and pandemic situational awareness damaging—such as if a highly transmissible, highly vir- in LMICs are described in box 17.2. ulent influenza virus were to emerge. Especially in LMICs, intensive care unit (ICU) beds and therapies for acute respiratory distress syndrome are in short supply, which Preventing and Extinguishing Pandemic Sparks could lead to many casualties (Osterholm 2005). Although most pandemic preparedness activities focus on reducing morbidity and mortality after a pandemic has spread widely, certain activities may prevent and Situational Awareness contain pandemic sparks before they become a wider Situational awareness—in the context of pandemic threat. At the core of pandemic prevention is the concept preparedness—can be defined as having an accurate, of One Health, an approach that considers human up-to-date view of potential or ongoing infectious dis- health, animal health, and the environment to be funda- ease threats (including through traditional surveillance mentally interconnected (Zinsstag and others 2005).1 in humans and animals) and the resources (human, Activities that focus on understanding and controlling financial, informational, and institutional) available to zoonotic pathogens may prevent spillover events and manage those threats (ASPR 2014). Situational awareness subsequent pandemics (Morse and others 2012). is a crucial activity at all stages of a pandemic, including To understand the etiology of pandemics, impor- prepandemic, spark, and spread periods. It requires the tant One Health activities include the surveillance of support of health care resources (such as hospitals, doc- zoonotic pathogens of pandemic potential at the tors, and nurses), diagnostic infrastructure, and commu- human-animal interface, the modeling of evolutionary nications systems. It also requires the population to have dynamics, the risk assessments of zoonotic pathogens, access to and trust in the health care system. and other methods of understanding the interplay Situational awareness supports policy decisions by between environmental changes and pathogen tracking if and where disease transmission is occurring, emergence (Paez-Espino and others 2016; Wolfe and Pandemics: Risks, Impacts, and Mitigation 327 Box 17.2 Situational Awareness Constraints in Low- and Middle-Income Countries Perhaps the greatest challenge in epidemic and where health system gaps are significant, monitor- pandemic response is the timely identification and ing unofficial sources of information, including notification of the first pandemic case. However, rumors, may be useful (Samaan and others 2005). low- and middle-income countries are substantially Even once a potentially unusual or significant case slower than high-income countries to identify and has been identified, delays can be caused by low communicate infectious disease outbreaks (Chan statistical capacity, low data management capacity, and others 2010). In most outbreaks, the first (or and low communication capacity among local front- index) case is found retrospectively. Reporting delays line health workers. Delays also can arise from how result from multiple factors, which are discussed surveillance and reporting systems are designed— here. Moreover, the epidemiological characteristics for example, if health workers routinely report of the index case often are difficult to ascertain, potentially significant cases at the end of the month particularly in settings with limited diagnostic and rather than when they are identified. laboratory capacity. Another constraint arises from inconsistencies in Patients infected with potentially pandemic patho- real-time reporting of data. During an outbreak gens may present with nonspecific symptoms, response, national and regional health authorities making discriminating between endemic and novel must have strong relationships with local health or significant pathogens difficult unless differential providers to understand how data are generated and diagnostic tools are available. Gaps in health system reported at the clinical level. Robust monitoring and access and surveillance system coverage also ham- data validation procedures, such as the use of global per identification and reporting. In such cases, an positioning systems and case-based systems, along incipient epidemic will be identified only after with positive incentives for correct reporting, may sufficient deaths have occurred to draw the atten- help to alleviate such problems (Mancini and others tion of health authorities. Particularly in areas 2014). others 2005). For example, the PREDICT project of of basic information (such as how the pathogen is trans- the U.S. Agency for International Development mitted, guidance on managing patient care, high- (USAID) has invested a significant amount of resources risk practices, and protective behavioral measures) can in understanding and characterizing zoonotic risk rapidly and significantly reduce the transmission of (Anthony and others 2013).2 disease. Countries can focus their spark mitigation efforts on The way in which risk communications are framed policies designed to control animal reservoirs; monitor and transmitted matters a great deal; they must be clear, high-risk populations such as people working at the simple, timely, and delivered by credible messengers. animal interface (for example, those involved in animal Factors such as literacy rates, cultural sensitivities, famil- husbandry, animal slaughter, and so on); and maintain iarity with scientific principles (such as the germ theory robust animal health infrastructure, biosecurity, and of disease), and reliance on oral versus written traditions veterinary public health capacities (Jonas 2013; Pike and all have implications for how messages should be others 2010; Watts 2004; Yu and others 2014). designed and delivered (Bedrosian and others 2016). Public health officials also need to identify and address misinformation, rumors, and anxieties. This Risk Communications can be a significant challenge. During the 2014 West Risk communications can play a significant role in the Africa Ebola epidemic, many communities reached for control of an emerging epidemic or pandemic by pro- culturally familiar explanations of disease transmission viding information that people can use to take protective and rejected disease control practices that clashed with and preventive action (WHO 2013c). The dissemination their traditional healing and burial practices (Roca and 328 Disease Control Priorities: Improving Health and Reducing Poverty others 2015). Still other individuals spread rumors Curtailing Interactions between Infected and about the source of the infection; for example, in Liberia Uninfected Populations some community leaders claimed that the disease was The methods for curtailing interactions between infected created by the government (Epstein 2014). and uninfected populations include patient isolation, Rumors can impede disease control and can be quarantine, social distancing practices, school closures, amplified by mistrust of government officials, which is use of personal protective equipment, and travel a significant challenge in LMICs with high levels of restrictions. corruption or legacies of violent conflict and social The practice of quarantine began in the fourteenth division. Research has found that in unstable contexts, century in response to the Black Death and continues people tend to believe rumors that confirm their today (Mackowiak and Sehdev 2002). Quarantine and preexisting beliefs and anxieties (Greenhill and social distancing (such as the prohibition of mass Oppenheim 2017). This finding suggests that counter- gatherings) during the 1918 influenza pandemic reduced ing rumors with facts alone will not be sufficient. Risk spread and mortality rates, particularly when imple- communications need to be both factual and empa- mented in the early stages of the pandemic (Bootsma and thetic, addressing unfolding events and underlying Ferguson 2007; Hollingsworth, Ferguson, and Anderson fears through the lens of community experiences, his- 2006). During SARS and Ebola outbreaks, health agen- tories, and perceptions. cies and hospitals limited disease spread by isolating The effectiveness of risk communications is diffi- symptomatic patients, quarantining patient contacts, and cult to measure. However, previous risk communica- improving hospital infection control practices (Cohen tion efforts have brought forth overarching themes and others 2016; Twu and others 2003). During the 2003 that may be beneficial during the next epidemic or SARS pandemic, none of the health care workers in hos- pandemic. One notable model comes from a Nipah pitals in Hong Kong SAR, China, who reported appropri- virus outbreak in Bangladesh in 2010. In that out- ate and consistent use of masks, gloves, gowns, and hand break, investigators found that messages about the washing (as recommended under droplet and contact sources of infection and potential strategies to reduce precautions) were infected (Seto and others 2003). risk were more effective when conveyed by trusted Travel restrictions are sometimes implemented by local leaders and in terms that were relevant and governments to curtail disease spread. Fear and lack of grounded in the shared experiences of the affected scientific understanding may motivate the imposition of community (Parveen and others 2016). travel restrictions (Flahault and Valleron 1990). As such, these measures are sometimes implemented for inappro- priate pathogens or too late to contain an outbreak and Reducing Pandemic Spread can cause substantial economic damage and public anx- Once a pandemic has begun in earnest, public health iety. Travel restrictions are more beneficial for pathogens efforts often focus on minimizing its spread. Limiting that do not have a significant asymptomatic carrier state the spread of a pandemic can help to reduce the number and have a relatively long incubation period (for exam- of total people who are infected and thus also mitigate ple, SARS and Ebola). However, such restrictions may be some of the indirect health and economic effects. of limited efficacy for influenza pandemics unless initi- Strategies to minimize pandemic spread include the fol- ated when there are fewer than 50 cases at the spark site lowing (Ferguson and others 2005): (Ferguson and others 2005). • Curtailing interactions between infected and unin- Reducing Infectiousness and Susceptibility fected populations: for example, through patient Vaccines, antibiotics, and antiviral drugs can play a isolation, quarantine, social distancing practices, and critical role in mitigating a pandemic by reducing the school closures infectiousness of symptomatic patients and the sus- • Reducing infectiousness of symptomatic patients: for ceptibility of uninfected individuals. Antivirals may example, through antiviral and antibiotic treatment reduce influenza transmission, although the extent of and infection control practices their effectiveness is unclear (Ferguson and others • Reducing susceptibility of uninfected individuals: for 2005; Jefferson and others 2014). A systematic review example, through vaccines. of clinical trial data among treated adults showed that oseltamivir reduced the duration of influenza symp- During the prepandemic period, plans for imple- toms by 17 hours, but prophylaxis trials found no sig- menting those measures should be developed and tested nificant reduction of transmission (Jefferson and through simulation exercises. others 2014). Pandemics: Risks, Impacts, and Mitigation 329 If available, vaccines can reduce susceptibility. Care and Treatment to Reduce the Severity of Significant efforts have focused on speeding up vaccine Pandemic Illness development and scaling up production. However, the During a pandemic, health authorities work to reduce availability of vaccines—particularly in LMICs— the severity of illness through patient care and treat- depends on the affected area’s capacity for distribution ment, which can help decrease the likelihood of severe (including the scale and integrity of the cold chain), its outcomes such as hospitalizations and deaths. Treatments capacity for last-mile delivery to rural areas, and the may range from nonspecific, supportive care to dis- population’s willingness to adopt the vaccine. Vaccination ease-specific drugs. During the prepandemic period, strategies targeting younger populations may be espe- plans to implement these measures should be developed cially beneficial, in part because influenza transmissibil- and tested through simulation exercises. ity is higher among younger populations during Maintaining supportive care during an epidemic or pandemics (Miller and others 2008). pandemic can improve mortality rates by alleviating the The effectiveness of antivirals, antibiotics, and vac- symptoms of disease. During the 2014 West Africa Ebola cines in reducing spread diminishes if the pandemic is epidemic, for example, evidence suggests that earlier already global, if LMICs cannot afford adequate vaccine case identification, supportive care, and rehydration stocks for their populations, or if specific populations therapy modestly reduced mortality (Walker and Whitty (for example, the poor or the socially vulnerable) cannot 2015). Indeed, despite the unavailability of antivirals or access vaccines. Additionally, pandemics may be caused vaccines, efforts to engage communities with added by a pathogen without an available vaccine or efficacious medical supplies and trained clinicians decreased the biomedical therapy. Efforts to improve the vaccine case-fatality ratio moderately as more patients trusted, development pipeline are underway (box 17.3). sought, and received clinical care (Aylward and others 2014). Box 17.3 Vaccine Research and Development to Meet Pandemic Threats Current vaccine research, development, and pro- global influenza pandemics, quicken the produc- duction time lines are not conducive to quick tion of vaccines, and research a universal influenza responses to pandemic threats. For example, despite vaccine (Nannei and others 2016). Egg- biomedical advances, most influenza vaccines are independent cell culture platforms also have produced through vaccine platforms that rely on become a reality: in 2013 the U.S. Food and Drug the availability of embryonated chicken eggs and Administration approved an influenza vaccine can take several months to produce (Reperant, produced in insect cell lines (Milián and Kamen Rimmelzwaan, and Osterhaus 2014). Vaccines that 2015). are in development may take decades to become available for human use. For example, Ebola vac- In preparation for a noninfluenza pandemic, the cines were in development for more than a decade, public-private Coalition for Epidemic Preparedness with the first vaccine approved for clinical use Innovations (CEPI) is building a bank of potential only in 2015 (Henao-Restrepo and others 2016; vaccines for viral diseases, such as SARS and MERS Richardson and others 2010). (Middle East respiratory syndrome), that are not currently of commercial interest. CEPI’s goal is to Several areas of active research seek to hasten and focus on the development or licensure and manu- strengthen vaccine development. Of note is the facturing of high-potential viral vaccines through World Health Organization’s Global Action Plan early-stage human trials and to purchase small for Influenza Vaccines, whose mission, in part, is to stockpiles to mitigate the next pandemic (Mullard increase the capacity to produce vaccines for 2016). 330 Disease Control Priorities: Improving Health and Reducing Poverty Medical supplies that may be needed for supportive care • Prepandemic vaccines may not be closely matched to during a pandemic include hospital beds, disinfectants, the pathogen causing the pandemic. ICU supplies (such as ventilators), and personal protec- • The optimal size of a stockpile can be challenging to tive equipment (WHO 2015b). determine. Medical interventions for pandemic influenza include • Stockpiles need to be refreshed regularly, because antiviral drugs and antibiotics to treat bacterial coinfec- pharmaceuticals and equipment can reach expiration tions. Antivirals especially may reduce mortality when dates. given within 48 hours of symptom onset (Domínguez- • Robust health systems and channels for disseminat- Cherit and others 2009; Jain and others 2009). However, ing and using the stockpiles also must exist. because of delays in case identification and antiviral deployment (as discussed in box 17.2), LMICs may Boosting local production capacity for necessary sup- experience only limited benefits from antiviral drugs. plies may be a viable strategy for pandemic preparedness and may circumvent some of the challenges associated with amassing stockpiles. Potential for Scaling Up The 2009 influenza pandemic demonstrated how The term scaling up refers to the expansion of health scaling up can affect the success rate of a mass vaccina- intervention coverage (Mangham and Hanson 2010). In tion campaign (table 17.5). Vaccination rates increased the context of pandemic preparedness, successfully scal- according to country income level, suggesting that vacci- ing up requires health systems to expand services to nation campaigns were most successful in HICs, likely accommodate rapid increases in the number of sus- because of the size of their stockpiles, increased manu- pected cases. Scaling up is facilitated by surge capacity facturing capacity for vaccines, increased availability of (the ability to draw on additional clinical personnel, vaccines, and more streamlined logistics in vaccine logisticians, and financial and other resources) as well as deployment. preexisting operational relationships and plans linking Building local capacity to scale up is challenging, government, nongovernmental organizations, and the especially in LMICs. The biggest challenges include private sector. Ultimately, scaling up consists of having infrastructural gaps (such as weak road, transporta- both local surge capacity and the absorptive capacity to tion, and communications networks) and shortfalls in accept outside assistance. human resources (such as logisticians, epidemiolo- Local capacity building is vital, and some capacities gists, and clinical staff). Bilateral and multilateral aid may have particularly important positive externalities organizations have channeled substantial funding into during outbreaks. During the 2014 Ebola importation building and sustaining local technical capacities in into Nigeria, surge capacity that existed because of polio LMICs. This type of investment is critically important. eradication efforts contributed to a more successful out- But, particularly in LMICs with weak health system break response (Yehualashet and others 2016). Key ele- capacity, progress in expanding local surge capacity ments included national experience running an likely will be slow. emergency operations center and the use of global posi- Another key component of scaling up, especially in tioning systems to support contact tracing (Shuaib and LMICs, is the ability to use external assistance effectively. others 2014; WHO 2015a). Stockpiling of vaccines, medicines (including antibi- otics and antivirals), and equipment (such as masks, Table 17.5 Vaccination Rates during the 2009 Influenza gowns, and ventilators) also can be useful for building Pandemic, by Country Income Level local surge capacity (Dimitrov and others 2011; Jennings Country income Number of countries Share of population and others 2008; Morens, Taubenberger, and Fauci 2008; levela with data vaccinated (%) Radonovich and others 2009). During a pandemic, health systems can tap into stockpiles more quickly than they Low-income 13 5.7 can procure supplies from external sources or boost pro- Middle-income 42 8.5 duction. However, there are five important consider- High-income 31 16.8 ations for keeping stockpiles: Sources: Mihigo and others 2012; Tizzoni and others 2012; WHO 2013b. a. Income groups follow World Bank income classifications for fiscal 2018, based on estimates of • Building a stockpile requires significant up-front 2016 gross national income per capita and calculated using the World Bank Atlas method: low-income (US$1,005 or less), middle-income (US$1,006–US$12,235), and high-income costs, which can be especially prohibitive for LICs (US$12,236 or more). For further explanation, see https://datahelpdesk.worldbank.org (Oshitani, Kamigaki, and Suzuki 2008). /knowledgebase/articles/906519-world-bank-country-and-lending-groups. Pandemics: Risks, Impacts, and Mitigation 331 During the 2014 West Africa Ebola epidemic, a surge shocks to public health and public finances. Risk transfer of foreign clinicians, mobile medical units, and epidemi- mechanisms (such as specialized insurance facilities) ologists and other public health personnel was required offer an additional tool to manage this risk. to bolster limited local resources. LMICs can improve Risk-based insurance products are increasingly systems to facilitate and coordinate surges of foreign deployed in LMICs to pay for remediation and recon- support in the following ways: struction costs following natural catastrophes such as hurricanes, floods, and droughts (ARC 2016; IFRC • Streamline customs processes for critical medical 2016). Insurance products for epidemics and pandem- supplies and drugs. ics require specific characteristics. First, insurance • Establish mechanisms to coordinate the deployment policies should be designed to release discretionary and operations of foreign medical teams. funds early in the course of an outbreak. In situations • Build mechanisms to coordinate between military where financing poses a constraint to mobilizing per- and humanitarian units involved in crisis response. sonnel, drugs, or other supplies, payouts can be used to mobilize a public health response and mitigate further Even so, local absorptive capacity (that is, the ability spread of disease, reducing the potential health and to channel and use foreign assistance effectively) has its economic impacts of the pandemic. Second, because limits. Constraints in bureaucratic capacity, financial pandemics do not stay contained in national borders, controls, logistics, and infrastructure all are likely to be a strong case can be made for mobilizing bilateral and most severe in the countries that most need foreign multilateral financing of LMICs’ insurance premiums assistance to manage infectious disease crises. as a cost-effective way to improve global preparedness Furthermore, although external assistance is a and support mitigation efforts. Third, risk transfer viable strategy during localized epidemics, it has lim- systems require the availability of rigorously and itations that are likely to arise during large-scale pan- transparently compiled data to trigger a payout. In the demics. First, supply constraints exist, including limits context of pandemic insurance, the development of to the number of medical personnel (especially those risk transfer systems requires countries to build the with crisis response and infectious disease competen- following capacities, among others: cies) and the number of specialized resources (such as integrated mobile medical clinics available for • Robust surveillance data to identify when an out- deployment). break has reached sufficient scale to require the Second, during a severe pandemic, countries are release of funds likely to use such resources locally before providing • Laboratory capacity to confirm the causative agent medical assistance abroad. The global humanitarian sys- • Predefined contingency and response plans to spend tem provides a critical reservoir of crisis response capac- the funds effectively upon their release. ity and shock absorption. However, the humanitarian system currently is straining under the pressure of other Insurance facilities can create positive incentives for crises, including upsurges in violent conflict (Stoddard LMICs to invest in planning and capacity building. and others 2015). A severe epidemic or pandemic can Insurance mechanisms may have other positive external- quickly outstrip international resources. Médecins Sans ities: most notably, the potential release of funds may Frontières (Doctors Without Borders), an international provide a strong incentive for the timely reporting of health organization with deep experience providing surveillance data. However, insurance facilities also may Ebola treatment, found itself “pushed to the limits and introduce perverse incentives (including incentives to beyond” during the 2014 West Africa Ebola epidemic distort surveillance data) and potential moral hazards (MSF 2015). (such as permitting riskier activities). These incentive problems may be mitigated in the design of the risk transfer mechanism, such as by providing coverage only Risk Transfer Mechanisms when minimum requirements for surveillance accuracy As with any other type of natural disaster, the risk from are met, by having preset phased triggers for payouts, pandemics cannot be eliminated. Despite prevention and by including incentive payouts for successfully con- efforts, pandemics will continue to occur and will at taining an outbreak. times overwhelm the systems that have been put in place Relative to investments in basic health provision, to mitigate their health, societal, and economic effects. building capacity in infectious disease surveillance sys- The residual risk may be significant, particularly for tems and other dimensions of pandemic preparedness LMICs that lack the resilience or resources to absorb has uncertain and potentially distant benefits. In LICs 332 Disease Control Priorities: Improving Health and Reducing Poverty where near-term health needs are acute, this can compli- Figure 17.3 Hypothetical Pandemic Preparedness Budget and cate the political and economic logic for investing in Response Shortfall, Which Could Be Managed via Risk Transfer pandemic preparedness (Buckley and Pittluck 2016). Mechanisms The use of catastrophe modeling tools (such as EP curves) can clarify the benefit-cost rationale and the relevant time horizon for investments in preparedness, and it can inform the design and financial structure of pandemic insurance policies. Response Figure 17.3 shows a country’s hypothetical pandemic shortfall US$75 million preparedness budget allocation and the portion of risk Estimated transfer in estimated total costs of spread response. total spread In this example, a country has a total budget of US$100 response costs million to cover all aspects of pandemic preparedness US$125 million during the prepandemic, spark, and spread periods. After allocating half of the funds for prepandemic and spark response activities, US$50 million is left for pan- Spread response demic spread response. On the basis of its risk tolerance, US$50 million Available the country makes a decision to manage its risk at the budget for 3 percent annual probability point on its EP curve. pandemic Modeling estimates indicate that a successful response to preparedness Spark response and response a pandemic at this level would require at least US$125 US$30 million US$100 million million, which would fund spread response activities, shown in box 17.1. Because only US$50 million is left after allocation to prepandemic and spark response Prepandemic US$20 million activities, this would leave a shortfall of US$75 million. Some or all of this shortfall could be offloaded to Source: Metabiota. another entity, such as a catastrophe risk insurance pool, Note: Numbers are provided solely for illustrative purposes. which would give the country access to a payout during a pandemic. Innovations in pandemic financing have been devel- (clinicians, community health workers), personal pro- oped in response to the significant burden that a pan- tective equipment and other medical equipment con- demic can place on a country’s financial resources. One sumables, and vaccines and therapeutics, from either such innovation is the World Bank’s Pandemic domestic or international resources. Emergency Financing Facility (PEF) (Katz and Seifman 2016).3 A type of disaster risk pool, the PEF provides poorly resourced countries with an infusion of funds to Adequacy of Evidence on Pandemics in LMICs help with the costs of response in the early stages of an Much of the available data regarding pandemics (includ- epidemic or pandemic. The maximum total coverage ing the morbidity and mortality impacts of historical over a three-year period is US$500 million. Notably, the pandemics) and the effectiveness of different prepared- US$500 million coverage is much lower than the esti- ness efforts and interventions come from HICs and mated US$3.8 billion cost of the multinational response upper-middle-income countries. Understanding of the to the 2014 West Africa Ebola epidemic (USAID and prevalence of risk drivers, especially regarding spark risk, CDC 2016). Because the PEF is designed to trigger early has improved markedly in both high- and low-income in an outbreak, the anticipated funding is less than contexts. However, gaps in surveillance and reporting would be required for a full-fledged response once a infrastructure in LMICs mean that, during a pandemic, widespread pandemic is under way. many cases may never be detected or reported to the Risk transfer mechanisms such as insurance offer an appropriate authorities (Katz and others 2012). injection of financial resources to help insured parties Particularly in LICs, empirical data on outbreak occur- rapidly scale up disease response activities. As such, rences may be biased downward systematically. the utility of risk transfer mechanisms depends, in large Additionally, the means to disseminate collected data part, on the absorptive capacity of the insured party. rapidly may not exist. For example, data may be kept in A country must have the ability to use insurance pay- paper archives, so resource-intensive digitization may be outs effectively to access additional human resources required to analyze and report data to a wider audience. Pandemics: Risks, Impacts, and Mitigation 333 Data dissemination challenges are further compounded global health spending found that roughly 1 percent of by a publication bias that results in overrepresentation of global ODA spending on health in 2013 (approximately HICs in the scientific literature (Jones and others 2008). US$204 million) focused specifically on pandemic pre- paredness (Schäferhoff and others 2015). Other, non- ODA spending on pandemic preparedness is similarly SUMMARY OF PANDEMIC INTERVENTION difficult to measure but likely to be significant; in 2013, COSTS AND COST-EFFECTIVENESS the U.S. Department of Defense spent roughly US$256 Few data are available regarding costs and cost- million on efforts to build global biosurveillance and effectiveness of pandemic preparedness and response response capacities (KFF 2014). measures, and they focus almost exclusively on HICs. Globally, the current funding for pandemic prepared- The available data suggest that the greatest cost-related ness and response falls short of what is needed. In 2016, benefits in pandemic preparedness and response are the international Commission on a Global Health Risk realized from early recognition and mitigation of dis- Framework for the Future recommended an additional ease—that is, catching and stopping sparks before they US$4.5 billion annual global investment for upgrading spread. Costs can be reduced if action is taken before an pandemic preparedness at the country level, for funding outbreak becomes a pandemic. Similarly, once a infectious disease research and development efforts, and pandemic has begun, preventing illness generally is more for establishing or replenishing rapid-response financing cost-effective than treating illness, especially because mechanisms such as the World Bank’s PEF (Sands, hospitalizations typically have the highest direct cost per Mundaca-Shah, and Dzau 2016). person. High costs also may occur as a result of interven- Costs for efforts associated with prepandemic pre- tions (such as quarantines and school closures) that lead paredness activities also are not well quantified, although to economic disruption. These interventions may be investment in One Health activities is likely to be cost- more cost-effective during a severe pandemic. effective (World Bank 2012). The USAID PREDICT project has estimated that discovery and detection of the majority of zoonotic viruses would cost US$1.6 billion Program and Health System Costs (Anthony and others 2013). The Global Virome Project, No systematic time-series data exist on global spending a more comprehensive study aiming to characterize on pandemic preparedness, and arriving at an exact more than 99 percent of the world’s viruses, is estimated figure is complicated by the fact that many investments in to cost US$3.4 billion over 10 years (Daszak and others building basic health system capacity also support core 2016). Building on efforts to identify and describe the dimensions of pandemic preparedness. An analysis of ecology of potential pandemic viruses, the Coalition for Figure 17.4 Unit Costs for Selected Influenza Pandemic Response Activities 3,000 $2,470 2,500 Unit cost (2012 US$) 2,000 1,500 1,000 $699 500 $139 $170 $7 $9 $11 $18 $26 0 Over‐the- Prescription Antiviral Vaccination Antiviral Physician Emergency Hospitalization Intensive counter drugs prophylaxis (per dose) therapy (per visit) department (per day) care unit drugs excluding (per course) (per course) (per visit) (per day) (per course) antivirals (per course) Response Source: Based on Lugnér and Postma 2009. Note: Includes studies from France, Israel, the Netherlands, Singapore, the United Kingdom, and the United States. 334 Disease Control Priorities: Improving Health and Reducing Poverty Epidemic Preparedness Innovations (CEPI) estimated a and subsidies could push drug costs down even more. cost of US$1 billion over five years to develop vaccine Conversely, hospital care has the highest unit costs. Costs candidates against known emerging infectious diseases per day of hospitalization (especially those with ICU (for example, Ebola virus) and to build technology plat- involvement) can add up quickly when aggregated at forms and production facilities to accelerate vaccine the national level. However, these medical care costs response to outbreaks of known or unknown pathogens are potentially bounded by capacity limits (such as a finite (Brende and others 2017). number of hospital beds), especially during more severe Instituting response measures after a pandemic has pandemics. begun can be expensive, with most of the direct cost Pandemic severity itself can play a role in the drivers borne by the health care sector, although response costs of cost and the effects of mitigation efforts. One study typically are not reported in a cohesive manner. As based on modeling simulations in an Australian popula- noted, the response to the 2014 West Africa Ebola epi- tion found that, in low-severity pandemics, most costs demic cost more than US$3.8 billion, including dona- borne by the larger economy (not just the health care tions from several countries (USAID and CDC 2016). system) come from productivity losses related to illness Additionally, the World Bank Group mobilized US$1.6 and social distancing. In higher-severity pandemics, the billion from the International Development Association largest drivers of costs are hospitalization costs and pro- and the International Finance Corporation to stimulate ductivity loss because of deaths (Milne, Halder, and economic recovery in the three worst-affected coun- Kelso 2013). tries of Guinea, Liberia, and Sierra Leone (World Bank 2016). Taken together, at US$5.4 billion, these values amount to a cost of US$235 per capita for these three Costs per Death Prevented countries. Figure 17.5 depicts a compilation of data from 18 When total costs for response are not available, unit scientific publications that examined costs and benefits costs for response activities provide valuable insights. associated with response during the 2009 influenza Figure 17.4 shows estimated unit costs for selected pandemic. The lowest costs per deaths prevented were response measures, based on modeling studies for pan- found for contact tracing, face masks, and surveillance. demic influenza in HICs. Vaccinations and medicines have Pharmaceutical interventions such as vaccines and the lowest unit costs; in LMICs, large-scale purchasing antiviral therapies were in the midrange. Figure 17.5 Health Care System and Economic Costs per Death Prevented for Selected Interventions during the 2009 Influenza Pandemic School closure 9,860,000 Quarantine 2,210,000 Antiviral therapy 1,770,000 Intervention measure Social distancing 1,640,000 Antiviral stockpile 519,000 Vaccination 297,000 Surveillance 3,770 Face masks 2,320 Contact tracing 2,260 0 2 million 4 million 6 million 8 million 10 million 12 million Costs (2012 US$) per death prevented Source: Based on data from Pasquini-Descomps, Brender, and Maradan 2016. Note: Includes studies from Australia, Brazil, Canada, China, Singapore, Sweden, the United Kingdom, and the United States. Pandemics: Risks, Impacts, and Mitigation 335 Measures that decreased person-to-person contact, the cost per death prevented could decrease for some including social distancing, quarantine, and school clo- interventions, such as school closures. sures, had the greatest cost per death prevented, most • Results are sensitive to assumptions about the value likely because of the amount of economic disruption of a prevented death and estimated costs of different caused by those measures. Social distancing includes interventions. avoidance of large gatherings and public places where • The data cover only pandemics caused by influenza. economic activities occur. School closures often lead to For pandemics caused by other types of pathogens, lost productivity because they cause workplace absen- the cost-utility values may be different, and not all teeism among caretakers of school-age children. intervention measures may be available. Macroeconomic model simulations also have identified school closures as a potential source of GDP loss during Data on antiviral stockpiles provide some insight into a moderately severe pandemic (Smith and others 2009). how the cost utility of pandemic preparedness efforts The information shown in figure 17.5 is subject to may vary by country income level. Figure 17.6 shows the several caveats: cost utility of antiviral stockpiling by country income level, based on simulation studies. • The data come from only a few studies covering a A more recent study found that antiviral stockpiling handful of countries. in Cambodia (a lower-middle-income country) would • Cost-utility analyses of pandemic preparedness and cost between US$3,584 and US$115,168 per death pre- response for LMICs are rare. Because the underlying vented; however, this result is highly sensitive to assump- data for these studies were drawn primarily from tions about the timing between pandemics (Drake, HICs, the estimates may not accurately represent Chalabi, and Coker 2015). the relative benefit-cost of interventions in LMICs. Although based only on a handful of countries, the For example, in countries with high unemployment results suggest that antiviral stockpiling in LICs has an and underemployment, school closures may not lead extremely high cost per death prevented, whereas coun- to increased workforce absenteeism and thus might tries at other income levels are clustered within much have a lower cost per death prevented. lower ranges. Antiviral stockpiling is not cost-effective • The 2009 influenza pandemic is considered a relatively or feasible for LICs, primarily because of the high cost mild pandemic. In a more severe influenza pandemic, of antiviral agents. For stockpiling to be a cost-effective Figure 17.6 Cost Utility of Antiviral Stockpiling for Pandemic Influenza Preparedness, by Share of Population Covered and Country Income Level, 2011 High-income Country income level Upper-middle income Lower-middle income Low-income 0 1 million 2 million 3 million 4 million 5 million Costs (2012 US$) per death prevented 10% of population covered 20% of population covered 30% of population covered Source: Based on data from Carrasco and others 2011. Note: Includes data from one low-income country (Zimbabwe), three lower-middle-income countries (Guatemala, India, and Indonesia), two upper-middle-income countries (Brazil and China), and four high-income countries (New Zealand, Singapore, the United Kingdom, and the United States). 336 Disease Control Priorities: Improving Health and Reducing Poverty strategy for LICs, almost all of the costs would have to as effective in reducing transmission as having vaccines be subsidized. The associated costs also may be reduced available and distributed earlier in the pandemic. by the increased availability of generic antiviral drugs. Allocation of limited resources (by creating priority Additionally, the efficacy of antivirals is not assured, groups for vaccines and antivirals) is an important particularly for LICs, which may not be able to iden- consideration during a pandemic. Modeling studies tify cases early enough to administer antivirals from the 2009 influenza pandemic investigated the efficaciously. most cost-effective strategies for allocating vaccines. Those studies found that vaccinating high-risk individuals was more cost-effective than prioritizing Cost-Effectiveness children. Favoring children decreased the overall infec- Pérez Velasco and others (2012) synthesized informa- tion rate, but high-risk individuals were the predomi- tion from 44 studies that contained economic evalua- nant drivers of direct costs during the pandemic, tions of influenza pandemic preparedness and response because they were more likely to be hospitalized (Lee strategies in HICs (figure 17.7). In their analysis, the and others 2010). However, these studies did not following interventions among the general population account for the indirect costs of school closures and had the potential to provide cost savings: vaccines, anti- absenteeism. Consideration of these factors could reveal viral treatment, social distancing, antiviral prophylaxis increased cost savings from vaccinating children. plus antiviral treatment, and vaccines plus antiviral Another key question for benefit-cost analyses treatment. The cost savings from antiviral drugs found related to pandemics is the extent to which stockpiles in this study are likely to be diminished in LMICs, as of vaccines, antiviral drugs, and protective equip- inability to deploy antivirals in a timely manner poses a ment should be assembled in advance of a pandemic. serious challenge to their efficacious use. Vaccines for a novel influenza virus can take several Depending on the characteristics of a pandemic and months to develop, and vaccines for other pathogens how mitigation efforts are implemented, some mitiga- (for example, Ebola and Zika) can take even longer to tion strategies could become highly cost-ineffective. For develop. Studies have examined the cost-effectiveness example, a costly vaccination campaign that is carried of stockpiling prepandemic vaccines that have lower out in an area well after a pandemic peaks is not nearly efficacy than reactive vaccines but can be deployed Figure 17.7 Cost-Effectiveness of Selected Interventions for Pandemic Influenza Preparedness and Response in High-Income Countries School closure 1,630 Quarantine of household contacts 1,310 Intervention measure Vaccine 490 AV therapy 147 Vaccine + AV therapy 139 AV prophylaxis + AV therapy 49 AV prophylaxis + school closure 0.4 Vaccine + school closure 0.3 0 200,000 400,000 600,000 800,000 1 million 1.2 million 1.4 million 1.6 million 1.8 million ICER (2012 US$ per death prevented) More cost-effective Less cost-effective Source: Based on Pérez Velasco and others 2012. Note: AV = antiviral; ICER = incremental cost-effectiveness ratio. Pandemics: Risks, Impacts, and Mitigation 337 more quickly. One study found that cost savings can probabilistic modeling and EP curves, can quantify the be obtained as long as prepandemic vaccines have at potential pandemic risks facing each country and clar- least 30 percent efficacy. However, cost-effectiveness ify the benefit-cost case for investing in pandemic differs by pandemic severity and the percentage preparedness. of the population that receives the vaccine dur- No single, optimal response to a public health emer- ing the vaccination campaign (Halder, Kelso, and gency exists; strategies must be tailored to the local Milne 2014). context and to the severity and type of pandemic. Antiviral drugs to fight pandemic influenza also can However, overarching lessons emerge after multiple be stockpiled ahead of time. However, the optimal num- regional epidemics and global pandemics. For example, ber of doses to stockpile depends on factors including because of their high spark and spread risks, many the effectiveness of concurrent interventions and the LMICs would benefit most from building situational likelihood of antiviral wastage on noninfluenza respira- awareness and health care coordination capacity; public tory infections (Greer and Schanzer 2013). health response measures are far more cost-effective Most pandemic-related benefit-cost studies focus if they are initiated quickly and if scarce resources are on pharmaceutical interventions for high-income and targeted appropriately. upper-middle-income countries. The studies have Building pandemic situational awareness is complex, largely neglected the question of how to allocate requiring coordination across bureaucracies, across the strained resources in low- and lower-middle-income public and private sectors, and across disciplines with countries. Furthermore, few evaluations have been different training and different norms (including epi- conducted of the cost-effectiveness of general invest- demiology, clinical medicine, logistics, and disaster ment in health systems, infrastructure, and capacity response). However, an appropriately sized and trained building as a means to achieve pandemic preparedness health workforce (encompassing doctors, nurses, epi- (Drake, Chalabi, and Coker 2012). demiologists, veterinarians, laboratorians, and others) that is supported by adequate coordination systems is a fundamental need—the World Health Organization has CONCLUSIONS AND RECOMMENDATIONS FOR recommended a basic threshold of 23 skilled health pro- PRIORITIZING INVESTMENTS TO MITIGATE fessionals per 10,000 people (WHO 2013a). PANDEMIC RISK IN RESOURCE-LIMITED Increasing the trained health workforce also will increase the capacity to detect whether any particular SETTINGS population (for example, human, farm animal, or wild- Preparing for a pandemic is challenging because of a life) is suffering from a pathogen with high pandemic multitude of factors, many of which are unique among risk. Increasing the health workforce also will improve natural disasters. Pandemics are rare events, and the risk the overall resiliency of the health system, an improve- of occurrence is influenced by anthropogenic changes in ment that can be applied to any emergency that results the natural environment. In addition, accountability for in morbidity and mortality shocks. preparedness is diffuse, and many of the countries at Additionally, building situational awareness will greatest risk have the most limited capacity to manage require sustained investment in infectious disease sur- and mitigate pandemic risk. veillance, crisis management, and risk communications Unlike most other natural disasters, pandemics do systems. Investments in these capacities are likely to not remain geographically contained, and damages can surge after pandemic or epidemic events and then abate be mitigated significantly through prompt intervention. as other priorities emerge. Hence, stable investment to As a result, there are strong ethical and global health build sustained capacity is critical. imperatives for building capacity to detect and respond Risk transfer mechanisms such as catastrophe risk to pandemic threats, particularly in countries with weak pools offer a viable strategy for countries to manage preparedness and high spark and spread risk. pandemic risk. Further developing these mechanisms Investments to improve pandemic preparedness will allow countries to offload portions of pandemic risk may have fewer immediate benefits, particularly rela- and response that are beyond their immediate budgetary tive to other pressing health needs in countries with capacity. For this reason, risk transfer solutions should heavy burdens of endemic disease. Therefore, charac- be designed with the needs and constraints of LMICs in terizing pandemic risk and identifying gaps in pan- mind. However, countries must have predefined contin- demic preparedness are essential for prioritizing and gency and response plans as well as the absorptive capac- targeting capacity-building efforts. Thinking about ity to use the emergency financing offered by such risks in terms of frequency and severity, notably using solutions. Broad and effective use of pandemic insurance 338 Disease Control Priorities: Improving Health and Reducing Poverty will require parallel investments in capacity building and 2. PREDICT, a project of USAID’s Emerging Pandemic emergency response planning. Threats Program, was initiated in 2009 to strengthen Finally, researchers must address the significant global capacity for detection and discovery of zoonotic knowledge gaps that exist regarding LMICs’ pandemic viruses with pandemic potential. Working with partners in 31 countries, PREDICT is building platforms for conduct- preparedness and response. Improving the tracking of ing disease surveillance and for identifying and monitoring spending and aid flows specifically tied to pandemic pathogens that can be shared between animals and people. prevention and preparedness is vital to tracking gaps Using the One Health approach, the project is investigat- and calibrating aid flows for maximum efficiency. ing the behaviors, practices, and ecological and biological Systematic data on response costs in low-income set- factors driving the emergence, transmission, and spread tings are scarce, including data regarding spending on of disease. For more information, see the project website, clinical facilities, supplies, human resources, and http://www.vetmed.ucdavis.edu/ohi/predict/. response activities such as quarantines. Bridging these 3. For more information about the PEF, see the brief on the data gaps can improve pandemic preparedness planning World Bank website, “Pandemic Emergency Financing and response through evidence-based decision mak- Facility: Frequently Asked Questions,” http://www ing and support efforts to prevent and mitigate epidem- . worldbank .org/en/topic/pandemics/brief / pandemic -emergency-facility-frequently-asked-questions. ics and pandemics. ACKNOWLEDGMENTS REFERENCES The authors would like to thank Ron Waldman, Dean Achonu, C., A. Laporte, and M. A. Gardam. 2005. “The Financial Jamison, Steve Morse, Peter Sands, Mukesh Chawla and Impact of Controlling a Respiratory Virus Outbreak in a the Secretariat of the International Working Group on Teaching Hospital: Lessons Learned from SARS.” Canadian Financing Preparedness, Kimberly Dodd, Mary Guttieri, Journal of Public Health 96 (1): 52–54. Anthony, S. J., J. H. Epstein, K. A. Murray, I. Navarrete-Macias, Patrick Ayscue, Damien Joly, Sarah Barthel, Joseph C. M. Zambrana-Torrelio, and others. 2013. “A Strategy to Krilanovich, Jonathan Koshi, Mike Gahan, Kierste Miller, Estimate Unknown Viral Diversity in Mammals.” MBio 4 Cathine Lam, Catherine Planey, Nicole Stephenson, (5): e00598-13. Volodymyr Serhiyenko, Ming Yii Goh, Alessandro Arabi, Y. M., H. H. Balkhy, F. G. Hayden, A. Bouchama, T. Luke, Vespignani, Qian Zhang, Matteo Chinazzi, Ana Pastore y and others. 2017. “Middle East Respiratory Syndrome.” Piontti, Marco Schäferhoff, and Jessica Kraus for their New England Journal of Medicine 376 (6): 584–94. valuable technical and editorial contributions to the ARC (African Risk Capacity). 2016. “African Risk Capacity chapter. Strategic Framework 2016–2020.” Strategy and policy doc- ument, ARC, Johannesburg, South Africa. Arimah, B. C. 2010. “The Face of Urban Poverty: Explaining NOTES the Prevalence of Slums in Developing Countries.” In Urbanization and Development: Multidisciplinary This chapter uses World Bank Income Classifications for 2018 Perspectives , edited by Jo Beall, Basudeh Guha- as follows, based on estimates of gross national income (GNI) Khasnobis, and Ravi Kanbur, 143–64. Oxford: Oxford per capita for 2015: University Press. ASPR (Assistant Secretary for Preparedness and Response). • Low-income countries (LICs) = US$1,005 or less 2014. “Public Health and Medical Situational Awareness • Middle-income countries (MICs) are subdivided: Strategy.” Strategy document for situational awareness (a) lower-middle-income = US$1,006 to US$3,995 implementation plan, U.S. Department of Health and (b) upper-middle-income countries (UMICs) = US$3,996 Human Services, Washington, DC. to US$12,235 Aylward, B., P. Barboza, L. Bawo, E. Bertherat, P. Bilivogui, and • High-income countries (HICs) = US$12,236 or more. others. 2014. “Ebola Virus Disease in West Africa—The First 9 Months of the Epidemic and Forward Projections.” 1. One Health considers individual, community, and animal New England Journal of Medicine 371 (16): 1481–95. health as interconnected and requires the collaboration of Barden-O’Fallon, J., M. A. Barry, P. Brodish, and J. Hazerjian. human, animal, and environmental health professionals 2015. “Rapid Assessment of Ebola-Related Implications to recognize and alleviate the problems on one level to for Reproductive, Maternal, Newborn, and Child Health reduce the downstream health effects on another level Service Delivery and Utilization in Guinea.” PLoS (for example, rabies in animals and humans). For more Currents Outbreaks (August): 7. doi:10.1371/currents information, see the U.S. Centers for Disease Control .outbreaks.0b0ba06009dd091bc39ddb3c6d7b0826. and Prevention’s webpage, https://www.cdc.gov/onehealth Barrett, R., and P. J. Brown. 2008. “Stigma in the Time of /basics/index.html. Influenza: Social and Institutional Responses to Pandemic Pandemics: Risks, Impacts, and Mitigation 339 Emergencies.” Journal of Infectious Diseases 197 (Suppl 1): Colizza, V., A. Barrat, M. Barthelemy, A. J. Valleron, and S34–S37. A. Vespignani. 2007. “Modeling the Worldwide Spread Bedrosian, S. R., C. E. Young, L. A. Smith, J. D. Cox, C. Manning, of Pandemic Influenza: Baseline Case and Containment and others. 2016. “Lessons of Risk Communication and Interventions.” PLoS Medicine 4 (1): 95–110. Health Promotion—West Africa and United States.” Daszak, P., D. Carroll, N. Wolfe, and J. Mazet. 2016. “The Global Morbidity and Mortality Weekly Report (MMWR) Virome Project.” International Journal of Infectious Diseases Supplements 65 (3): 68–74. 53 (Suppl): 36. Bootsma, M. C. J., and N. M. Ferguson. 2007. “The Effect Dawood, F. S., A. D. Iuliano, C. Reed, M. I. Meltzer, D. K. Shay, of Public Health Measures on the 1918 Influenza and others. 2012. “Estimated Global Mortality Associated Pandemic in U.S. Cities.” Proceedings of the National with the First 12 Months of 2009 Pandemic Influenza Academy of Sciences of the United States of America 104 A H1N1 Virus Circulation: A Modelling Study.” The Lancet (18): 7588–93. Infectious Diseases 12 (9): 687–95. Brattberg, E., and M. Rhinard. 2011. “Multilevel Governance de Wit, E., D. Falzarano, C. Onyango, K. Rosenke, A. Marzi, and and Complex Threats: The Case of Pandemic Preparedness others. 2016. “The Merits of Malaria Diagnostics during an in the European Union and the United States.” Global Ebola Virus Disease Outbreak.” Emerging Infectious Diseases Health Governance 5 (1): 1–21. 22 (2): 323–26. Brende, B., J. Farrar, D. Gashumba, C. Moedas, T. Mundel, and DeWitte, S. N. 2014. “Mortality Risk and Survival in the others. 2017. “CEPI—A New Global R&D Organisation Aftermath of the Medieval Black Death.” PLoS One 9 (5): for Epidemic Preparedness and Response.” The Lancet 389 e96513. (10066): 233–35. Diamond, J. 2009. Guns, Germs, and Steel: The Fates of Human Brundage, J. F., and G. D. Shanks. 2008. “Deaths from Bacterial Societies. New York: Norton. Pneumonia during 1918–19 Influenza Pandemic.” Emerging Dimitrov, N., S. Goll, N. Hupert, B. Pourbohloul, and L. Meyers. Infectious Diseases 14 (8): 1193–99. 2011. “Optimizing Tactics for Use of the U.S. Antiviral Buckley, G. J., and R. E. Pittluck. 2016. Global Health Risk Strategic National Stockpile for Pandemic Influenza.” PloS Framework: Pandemic Financing: Workshop Summary. One 6 (1): e16094. Washington, DC: National Academies Press. Dixon, S., S. McDonald, and J. Roberts. 2001. “AIDS and Burns, A., D. Van der Mensbrugghe, and H. Timmer. Economic Growth in Africa: A Panel Data Analysis.” Journal 2006. “Evaluating the Economic Consequences of of International Development 13 (4): 411–26. Avian Influenza.” Working Paper 47417, World Bank, Domínguez-Cherit, G., S. E. Lapinsky, A. E. Macias, R. Pinto, Washington, DC. L. Espinosa-Perez, and others. 2009. “Critically Ill Patients Carrasco, L. R., V. J. Lee, M. I. Chen, D. B. Matchar, with 2009 Influenza A(H1N1) in Mexico.” Journal of the J. P. Thompson, and others. 2011. “Strategies for Antiviral American Medical Association 302 (17): 1880–87. Stockpiling for Future Influenza Pandemics: A Global Drake, T. L., Z. Chalabi, and R. Coker. 2012. “Cost-Effectiveness Epidemic-Economic Perspective.” Journal of the Royal Analysis of Pandemic Influenza Preparedness: What’s Society Interface 8 (62): 1307–13. Missing?” Bulletin of the World Health Organization 90 Chan, E. H., T. F. Brewer, L. C. Madoff, M. P. Pollack, (12): 940–41. A. L. Sonricker, and others. 2010. “Global Capacity for ———. 2015. “Buy Now, Saved Later? The Critical Impact Emerging Infectious Disease Detection.” Proceedings of the of Time-to-Pandemic Uncertainty on Pandemic Cost- National Academy of Sciences of the United States of America Effectiveness Analyses.” Health Policy and Planning 30 (1): 107 (50): 21701–6. 100–10. Chandy, L., H. Kato, and H. Kharas, eds. 2015. The Last Mile Elbe, S. 2002. “HIV/AIDS and the Changing Landscape of War in Ending Extreme Poverty. Washington, DC: Brookings in Africa.” International Security 27 (2): 159–77. Institution Press. Epstein, H. 2014. “Ebola in Liberia: An Epidemic of Rumors.” Charu, V., G. Chowell, L. S. Palacio Mejia, S. Echevarría-Zuno, New York Review of Books 61 (20): 91–95. V. H. Borja-Aburto, and others. 2011. “Mortality Burden of Evans, D. K., M. Goldstein, and A. Popova. 2015. “Health-Care the A/H1N1 Pandemic in Mexico: A Comparison of Deaths Worker Mortality and the Legacy of the Ebola Epidemic.” and Years of Life Lost to Seasonal Influenza.” Clinical The Lancet Global Health 3 (8): e439–e440. Infectious Diseases 53 (10): 985–93. Falcone, R. E., and A. Detty. 2015. “The Next Pandemic: Hospital Chisholm, H. 1911. “Cholera.” Encyclopedia Britannica 11 (6): Response.” Emergency Medical Reports 36 (26): 1–16. 265–66. Fan, V. Y., D. T. Jamison, and L. S. Summers. 2016. “The Cohen, N. J., C. M. Brown, F. Alvarado-Ramy, H. Bair-Drake, Inclusive Cost of Pandemic Influenza Risk.” NBER Working G. A. Benenson, and others. 2016. “Travel and Border Paper 22137, National Bureau of Economic Research, Health Measures to Prevent the International Spread of Cambridge, MA. Ebola.” Morbidity and Mortality Weekly Report (MMWR) Farmer, P. 1996. “Social Inequalities and Emerging Infectious Supplements 65 (3): 57–67. Diseases.” Emerging Infectious Diseases 2 (4): 259–69. Cohn, S. K. 2007. “The Black Death and the Burning of Jews.” Fearon, J. D., and D. D. Laitin. 2003. “Ethnicity, Insurgency, and Past and Present 196 (1): 3–36. Civil War.” American Political Science Review 97 (1): 75–90. 340 Disease Control Priorities: Improving Health and Reducing Poverty Ferguson, N. M., D. A. T. Cummings, S. Cauchemez, C. Fraser, Jain, S., L. Kamimoto, A. M. Bramley, A. M. Schmitz, S. Riley, and others. 2005. “Strategies for Containing an S. R. Benoit, and others. 2009. “Hospitalized Patients with Emerging Influenza Pandemic in Southeast Asia.” Nature 2009 H1N1 Influenza in the United States, April–June 437 (7056): 209–14. 2009.” New England Journal of Medicine 361 (20): 1935–44. Fischer, J. E., and R. Katz. 2013. “Moving Forward to 2014: Jefferson, T., M. Jones, P. Doshi, E. A. Spencer, I. Onakpoya, Global IHR (2005) Implementation.” Biosecurity and and others. 2014. “Oseltamivir for Influenza in Adults and Bioterrorism: Biodefense Strategy, Practice, and Science 11 Children: Systematic Review of Clinical Study Reports (2): 153–56. and Summary of Regulatory Comments.” British Medical Flahault, A., and A. J. Valleron. 1990. “HIV and Travel, Journal 348 (April): g2545. No Rationale for Restrictions.” The Lancet 336 (8724): Jennings, L. C., A. A. Monto, P. K. Chan, T. D. Szucs, and K. 1197–98. G. Nicholson. 2008. “Stockpiling Prepandemic Influenza Fraser, C., S. Riley, R. M. Anderson, and N. M. Ferguson. Vaccines: A New Cornerstone of Pandemic Preparedness 2004. “Factors That Make an Infectious Disease Outbreak Plans.” The Lancet Infectious Diseases 8 (10): 650–58. Controllable.” Proceedings of the National Academy of Johnson, N. P. A. S., and J. Mueller. 2002. “Updating the Sciences of the United States of America 101 (16): 6146–51. Accounts: Global Mortality of the 1918–1920 ‘Spanish’ Frieden, N. M. 1977. “The Russian Cholera Epidemic, 1892–93, Influenza Pandemic.” Bulletin of the History of Medicine 76 and Medical Professionalization.” Journal of Social History (1): 105–15. 10 (4): 538. Jonas, O. B. 2013. “Pandemic Risk.” Background paper for World Fullam, J. D., and N. Madhav. 2015. “Quantifying Pandemic Development Report 2014: Risk and Opportunity; Managing Risk.” The Actuary 12 (1): 29–34. Risk for Development, World Bank, Washington, DC. Gilbert, M., N. Golding, H. Zhou, G. R. W. Wint, T. P. Robinson, Jones, D. S. 2006. “The Persistence of American Indian Health and others. 2014. “Predicting the Risk of Avian Influenza A Disparities.” American Journal of Public Health 96 (12): H7N9 Infection in Live-Poultry Markets across Asia.” 2122–34. Nature Communications 5 (May): 1–7. Jones, K. E., N. G. Patel, M. A. Levy, A. Storeygard, D. Balk, Greenhill, K., and B. Oppenheim. 2017. “Rumor Has It: The and others. 2008. “Global Trends in Emerging Infectious Adoption of Unverified Information in Conflict Zones.” Diseases.” Nature 451 (7181): 990–93. International Studies Quarterly 61 (3). Jun, K. 2015. “MERS Outbreak Prompts South Korean Stimulus Greer, A. L., and D. Schanzer. 2013. “Using a Dynamic Model Package.” Wall Street Journal, June 25. to Consider Optimal Antiviral Stockpile Size in the Face of Katz, M. A., B. D. Schoub, J. M. Heraud, R. F. Breiman, Pandemic Influenza Uncertainty.” PLoS One 8 (6): e67253. M. K. Njenga, and others. 2012. “Influenza in Africa: Halder, N., J. K. Kelso, and G. J. Milne. 2014. “A Model- Uncovering the Epidemiology of a Long-Overlooked Based Economic Analysis of Pre-Pandemic Influenza Disease.” Journal of Infectious Diseases 26 (Suppl 1): S1–S4. Vaccination Cost-Effectiveness.” BMC Infectious Diseases Katz, R. 2009. “Use of Revised International Health Regulations 14 (1): 266–85. during Influenza A (H1N1) Epidemic, 2009.” Emerging Henao-Restrepo, A. M., A. Camacho, I. M. Longini, C. H. Watson, Infectious Diseases 15 (8): 1165–70. W. J. Edmunds, and others. 2016. “Efficacy and Effectiveness Katz, R., and R. Seifman. 2016. “Opportunities to Finance of an rVSV-Vectored Vaccine in Preventing Ebola Virus Pandemic Preparedness.” The Lancet Global Health 4 (11): Disease: Final Results from the Guinea Ring Vaccination, e782–e783. Open-Label, Cluster-Randomised Trial (Ebola Ça Suffit!).” Kavet, J. 1977. “A Perspective on the Significance of Pandemic The Lancet 389 (10068): 505–18. Influenza.” American Journal of Public Health 67 (11): Herstein, J. J., P. D. Biddinger, C. S. Kraft, L. Saiman, S. G. Gibbs, 1063–70. and others. 2016. “Initial Costs of Ebola Treatment Centers Kelly, H. A., P. C. Priest, G. N. Mercer, and G. K. Dowse. 2011. “We in the United States.” Emerging Infectious Diseases 22 (2): Should Not Be Complacent about Our Population-Based 350. Public Health Response to the First Influenza Pandemic of Hollingsworth, T. D., N. M. Ferguson, and R. M. Anderson. the 21st Century.” BMC Public Health 11 (1): 78. 2006. “Will Travel Restrictions Control the International Keogh-Brown, M. R., and R. D. Smith. 2008. “The Economic Spread of Pandemic Influenza?” Nature Medicine 12 (5): Impact of SARS: How Does the Reality Match the 497–99. Predictions?” Health Policy 88 (1): 110–20. Hooghe, L., and G. Marks. 2003. “Unraveling the Central State, KFF (Henry J. Kaiser Family Foundation). 2014. “The U.S. but How? Types of Multi-Level Governance.” American Government and Global Emerging Infectious Disease Political Science Review 97 (2): 233–43. Preparedness and Response.” Fact sheet, December 8, ICG (International Crisis Group). 2015. The Politics behind the http://kff.org/global-health-policy/fact-sheet/the-u Ebola Crisis. Crisis Group Africa Report 232, International -s- government-global-emerging-infectious-disease Crisis Group, Brussels, October 28. -preparedness-and-response/. IFRC (International Federation of Red Cross and Red Crescent Kim, Y.-W., S.-J. Yoon, and I.-H. Oh. 2013. “The Economic Societies). 2016. World Disasters Report, Resilience: Saving Burden of the 2009 Pandemic H1N1 Influenza in Korea.” Lives Today, Investing for Tomorrow. Geneva: IFRC. Scandinavian Journal of Infectious Diseases 45 (5): 390–96. Pandemics: Risks, Impacts, and Mitigation 341 Kozlowski, R. T., and S. B. Mathewson. 1997. “A Primer on Morens, D. M., J. K. Taubenberger, G. K. Folkers, and A. S. Fauci. Catastrophe Modeling.” Journal of Insurance Regulation 15 2010. “Pandemic Influenza’s 500th Anniversary.” Clinical (3): 322. Infectious Diseases 51 (12): 1442–44. Lee, B. Y., S. T. Brown, G. W. Korch, P. C. Cooley, R. K. Morse, S. S. 1995. “Factors in the Emergence of Infectious Zimmerman, and others. 2010. “A Computer Simulation Diseases.” Emerging Infectious Diseases 1 (1): 7–15. of Vaccine Prioritization, Allocation, and Rationing dur- Morse, S. S., J. A. K. Mazet, M. Woolhouse, C. R. Parrish, ing the 2009 H1N1 Influenza Pandemic.” Vaccine 28 (31): D. Carroll, and others. 2012. “Prediction and Prevention 4875–79. of the Next Pandemic Zoonosis.” The Lancet 380 (9857): Lugnér, A. K., and M. J. Postma. 2009. “Mitigation of Pandemic 1956–65. Influenza: Review of Cost-Effectiveness Studies.” Expert MSF (Médecins Sans Frontières). 2015. Pushed to the Limit and Review of Pharmacoeconomics and Outcomes Research 9 (6): Beyond: A Year into the Largest Ever Ebola Outbreak. Report, 547–58. MSF, London. Mackowiak, P. A., and P. S. Sehdev. 2002. “The Origin of Mullard, A. 2016. “New Vaccine Coalition Targets Epidemics.” Quarantine.” Clinical Infectious Diseases 35 (9): 1071–72. Nature Reviews Drug Discovery 15 (10): 669. Mancini, S., M. E. Coldiron, A. Ronsse, B. K. Ilunga, K. Porten, Murphy, F. A. 1998. “Emerging Zoonoses.” Emerging Infectious and others. 2014. “Description of a Large Measles Epidemic Diseases 4 (3): 429–35. in Democratic Republic of Congo, 2010–2013.” Conflict and Murray, C. J., A. D. Lopez, B. Chin, D. Feehan, and K. H. Hill. Health 8 (1): 9. 2006. “Estimation of Potential Global Pandemic Influenza Mangham, L. J., and K. Hanson. 2010. “Scaling Up in Mortality on the Basis of Vital Registry Data from the International Health: What Are the Key Issues?” Health 1918–20 Pandemic: A Quantitative Analysis.” The Lancet Policy and Planning 25 (2): 85–96. 368 (9554): 2211–18. Mathews, J. D., J. M. Chesson, J. M. McCaw, and J. McVernon. Nannei, C., S. Goldin, G. Torelli, H. Fatima, K. Kumar, and 2009. “Understanding Influenza Transmission, Immunity, others. 2016. “Stakeholders’ Perceptions of 10 Years of the and Pandemic Threats.” Influenza and Other Respiratory Global Action Plan for Influenza Vaccines (GAP)—Results Viruses 3 (4): 143–49. from a Survey.” Vaccine 34 (45): 5393–99. McCoy, T. 2014. “Why the Brutal Murder of Several Ebola Neiderud, C.-J. 2015. “How Urbanization Affects the Workers May Hint at More Violence to Come.” Washington Epidemiology of Emerging Infectious Diseases.” Infection Post, September 19. Ecology and Epidemiology 5: 27060. doi:/10.3402/iee McKibbin, W. J., and A. A. Sidorenko. 2006. “Global .v5.27060. Macroeconomic Consequences of Pandemic Influenza.” Oppenheim, B., M. D. Gallivan, N. Madhav, N. Brown, Analysis, Lowy Institute for International Policy, Sydney, V. Serhiyenko, and others. 2017. “Global Preparedness for the Australia. Next Pandemic.” Unpublished manuscript. Mihigo, R., C. V. Torrealba, K. Coninx, D. Nshimirimana, Oshitani, H., T. Kamigaki, and A. Suzuki. 2008. “Major Issues M. P. Kieny, and others. 2012. “2009 Pandemic Influenza A and Challenges of Influenza Pandemic Preparedness in Virus Subtype H1N1 Vaccination in Africa—Successes and Developing Countries.” Emerging Infectious Diseases 14 (6): Challenges.” Journal of Infectious Diseases 206 (Suppl 1): 875–80. S22–S28. Osterholm, M. T. 2005. “Preparing for the Next Pandemic.” Milián, E., and A. A. Kamen. 2015. “Current and Emerging New England Journal of Medicine 352 (18): 1839–42. Cell Culture Manufacturing Technologies for Influenza Paez-Espino, D., E. A. Eloe-Fadrosh, G. A. Pavlopoulos, Vaccines.” BioMed Research International 2015 (1): 1–11. A. D. Thomas, M. Huntemann, and others. 2016. “Uncovering Miller, M. A., C. Viboud, D. R. Olson, R. F. Grais, M. A. Rabaa, Earth’s Virome.” Nature 536 (7617): 425–30. and others. 2008. “Prioritization of Influenza Pandemic Park, J., and J. Kim. 2015. “Hong Kong Sets ‘Serious’ Response Vaccination to Minimize Years of Life Lost.” Journal of to South Korea’s MERS Outbreak.” Reuters, June 8. Infectious Diseases 198 (3): 305–11. Parpia, A. S., M. L. Ndeffo-Mbah, N. S. Wenzel, and A. P. Milne, G. J., N. Halder, and J. K. Kelso. 2013. “The Cost-Effectiveness Galvani. 2016. “Effects of Response to 2014–2015 Ebola of Pandemic Influenza Interventions: A Pandemic Severity Outbreak on Deaths from Malaria, HIV/AIDS, and Based Analysis.” PLoS One 8 (4): e61504. Tuberculosis, West Africa.” Emerging Infectious Diseases 22 Moon, S., D. Sridhar, M. A. Pate, J. K. Jha, C. Clinton, (3): 433–41. and others. 2015. “Will Ebola Change the Game? Ten Parveen, S., M. S. Islam, M. Begum, M. Alam, H. M. S. Sazzad, Essential Reforms before the Next Pandemic. The Report and others. 2016. “It’s Not Only What You Say, It’s Also of the Harvard–LSHTM Independent Panel on the Global How You Say It: Communicating Nipah Virus Prevention Response to Ebola.” The Lancet 386 (10009): 2204–21. Messages during an Outbreak in Bangladesh.” BMC Public Morens, D. M., J. K. Taubenberger, and A. S. Fauci. 2008. Health 16 (1): 726–37. “Predominant Role of Bacterial Pneumonia as a Cause of Pasquini-Descomps, H., N. Brender, and D. Maradan. 2016. Death in Pandemic Influenza: Implications for Pandemic “Value for Money in H1N1 Influenza: A Systematic Review Influenza Preparedness.” Journal of Infectious Diseases 198 of the Cost-Effectiveness of Pandemic Interventions.” Value (7): 962–70. in Health 20 (6): 819–27. 342 Disease Control Priorities: Improving Health and Reducing Poverty Pathmanathan, I., K. A. O’Connor, M. L. Adams, C. Y. Rao, Sands, P., C. Mundaca-Shah, and V. J. Dzau. 2016. “The P. H. Kilmarx, and others. 2014. “Rapid Assessment of Ebola Neglected Dimension of Global Security—A Framework Infection Prevention and Control Needs—Six Districts, for Countering Infectious-Disease Crises.” New England Sierra Leone, October 2014.” Morbidity and Mortality Journal of Medicine 374 (March): 1281–87. Weekly Report (MMWR) 63 (49): 1172–74. Schäferhoff, M., S. Fewer, J. Kraus, E. Richter, L. H. Summers, Pérez Velasco, R., N. Praditsitthikorn, K. Wichmann, and others. 2015. “How Much Donor Financing for Health A. Mohara, S. Kotirum, and others. 2012. “Systematic Is Channelled to Global versus Country-Specific Aid Review of Economic Evaluations of Preparedness Strategies Functions?” The Lancet 386 (10011): 2436–41. and Interventions against Influenza Pandemics.” PloS One Seto, W. H., D. Tsang, R. W. H. Hung, T. Y. Ching, T. K. Ng, and 7 (2): e30333. others. 2003. “Effectiveness of Precautions against Droplets Person, B., F. Sy, K. Holton, B. Govert, A. Liang, and others. and Contact in Prevention of Nosocomial Transmission of 2004. “Fear and Stigma: The Epidemic within the SARS Severe Acute Respiratory Syndrome (SARS).” The Lancet Outbreak.” Emerging Infectious Diseases 10 (2): 358–63. 361 (9368): 1519–20. Pike, B. L., K. E. Saylors, J. N. Fair, M. Lebreton, U. Tamoufe, and Shuaib, F., R. Gunnala, E. O. Musa, F. J. Mahoney, O. Oguntimehin, others. 2010. “The Origin and Prevention of Pandemics.” and others. 2014. “Ebola Virus Disease Outbreak—Nigeria, Clinical Infectious Diseases 50 (12): 1636–40. July–September 2014.” Morbidity and Mortality Weekly Platt, C. 2014. King Death: The Black Death and Its Aftermath in Report (MMWR) 63 (39): 867–72. Late-Medieval England. Oxon, U.K.: Routledge. Simonsen, L., P. Spreeuwenberg, R. Lustig, R. J. Taylor, D. M. Porta, M., ed. 2014. A Dictionary of Epidemiology. 6th ed. Fleming, and others. 2013. “Global Mortality Estimates for Oxford: Oxford University Press. the 2009 Influenza Pandemic from the GLaMOR Project: Price-Smith, A. T. 2001. The Health of Nations: Infectious A Modeling Study.” PLoS Medicine 10 (11): e1001558. Disease, Environmental Change, and Their Effects on Siu, J. Y. M. 2015. “Influence of Social Experiences in Shaping National Security and Development. Cambridge, MA: MIT Perceptions of the Ebola Virus among African Residents Press. of Hong Kong during the 2014 Outbreak: A Qualitative ———. 2009. Contagion and Chaos: Disease, Ecology, and Study.” International Journal for Equity in Health 14 (1): 88. National Security in the Era of Globalization. Cambridge, Smith, R. D., M. R. Keogh-Brown, T. Barnett, and J. Tait. 2009. MA: MIT Press. “The Economy-Wide Impact of Pandemic Influenza on Pritchett, L., M. Woolcock, and M. Andrews. 2013. “Looking the UK: A Computable General Equilibrium Modelling Like a State: Techniques of Persistent Failure in State Experiment.” British Medical Journal 339 (November): Capability for Implementation.” Journal of Development b4571. Studies 49 (1): 1–18. Smolinsky, M. S., M. A. Hamburg, and J. Lederberg, eds. 2003. Radonovich, L. J., P. D. Magalian, M. K. Hollingsworth, and Microbial Threats to Health: Emergence, Detection, and G. Baracco. 2009. “Stockpiling Supplies for the Next Response. Washington, DC: National Academies Press. Influenza Pandemic.” Emerging Infectious Diseases 15 (6): e1. Stoddard, A., A. Harmer, K. Haver, G. Taylor, and P. Harvey. Reperant, L. A., G. F. Rimmelzwaan, and A. D. M. E. Osterhaus. 2015. The State of the Humanitarian System: 2015 Edition. 2014. “Advances in Influenza Vaccination.” F1000 Prime London: Active Learning Network for Accountability Reports 6: 47. and Performance in Humanitarian Action/Overseas Richardson, J. S., J. D. Dekker, M. A. Croyle, and G. P. Kobinger. Development Institute. 2010. “Recent Advances in Ebolavirus Vaccine Development.” Thomas, M. R., G. Smith, F. H. G. Ferreira, D. Evans, M. Human Vaccines 6 (6): 439–49. Maliszewska, and others. 2015. “The Economic Impact of Roca, A., M. O. Afolabi, Y. Saidu, and B. Kampmann. 2015. Ebola on Sub-Saharan Africa: Updated Estimates for 2015.” “Ebola: A Holistic Approach Is Required to Achieve Working Paper 93721, World Bank, Washington, DC. Effective Management and Control.” Journal of Allergy and Tilman, D., and M. Clark. 2014. “Global Diets Link Clinical Immunology 135 (4): 856–67. Environmental Sustainability and Human Health.” Nature Rosello, A., M. Mossoko, S. Flasche, A. J. Van Hoek, P. Mbala, 515 (7528): 518–22. and others. 2015. “Ebola Virus Disease in the Democratic Tizzoni, M., P. Bajardi, C. Poletto, J. J. Ramasco, D. Balcan, and Republic of the Congo, 1976–2014.” eLife 2015 (4): e09015. others. 2012. “Real-Time Numerical Forecast of Global Rubinson, L., R. Mutter, C. Viboud, N. Hupert, T. Uyeki, and Epidemic Spreading: Case Study of 2009 A/H1N1pdm.” others. 2013. “Impact of the Fall 2009 Influenza A(H1N1) BMC Medicine 10 (1): 165. pdm09 Pandemic on US Hospitals.” Medical Care 51 (3): Toole, M. J., and R. J. Waldman. 1990. “Prevention of Excess 259–65. Mortality in Refugee and Displaced Populations in Samaan, G., M. Patel, B. Olowokure, M. C. Roces, H. Oshitani, Developing Countries.” Journal of the American Medical and others. 2005. “Rumor Surveillance and Avian Influenza Association 263 (24): 3296–302. H5N1.” Emerging Infectious Diseases 11 (3): 463–66. Twu, S-J., T-J. Chen, C-J. Chen, S. J. Olsen, L-T. Lee, and oth- Sands, P., A. El Turabi, P. A. Saynisch, and V. J. Dzau. 2016. ers. 2003. “Control Measures for Severe Acute Respiratory “Assessment of Economic Vulnerability to Infectious Syndrome (SARS) in Taiwan.” Emerging Infectious Diseases Disease Crises.” The Lancet 388 (10058): 2443–48. 9 (6): 718–20. Pandemics: Risks, Impacts, and Mitigation 343 Tyler, T. 2016. “IATA 2016 Annual Review.” Annual review pub- Wang, M-D., and A. M. Jolly. 2004. “Changing Virulence of the lication, International Air Transport Association (IATA), SARS Virus: The Epidemiological Evidence.” Bulletin of the Montreal. World Health Organization 82 (7): 547–48. UNDESA (United Nations Department of Economic and Watts, J. 2004. “China Culls Wild Animals to Prevent New Social Affairs). 2015. “World Population Prospects: The SARS Threat.” The Lancet 363 (9403): 134. 2015 Revision, Key Findings and Advance Tables.” Report WHO (World Health Organization). 1978. “Ebola ESA/P/WP.241, UNDESA, New York. Haemorrhagic Fever in Zaire, 1976. Report of an UNDP (United Nations Development Programme). 2014. International Commission.” Bulletin of the World Health “Assessing the Socio-Economic Impacts of Ebola Virus Organization 56 (2): 271–93. Disease in Guinea, Liberia, and Sierra Leone: The Road to ———. 2005. International Health Regulations. Geneva: WHO. Recovery.” Synthesis report, UNDP, New York. ———. 2010. “What Is a Pandemic?” WHO, February 24. ———. 2017. “A Socio-Economic Impact Assessment of the http://www.who.int/csr/disease/swineflu/frequently_asked Zika Virus in Latin America and the Caribbean: With _questions/pandemic/en/. a Focus on Brazil, Colombia, and Suriname.” Synthesis ———. 2013a. “Global Health Workforce Shortage to Reach report, UNDP, New York. 12.9 Million in Coming Decades.” News release, November USAID (U.S. Agency for International Development) and 11. CDC (Centers for Disease Control and Prevention). 2016. ———. 2013b. Global Survey on National Vaccine “West Africa—Ebola Outbreak, Fact Sheet #11, Fiscal Year Deployment and Vaccination Plans for Pandemic A(H1N1) (FY) 2016.” Fact sheet, USAID, CDC, June 24. 2009 Vaccine—2010. Report of survey findings, WHO, U.S. Department of Health and Human Services. 2005. “HHS Geneva. Pandemic Influenza Plan.” Strategic, guidance, and opera- ———. 2013c. “IHR Core Capacity Monitoring Framework: tional planning document, U.S. Department of Health and Checklist and Indicators for Monitoring Progress in the Human Services, Washington, DC. Development of IHR Core Capacities in States Parties.” U.S. Homeland Security Council. 2006. “Implementation International Health Regulations (2005) document, Plan for the National Strategy for Pandemic Influenza.” Reference WHO/HSE/GCR/2013.2, WHO, Geneva. Plan document, U.S. Homeland Security Council, ———. 2014. “Summary of States Parties 2013 Report on Executive Office of the President of the United States, IHR Core Capacity Implementation: Regional Profiles.” Washington, DC. International Health Regulations (2005) document, Van Boeckel, T. P., S. Gandra, A. Ashok, Q. Caudron, B. T. Reference WHO/HSE/GCR/2014.10, WHO, Geneva. Grenfell, and others. 2014. “Global Antibiotic Consumption ———. 2015a. “Successful Ebola Responses in Nigeria, Senegal, 2000 to 2010: An Analysis of National Pharmaceutical Sales and Mali.” In One Year into the Ebola Epidemic: A Deadly, Data.” The Lancet Infectious Diseases 14 (8): 742–50. Tenacious, and Unforgiving Year. Paper Series, January 2015, Van Boeckel, T. P., W. Thanapongtharm, T. Robinson, WHO, Geneva. C. M. Biradar, X. Xiao, and others. 2012. “Improving Risk ———. 2015b. “WHO Model List of Essential Medicines, 19th Models for Avian Influenza: The Role of Intensive Poultry ed.” WHO, Geneva. Farming and Flooded Land during the 2004 Thailand ———. 2016a. Ebola Situation Report. Weekly data report, Epidemic.” PLoS One 7 (11): e49528. April 15. Viboud, C., M. Miller, D. R. Olson, M. Osterholm, and ———. 2016b. “Influenza (Seasonal).” Fact sheet, November. L. Simonsen. 2010. “Preliminary Estimates of Mortality http://www.who.int/mediacentre/factsheets/fs211/en/. and Years of Life Lost Associated with the 2009 A/H1N1 ———. 2016c. “Pandemic Influenza Preparedness Framework Pandemic in the US and Comparison with Past Influenza Partnership Contribution: Annual Report 2015.” Document Seasons.” PLoS Currents 20 (March): RRN1153. WHO/OHE/PED/2016.01, Pandemic Influenza Preparedness Viboud, C., L. Simonsen, R. Fuentes, J. Flores, M. A. Miller, and (PIP) Secretariat, WHO, Geneva. others. 2016.“Global Mortality Impact of the 1957–1959 Influenza ———. 2017. Situation Report: Zika Virus, Microcephaly, Pandemic.” Journal of Infectious Diseases 212 (11): 738–45. Guillain-Barré Syndrome. Weekly data report, February 2. Waggoner, J. J., and B. A. Pinsky. 2016. “Zika Virus: Diagnostics Wolfe, N. D., P. Daszak, A. M. Kilpatrick, and D. S. Burke. for an Emerging Pandemic Threat.” Journal of Clinical 2005. “Bushmeat Hunting, Deforestation, and Prediction Microbiology 54 (4): 860–67. of Zoonotic Disease.” Emerging Infectious Diseases 11 (12): Walker, N. F., and C. J. Whitty. 2015. “Tackling Emerging 1822–27. Infections: Clinical and Public Health Lessons from the Wolfe, N. D., C. P. Dunavan, and J. Diamond. 2007. “Origins West African Ebola Virus Disease Outbreak.” Clinical of Major Human Infectious Diseases.” Nature 447 (7142): Medicine 15 (5): 457–60. 279–83. Wallinga, J., and P. Teunis. 2004. “Different Epidemic Curves for Wolicki, S. B., J. B. Nuzzo, D. L. Blazes, D. L. Pitts, J. K. Iskander, Severe Acute Respiratory Syndrome Reveal Similar Impacts and others. 2016. “Public Health Surveillance: At the Core of of Control Measures.” American Journal of Epidemiology the Global Health Security Agenda.” Health Security 14 (3): 160 (6): 509–16. 185–88. 344 Disease Control Priorities: Improving Health and Reducing Poverty Woolhouse, M. E. J., and S. Gowtage-Sequeria. 2005. “Host Yehualashet, Y. G., P. Mkanda, A. Gasasira, T. Erbeto, A. Onimisi, Range and Emerging and Reemerging Pathogens.” Emerging and others. 2016. “Strategic Engagement of Technical Surge Infectious Diseases 11 (12): 1842–47. Capacity for Intensified Polio Eradication Initiative in World Bank. 2012. People, Pathogens, and Our Planet, Volume Nigeria, 2012–2015.” Journal of Infectious Diseases 213 2: The Economics of One Health. Report 69145, World Bank, (Suppl 3): S116–S23. Washington, DC. Yu, H., J. T. Wu, B. J. Cowling, Q. Liao, V. J. Fang, and others. ———. 2014. “The Economic Impact of the 2014 Ebola 2014. “Effect of Closure of Live Poultry Markets on Poultry- Epidemic: Short and Medium Term Estimates for Guinea, to-Person Transmission of Avian Influenza A H7N9 Virus: Liberia, and Sierra Leone.” Working Paper 90748, World An Ecological Study.” The Lancet 383 (9916): 541–48. Bank, Washington, DC. Zell, R. 2004. “Global Climate Change and the Emergence ———. 2016. “World Bank Group Ebola Response Fact Sheet.” /Re-emergence of Infectious Diseases.” International Fact sheet, World Bank, April 6. http://www.worldbank.org Journal of Medical Microbiology Supplements 293 (Suppl /en/topic/health/brief/world-bank-group-ebola-fact-sheet. 37): 16–26. Yamey, G., M. Schäferhoff, O. K. Aars, B. Bloom, D. Carroll, and Zinsstag, J., E. Schelling, K. Wyss, and M. B. Mahamat. 2005. others. 2017. “Financing of International Collective Action “Potential of Cooperation between Human and Animal for Epidemic and Pandemic Preparedness.” The Lancet Health to Strengthen Health Systems.” The Lancet 366 Global Health 5 (8): e742–e744. (9503): 2142–45. Pandemics: Risks, Impacts, and Mitigation 345 Chapter 18 The Loss from Pandemic Influenza Risk Victoria Y. Fan, Dean T. Jamison, and Lawrence H. Summers INTRODUCTION the consequences of a range of pandemic severities (mild, moderate, severe, and ultra) and estimated income The 2014–16 Ebola virus outbreak in West Africa losses exceeding 12 percent of gross national income reminded the world that enormous economic and (GNI) worldwide and exceeding 50 percent in some low- human losses result from the uncontrolled spread of a and middle-income countries (LMICs). deadly infection. Less noticed was the likelihood that a pandemic with characteristics similar to the 1918 influ- enza pandemic would have killed about 10 times as Value of Lives Lost and Illness Suffered many people in Liberia, Guinea, and Sierra Leone as did The second major dimension of loss from a pandemic Ebola. The global death total from such a pandemic lies in the intrinsic value of lives prematurely lost and of could be 2,500 times higher than the World Health illness suffered. Efforts to measure the dollar value of Organization’s (WHO) estimate of 11,300 deaths from losses associated with premature mortality and illness Ebola through March 16, 2016 (WHO 2016a). remain imperfect. Nevertheless, extensive empirical findings appear in the economics literature, particularly for losses from premature mortality (Hammitt and Economic Loss Robinson 2011; Lindhjem and others 2011; Viscusi In addition to the enormous loss in terms of human 2014). Although the valuation of a change in mortality suffering, an important dimension of loss lies in a pan- appears most frequently in the environmental demic’s effect on income. Premature deaths reduce the economics literature, the report of the Lancet size of the labor force, illness leads to absenteeism and Commission on Investing in Health—“Global Health reduced productivity, resources flow to treatment and 2035: A World Converging within a Generation,” or control measures, and individual and societal measures Global Health 2035—systematically applied these to reduce disease spread can seriously disrupt economic methods to global health (Jamison and others 2013; activity. The World Bank has generated estimates of OECD 2014). This chapter estimates the magnitude of these losses (Burns, Mensbrugghe, and Timmer 2008; this dimension of loss from pandemic influenza using Jonas 2013) and found that a pandemic of the severity of standard methods. that in 1918 could reduce global gross domestic product This chapter assesses the expected annual loss from a (GDP) by about 5 percent and that the disruptive effects pandemic with risk r, expressed as a percentage of the of avoiding infection would account for about 60 percent annual probability of a pandemic, and severity s, of that total. McKibbin and Sidorenko (2006) examined expressed as the fraction of the world population that Corresponding author: Victoria Y. Fan. Myron B. Thompson School of Social Work, University of Hawai’i at Mãnoa, Honolulu, Hawai’i, U.S.A.; vfan@hawaii.edu. 347 dies from the pandemic. It uses the historical and mod- In the world’s 2015 population of 7.35 billion, 1 SMU eling literatures to generate expected values of r and s, corresponds to 735,000 deaths. Seasonal influenza causes and it uses those values to generate estimates of mortal- about 250,000 to 500,000 deaths per year (WHO 2016b). ity and its associated losses. The estimated loss is relative We define severe pandemics as having mortality rates of to the counterfactual of no risk (r = 0). Box 18.1 places 10 SMU or greater, and moderately severe pandemics as the results of our research into context. having a severity less than 10 SMU. The historical record suggests that the 1918 influenza pandemic was an outlier, with unusual circumstances, REVIEW OF HISTORICAL PANDEMIC RISK including the co-occurrence of World War I. No other influenza pandemic had such devastatingly high mortal- AND SEVERITY ity rates. The 1918 influenza pandemic had an estimated The literature defines pandemic “severity” in different 20 million to 50 million (or more) excess deaths from ways. We define it in terms of mortality only. Paules and 1918 to 1920, most of which were concentrated in 1918. Fauci (2017) point to long-term morbidity and disability In 1918, 20 million deaths would constitute 1.1 percent consequences of a range of potential pandemic patho- of the world’s population. In addition to the severe pan- gens. Global Health 2035 appendix 4 introduced the term demic of 1918, the sparse record suggests that 12 to 17 standardized mortality unit (SMU) in which 1 SMU is other pandemics have occurred since 1700. Of these, we 10−4. For example, the pandemic of 1957–58 had a global identify six as having substantial excess mortality, with death rate of 3 SMU (or 0.03 percent of the population). mortality rates in the range of 3–8 SMU (table 18.1). Box 18.1 Research in Context Evidence before This Study increased mortality. It uses an expected value frame- We searched PubMed and Google Scholar for all work to estimate losses from an uncertain and studies on influenza epidemics and pandemics. We rare event over time. Past work found that income also searched libraries at Harvard University and losses (US$80 billion per year) are much lower the University of Hawai‘i for historical documents than the losses from increased mortality (US$490 and life tables. Studies were restricted to those with billion per year). We further analyzed economic abstracts in English. losses of national income levels by world regions and conducted sensitivity analyses on the value of a Our review showed a wide range in the estimates of statistical life. deaths caused by the 1918 influenza pandemic. We found three studies that examined loss in national income from influenza pandemics of varying Implications of All of the Available Evidence severity. A substantial literature exists that estimates Estimates of intrinsic loss substantially exceed the monetary value of mortality risk—the value of a previous estimates of income loss. As significant as statistical life—but we found only one paper in that the direct effect of a pandemic on income appears literature that estimates the loss from elevated to be, we conclude that intrinsic losses far exceed mortality associated with pandemics. Integrative the income losses. This finding points to the need estimates of the magnitude of pandemic risk were for more attention to pandemic risk in public pol- found in only two sources, both partially icy and to the value of enhanced understanding proprietary. of both the magnitude and the consequences of pandemic risk. Low- and middle-income countries Added Value of This Study would suffer more than high-income countries in This study provides the first assessment of the mortality losses. Further studies to investigate the expected value of losses from pandemic influenza potential losses from pandemics from other causes and, specifically, the value of intrinsic losses from are ongoing. 348 Disease Control Priorities: Improving Health and Reducing Poverty Table 18.1 Worldwide Mortality from Selected Influenza Pandemics, 1700–2000a Estimated worldwide pandemic- Severity, s related deaths Estimated world population (fraction of world population Year (millions) (millions) killed, SMU)b 1729c 0.4 720 6 c 1781–82 0.7 920 8 1830–33c 0.8 1,150 7 1898–1900c 1.2 1,630 7 c,d 1918–20 20.0–50.0 1,830 110–270 1957–58c 1.0 2,860 3 c 1968–69 1.0–2.0 3,540 3–6 Note: SMU = standardized mortality unit. a. The table includes pandemics dating from 1700 to 2000 for which severity could be ascertained from the literature. Morens and Fauci (2004) and Morens and Taubenberger (2011) identify 12 to 17 pandemics in the period from 1700 to 2000, but many of those resulted in lower mortality than those in this table (or had mortality levels that could not be ascertained). b. The SMU represents a 10−4 mortality risk and is used to represent small numbers as integers. For example, the 1729 pandemic led to an elevation in mortality of 0.06 percent of the world’s population, which is more conveniently expressed as 6 SMU. In the world’s 2015 population, 1 SMU corresponds to 735,000 deaths. c. See Potter (2001). d. See Beveridge (1991); Ghendon (1994); Johnson and Mueller (2002). Although the world may be expected to experience rather to select plausible values from that literature to moderately severe to severe pandemics several times define reference cases. With Taubenberger and others each century, there is consensus among influenza experts (2007), we emphasize the uncertainty inherent both in that an event on the very severe scale of the 1918 pan- the history and in projections drawn from it. In light of demic may be plausible but remains historically and this literature and its attendant uncertainty, we develop biologically unpredictable (Taubenberger, Morens, and and report results for two representative levels of sever- Fauci 2007). A modeling exercise conducted for the ity. Table 18.2 defines the severity levels we use and indi- insurance industry concluded that 100 to 200 years cates the levels of annual risk assigned to them. Box 18.2 would pass before a 1918-type pandemic returned, but provides the background to the calculation of expected the exercise acknowledged major uncertainty (Madhav severity that table 18.2 summarizes. 2013). Although a biological replica of the 1918 influ- enza pandemic would result in lower mortality rates METHODS than those that occurred in 1918 (Madhav 2013), other studies point to the possibility that exceptionally trans- The effort proceeds in two steps. First, information on missible and virulent viruses could lead to global death pandemic severity is used to generate increases in rates substantially higher than in 1918 (McKibbin and age-specific death rates for the world and for each of the Sidorenko 2006; Osterholm 2005). World Bank’s four income groups of countries. Second, In general, lower-income areas of the world suffered the literature on valuation of changes in mortality rates disproportionately in 1918; in particular, India suffered is used to generate estimates of the age-specific losses a major share of global pandemic mortality (Davis from mortality increase and, by extension, of total loss. 1968). Similarly, Madhav (2013) and Morens and Fauci We begin by estimating the change in a population’s (2007) argue that a modern epidemic would dispropor- age-specific mortality rate for the two severity refer- tionately affect poor countries. However, China’s mortal- ence cases. Estimates of the age-specific excess mortal- ity rate in 1918 was low, probably because of lower case ity rates of different populations from the 1918 fatality rates rather than lower incidence rates (Cheng pandemic are consistent in their form of a unique and Leung 2007). This finding points to the possibility of inverted U-shaped distribution, whereby adults ages heterogeneity between countries of comparable national 15 to 60 years experienced elevated rates compared to income levels in a modern pandemic. elderly persons (greater than age 60 years) (Luk, Gross, This chapter does not seek to provide a new review of and Thompson 2001; Murray and others 2006). We use the literature on mortality in previous pandemics but the specific U.S. data for age distribution of excess The Loss from Pandemic Influenza Risk 349 Table 18.2 Worldwide Pandemic Risk: Two Representative Scenarios, 2015 Moderately severe pandemic Severe pandemic Parameter (< 10 SMU)a (> 10 SMU)b Any pandemic a 1. Annual probability, r (%) 2.0 1.6 3.6 2. Return time, 1/r (years) 50 63 28 3. Average severity (SMU)c 2.5 58 27 d 4. Expected severity, s (SMU) 0.05 0.93 0.98 Note: SMU = standardized mortality unit. a. The SMU represents a 10−4 mortality risk and is used to represent small numbers as integers. For example, the 1729 pandemic led to an elevation in mortality of 0.06 percent of the world’s population, which is more conveniently expressed as 6 SMU. In the world’s 2015 population, 1 SMU corresponds to 735,000 deaths. b. These severity states are mutually exclusive. Hence, the annual probability of any pandemic is [1 – (1 – 0.2) (1 – 0.016)] = 3.6% c. The average severity of a pandemic in a given severity range is the expected value of severity given that a pandemic did in fact occur in that range. For example, 2.5 SMU is the expected severity given that a pandemic of severity s < 10 SMU has occurred. d. “Expected severity” is average severity multiplied by the probability of occurrence [s = row (3) × row (1)]. Box 18.2 Estimating Pandemic Severity and Risk Following its usage in the insurance industry, we of moderately severe pandemics. In particular, we define risk, r(s), in terms of “exceedance probability,” assume two such pandemics per century in this the annual probability of a pandemic having a sever- severity range with the average severity of 2.5 SMU ity exceeding s. Again following insurance industry globally. The expected annual severity of moderately usage, the “return time” for s is the expected number severe pandemics is 0.02 × 2.5 = 0.05 SMU, corre- of years before a pandemic of at least severity s will sponding to just over 35,000 expected annual deaths occur. If t(s) is the return time, then t(s) = r(s)−1. For worldwide. example, if the annual probability of a pandemic of We turn next to equation B18.2.2 to estimate the con- severity at least s is 1%, then its return time will be tributions to expected severity from pandemic sever- 100 years. ity greater than 10 SMU worldwide (or 4 SMU in the If we had access to a function r(s) showing exceed- United States). Let s*(x) be the contribution of ance probability as a function of severity, our analy- pandemic severity greater than x to expected pan- sis could proceed using the expected value of severity demic severity. Information available from AIR and of all pandemics. Because r(s) is the complementary its Pandemic Flu Model (AIR Worldwide 2014) allows cumulative of the density for s, we would have calibration of r(s) for the United States with s > 4: ∞ s*(4) = ∫ ∞ 4 r(s)ds. (B18.2.2) Expected value of s = ∫ 0 r(s)ds. (B18.2.1) (Available data allow us to calibrate only an exceed- Modeled estimates of the function r(s) are not (pub- ance probability function, r(s), for the United States. licly) available, so we approximated in two steps. We Hence, we start with that and translate to world label pandemics with global s > 10 SMU as “severe.” values from severity ratios available in Madhav (As defined in the text, 1 SMU corresponds to a 10−4 [2013].) The calibration points to a very fat-tailed mortality risk.) We label pandemics with global distribution. The hyperbolic family of complemen- s < 10 as moderately severe. For the first step in our tary cumulative distributions provides natural can- assessment of expected severity, we use recent history didates for r(s), and we parameterize the hyperbolic as a straightforward guide to frequency and severity in terms of its expectation and the fatness of its tail box continues next page 350 Disease Control Priorities: Improving Health and Reducing Poverty Box 18.2 (continued) (see Jamison and Jamison 2011, table 2, in the for- Calibrating an exponential as we did for the mally identical context of discounting). Thus, hyperbolic—so that the contribution to expected severity of severity > 4 SMU is equal to 0.18—gives r(s) = [1 + m(1 – f )s] – [1 + 1/(1 – f)] (B18.2.3) r(s) = e0−0.57s, and a return time for a 1918-type pan- demic of 150 years, quite close to the 175 years of where 1/m is the expected value of s, and f indicates equation B18.2.3. However, for s = 4 in the United the fatness of the tail (smaller values imply a fatter States (over 7 million deaths worldwide), the expo- tail). Our calibration yields a value of m = 1.8 and nential gives an unrealistic return time of only 10 years f = –2. Hence, s*(0) = 1/1.8 = 0.56. s*(4) is given as whereas equation B18.2.3 gives 63 years. AIR (AIR Worldwide 2014) estimates that an extreme pandemic s*(4) = 0.56 – ∫ 0 4 (1 + 3ms )−1.33 ds , (B18.2.4) with s = 30 in the United States (and perhaps 100 mil- lion deaths worldwide) has a return time of 1,000 years, and the integral is approximately 0.38. (For small and equation 2.1.3 gives 875 years. The exponential values of s, equation B18.2.3 substantially overesti- would give 27 million years. mates r when the equation for r(s) has been cali- brated to fit larger values of s and thus the need for Clearly, uncertainty surrounds the numbers we use this two-step procedure.) Hence, s*(4) = 0.56 − 0.38 = to reflect the likelihood of pandemics of varying 0.18, which is the contribution to expected severity levels of severity. In particular, we point to recent in the United States of severity levels > 4 SMU. We estimates (Madhav and others 2018) of exceedance infer global severity from the severity in the United probability and pandemic risk that use methods States using the approach described in the main text. similar to those of AIR but come to a substantially smaller number of expected annual deaths. However, Madhav (2013), using the AIR model, estimates that a our numbers represent conservative choices that are 1918-type pandemic would kill 21 million to 33 million broadly consistent with historical experience and people in today’s world. She reports a mid-range modeling parameters. Substantially greater severities severity for the United States of such a pandemic of and likelihoods have been discussed by Madhav 8.8 SMU with a return time of 100 to 200 years. (2013) and colleagues elsewhere in the literature Equation B18.2.3 predicts that the return time for a (Bruine De Bruin and others 2006; McKibbin and pandemic of at least that severity is about 175 years. Sidorenko 2006; Osterholm 2005). As Morens and Our calibrated value of −2 for f, the tail fatness Taubenberger (1977, 277) stated, “With human parameter in equation 2.1.3, suggests that the distri- influenza the only certain thing seems to be uncer- bution of exceedance probabilities is very fat tailed tainty.” We would slightly modify that statement to indeed. An exponential distribution for r(s) could assert the virtual certainty that, “sooner or later, the be considered to be neither fat nor thin tailed. world will again suffer a severe pandemic.” deaths to generate age distributions for the world, have age distributions of death like those of the 1918 adjusting for greater absolute increases elsewhere (Luk, pandemic. Gross, and Thompson 2001). The fatality rate among Using the age distributions of populations and the young adults, although high in the 1918 influenza pan- life tables from the World Population Prospects of the demic, was relatively low in the 1957 and 1968 epidem- United Nations Population Division (2015), we calculate ics (Simonsen and others 1998). We also use an excess deaths and the estimated reduction in life expec- alternative and more typical distribution of excess tancy based on these age-specific mortality rates mortality, where young children and elderly persons (Preston, Heuveline, and Guillot 2000). Table 18.3 shows are disproportionally affected, as well as a combination the results for our severity categories. Our expected of the two, assuming the same proportional increase in annual pandemic death total across both severities is mortality for all age groups. Our final calculations are 720,000 (or 1.2 percent of the number of deaths in based on the assumption that moderately severe pan- 2015), resulting in a decrease in life expectancy at birth demics will have age distributions like those of the 1957 by 0.3–0.4 years in low-income countries (LICs) and and 1968 pandemics, whereas severe pandemics will LMICs. The Loss from Pandemic Influenza Risk 351 Table 18.3 Expected Annual Influenza Pandemic Deaths, by Country Income Group, 2015a Income levelb Parameter Low Lower-middle Upper-middle High World 1. Population (millions) 640 2,900 2,400 1,400 7,350 2. Moderately severe pandemics 2.1. Relative pandemic severityc 4 3 2 1 n.a. 2.2. Expected annual pandemic-related mortality rate 0.08 0.06 0.04 0.02 0.05 (SMU) 2.3. Expected excess deaths per year [x = (1) × (2.2)] 5,100 18,000 9,600 2,800 37,000 3. Severe pandemics (all severities combined) 3.1. Relative pandemic severityc 10 7 4 1 n.a. 3.2. Expected annual pandemic-related mortality rate 1.8 1.26 0.72 0.18 0.93 (SMU) 3.3. Expected excess deaths per year [x = (1) × (3.2)] 120,000 370,000 170,000 25,000 680,000 4. Expected totals 4.1. Expected mortality rate (SMU) 1.9 1.3 0.76 0.2 0.98 4.2. Expected excess deaths per year [x = (2.3) + (3.3)] 125,000 390,000 180,000 28,000 720,000 (430,000–1,000,000)a Note: n.a. = not applicable. In the “World” column, the rows on pandemic severity are not applicable because this column incorporates all pandemic severities. a. Very substantial uncertainty adheres to all data in panels 2–4 of this table. We judge that ± 40 percent reasonably reflects this uncertainty. AIR’s (AIR Worldwide 2014) mortality estimates for a 1918-type pandemic occurring today are given ± 22 percent, and we have amplified that somewhat (Madhav 2013). Rather than report a ± 40 percent range, this table reports only our point estimates except for our estimate of total annual expected deaths where we state the range. b. We use the World Bank’s income level classification of countries (World Bank 2015). c. Relative severity indicates severity in each income group relative to the high-income group. This ratio is assumed to be different for each level of severity. Our estimates for severe pandemics come from AIR (Madhav 2013). AIR estimates a narrow range of mortality rates across high-income countries (6–11 SMU) for its model of a 1918-type pandemic, and the relative severities we indicate are consistent with the HIC rates and AIR’s estimate of 21 million to 33 million deaths globally in such a pandemic. Evidence for the moderate pandemics of 1957–58 and 1968–69 suggest a more compressed range for these less severe pandemics, and our relative severity numbers in row 2.1 reflect this. Alternative estimates of relative severity in lower-income countries are lower than those for AIR that we use, resulting in a lower estimated global death total. Next, we place dollar values on the changes in mortal- RESULTS ity rates. Our specific calculations followed the methods used in Global Health 2035 (Jamison and others 2013). Table 18.4 shows the results of our calculation of the We defined levels of our valuation metric v of 0.7, 1.0, value of intrinsic loss from pandemic risk, using values 1.3, and 1.6 percent of income per capita per SMU of of v of 0.7–1.6 percent of GNI per SMU, depending on mortality increase, that is, per 1/10,000 increase in mor- income category. We stress that these are expected tality risk for one year for countries in each of the World annual values of loss associated with the indicated risks Bank’s four income groups of countries: 0.7 percent was of pandemics in the severity ranges we have chosen. used for LICs; 1.0 percent for lower-middle-income Expected losses from an actual severe pandemic would countries; 1.3 percent for upper-middle-income countries be about 60 times as large. The World Bank expresses (UMICs); and 1.6 percent for HICs. In calculating the income loss figures as expected annual values but uses value of change in mortality at age, we used as a reference different values for annual pandemic risk. the literature’s value as a fraction of GNI per capita for Table 18.4 shows our estimate of the expected annual age 35 years. This amount was adjusted up or down for loss for the world as a whole from the intrinsic loss from ages other than 35 years in proportion to the ratio of life pandemic risk to be -0.6 percent of global income or expectancies at those ages to life expectancy at age 35 about US$490 billion per year. Loss varies by income years. Hence, for a given level of overall mortality, the group, from a little over 0.3 percent in HICs to 1.6 percent value of mortality loss will depend on which of the age in lower-middle-income countries. distributions of excess pandemic mortality described is Although the direct effect of a pandemic on income assumed. appears to be significant, we conclude that intrinsic losses 352 Disease Control Priorities: Improving Health and Reducing Poverty Table 18.4 Value of Mortality Losses from Pandemic Risk, by Country Income Group, 2015 (age-dependent VSMU) Income levela Parameter Low Lower-middle Upper-middle High World 1. Economic parameters 1.1 Income, Y (trillions of 2013 US$) 0.7 6.0 20.0 54.0 80.0 1.2 Per person income, y (2013 US$) 780 2,300 8,200 41,000 11,000 1.3 v (%)b 0.7 1.0 1.3 1.6 n.a. c 2. Losses from pandemic 2.1 Expected annual value of mortality loss, C (billions of −7 −100 −200 −180 −490 2013 US$)d (−290 to −690) 2.2 Annual mortality loss, c [as % of income = (2.1) ÷ (1.1)] −1.1 −1.6 −1.0 −0.34 −0.62 (−0.37 to −0.87) Note: n.a. = not applicable; VSMU = value of a standardized mortality unit. In the “World” column, row 1.3 on pandemic severity is not applicable because this column incorporates all pandemic severities. a. We use the World Bank’s income data and income level classification of countries (World Bank 2015). b. We use “v” to denote the value of a 1-in-10,000 risk of death, expressed as a percentage of per capita GNI. The dominant position in the literature is that lower-income countries should have lower values for v (Hammitt and Robinson 2011). The literature provides weak quantitative guidance on how v should vary with y, if at all, and the numbers we have chosen should be viewed as reasonable assumptions within the spirit of the literature. c. Very substantial uncertainty adheres to these cost estimates (see note a, table 18.3). We judge that ± 40 percent reasonably reflects this uncertainty but report that range for our estimates of worldwide costs only. d. For any given value of s, our calculation of the value of intrinsic loss from a pandemic depends on the age distribution of deaths from the pandemic, and the calculations reported here use different age distributions for pandemics of different severities. In particular, for moderately severe pandemics, we assume an older age distribution of deaths, typical of such pandemics. For severe pandemics, we assume the younger age distribution of deaths that characterized the 1918 pandemic. far exceed the loss from lost income. We referred to esti- (Moore and Diaz 2015). In contrast to the modest num- mates in the literature of the income loss from pandem- ber of studies on potential pandemic loss, there are hun- ics of differing levels of severity (Burns, Mensbrugghe, dreds of studies on the cost of climate change and the and Timmer 2008; Jonas 2013; McDonald and others social cost of carbon (Pizer and others 2014; Tol 2013). 2008; McKibbin and Sidorenko 2006). Though our sever- Global carbon dioxide emissions were on the order of ity categories differ from theirs, the values of 1 percent of 36,000 million tons in 2013, containing 6,200 million global income from a moderately severe pandemic and tons of carbon (Global Carbon Project 2015). Estimates 4 percent from a severe pandemic are consistent with esti- of the social cost of carbon vary widely, but if it were mates in the literature. Using our estimates of the annual around US$120 per ton, then the cost of carbon dioxide probabilities of such pandemics (table 18.2), we find emissions in 2013 would be about 1 percent of world expected annual income losses globally of US$16 billion income; US$120 per ton is within the range of available for moderately severe pandemics and US$64 billion for estimates (Nordhaus 2010; Tol 2013). One must add the severe pandemics, for a cost of approximately US$80 billion losses from carbon in carbon dioxide to the losses from per year. Table 18.4 shows an expected annual value of methane, which are likely to be substantial (Smith and mortality loss from pandemics of US$490 billion, of others 2013). The synthesis of the 2014 report of the which 95 percent is from severe pandemics. (See annex Intergovernmental Panel on Climate Change (IPCC) 18A for further details on research methods.) assessed the literature and estimated that global economic losses for warming of 2.5°C higher than pre-industrial levels range from 0.2 to 2.0 percent of income (Pachauri DISCUSSION and others 2014). In comparison, our expected annual Expected annual pandemic losses appear substantial. intrinsic loss from pandemic risk (at 0.7 percent of global Comparing the loss from pandemic risk with losses income) lies 25 percent higher than the low end of the from climate change is instructive. As with pandemic range of the IPCC’s estimated range for global warming. risk, much uncertainty is attached both to the magni- Although most studies of the cost of climate change tude of future climate change and to the possible losses fail to include the intrinsic loss of increased mortality The Loss from Pandemic Influenza Risk 353 risk, the effect of doing so may be modest. The IPCC heart disease using methods closely related to ours. Far report anticipates increased risks, with very high confi- more studies assess the losses from specific environmen- dence, of ill health owing to heat waves and fires, tal risk factors (OECD 2014). undernutrition from diminished food production in poor regions, and increased foodborne and waterborne diseases and some vectorborne and infectious diseases SENSITIVITY TESTS AND LIMITATIONS (Pachauri and others 2014). Modest reductions in Sensitivity to Assumptions cold-related mortality and morbidity will be offset by the magnitude and severity of the increased risks. The methods used to value mortality risk have limitations. Although the IPCC presents scenarios of health risks, The valuation of health risks—including fatalities, illness, the aggregate effect of climate change on mortality was and injuries—is inherently difficult because money is not summarized. However, the gradual nature of often an ineffective substitute for dimensions of human warming allows time for costly adaptations that could well-being. In practice, however, these estimates are be expected to reduce the mortality consequences. A obtained from ex post observations of the labor market recent paper points to potentially important mortality and reflect the way people differentially value and trade off reductions in the United States resulting from efforts to very small fatality risks for income. Substantial variation keep U.S. emissions consistent with global warming of exists both in the estimated value of a small mortality risk 2°C (Shindell, Lee, and Faluvegi 2016). These benefits at a given age in the United States and in the way the val- appear to flow almost entirely from reduced pollution uation (v) should vary across ages and countries (Hammitt rather than slower atmospheric warming. Most health and Robinson 2011; Lindhjem and others 2011). Our losses from climate change are then likely to be calculations to test the sensitivity of our results to this included in the income losses from adaptation rather alternative assumption found a change of only about than included separately. 5 percent in our headline number of US$490 billion. Another useful comparator for pandemic risk lies in Hammitt and Robinson (2011) have assembled the deaths from selected alternative causes. The expected evidence that the value of mortality risk as a percentage annual number of pandemic influenza deaths for 2015 of income in LICs may be less than for HICs. Global in our reference cases is 720,000 (table 18.3). One might Health 2035 did not include this potential effect in its reasonably add 300,000 deaths per year from seasonal calculations (Jamison and others 2013). This chapter influenza to this number for a total of over 1 million does include an adjustment for this effect, which leaves deaths (WHO 2016b). In comparison, Mathers (2018) estimates of losses in HICs unchanged but reduces our reports new WHO estimates for the diseases of compa- estimated cost for the world as a whole. We assessed the rable magnitude for 2015 (table 18.5). sensitivity of our results to alternative assumptions on Earlier studies have estimated losses from disease that this point and others and concluded our main findings included valuation of mortality consequences. Ozawa to be robust to the specific assumptions made. and others (2011), for example, estimated the losses from vaccine-preventable diseases, and Watkins and Daskalakis (2015) estimated burdens from rheumatic Limitations A key limitation of this study is its use of historical mor- tality estimates and modeled estimates from various Table 18.5 Causes of Death with Magnitude sources to estimate pandemic risk. As we have noted Comparable to Expected Deaths from Pandemic Influenza, 2015 throughout, the estimates we use for pandemic risk, r, and severity, s, remain subject to substantial inherent uncer- Cause of death Magnitude of deaths tainty. Although the AIR modeling efforts (Madhav 2013) Tuberculosis 1.4 million on which we rely explicitly account for potentially increased risks associated with increased air travel and HIV/AIDS 1.1 million mobility of persons and goods, as well as increased urban- Maternal mortality 0.3 million ization, we lacked access to the full results of that study. Cancers 8.8 million Similarly, whereas AIR attempted to account for decreased Ischemic heart disease 7.9 million risks associated with increased incomes, schooling, and access to health care services—including vaccination, Stroke 6.2 million antiviral medications, improved infection control, Source: Mathers 2018. Note: HIV/AIDS = human immunodeficiency virus/acquired immune deficiency increased surveillance, and real-time communications— syndrome. we could use that information only indirectly. 354 Disease Control Priorities: Improving Health and Reducing Poverty A modeling effort separate from that by AIR uses pandemic deaths per year is 720,000. The expected annual similar methods but different assumptions, resulting in a intrinsic cost that results for the world is US$490 billion, smaller expected annual mortality, although in the same or 0.6 percent of global income. In comparison, the IPCC broad range (Madhav and others 2018). In contrast to estimates that the likely cost of global warming falls in the the robustness of our conclusions with respect to how range 0.2 to 2.0 percent of global income annually. to value mortality risk, our findings respond sensitively Posner (2004) has argued that economics and the to how we model r and s. Increased global temperature social sciences generally fail to pay adequate attention to may reduce the case fatality rates of influenza, but it may potentially catastrophic events, although literature is also increase the transmissibility of the virus. Population- emerging (Barro and Jin 2011; Pike and others 2014; level immunity against a particular influenza strain likely Pindyck and Wang 2013). Concluding that the academic varies by region and by age distribution, although the and policy attention provided to pandemic risk falls well extent of that variation is not known. In 1918, a few short of a sensibly estimated comparison of that risk with countries did not experience the typical inverted its consequences is reasonable. However, recent trends U-shaped distribution of excess age-specific mortality are encouraging. As he prepared to host the G-7 (Group from influenza. In Mexico, elderly persons were not of Seven) in 2016, Japanese Prime Minister Shinzo Abe spared from excess mortality in contrast to those in the placed high priority on dealing with health crises (Abe United States, although its working-age population suf- 2015). German Prime Minister Angela Merkel, as host for fered as significantly as those in other regions. (Chowell the meeting of the G-20 (Group of Twenty) in Hamburg and others 2010). In China, mortality rates were low at in June 2017, maintained this high-level interest by all ages. The characteristics of new pandemic viral including specific attention to pandemic preparedness. strains depend on poorly understood patterns of immu- Hosted for the WHO and the World Bank by the U.S. nity and the complex and poorly understood process of National Academy of Medicine, a recent Commission on viral evolution and genetic re-assortment in dynamic a Global Health Risk Framework for the Future pointed ecosystems (Morens, Folkers, and Fauci 2004). to practical and significant financial and organizational An additional limitation of this study is its omission of steps to improve pandemic preparedness and response an estimated value of the intrinsic undesirability of non- (GHRF Commission 2016; Sands and others 2016). fatal illness or of pandemic fear—significant characteris- Despite these encouraging indicators, Moon and others tics of population response to SARS (severe acute (2017) have concluded that inadequate action followed respiratory syndrome) in Taiwan, China (Liu and others the warning from the Ebola virus in West Africa. 2005). The high media salience and associated fear may In chapter 17 of this volume, Madhav and others (2018) also lead populations to overreact to mild pandemics, assess the costs and probable effects of investments to increasing the effect beyond what might be considered reduce the likelihood or potential severity of a pandemic. optimal (Brahmbhatt and Dutta 2008). The economics These investments could range from research and develop- literature currently provides value estimates almost ment to a universal influenza vaccine to much-enhanced entirely for mortality risk. However, when appropriate surveillance to pre-investment in manufacturing capacity valuations of illness and fear become available, our results for drugs and vaccines (Varney and others 2017). Important may be shown to be underestimates for this reason. investments along these lines are indeed being made. A final limitation of this study is its estimation of Given this chapter’s estimate of the intrinsic expected loss losses from only pandemic influenza risk. Further work from pandemic risk, the economic benefits of further should extend this analysis to at least 11 additional patho- investments are likely to substantially exceed their cost. gens that the WHO regards as known potential causes of pandemics or epidemics (Brende and others 2017). Including most other known pathogens may increase the ANNEX risk to about 50 percent over that from influenza alone The annex in this chapter is as follows. It is available at (personal communication, J. Douglas Fullam). http://www.dcp-3.org/DCP. CONCLUSIONS • Annex 18.A. Materials and Methods World Bank studies estimate approximately 5 percent of global income as the probable income loss from a pan- ACKNOWLEDGMENTS demic as severe as that of 1918. This chapter estimates the value of intrinsic loss from the excess deaths from poten- The authors thank Branden Nakamura (University of tial pandemics. Our estimate of the expected number of Hawai‘i) and Jennifer Nguyen (Susan G. Komen Foundation) The Loss from Pandemic Influenza Risk 355 for valuable research assistance. Kristie Ebi (University of Journal of Infectious Diseases 11 (4): 360–64. doi:10.1016/j Washington) provided guidance to the literature on carbon .ijid.2006.07.009. emission levels and their costs. Julian Jamison and Olga Chowell, G., C. Viboud, L. Simonsen, M. A. Miller, and Jonas of the World Bank provided helpful suggestions. The R. Acuna-Soto. 2010. “Mortality Patterns Associated with the 1918 Influenza Pandemic in Mexico: Evidence for a Bill & Melinda Gates Foundation provided partial financial Spring Herald Wave and Lack of Preexisting Immunity in support for this research through grants to the University of Older Populations.” Journal of Infectious Diseases 202 (4): California, San Francisco, for the Commission on Investing 567–75. doi:10.1086/654897. in Health, Phase 3 and to the University of Washington for Davis, K. 1968. The Population of India and Pakistan. New York: the Disease Control Priorities Network. Russell and Russell. Ghendon, Y. 1994. “Introduction to Pandemic Influenza through History.” European Journal of Epidemiology 10 (4): 451–53. NOTE GHRF Commission (Commission on a Global Health Risk Framework for the Future). 2016. The Neglected Dimension World Bank Income Classifications as of July 2014 are as fol- of Global Security: A Framework to Counter Infectious Disease lows, based on estimates of gross national income (GNI) per Crises. Washington, DC: National Academies Press. http:// capita for 2013: www.nap.edu/catalog/21891/the-neglected-dimension-of -global-security-a-framework-to-counter. • Low-income countries (LICs) = US$1,045 or less Global Carbon Project. 2015. “Global Carbon Atlas: • Middle-income countries (MICs) are subdivided: CO2 Emissions.” http://www.globalcarbonatlas.org/?q=en (a) lower-middle-income = US$1,046 to US$4,125 /emissions. (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 Hammitt, J. K., and L. A. Robinson. 2011. “The Income • High-income countries (HICs) = US$12,746 or more. Elasticity of the Value per Statistical Life: Transferring Estimates between High and Low Income Populations.” Journal of Benefit-Cost Analysis 2 (01): 1–29. doi:10.2202 REFERENCES /2152-2812.1009. Jamison, D. T., and J. Jamison. 2011. “Characterizing the Abe, S. 2015. “Japan’s Vision for a Peaceful and Healthier Amount and Speed of Discounting Procedures.” Journal World.” The Lancet 386 (10011): 2367–69. of Benefit-Cost Analysis 2 (2). doi:10.2202/2152-2812.1031. AIR Worldwide. 2014. “The AIR Pandemic Flu Model.” AIR Jamison, D. T., L. H. Summers, G. Alleyne, K. J. Arrow, Worldwide, September 3. http://www.air-worldwide.com S. Berkley, and others. 2013. “Global Health 2035: A World /Publications/AIR-Currents/2014 / The- AIR-Pandemic Converging within a Generation.” The Lancet 382 (9908): -Flu-Model/. 1898–955. doi:10.1016/S0140-6736(13)62105-4. Barro, R. J., and T. Jin. 2011. “On the Size Distribution of Johnson, N. P. A. S., and J. Mueller. 2002. “Updating the Macroeconomic Disasters.” Econometrica 79 (5): 1567–89. Accounts: Global Mortality of the 1918–1920 ‘Spanish’ doi:10.3982/ECTA8827. Influenza Pandemic.” Bulletin of the History of Medicine Beveridge, W. I. 1991. “The Chronicle of Influenza Epidemics.” 76 (1): 105–15. doi:10.1353/bhm.2002.0022. History and Philosophy of the Life Sciences 13 (2): 223–34. Jonas, O. B. 2013. “Pandemic Risk.” World Development Brahmbhatt, M., and A. Dutta. 2008. “On SARS Type Economic Report Background Paper, World Bank, Washington, Effects during Infectious Disease Outbreaks.” Policy DC. https://openknowledge.worldbank.org/bitstream Research Working Paper 4466, World Bank, Washington, /handle/10986/16343/WDR14_bp_Pandemic_Risk_Jonas DC. http://dx.doi.org/10.1596/1813-9450-4466. .pdf?sequence=1&isAllowed=y. Brende, B., J. Farrar, D. Gashumba, C. Moedas, T. Mundel, Lindhjem, H., S. Navrud, N. A. Braathen, and V. Biausque. 2011. and others. 2017. “CEPI-A New Global R&D Organisation “Valuing Mortality Risk Reductions from Environmental, for Epidemic Preparedness and Response.” The Lancet 389 Transport, and Health Policies: A Global Meta-Analysis of (10066): 233–35. doi:10.1016/S0140-6736(17)30131-9. Stated Preference Studies.” Risk Analysis 31 (9): 1381–407. Bruine De Bruin, W., B. Fischhoff, L. Brilliant, and doi:10.1111/j.1539-6924.2011.01694.x. D. Caruso. 2006. “Expert Judgments of Pandemic Liu, J.-T., J. K. Hammitt, J.-D. Wang, and M.-W. Tsou. 2005. Influenza Risks.” Global Public Health 1 (2): 179–94. “Valuation of the Risk of SARS in Taiwan.” Health Economics doi:10.1080/17441690600673940. 14 (1): 83–91. doi:10.1002/hec.911. Burns, A., D. Mensbrugghe, and H. Timmer. 2008. “Evaluating Luk, J., P. Gross, and W. W. Thompson. 2001. “Observations the Economic Consequences of Avian Influenza.” Working on Mortality during the 1918 Influenza Pandemic.” Clinical Paper 47417, World Bank, Washington, DC. http:// Infectious Diseases 33 (8): 1375–78. doi:10.1086/322662. documents.worldbank.org/curated/en/2006/06/10247442 Madhav, N. 2013. “Modelling a Modern-Day Spanish Flu /evaluating-economic-consequences-avian-influenza. Pandemic.” AIR Worldwide. February 21. http://www Cheng, K. F., and P. C. Leung. 2007. “What Happened in .air-worldwide.com/publications/air-currents/2013/modeli China during the 1918 Influenza Pandemic?” International ng-a-modern-day-spanish-flu-pandemic/. 356 Disease Control Priorities: Improving Health and Reducing Poverty Madhav, N., B. Oppenheim, M. Gallivan, P. Mulembakani, Ozawa, S., M. L. Stack, D. M. Bishai, A. Mirelman, I. K. Friberg, E. Rubin, and others. 2018. “Pandemics: Risks, Impacts, and and others. 2011. “During the ‘Decade Of Vaccines’, the Lives of Mitigation.” In Disease Control Priorities (third edition): 6.4 Million Children Valued at $231 Billion Could Be Saved.” Volume 9, Disease Control Priorities: Improving Health and Health Affairs 30 (6): 1010–20. doi:10.1377/hlthaff.2011.0381. Reducing Poverty, edited by D. T. Jamison, H. Gelband, Pachauri, R., M. Allen, V. Barros, J. Broome, W. Cramer, and S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and others, eds. 2014. “Climate Change 2014: Synthesis Report. R. Nugent. Washington, DC: World Bank. Contribution of Working Groups I, II and II to the Fifth Mathers, C., G. Stevens, D. Hogan, A. Mahanani, and J. Ho. Assessment Report of the Intergovernmental Panel on 2018. “Global and Regional Causes of Death: Patterns and Climate Change.” Intergovernmental Panel on Climate Trends, 2000–15.” In Disease Control Priorities (third edi- Change, Geneva. http://ar5-syr.ipcc.ch/. tion): Volume 9, Disease Control Priorities: Improving Health Paules, C. I., and A. S. Fauci. 2017. “Emerging and Reemerging and Reducing Poverty, edited by D. T. Jamison, H. Gelband, Infectious Diseases: The Dichotomy between Acute S. Horton, P. Jha, R. Laxminarayan, C. N. Mock, and Outbreaks and Chronic Endemicity.” Journal of the R. Nugent. Washington, DC: World Bank. American Medical Association 317 (7): 691–92. doi:10.1001/ McDonald, M., K.-B. Scott, W. J. Edmunds, P. Beutels, and jama.2016.21079. R. D. Smith. 2008. “The Macroeconomic Costs of a Global Pike, J., T. Bogich, S. Elwood, D. C. Finnoff, and P. Daszak. Influenza Pandemic.” Paper presented at the Global Trade 2014. “Economic Optimization of a Global Strategy to Analysis Project 11th Annual Conference on Global Address the Pandemic Threat.” Proceedings of the National Economic Analysis, “Future of Global Economy,” Helsinki, Academy of Sciences of the United States of America 111 (52): June. http://www.gtap.agecon.purdue.edu/resources 18519–23. doi:10.1073/pnas.1412661112. /download/3828.pdf. Pindyck, R. S, and N. Wang. 2013. “The Economic and Policy McKibbin, W. J., and A. Sidorenko. 2006. “Global Consequences of Catastrophes.” American Economic Journal: Macroeconomic Consequences of Pandemic Influenza.” Economic Policy 5 (4): 306–39. doi:10.1257/pol.5.4.306. Lowy Institute for International Policy, Sydney, Australia. Pizer, W., M. Adler, J. Aldy, D. Anthoff, M. Cropper, and others. Moon, S., J. Leigh, L. Woskie, F. Checchi, V. Dzau, and others. 2014. “Using and Improving the Social Cost of Carbon.” 2017. “Post-Ebola Reforms: Ample Analysis, Inadequate Science 346 (6214): 1189–90. Action.” British Medical Journal 356 (January): j280. Posner, R. A. 2004. Catastrophe: Risk and Response. Oxford, doi:10.1136/bmj.j280. U.K.: Oxford University Press. Moore, F. C., and D. B. Diaz. 2015. “Temperature Impacts Potter, C. W. 2001. “A History of Influenza.” Journal of Applied on Economic Growth Warrant Stringent Mitigation Microbiology 91 (4): 572–79. Policy.” Nature Climate Change 5 (2): 127–31. doi:10.1038/ Preston, S., P. Heuveline, and M. Guillot. 2000. Demography: nclimate2481. Measuring and Modeling Population Processes. Malden, MA: Morens, D. M., and A. S. Fauci. 2007. “The 1918 Influenza Wiley-Blackwell. Pandemic: Insights for the 21st Century.” Journal of Sands, P., A. El Turabi, P. A. Saynisch, and V. J. Dzau. 2016. Infectious Diseases 195 (7): 1018–28. doi:10.1086/511989. “Assessment of Economic Vulnerability to Infectious Disease Morens, D. M., G. K. Folkers, and A. S. Fauci. 2004. “The Crises.” The Lancet 388 (10058): 2443–48. doi:10.1016/ Challenge of Emerging and Re-Emerging Infectious S0140-6736(16)30594-3. Diseases.” Nature 430 (6996): 242–49. doi:10.1038/ Shindell, D. T., Y. Lee, and G. Faluvegi. 2016. “Climate and nature02759. Health Impacts of US Emissions Reductions Consistent Morens, D. M., and J. K. Taubenberger. 2011. “Pandemic with 2°C.” Nature Climate Change 6: 503–507. doi:10.1038/ Influenza: Certain Uncertainties.” Reviews in Medical nclimate2935. Virology 21 (5): 262–84. doi:10.1002/rmv.689. Simonsen, L., M. J. Clarke, L. B. Schonberger, N. H. Arden, Murray, C. J. L., A. D. Lopez, B. Chin, D. Feehan, and K. H. Hill. N. J. Cox, and others. 1998. “Pandemic versus Epidemic 2006. “Estimation of Potential Global Pandemic Influenza Influenza Mortality: A Pattern of Changing Age Mortality on the Basis of Vital Registry Data from the Distribution.” Journal of Infectious Diseases 178 (1): 53–60. 1918–20 Pandemic: A Quantitative Analysis.” The Lancet Smith, K. R., Manish A. Desai, J. V. Rogers, and R. A. Houghton. 368 (9554): 2211–18. doi:10.1016/S0140-6736(06)69895-4. 2013. “Joint CO2 and CH4 Accountability for Global Nordhaus, W. D. 2010. “Economic Aspects of Global Warming Warming.” Proceedings of the National Academy of Sciences in a Post-Copenhagen Environment.” Proceedings of 110 (31): E2865–74. the National Academy of Sciences 107 (26): 11721–26. Taubenberger, J. K., D. M. Morens, and A. S. Fauci. 2007. “The doi:10.1073/pnas.1005985107. Next Influenza Pandemic: Can It Be Predicted?” Journal OECD (Organisation for Economic Co-operation and of the American Medical Association 297 (18): 2025–27. Development). 2014. The Cost of Air Pollution: Health doi:10.1001/jama.297.18.2025. Impacts of Raod Pollution. Paris: OECD. http://www.oecd Tol, R. S. J. 2013. “Climate Change. CO2 Abatement.” In Global -ilibrary.org/content/book/9789264210448-en. Problems, Smart Solutions: Costs and Benefits, edited Osterholm, M. T. 2005. “Preparing for the Next Pandemic.” by B. Lomborg, 186–91. Cambridge, U.K.: Cambridge New England Journal of Medicine 352 (18): 1839–42. University Press. The Loss from Pandemic Influenza Risk 357 UNPD (United Nations Population Division). 2015. World World Bank. 2015. World Development Indicators database. Population Prospects, the 2015 Revision—Key Findings and http://data.worldbank.org/products/wdi. Advance Tables. New York: United Nations Department of WHO (World Health Organization). 2016a. “Ebola Situation Economic and Social Affairs. http://esa.un.org/unpd/wpp/. Report—16 March 2016.” http://apps.who.int/ebola Viscusi, W. K. 2014. “The Value of Individual and Societal Risks to /current-situation/ebola-situation-report-16-march-2016. Life and Health.” In Handbooks in Economics: Vol. 1, Economics ———. 2016b. “Influenza (Seasonal).” Fact Sheet, November, of Risk and Uncertainty, edited by M. Machina and W. K. WHO, Geneva. http://www.who.int/mediacentre/factsheets Viscusi, 385–452. Amsterdam: North-Holland. http://www /fs211/en/. .sciencedirect.com/science/article/pii/B9780444536853000076. Yamey, G., M. Schäferhoff, O. K. Aars, B. Bloom, D. Carroll, Watkins, D. A., and A. Daskalakis. 2015. “The Economic Impact and others. 2017. “Financing of International Collective of Rheumatic Heart Disease in Developing Countries.” The Action for Epidemic and Pandemic Preparedness.” The Lancet Global Health 3: S37. Lancet Global Health. doi:10.1016/S2214-109X(17)30203-6. 358 Disease Control Priorities: Improving Health and Reducing Poverty Chapter 19 Fiscal Instruments for Health in India Amit Summan, Nicholas Stacey, Karen Hofman, and Ramanan Laxminarayan INTRODUCTION • The narrow fiscal space for health care, even in countries with relatively high growth rates, is a con- It has been recognized for some time that the primary sequence of a low tax base and constrains health care determinants of population health and health inequalities, spending by national and state governments. In India, particularly in low- and middle-income countries (LMICs), although government health expenditures as a pro- lie outside of the health care system (CSDH 2008). These portion of total government expenditures are compa- determinants include individual-level factors—such as rable to similar countries, they lag when measured as access to clean water and sanitation, nutrition, and antena- a proportion of gross domestic product (GDP). tal care—as well as environmental-level factors—such as • Countries seeking to transition to UHC have weak pollution, walkability of neighborhoods, rates of open health care systems that are challenged in delivering defecation, and tariffs on food imports and exports. quality health care coverage even when additional Exposure to these hazardous risk factors is the resources are available. India and South Africa are primary contributor to adverse health outcomes, which examples of countries where the health care system increase resource demands on health care systems and serves a fairly small proportion of the population; increase private and public health expenditures. The large segments are excluded from even basic health impetus for universal health coverage (UHC) in coun- coverage. tries as diverse as Brazil, India, and South Africa has run up against the barrier of these broader determinants that Despite the recognition that social determinants exer- hinder efforts to improve health. There are three addi- cise a significant influence on population health in tional challenges to UHC: LMICs as direct interventions in the health sector, there remains a limited understanding of how existing fiscal • The economic slowdown has significantly reduced policy instruments available to governments in LMICs growth rates and government revenues in LMICs. can be leveraged to improve health. Annual growth rates in Brazil, the Russian Federation, This chapter presents the analytic framework India, China, and South Africa (BRICS) were a pop- for assessing the potential of fiscal instruments to ulation weighted average of nearly two percentage improve population health. We describe the applica- points lower during 2011–15 than during the pre- tion of this method to specific interventions in India vious decade (World Bank and IHME 2016). As a and discuss the implications of these policy changes. result, government expenditures and the ability to The goal is to inform policies at ministries of finance increase spending on health care have tightened. that have an effect on health, either through new Corresponding author: Ramanan Laxminarayan, Center for Disease Dynamics, Economics and Policy, Washington, DC; ramanan@cddep.org. 359 policies or by examining existing policies that affect for twice the direct health expenditures of the roughly important health risk factors. US$18 billion spent by the state and central governments on health. Tax and tariff policies are also important and can potentially modify health when applied to commod- ities that potentially affect health adversely, including ROLE OF FISCAL POLICY INTERVENTIONS alcohol, tobacco, salt, sugar, and trans fats. Current levels Fiscal measures, including tax and subsidy reforms, offer of taxes and subsidies for key influencers of health are an appealing complementary opportunity to improve described in table 19.2. health without reliance on additional budgetary alloca- Fiscal policies can also implicitly influence health and tions to ministries of health. In India (table 19.1), increase public usage of health systems by modifying subsidies for food, fertilizer, and petroleum—three com- incentives for treatment of illness, prevention of illness, modities that can have large direct and indirect health and promotion of healthy lifestyles. Additionally, fiscal impacts—total US$42 billion and together account policies can be used to influence the large portion of Table 19.1 Current National Accounts for India: Combined Revenue and Capital Expenditures and Receipts for Central and State Governments Item US$ (Rs 65 = US$1) GDP (percent) GDP at current market prices (BE) 2.17 trillion 100.00 Revenue receipts (BE) 437 billion 20.15 Revenue expenditures (BE), including 488 billion 22.51 • Interest payments 67 billion 4.75 • Food subsidy 20 billion 0.92 • Fertilizer subsidy 11 billion 0.52 • Petroleum subsidy 5 billion 0.21 • Health expenditures (includes medical and public health, 21 billion 0.96 water supply, sanitation, and family welfare) • Defense 23 billion 1.07 Total capital expenditures, including loans and advances 95 billion 4.42 Total expenditures (revenue + capital) 583 billion 26.89 Source: Ministry of Finance 2016. Note: BE = budget estimate; GDP = gross domestic product; Rs = Indian rupees. Table 19.2 Current Levels of Taxes/Subsidies and Health Risk Factors and Outcomes in India Commodity Outcome Risk factor Instrument Level of tax/subsidy in India a Cigarettes Cancers, heart disease Smoking, chewing Tax 33% plus Rs 2076 per thousand cigarettes tobacco (Central Board of Excise and Customs 2017) Alcohol Road traffic accidents, Drunk driving, Tax Rates vary by state and product, including cancers, liver disease, STIs unsafe sex prohibition in five states Condoms STIs Unsafe sex Subsidy Free condoms for high-risk groups (Ministry of Health and Family Welfare 2016a) Vaccines Infectious diseases Measles, Subsidy Under Universal Immunization Programme, 10 free pneumococcal vaccines provided against VPDs (Ministry of disease, other VPDs Health and Family Welfare 2016b) table continues next page 360 Disease Control Priorities: Improving Health and Reducing Poverty Table 19.2 Current Levels of Taxes/Subsidies and Health Risk Factors and Outcomes in India (continued) Commodity Outcome Risk factor Instrument Level of tax/subsidy in India Essential drugs for HIV, TB, malaria, bacterial Lack of treatment Subsidy 100% for ARTs (Ministry of Health and Family infectious disease infections Welfare 2016c); 100% for TB DOTS (Ministry of Health and Family Welfare 2016d) TB rapid diagnostics TB Lack of TB diagnosis Subsidy GeneXpert ceiling price of Rs 2,000 for private clinics receiving reduced pricing (Initiative for Promoting Affordable Quality TB Tests 2013) Salt Stroke Hypertension Tax None (Central Board of Excise and Customs 2017) Sugar-sweetened Cancer, heart disease, Obesity Tax 40% (Central Board of Excise and Customs 2017) beverages diabetes Trans fats Heart disease, diabetes Obesity Tax None Diesel COPD Air pollution Tax 18.6–27% (varies by state) LPG to substitute for TB, COPD Air pollution (reduced) Subsidy Rs 568 per 14.2 kg cylinder (or Rs 40/kg) solid cooking fuels Note: ARTs = antiretroviral therapies; COPD = chronic obstructive pulmonary disease; DOTs = directly observed treatment, short course; HIV = human immunodeficiency virus; kg = kilogram; LPG = liquefied petroleum gas; Rs = Indian rupees; STIs = sexually transmitted diseases; TB = tuberculosis; VPDs = vaccine preventable diseases. a. Cigarettes not exceeding 65 mm in length. health prevention expenditures still occurring in the that are not taken into consideration by those who con- private sector that are not directly paid for or monitored sume them. In the case of alcohol and cigarettes, the by the government. The government’s role can be to externalities are negative—consumption of these goods encourage uptake of preventive health services using causes secondhand smoke or fires (cigarettes) and drunk direct subsidy policies that are similar to the production driving accidents (alcohol). In the case of condoms and level subsidy for antimalarial artemisinin-combination vaccines, the externalities are positive because of reduc- therapies initiated under the Affordable Medicines tions in the transmission of infections. Taxes can be Facility-malaria (AMFm) financing mechanism. Fiscal levied to facilitate a socially optimal level of consump- policies are practical alternatives to regulation, particu- tion of commodities with negative externalities; larly in areas where regulation is challenged by the num- subsidies can be used for commodities with positive ber of actors. For example, subsidies for micronutrient externalities. Paternalistic preferences—where the state’s fortification of food commodities may be more effective desire to improve societal welfare supersedes the individ- than compulsory fortification when there are many pro- ual’s preferences—over health outcomes for other ducers and it is difficult to enforce compliance (Chow, households are a common, although contentious, justifi- Klein, and Laxminarayan 2010). Fiscal policies can also cation for government intervention. Paternalistic prefer- be more effective than regulation in modifying incen- ences recognize that the social marginal benefit from tives. For example, a package of regulatory interventions better health exceeds the private marginal benefit in the to reduce carbon emissions—efficiency standards for case of a positive consumption externality, thereby off- buildings, fuel efficiency standards for vehicles, and a setting the distortion created by the subsidy instrument carbon ceiling for energy production—could encourage (Browning 1999.) the substitution of alternative energy sources and reduc- However, the optimal tax on a commodity may tions in emissions intensity through greater efficiency; exceed any amount that might be justified on externality however, these regulations would still fail to reduce fuel grounds alone if the commodity is a weaker substitute demand (Parry and others 2014). for leisure than the average consumption good; the opti- mal tax rises further the more inelastic the demand for the taxed commodity (Sandmo 1976).1 Taxing leisure ANALYTIC FRAMEWORK items—such as tobacco or alcohol—would discourage The consumption of commodities such as alcohol, ciga- their use during leisure activities and consequently rettes, condoms, and vaccines involves external effects increase the labor supply. If these taxes offset labor taxes, Fiscal Instruments for Health in India 361 which distort labor and leisure decisions, they would calculate premature deaths averted and years of life increase welfare. Therefore, a tax on individual products gained (YLG). A lag factor is used to incorporate the can increase welfare, but this will further depend on delay in change in exposure to change in risk and to whether tax-neutrality is specified in legislation. Because account for the irreversibility of the effects of some extra tax revenues could end up funding more public exposures. Incorporating fertility rates and trends in spending rather than other tax reductions, the fiscal future mortality rates, we project the difference in the rationale for higher taxes may be undermined and number of deaths and YLG over 15 years. We estimate would have to be evaluated under alternative possibili- changes in health expenditures (both private and pub- ties for recycling of the revenues. In previous work, lic, except for tuberculosis diagnostic tools subsidies we estimated that the optimal tax on alcohol exceeds the where only private expenditures are estimated) and level warranted on externality grounds by between government receipts. To capture uncertainty, we con- 59 and 126 percent, because of the revenue-raising com- duct Monte Carlo simulations with 1,000 iterations at ponent of the optimal tax (Parry, Laxminarayan, and the 95 percent confidence interval on relative risk West 2009). and elasticity parameters. To assess the health and economic effects of tax The outcomes of taxation will significantly depend and subsidy interventions in India, we use simple on the elasticity of demand. If demand is inelastic, a macrosimulation spreadsheet-based simulation models. higher tax will cause only a small fall in demand. Most Taxes reduce consumption of the taxed good (or of the tax will be passed on to consumers. When increase it in the case of a subsidy—a negative tax), demand is inelastic, governments will see a significant which changes exposure to risk factors within the increase in tax revenue. However, if demand is elastic, affected populations. We employ statistical parameters the tax will be effective in reducing demand for the com- called elasticities to estimate the change in consump- modity, which is helpful in reducing its adverse health tion caused by changes in prices. We assume full pass- impact but may be less effective in raising revenue. through of the tax to the consumer and zero tax Table 19.3 summarizes the evidence on price elasticity of evasion, except for the alcohol tax intervention. We demand for various categories of health-impacting employ a lagged population impact factor, which esti- commodities. The next section presents the results of mates the proportional reduction in risk from changing our fiscal policy simulations and complementary policy risk factor exposure, in conjunction with life tables to recommendations. Table 19.3 Price Elasticity of Demand for Various Commodities That Influence Health Commodity Elasticity Country Year Source Link Tobacco (bidis) −0.89 India 1999–2000 John (2008) http://heapol.oxfordjournals.org/content/23/3/200 /T5.expansion.html Alcohol −0.9495 India 1999–2000 John (2005) http://oii.igidr.ac.in:8080/jspui /bitstream/2275/24/1/WP-2005-003.pdf Condoms −0.5 to −0.1 Pakistan 1998 Matheny (2004) http://www.guttmacher.org/pubs /journals/3013404.html −0.29 to -2.68 Bangladesh, 1995 Matheny (2004) http://www.guttmacher.org/pubs Haiti, Pakistan /journals/3013404.html Vaccines 0 to −1.07 Japan 2001–02 and Kondo, Hoshi, and http://www-ncbi-nlm-nih-gov.ezproxy.bu (influenza) 2004–05 Okubo (2009) .edu/pubmed/?term=Does+subsidy+wo rk%3F+Price+elasticity+of+demand+for +influenza+vaccination+among+the +elderly+in+Japan Essential drugs for −1.9 (Indinavir Morocco 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/767 treating infectious for HIV) McGuire (2014) diseases −1.2 (Nevirapine Lebanon, 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/768 for HIV) Morocco McGuire (2014) table continues next page 362 Disease Control Priorities: Improving Health and Reducing Poverty Table 19.3 Price Elasticity of Demand for Various Commodities That Influence Health (continued) Commodity Elasticity Country Year Source Link −1.4 (Streptomycin Morocco 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/769 for TB) McGuire (2014) −1.4 (Benzathine Morocco 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/770 benzylpenicillin) McGuire (2014) −1.1 to −1.9 Lebanon, 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/771 (Zidovudine for HIV) Malaysia McGuire (2014) −1.0 to −1.7 various 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/772 (Ceftriaxone) countries McGuire (2014) −1.0 to −1.6 various 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/773 (Ciprofloxacin) countries McGuire (2014) −1.5 to −1.2 Syria, Tunisia 2003 Srivastava and http://www.biomedcentral.com/1471-2458/14/774 (Co-trimoxazole) McGuire (2014) Sugar-sweetened −0.94 India 2009/10 Basu and others http://www.plosmedicine.org/article/info%3Adoi beverages (2014) %2F10.1371%2Fjournal.pmed.1001582 −1.09 Mexico 1998–99 Barquera and others http://jn.nutrition.org/content/138/12/2454.long (2008) −0.85 Brazil 2005–06 Claro and others https://www.ncbi.nlm.nih.gov/pmc/articles (2012) /PMC3490548/ Grains (rice) −0.247 India 1983–2004 Kumar and others http://ageconsearch.umn.edu/bitstream (2011) /109408/2/1-P-Kumar.pdf Grains (wheat) −0.34 India 1983–2004 Kumar and others http://ageconsearch.umn.edu/bitstream (2011) /109408/2/1-P-Kumar.pdf Trans fats −0.48 USA 1938–2007 Andreyeva, Long, and http://www.ncbi.nlm.nih.gov/pmc/articles Brownell (2010) /PMC2804646/ Palm oil −1.24 USA 1991/92– Kojima, Parcell, and http://ageconsearch.umn.edu/bitstream 2010/11 Cain (2014) /162472/2/A%20Demand%20Model%20of% 20the%20Wholesale%20Vegetable%20Oils% 20Market%20in%20the%20U.S.A%20(Revised% 20in%20March%202014)%20(1).pdf Diesel −0.55 Korea. Rep. 1986–2011 Lim and others (2012) http://www.mdpi.com/1996-1073/5/12/5055 (long-run value) LPG to substitute −0.92 to −1.05 India 1998–99 Gundimeda and http://www.eaber.org/node/22501 for solid cooking Köhlin (2006) fuels Note: A price elasticity of demand greater than −1 is considered elastic and less than −1 is considered inelastic. A price elasticity of demand equal to −1 would mean a 1 percent change in price results in a 1 percent change in demand. Inelastic goods tend to have fewer substitutes (gasoline), constitute a small percentage of expenditures (salt), or may be necessary for survival (for example, food). Bidi = a small, thin, hand-rolled cigarette made in Southeast Asian countries. HIV = human immunodeficiency virus; LPG = liquefied petroleum gas; TB = tuberculosis. FISCAL POLICY ANALYSIS AND subsidies for sugar, cooking fuels, and tuberculosis diag- RECOMMENDATIONS nostic tools. This section discusses the main results of our fiscal policy interventions and presents complemen- We explore both fiscal policies that have been adopted tary policy recommendations. In many cases, the success widely and those that have been introduced only recently. of the tax and subsidy policy can be strengthened by These include taxes on alcohol, tobacco, coal, transpor- implementing these complementary policies. Results tation fuels, and sugar-sweetened beverages (SSBs) and from the models are highlighted in table 19.4. Fiscal Instruments for Health in India 363 Table 19.4 Results Summary—Health and Economic Effects of India’s Main Fiscal Policies, 2017–2032 Tax revenue Decreased health Intervention YLG Discounted YLG Deaths averted gains expenditures area Product Intervention (thousands) (thousands) (thousands) (US$,a millions) (US$, thousands) Tobacco Bidi 20% price increase 23,082 17,038 3,561 3,998 87,322 (200% tax increase) (13,742–33,131) (10,203–24,427) (2,020–5,231) (3,345–4,521) (63,692–114,307) Cigarette 50% price increase 7,108 5,410 851 16,200 40,743 (90% tax increase) (3,695–11,577) (2,803–8,846) (449–1,359) (11,597–21,081) (27,230–53,846) Alcohol Country 20% price increase 300 206 35 12,977 81,002 Liquor (170% tax increase) (114–482) (74–339) (13–56) (12,303–13,554) (60,307–114,769) Foreign 20% price increase 76 58 9 24,828 63,127 Liquor (95% tax increase) (−4–170) (−3–130) (−1–20) (24,286–25,292) (49,538–77,230) Cooking fuel LPG 25% of WQ 1 and 2 25,839 67,633 12,197 0 399,548 households receive (2,515–170,956) (1,888–127,989) (331–23,552) (149,692–564,000) LPG subsidy Fossil fuels Diesel Rs 2.38/liter annual 86 64 13 268,508 544 tax increase (46–135) (34–100) (7–21) (223,824–308,654) (77–1,430) Gasoline Rs 1.54/liter annual 26 20 5 146,170 30 tax increase (12–41) (9–30) (2–7) (123,655–166,502) (4–69) Coal Rs 100 annual levy 419 307 82 164,223 51,754 increase over 15 (216–607) (158–444) (42–118) (157,008–171,320) (8,153–113,692) years Food Sugar Removal of public 5,570 4,331 437 10,385 27,278 distribution of sugar (2,380–8,790) (1,850–6,835) (174–704) (17,538–40,153) subsidy SSB Tax 20% price increase 267 200 41 74,277 2,559 (114% tax increase) (109–434) (82–325) (17–68) (73,061–75,704) (1,692–3,846) Tuberculosis GeneXpert Replace 1 million 5,463 4,131 704 0 105,287 diagnostic tools sputum smear tests (3,610–7,463) (2,730–5,642) (464–962) (–83,384–284,769) with GeneXpert annually Note: LPG = liquified petroleum gas; Rs = Indian rupees; SSB = sugar-sweetened beverage; WQ = wealth quintile; YLG = years of life gained. a. US$1 = Rs 65. Taxation harmonization in taxes across states until July 2017, Taxation of tobacco (cigarettes and bidis [small, thin, when the goods and services tax (GST) harmonized hand-rolled cigarettes made in Southeast Asian coun- tobacco tax. In recent years, including the increase in tax tries]), alcohol (country liquor and foreign liquor), fossil due to GST implemenation, the real tax increase on cig- fuels (diesel, petrol, and coal), and SSBs are discussed in arettes has been small, and bidi taxes remain significantly this section. lower than levels recommended by the World Health Organization (WHO). Tobacco Our simulations focus on increased taxation of Tobacco taxation is one of India’s most familiar and smoked tobacco products. Our modeling suggests widely used health-directed fiscal policies. In 2016, that increasing the bidi tax by 200 percent could lead to roughly 29 percent of Indian adults used tobacco in 23.0 (95% confidence interval [CI]: 13.8–33.1) million some form (smoked or smokeless) (Ministry of Health YLG over 15 years and an increase in government tax and Family Welfare 2017). Over 900,000 lives are lost revenues by US$3.9 (CI: $3.3–$4.5) billion. Health prematurely each year from tobacco-related diseases expenditures can decrease by US$87 (CI: $63–$114) (IHME 2015). The Indian federal government and the million from the bidi tax increase. Increasing the ciga- states taxed tobacco products, with significant lack of rette tax by 90 percent can lead to 7.1 (CI: 3.6–11.6) 364 Disease Control Priorities: Improving Health and Reducing Poverty million YLG over 15 years and an increase in govern- is illicitly produced. Increasing Indian liquor taxa- ment tax revenues of $16.2 (CI: $11.6–$21.0) billion. tion and foreign liquor taxation by 170 percent and Health expenditures can decrease by US$40.7 (CI: 95 percent, respectively, could result in 300,000 (CI: $27.2–$53.8) million. Our estimates of health effects 114,000–482,000) YLG and 76,000 (CI: −4,000–170,000) ignore the harms of secondhand smoke, resulting in an YLG, respectively, over 15 years. Tax revenues can underestimation of possible health gains. Additional increase by US$13.0 (CI: $12.3–$13.5) billion from recommendations presented are directed at creating a country liquor taxation and by US$24.8 (CI: $24.3– more consistent tax structure throughout the country $25.3) billion from foreign liquor taxation, over 15 and stepping up the implementation of complementary years.2 Health expenditures can decrease by US$81 (CI: interventions. $60–$114) million from country liquor taxation and by The following additional interventions could further US$63 (CI: $49–$77) million over 15 years from foreign improve health and ensure success of taxation: liquor taxation. In our analysis of health effects, we exclude externalities, which would include individuals • Increase state and union territory National Tobacco killed by drunk drivers or alcohol-induced violence Control Programme (NTCP) fund transfers to be used against others, resulting in an underestimate of the for improved awareness and education campaigns, health effects. These health gains and any excess gains are more effective smoking cessation centers, and greater contingent on strong tax administration and control of enforcement of existing laws. illicit liquor production. We make four complementary • Link current tobacco taxes to inflation. recommendations: • Allocate funds to retrain bidi workers and tobacco farmers. • Formulate a national strategy on alcohol policy to • Remove tax exemptions for small bidi producers. guide state-level alcohol policy. • Remove price controls on tendu leaves (a plant native • Use alcohol tax revenue for research on alcohol to Asia that is used for making bidis). consumption patterns and unrecorded alcohol production. Traditional economic theories suggest that as taxes • Restrict the marketing of alcohol products to youth. increase, the incentives for smuggling and black market • Earmark alcohol tax revenues for strengthening activity increase (Cnossen 2006). Black market activity, enforcement to reduce the consumption of illicitly by its very nature, is difficult to gauge; for this reason, produced liquor. it is also difficult to measure black market activities that • Increase funding for alcohol addiction centers. involve tobacco products that are smoked (Blecher and others 2015). However, it is well documented that many In the case of alcohol taxation, the presence of a very countries have successfully implemented high levels of large illicit market for Indian-made liquor may challenge tobacco taxation without drastic increases in black the success of future tax increases and possibly exacer- market activity (WHO 2015). In our analysis, we only bate the current illicit liquor problem. Therefore, it is simulate tax levels consistent with WHO recommenda- necessary to first ensure that future tax increases do not tions for tobacco products that are smoked and addi- result in increased illicit liquor production by providing tionally recommend greater resources for India’s NTCP, greater monitoring of the alcohol market and increased which follows best practices for curtailing black market resources for tax administration. activity. Fossil Fuels Alcohol Fossil fuel taxes—on coal, diesel, and gasoline—are Alcohol taxation is another widely used health-directed designed to reduce air pollution and its massive deleteri- fiscal policy. In 2015, alcohol consumption in India was ous health consequences in India. Ambient particulate implicated in nearly 360,000 premature deaths (IHME matter pollution costs the Indian economy an estimated 2015). Alcohol taxes are levied at the state and central Rs 3.1 trillion per year, or 0.89 percent of GDP (World government levels and provide as much as 20 percent of Bank 2016). The two major sources of air pollution are state government income except in the states of Bihar, emissions from coal-fired power plants and vehicles. Gujarat, Manipur, Mizoram, and Nagaland, where alco- An annual increase of Rs 2.38 and Rs 1.54 per liter in hol is prohibited. Like tobacco, alcohol taxes are complex the diesel and gasoline taxes over 15 years could result and inconsistent. Alcohol regulation is further compli- in 86,000 (CI: 46,000–135,000) and 26,000 (CI: 12,000– cated by the presence of a large illicit liquor market; 41,000) YLG respectively, and an increase in aggre- some estimates suggest that up to 50 percent of alcohol gate tax revenues of US$414 (CI: $436–$474) billion. Fiscal Instruments for Health in India 365 Aggregate health expenditures could decrease by • Restrict advertisements for sugary beverages. US$574,000 (CI: $81,000–$1,494,000). Complementary • Subsidize healthier food options. recommendations include the following: • Allocate tax revenues for public transportation Subsidies investments. The analysis of the remaining health-directed fiscal • Implement toll roads or congestion charges. policies involve subsidies related to sugar, cooking fuels, • Establish new parking fines and enforce current fines. and tuberculosis diagnostic tools. • Facilitate the adoption of improved emission standards for vehicles. Public Distribution Sugar Subsidy • Reduce and control fuel adulteration. The first policy examined the reduction or elimination of the existing public distribution sugar subsidy. The Annually increasing the coal levy, which is now largely a past sugar subsidy under the public distribution system means of raising revenue for the National Clean Energy (PDS) (US$692 million annually) provided sugar subsi- Fund (NCEF), by Rs 100 over 15 years could prevent 82,000 dies to poor households. This year, the Indian govern- (CI: 42,000–118,000) premature deaths and result in ment announced it would not be funding the subsidy 419,000 (CI: 216,000–607,000) YLG while increasing tax and left this option to the states. Recently, however, the revenues by US$164 (CI: $157–$171) billion over 15 years. government has decided to provide sugar subsidies to Health expenditures could decrease by US$51 (CI: $8–$113) only the 25 million poorest families in the country. million. We only consider the health effects from changes in Historically, inclusion error has resulted in richer house- coal used for power generation and exclude the 30 percent holds also benefiting from the subsidy. Removal of the of coal used for other purposes, resulting in a conservative subsidy could result in 5.5 (CI: 2.3–8.8) million YLGs estimate of the possible health effects. Complementary over 15 years. Our estimates suggest health expenditures recommendations for coal taxation are as follows: could decrease by US$27.3 (CI: $17.5–$40.1) million. For our analysis, we have considered the effects of the • Increase the coal levy revenue allocation to the NCEF. intervention on body mass index (BMI) and added sugar • Increase transparency in the use of NCEF funds and consumption. Although individuals in the the lowest use them for the intended purposes. wealth quintile benefited from reduced added sugar con- • Prioritize NCEF allocations for improving the grid sumption, including potential reductions in BMIs, there infrastructure. are concens about the negative health consequences of • Allocate revenues to increase the efficiency of coal- reduced BMIs. Therefore, we recommend that poorer fired power plants to reduce emissions. households receive a replacement subsidy for healthy • Allocate coal levy revenues to expand continuous food products, such as fruits, vegetables, or grains, rather emissions monitoring systems in power plants. than a sugar subsidy, as the current policy has suggested. Although past PDS subsidies have sometimes failed to Sugar-Sweetened Beverages target their intended beneficiaries, some states have suc- An increase in the tax on SSBs could help to curb the cessfully implemented reforms in recent years that nascent obesity epidemic in India. An SSB tax was first encourage us to suggest greater subsidies that target the imposed in 2014 and was increased to 21 percent in 2017. poor in lieu of the sugar subsidy. For example, Bihar has This tax had not dampened demand sufficiently, and been able to decrease leaks (diversion of subsidized food following India’s Committee on Goods and Services Tax’s commodities to nonbeneficiaries) from 91 percent in recommendation, the tax was increased to 40 percent 2004 to 24 percent in 2011, with further improvements under the GST. We found that a tax increase of 114 percent, in the past few years, by tracking coupon use and better corresponding to a tax rate of 40 percent, could result in targeting households who would benefit most from 267,000 (CI: 109,000–434,000) YLG over 15 years and a reduced sugar subsidy (Dreze and Khera 2015). increase tax revenues by US$74 (CI: $73–$76) billion. Complementary measures to promote healthy diets are Health expenditures can decrease by $2.5 (CI: $1.7–$3.8) similar to those discussed in the SSB tax section. million. Complementary policies include the following: Cooking Fuel Subsidies • Conduct education and awareness campaigns on Improved targeting of the cooking fuel subsidy is mod- healthy diets. eled to estimate the effect of accelerated progress of • Label the sugar content of drinks clearly to make the current liquefied petroleum gas (LPG) subsidy. The nutritional information accessible to consumers. rationale for this government subsidy is to reduce the 366 Disease Control Priorities: Improving Health and Reducing Poverty number of households relying on biomass fuel for • Conduct public awareness campaigns on Revised indoor cooking, which takes a large toll on cardiovascu- National Tuberculosis Control Program (RNTCP), and lar and respiratory health. Unfortunately, as imple- publicize tuberculosis prevention and treatment options. mented, the subsidy has not greatly benefited the target • Engage private health care providers to improve their population—households in the lowest income quintile, diagnostic and treatment practices. particularly in rural areas—because of distribution chal- • Promote public-private alliances (PPAs), including lenges and preferences for biomass cooking. If 25 percent innovative schemes to incentivize notification and of households currently using biomass switched to LPG referral of patients to the RNTCP. next year, the result would be 25.8 (CI: 2.5–170.9) • Conduct periodic national surveys of tuberculosis million YLGs over 15 years. Health expenditures would prevalence and treatment practices. decrease by US$399 (CI: $149–$565) million. To ensure the success of the intervention, it will be critical to invest Increased taxes will necessitate increases in tax admin- in education and other behavioral change interventions istration resources. Our modeling results will be realized to increase uptake of the LPG subsidy. Uptake of the only if new taxes are actually collected. This may be more LPG subsidy can even be considered a greater challenge of a challenge for some items, such as alcohol, where than ensuring supply and accessibility, as has been additional resources must be employed to control illicit demonstrated in previous studies (Grossman 2012; liquor production. Our complementary policies suggest Hanna, Duflo, and Greenstone 2012). Employing inno- some of the ways in which the unintended negative con- vative behavior change interventions will help increase sequences can be mitigated and overall welfare gains demand for LPG cooking. maximized. These additional policies include assistance for affected workers and producers as they transition to Tuberculosis Diagnostic Tools alternative industries, investments in superior substitutes The final intervention is a subsidy for tuberculosis diag- (in the case of fuels, for example), and strengthening nostic tools. India has the highest tuberculosis burden in monitoring and enforcement of regulations. Deploying a the world. Progress in controlling tuberculosis has been portion of the tax revenues could fund these policies. hindered by poor diagnostic practices, related to long- Our analysis provides the lower bounds for the possi- standing problems in the Indian health care system— ble effects in three ways: mainly in the predominant private sector. A large proportion of the population chooses to use private • We focus only on mortality, excluding morbidity. sector providers, who deliver almost half of India’s • We do not consider externalities, except in the case tuberculosis services, many of which are of poor quality. of fossil fuel. Decreasing the price of accurate diagnostic technologies • We limit our analysis to health effects for older age (including removing existing import tariffs), particularly groups for many of our interventions because of the for the private sector, and giving private practitioners lack of health risk data for all age groups. incentives either to refer tuberculosis cases to the public sector for treatment or to improve their own treatment practices, would raise the overall quality of tuberculosis control. For example, provision of negotiated public DISCUSSION sector pricing for more accurate diagnostic tools, such as Health outcomes are determined by the complex interplay GeneXpert MTB/RIF (mycobacterium tuberculosis/ of social, economic, biological, and environmental factors, rifampicin) for India’s large private sector can increase which can be influenced through fiscal policies. Our demand for these tools. Our modeling suggests that report demonstrates that in times of fiscal exigency, taxa- replacing one million sputum smear tests annually with tion and subsidy reform for certain goods may deliver the accurate GeneXpert MTB/RIF test would decrease tremendous health gains while actually increasing govern- tuberculosis incidence by 26 (CI: 19–34) per 100,000 ment receipts. Even though challenged by large fiscal defi- people over 15 years and result in 5.4 (CI: 3.6–7.4) cits and insufficient outlays for health care, India has great million YLGs. Private health expenditures would scope to use complementary fiscal policies to improve decrease by US$105 (CI: -$84 –$284) million.3 both population health and fiscal health. The results of Complementary measures are as follows: the fiscal policy interventions modeled suggest that there are large potential health gains to be made from correcting • Enable reduced pricing for all accurate and approved market failures through tax and subsidy policies. diagnostic tools in the private sector. The gains in health are proportional to the changes in • Remove import duties on GeneXpert MTB/RIF. taxes or subsidies modeled, and the tax and subsidy Fiscal Instruments for Health in India 367 levels we chose to model were determined by a number outcomes of government policies, stakeholder engage- of factors, including the feasibility of uptake of the policy ment with relevant government departments and and the ability to administer and successfully enforce a affected populations will be crucial in the policy devel- tax or subsidy level. For tobacco, alcohol, and fossil fuels, opment process. theoretically, greater taxation could reduce health burdens by reducing exposure to the taxed product. However, tax officials may not have the resources to Limitations ensure the enforcement of a higher level of tax. For It is important to recognize the limitations of our example, half of alcohol consumption is currently illicit; models. First, the results rely heavily on a few central very large tax increases can further exacerbate this situa- parameters, such as relative risk and elasticity. We have tion, if greater resources are not devoted to ensure attempted to employ estimates that would be suitable for successful implementation of the tax and elimination of the Indian population; however, these estimates, partic- potential black market activity. ularly with respect to elasticity, are calculated for certain Other welfare effects need to be considered as well. populations in the past and may not be applicable to the In the short run, these may include reduced employ- populations in our study. Second, we only consider par- ment; the medium- and long-term effects may be on tial equilibrium effects of fiscal interventions and not the economic growth through a more productive labor force general equilibrium effects arising from the effect of or through effects on pension systems and health care these interventions on deficits, employment, growth, costs. In the case of fossil fuel taxation, the long-term and debt. Third, limitations in data do not allow us to effects may be on economic growth or the costs of goods. calculate health effects for all age groups, and we exclude Given the potential unintended consequences of our calculation of externality costs potentially leading to policies, it is critical that the complementary measures lower-bound estimates of health outcomes. Finally, our and the complete set of policy recommendations that consumption data for many interventions are based on accompany our tax and subsidy policy recommenda- household and individual surveys, which may not cap- tions be given as much importance as the tax or subsidy ture true consumption patterns, given the effects of recommendation itself. recall bias and underreporting. Tax and subsidy policies cannot be undertaken in isolation: they require complementary policies to realize the potential health and revenue gains that our model- CONCLUSIONS ing results suggest. Two themes that recur in the com- Although direct public health expenditures undoubt- plementary recommendations across the interventions edly play an integral role in determining population are (a) education and awareness and (b) monitoring health, health outcomes are determined by the complex and enforcement of taxes and regulations. Other com- interactions of social, economic, biological, and envi- plementary policies are more specific and focus on ronmental factors. A wide range of viable fiscal policy minimizing any potential adverse consequences of interventions could modify these proximate factors. policies—for example, by using the revenues or savings These are particularly useful when governments to invest in counseling and addiction services, alterna- find themselves unable to expand direct health care tive energy sources, and public transportation systems. expenditures. This chapter highlights that in times of These complementary measures involve revenue recy- fiscal exigency, reforming taxes and subsidies for certain cling into initiatives that may not be the purview of commodities may yield tremendous health gains while ministries of finance or excise departments. For exam- increasing government receipts. ple, the complementary policy recommendations for tobacco may involve the Ministry of Labour and Employment in retraining bidi workers or the Ministry NOTES of Education in conducting tobacco awareness cam- World Bank Income Classifications as of July 2014 are as paigns in schools. follows, based on estimates of gross national income (GNI) per A holistic view of health and its importance needs capita for 2013: to be adopted by all sectors of government. Subsidies by one department should not incentivize the use of • Low-income countries (LICs) = US$1,045 or less coal, for example, while another department pushes • Middle-income countries (MICs) are subdivided: for a coal levy. Coordination and communication will (a) lower-middle-income = US$1,046 to US$4,125 ensure that polices are consistent across departments. (b) upper-middle-income (UMICs) = US$4,126 to US$12,745 Given the complex and sometimes unanticipated • High-income countries (HICs) = US$12,746 or more. 368 Disease Control Priorities: Improving Health and Reducing Poverty 1. Recent literature on green tax swaps provides more insight Browning, E. K. 1999. “The Myth of Fiscal Externalities.” Public on this finding by decomposing two different links between Finance Review 27: 3–18. taxes on products or inputs and the broader fiscal system Central Board of Excise and Customs. 2017. “GST—Goods (for example, Bovenberg and Goulder [2002]; Parry and and Services Tax.” http://www.cbec.gov.in/htdocs-cbec/gst Oates [2000]). First is the efficiency gain from using new /central-tax-notfns-2017. revenue sources to reduce preexisting, distortionary taxes Chow, J., E. Y. Klein, E. Y., and R. Laxminarayan. 2010. “Cost- elsewhere in the economy. Second is a counteracting effect, Effectiveness of ‘Golden Mustard’ for Treating Vitamin A because of the impact of commodity taxes on driving up Deficiency in India.” PLoS ONE 5 (8): e12046. http://doi the general price level, thereby reducing real household .org/10.1371/journal.pone.0012046. wages and slightly reducing the overall level of labor Claro, R. M., R. B. Levy, B. M. Popkin, and C. A. Monteiro. supply. For the average good, the second effect dominates 2012. “Sugar-sweetened Beverage Taxes in Brazil.” American the former, so fiscal considerations warrant setting com- Journal of Public Health 102 (1): 178–83. http://doi modity taxes below (rather than above) marginal external .org/10.2105/AJPH.2011.300313. costs. However, the second effect is weaker, and possibly Cnossen, S. 2006. “Tobacco Taxation in the European Union.” reverses sign, when the commodity in question is a rela- FinanzArchiv/Public Finance Analysis 62 (2): 305–22. tively weak substitute (or complement) for leisure. Sgontz CSDH (Commission on Social Determinants in Health). 2008. (1993) discusses the efficiency gains from recycling alcohol Closing the Gap in a Generation: Health Equity through tax revenues in labor tax reductions. However, his partial Action on the Social Determinants of Health. Geneva: CSDH. equilibrium framework excludes impacts on labor supply Dreze, J., and R. Khera. 2015. “Understanding Leakages in the from the increase in price of alcohol relative to the price of Public Distribution System.” Economic & Political Weekly leisure. 1 (7): 39–42. 2. This assumes 50 percent of country liquor is shifted Grossman, D. 2012. J-PAL Policy Briefcase: Up in Smoke. from licit to illicit consumption, which has the same Cambridge, MA: Abdul Latif Jameel Poverty Action Lab. mortality risks as licit country liquor and does not get Gundimeda, H., and G. Köhlin. 2006. “Fuel Demand Elasticities taxed. for Energy and Environmental Policies: Indian Sample 3. This intervention assumes reduced public sector pricing for Survey Evidence.” EABER working paper 9/2006. East Asian GeneXpert for private firms, which the private sectors can Bureau of Economic Research, Canberra. http://www.eaber operate profitably. Health expenditure estimates assume .org/node/22501. access to reduced price GeneXpert for diagnosis and a shift Hanna, R., E. Duflo, and M. Greenstone. 2012. “Up in from private to public treatment. Smoke: The Influence of Household Behavior on the Long-Run Impact of Improved Cooking Stoves.” Faculty Research Working Paper Series 12–10, Department of Economics, Massachusetts Institute of Technology, REFERENCES Cambridge, MA. Andreyeva, T., M. W. Long, and K. D. Brownell. 2010. “The Initiative for Promoting Affordable Quality TB Tests (IPAQT). Impact of Food Prices on Consumption: A Systematic 2013. “Initiative for Promoting Affordable Quality TB Review of Research on the Price Elasticity of Demand for Tests,” Slide presentation, IPAQT, New Delhi. http://www Food.” American Journal of Public Health 100 (2): 216–222. .who.int/tb/careproviders/ppm/IPAQT.pdf. http://doi.org/10.2105/AJPH.2008.151415. IHME (Institute for Health Metrics and Evaluation). 2015. Barquera, S., L. Hernandez-Barrera, M. L. Tolentino, GBD Compare. IHME. http://vizhub.healthdata.org/gbd J. Espinosa, S. W. Ng, J. A. Rivera, and B. M. Popkin. -compare/patterns. 2008. “Energy Intake from Beverages Is Increasing among John, R. M. 2005. “Price Elasticity Estimates for Tobacco and Mexican Adolescents and Adults.” The Journal of Nutrition Other Addictive Goods in India.” Working Paper 117, 138 (12): 2454–61. http://doi.org/10.3945/jn.108.092163. eSocial Sciences. http://www.esocialsciences.org/Download Basu, S., S. Vellakkal, S. Agrawal, D. Stuckler, B. Popkin, and /repecDownload.aspx?fname=Document90782282005530 S. Ebrahim. 2014. “Averting Obesity and Type 2 Diabetes .3319818.pdf&fcategory=Articles&AId=117&fref=repec. in India through Sugar-Sweetened Beverage Taxation: ———. 2008. “Price Elasticity Estimates for Tobacco Products An Economic-Epidemiologic Modeling Study.” PLoS in India.” Health Policy and Planning 23 (3): 200–9. http:// Medicine 11 (1): e1001582. http://doi.org/10.1371/journal doi.org/10.1093/heapol/czn007. .pmed.1001582. Kojima, Y., J. L. Parcell, and J. S. Cain. 2014. “A Demand Model of Blecher, E., A. Liber, H. Ross, and J. Birckmayer. 2015. the Wholesale Vegetable Oils Market in the U.S.A.” Paper no. “Euromonitor Data on the Illicit Trade in Cigarettes.” 162472 presented at the Southern Agricultural Economics Tobacco Control 24 (1): 100–101. http://doi.org/10.1136 Association annual meeting, Dallas, TX, February 1–4. http:// /tobaccocontrol-2013-051034. econpapers.repec.org/paper/agssaea14/162472.htm. Bovenberg, A. L., and L. H. Goulder. 2002. “Environmental Kondo, M., S. Hoshi, and I. Okubo. 2009. “Does Subsidy Taxation and Regulation.” In Handbook of Public Economics Work? Price Elasticity of Demand for Influenza (first edition): Volume 3, edited by A. J. Auerbach and Vaccination among the Elderly in Japan.” Health Policy M. Feldstein, 1471–1545. Amsterdam: Elsevier. 91 (3): 269–76. Fiscal Instruments for Health in India 369 Kumar, P., A. Kumar, S. Parappurathu, and S. S. Raju. Parry, I. W. H., D. Heine, E. Lis, and S. Li. 2014. Getting Energy 2011. “Estimation of Demand Elasticity for Food Prices Right: From Principles to Practice. Washington, DC: Commodities in India.” Agricultural Economics Research International Monetary Fund. Review 24 (June): 1–14. http://www.indianjournals. Parry, I. W. H., R. Laxminarayan, and S. E. West. (2006) 2009. com/ijor.aspx?target=ijor:aerr&volume=24&issue “Fiscal and Externality Rationales for Alcohol Taxes.” =1&article=001. Discussion paper, Resources for the Future, Washington, Lim, K.-M., M. Kim, C. Kim, and S.-H. Yoo. 2012. “Short- DC. http://www.rff.org/files/sharepoint/WorkImages Run and Long-Run Elasticities of Diesel Demand in /Download/RFF-DP-06-51-REV.pdf. Korea.” Energies 5 (12): 5055–64. http://doi.org/10.3390 Parry, I. W. H., and W. E. Oates. 2000. “Policy Analysis in the /en5125055. Presence of Distorting Taxes.” Journal of Policy Analysis and Matheny, G. 2004. “Family Planning Programs: Getting Management 19 (4): 603–13. http://doi.org/10.1002/1520 the Most for the Money.” International Family Planning -6688(200023)19:4<603::AID-PAM5>3.0.CO;2-3. Perspectives 30 (3): 134. Sandmo, A. 1976. “Optimal Taxation.” Journal of Public Ministry of Finance, Government of India. 2016a. “Condom Economics 6: 37–54. Promotion Programme.” National Aids Control Sgontz, L. 1993. “Optimal Taxation: The Mix of Alcohol Organisation, Ministry of Health and Family Welfare, New and Other Taxes.” Public Finance Review 21 (3): 260–75. Dehli. http://naco.gov.in/condom-promotion-programme. Srivastava, D., and A. McGuire. 2014. “Analysis of Prices ———. 2016b. “Universal Immunization Programme.” News Paid by Low-Income Countries: How Price Sensitive Is release, Press Information Bureau, Ministry of Health and Government Demand for Medicines?” BMC Public Health Family Welfare, New Dehli, December. http://pib.nic.in 14 (1): 767. http://doi.org/10.1186/1471-2458-14-767. /newsite/PrintRelease.aspx?relid=154779. WHO (World Health Organization). 2015. WHO Report on the ———. 2016c. “ART Treatment.” National Aids Control Global Tobacco Epidemic, 2015: Raising Taxes on Tobacco, 103. Organisation, Ministry of Health and Family Welfare, New Geneva: WHO. Dehli. http://naco.gov.in/treatment. World Bank and IHME (Institute for Health Metrics and ———. 2016d. TB India 2016: Revised National TB Control Evaluation). 2016. The Cost of Air Pollution: Strengthening Programme Annual Status Report. Ministry of Health the Economic Case for Action. Washington, DC: World Bank. and Family Welfare, New Dehli. http://www.tbcindia.nic.in http://documents.worldbank.org/curated/en/7815214731 /showfile.php?lid=3180. 77013155/pdf/108141-REVISED-Cost-of-PollutionWebC ———. 2017. “Global Adult Tobacco Survey: GATS-2 India ORRECTEDfile.pdf. 2016–17 Highlights.” New Delhi: World Health Organization. World Bank. 2016. World Bank Development Indicators. http://www.searo.who.int/india/mediacentre/events/2017 Washington, DC: World Bank. http://data.worldbank.org /gats2_india.pdf?ua=1. /products/wdi. 370 Disease Control Priorities: Improving Health and Reducing Poverty DCP3 Series Acknowledgments Disease Control Priorities, third edition (DCP3) draws on discussion, and yearly in-person meetings. They have the global health knowledge of institutions and experts ensured quality and intellectual rigor at the highest from around the world, a task that required the efforts of order and have helped maximize the impact and useful- over 500 individuals, including volume editors, chapter ness of DCP3. The ACE members are listed separately in authors, peer reviewers, and research and staff assistants. this volume. The finalization of this series would not have been pos- The U.S. National Academy of Medicine, in collabo- sible without the intellectual vision, enduring support, ration with the InterAcademy Medical Panel, coordi- and invaluable contributions of these individuals. nated the peer review process for DCP3 chapters. Patrick We owe gratitude to the financial sponsor of this Kelley, Gillian Buckley, Megan Ginivan, Rachel Pittluck, effort: the Bill & Melinda Gates Foundation. The and Tara Mainero managed this effort and provided Foundation provided sole financial support of the critical and substantive input. Disease Control Priorities Network (DCPN), of which The World Bank provided exceptional guidance and DCP3 is a main product. Many thanks to Program support throughout the demanding production and Officers Kathy Cahill, Philip Setel, Carol Medlin, Damian design process. Within the World Bank, Carlos Rossel Walker, and (currently) David Wilson for their thought- and Mary Fisk oversaw the editing and publication of ful interactions, guidance, and encouragement over the the series and served as champions of DCP3. We also life of the project. We also thank program assistants thank Nancy Lammers, Rumit Pancholi, Deborah Karolyne Carloss and Christine VanderWerf at the Naylor, Elizabeth Forsyth, and Sherrie Brown for their Foundation for working tirelessly to organize and exe- diligence and expertise. Additionally, we thank Jose de cute several critical review meetings. Buerba, Mario Trubiano, Yulia Ivanova, and Chiamaka We are grateful to the University of Washington’s Osuagwu of the World Bank for providing professional Department of Global Health—and to successive chairs counsel on communications and marketing strategies. King Holmes and Judy Wasserheit—for creating a home We thank the many contractors and consultants who base for the DCP3 Secretariat, a base that provided provided support to specific volumes in the form of intellectual collaboration, logistical coordination, and economic analytical work, volume coordination, and administrative support. We thank those who worked chapter drafting: the Center for Disease Dynamics, behind the scenes within the department to ensure this Economics & Policy; Centre for Global Health Research; grant ran smoothly, including Athena Galdonez, Meghan Emory University; Evidence to Policy Initiative; Harvard Herman, Aimy Pham, and Ann Van Haney. T. H. Chan School of Public Health; Public Health We are tremendously appreciative of the wisdom and Foundation of India; QURE Healthcare; University of guidance provided by the 36 members of the DCP3 California, San Francisco; University of Waterloo; Advisory Committee to the Editors (ACE). Steered by University of Queensland; and the World Health Chair Anne Mills, the committee provided keen oversight Organization. and guidance on the scope and development of DCP3 We are grateful for the efforts of several institutions through chapter and document review, collaborative that contributed to the organization and execution of 371 key consultation meetings and conferences that were indispensable inputs into our cost and cost-effectiveness convened as part of the preparation of this series. These analyses. Stéphane Verguet added valuable guidance in institutions include the International Health Economics applying and improving the extended cost-effectiveness Association; National Cancer Institute; Pan American analysis method. Elizabeth Brouwer, Nazila Dabestani, Health Organization; University of California, Berkeley Shane Murphy, Zachary Olson, Jinyuan Qi, David A. School of Public Health; and the World Health Watkins, and Daphne Wu provided exceptional research Organization’s Eastern Mediterranean Regional Office. and analytic assistance, and often served as chapter Formulation of the main messages of this volume authors. Kristen Danforth provided crucial guidance on benefited from a Policy Forum convened in London, strategic organization and implementation. Brianne September 16, 2016, jointly by DCP and EMRO under Adderley served ably as Project Manager since the the leadership of Regional Director Emeritus Dr. Ala beginning. We owe her a very particular thanks. Jennifer Alwan. We are grateful to the participants in that Forum, Nguyen, Shamelle Richards, Jennifer Grasso, Sheri whose names are listed elsewhere in this volume. Sepanlou, and Tiffany Wilk contributed exceptional Finally, we thank the individuals who served as mem- project coordination support. The efforts of these indi- bers of the DCP3 Secretariat over the life of the project. viduals were absolutely critical to producing this series, In particular, we thank Carol Levin, who provided and we are thankful for their commitment. 372 DCP3 Series Acknowledgments Volume Editors Dean T. Jamison the Balsillie School of International Affairs there. Dean T. Jamison is Emeritus Professor in Global Health She has consulted for the World Bank, the Asian Sciences at the University of California, San Francisco, Development Bank, several United Nations agencies, and the University of Washington. He previously held and the International Development Research Centre, academic appointments at Harvard University and among others, in work conducted in over 20 low- and the University of California, Los Angeles. Prior to his middle-income countries. She led the work on nutrition academic career, he was an economist on the staff of the for the Copenhagen Consensus in 2008, when micronu- World Bank, where he was lead author of the World trients were ranked as the top development priority. She Bank’s World Development Report 1993: Investing in has served as associate provost of graduate studies at the Health. He serves as lead editor for DCP3 and was lead University of Waterloo, vice-president academic at editor for the previous two editions. He holds a PhD in Wilfrid Laurier University in Waterloo, and interim dean economics from Harvard University and is an elected at the University of Toronto at Scarborough. member of the Institute of Medicine of the U.S. National Academy of Sciences. He recently served as Co-Chair and Study Director of The Lancet’s Commission on Investing Prabhat Jha in Health. Prabhat Jha is the founding director of the Centre for Global Health Research at St. Michael’s Hospital. He holds Endowed and Canada Research Chairs in Global Hellen Gelband Health in the Dalla Lana School of Public Health at the Hellen Gelband is an independent global health policy University of Toronto. He is lead investigator of the expert. Her work spans infectious disease, particularly Million Death Study in India, which quantifies the cause malaria and antibiotic resistance, and noncommunica- of death and key risk factors in over two million homes ble disease policy, mainly in low- and middle-income over a 14-year period. He is also Scientific Director of the countries. She has conducted policy studies at Resources Statistical Alliance for Vital Events, which aims to expand for the Future, the Center for Disease Dynamics, reliable measurement of causes of death worldwide. His Economics & Policy, the (former) Congressional Office research includes the epidemiology and economics of of Technology Assessment, the Institute of Medicine of tobacco control worldwide. the U.S. National Academies, and a number of interna- tional organizations. Ramanan Laxminarayan Ramanan Laxminarayan is Director of the Center for Susan Horton Disease Dynamics, Economics & Policy in Washington, Susan Horton is Professor at the University of Waterloo DC. His research deals with the integration of epidemio- and holds the Centre for International Governance logical models of infectious diseases and drug resistance Innovation (CIGI) Chair in Global Health Economics in into the economic analysis of public health problems. 373 He was one of the key architects of the Affordable His main interests include the spectrum of injury control, Medicines Facility–malaria, a novel financing mecha- especially as it pertains to low- and middle-income coun- nism to improve access and delay resistance to antima- tries: surveillance, injury prevention, prehospital care, larial drugs. In 2012, he created the Immunization and hospital-based trauma care. He was President of the Technical Support Unit in India, which has been credited International Association for Trauma Surgery and with improving immunization coverage in the country. Intensive Care from 2013–15. He teaches at Princeton University. Rachel Nugent Charles N. Mock Rachel Nugent is Vice President for Global Charles N. Mock, MD, PhD, FACS, has training as both a Noncommunicable Diseases at RTI International. She trauma surgeon and an epidemiologist. He worked as a was formerly a Research Associate Professor and Principal surgeon in Ghana for four years, including at a rural Investigator of DCPN in the Department of Global hospital (Berekum) and at the Kwame Nkrumah Health at the University of Washington. Previously, she University of Science and Technology (Kumasi). In served as Deputy Director of Global Health at the Center 2005−07, he served as Director of the University of for Global Development, Director of Health and Washington’s Harborview Injury Prevention and Research Economics at the Population Reference Bureau, Program Center. He worked at the WHO headquarters in Geneva Director of Health and Economics Programs at the from 2007 to 2010, where he was responsible for develop- Fogarty International Center of the National Institutes of ing the WHO’s trauma care activities. In 2010, he returned Health, and senior economist at the Food and Agriculture to his position as Professor of Surgery (with joint appoint- Organization of the United Nations. From 1991–97, she ments as Professor of Epidemiology and Professor was associate professor and department chair in of Global Health) at the University of Washington. economics at Pacific Lutheran University. 374 Volume Editors Contributors Olusoji Adeyi Afsan Bhadelia World Bank, Washington, DC Harvard T. H. Chan School of Public Health, Boston, Massachusetts, United States Zipporah Ali Kenya Hospices and Palliative Care Association, Zulfiqar Bhutta Nairobi, Kenya Aga Khan University Hospital, Division of Women and Child Health, Karachi, Pakistan Silvia Allende Instituto Nacional de Cancerologia, Agnes Binagwaho Mexico City, Mexico Former Minister of Health, Kigali, Rwanda Ala Alwan David Bishai Department of Global Health, University of Washington, Johns Hopkins University Bloomberg School of Public Seattle, Washington, United States Health, Baltimore, Maryland, United States Shuchi Anand Robert E. Black Department of Medicine, Stanford University, Johns Hopkins University Bloomberg School of Public Palo Alto, California, United States Health, Baltimore, Maryland, United States Hector Arreola-Ornelas Mark Blecher Fundación Mexicana para la Salud, Mexico City, South Africa Treasury Department, Cape Town, Mexico South Africa Rifat Atun Barry R. Bloom Harvard T. H. Chan School of Public Health, Boston, Harvard T. H. Chan School of Public Health, Boston, Massachusetts, United States Massachusetts, United States Eran Bendavid Stanford University School of Medicine, Palo Alto, Edward Broughton California, United States USAID, Washington DC, United States Stefano Bertozzi Elizabeth Brouwer University of California, Berkeley, School of Public University of Washington, Seattle, Washignton, Health, Berkeley, California, United States United States Melanie Y. Bertram Donald A. P. Bundy World Health Organization, Geneva, Bill & Melinda Gates Foundation, London, Switzerland United Kingdom 375 Angela Y. Chang Kenneth A. Fleming Harvard T. H. Chan School of Public Health, Boston, Center for Global Health, National Cancer Institute, Massachusetts, United States Bethesda, Maryland, United States Dan Chisholm Mark Gallivan Department of Health System Financing, World Health Metabiota, San Francisco, California, United States Organization, Geneva, Switzerland Patricia Garcia Annie Chu Minister of Health, Lima, Peru Regional Office for the Western Pacific, World Health Organization, Manila, Philippines Atul Gawande Brigham and Women’s Hospital, Boston, Massachusetts, Alarcos Cieza United States World Health Organization, Geneva, Switzerland Thomas Gaziano Stephen Connor Harvard Medical School, Boston, Massachusetts, United National Palliative Care Research Center, Washington, States DC, United States Abdul Ghaffar Mark Cullen The Alliance for Health Policy and Systems Research, Stanford University School of Medicine, Palo Alto, Geneva, Switzerland California, United States Roger Glass Kristen Danforth Forgarty International Center, National Institutes of Department of Global Health, University of Health, Bethesda, Maryland, United States Washington, Seattle, Washington, United States Amanda Glassman Liliana De Lima Center for Global Development, Washington, DC, International Association for Hospice and Palliative United States Care, Houston, Texas, United States Eduardo González-Pier Nilanthi de Silva Center for Global Development, Washington, DC, University of Kelaniya, Colombo, Sri Lanka Untied States Haile T. Debas Glenda Gray Global Health Institute, University of California, San Chris Hani Baragwanath Hospital, Johannesburg, Francisco, San Francisco, California, United States South Africa Peter Donkor Brian Greenwood Kwame Nkrumah University of Science and London School of Hygiene & Tropical Medicine, Technology, Kumasi, Ghana London, United Kingdom Tarun Dua Liz Gwyther Department of Mental Health and Substance Abuse, University of Cape Town School of Public Health and World Health Organization, Geneva, Switzerland Family Medicine, Cape Town, South Africa Beverley Essue Demissie Habte University of Sydney Medical School, Sydney, Australia International Clinical Epidemiological Network, Addis Ababa, Ethiopia Victoria Y. Fan Myron B. Thompson School of Social Work, University Ednin Hamzah of Hawai’i at Ma¯ noa, Honolulu, Hawaii, United States Chief Executive, Hospice Malaysia,Kuala Lampur, Malaysia Xiagming Fang Jessica Ho Department of Applied Economics, China Agricultural Department of Information, Evidence, and Research, University, Beijing, China World Health Organization, Geneva, Switzerland John Flanigan Karen Hofman African Strategies for Advancing Pathology, Baltimore, University of Witwatersrand School of Public Health, Maryland, United States Johannesburg, South Africa 376 Contributors Dan Hogan Carol Levin Department of Health Statistics and Information University of Washington, Department of Global Systems, World Health Organization, Geneva, Health, Seattle, Washington, United States Switzerland Lai Meng Looi King K. Holmes University of Malaya, Kuala Lumpur, Malaysia Department of Global Health, University of Washington, Seattle, Washington, United States Emmanuel Luyirika African Palliative Care Association, Kampala, Uganda Guy Hutton WASH Section, United Nations Children’s Fund, Nita Madhav New York, New York, United States Metabiota, San Francisco, California, United States Stephen Jan Annet Mahanani The George Institute for Global Health, Department of Information, Evidence, and Research, Sydney, Australia World Health Organization, Geneva, Switzerland Quach Thanh Khanh Adel Mahmoud The Institute for Palliative Medicine, San Diego, Woodrow Wilson School of Public and International California, United States Affairs, Princeton University, Princeton, New Jersey, United States Felicia Knaul University of Miami, Institute for Elaine Marks Advanced Study of the Americas, Miami, Blindness and Deafness Prevention, Disability and Florida, United States Rehabilitation Unit, World Health Organization, Geneva, Switzerland Olive Kobusingye Makerere University Medical School, Colin Mathers Kampala, Uganda Department of Information, Evidence, and Research, World Health Organization, Geneva, Switzerland Eric L. Krakauer Massachusetts General Hospital, Boston, Massachusetts, Jean-Claude Mbanya United States Faculty of Medicine, University of Yaoundé I, Yaoundé, Cameroon Margaret E. Kruk Harvard T. H. Chan School of Public Health, Boston, Anthony R. Measham Massachusetts, United States World Bank (retired) Suresh Kumar Maria Elena Medina-Mora Institute of Palliative Medicine, Kerala, India National Institute of Psychiatry Ramón de la Fuente Muñiz, Mexico City, Mexico Modupe Kuti University of Ibadan, College of Medicine, Ibadan, Carol Medlin Nigeria Independent Consultant Xiaoxiao Kwete Oscar Mendez Harvard T. H. Chan School of Public Health, Boston, Fundación Mexicana para la Salud, Massachusetts, United States Mexico City, Mexico Tracey Laba University of Sydney Medical School, Anne Merriman Sydney, Australia Hospice Africa, Kampala, Uganda Peter Lachmann Anne Mills University of Cambridge, Cambridge, London School of Hygiene & Tropical Medicine, United Kingdom London, United Kingdom Nestor Lago Jody-Anne Mills University of Buenos Aires, Department of Pathology, World Health Organization, Geneva, Buenos Aires, Argentina Switzerland Contributors 377 Hoang Van Minh Nancy Padian Hanoi University of Public Health, Center for University of California, Berkeley School of Public Population Health Sciences, Hanoi, Vietnam Health, Berkeley, California, United States Jaime Montoya Vikram Patel Philippine Council for Health Research and London School of Hygiene & Tropical Medicine, Development, Taguig City, the Philippines London, United Kingdom Egide Mpanumusingo George C. Patton Butaro Hospital, Burera, Rwanda Murdoch Children’s Research Institute, Melbourne, Australia Prime Mulembakani Metabiota, Democratic Republic of the Congo John Peabody QURE Healthcare, San Francisco, California, Mahendra Naidoo United States National Cancer Institute Center for Global Health, Bethesda, Maryland, United States Pedro Perez-Cruz Departamento de Medicina Interna, Facultad de Diana Nevzrova Medicina, Pontificia Universidad Católica de Chile, Russian Health Ministry, Moscow, Russia Santiago, Chile Liu Peilong Dorairaj Prabhakaran Global Health Department, Peking University, Public Health Foundation of India, New Delhi, India Beijing, China Christopher Price Thi Kim Phoung Nguyen Tallahassee Memorial HealthCare, Tallahassee, Florida, World Health Organization, Hanoi, Vietnam United States Ole Frithjof Norheim Jinyuan Qi University of Bergen Department of Global Public Office of Population Research, Princeton University, Health and Primary Care, Bergen, Norway Princeton, New Jersey, United States Christian Ntizimira Lukas Radbruch Kibagabaga Hospital, Kigali, Rwanda Malteser International, Cologne, Germany Osondu Ogbuoji M. R. Rajagopal Harvard T. H. Chan School of Public Health, Boston, International Association for Hospice and Palliative Massachusetts, United States Care, New Delhi, India Zachary Olson Teri A. Reynolds University of California, Berkeley, School of Public World Health Organization, Geneva, Switzerland Health, Berkeley, California, United States Natalia Rodriguez Folashade Omokhodion Research Support for Global Health, University of University College Hospital, Ibadan, Nigeria Miami, Miami, Florida, United States Ben Oppenheim John-Arne Rottingen Metabiota, San Francisco, California, United States The Research Council of Norway, Oslo, Norway Toby Ord Kun Ru University of Oxford, Oxford, United Kingdom Allegheny General Hospital, Pittsburgh, Pennsylvania, United States Hibah Osman Lebanese Center for Palliative Care, Beirut, Lebanon Sevket Ruacan Koc University School of Medicine, Istanbul, Turkey Trygve Ottersen University of Oslo, Department of Community Andres Rubiano Medicine and Global Health, Oslo, Norway Ashoka Colombia, Neiva, Colombia 378 Contributors Edward Rubin Mark Stoltenberg Metabiota, San Francisco, California, United States Division of Palliative Care and Geriatric Medicine, Massachusetts General Hospital, Massachusetts, Rengaswamy Sankaranarayanan United States International Agency for Research on Cancer, Lyon, France Amit Summan Center for Disease Dynamics, Economics, & Policy, Hendry Sawe New Delhi, India Muhimbili University of Health and Allied Sciences, Emergency Medicine Department, Dar es Salaam, Lawrence H. Summers Tanzania Harvard Kennedy School, Cambridge, Massachusetts, United States Helen Saxenian World Bank (retired), Washington, DC, United States Neo Tapela Brigham and Women’s Hospital, Boston, Massachusetts, Marco Schäferhoff United States SEEK Development, Berlin, Germany Marleen Temmerman Jaime Sepulveda Aga Khan University East Africa, Nairobi, Kenya Global Health Sciences, University of California, San Francisco, San Francisco, California, Stephen Tollman United States University of Witwatersrand, Johannesburg, South Africa Melissa Sherry Johns Hopkins Bloomberg School of Public Health, Stéphane Verguet Baltimore, Maryland, United States Harvard T. H. Chan School of Public Health, Boston, Massachusetts, United States Riti Shimkhada Center for Health Policy Research, University of Damian Walker California Los Angeles, Los Angeles, California, Bill & Melinda Gates Foundation, Seattle, Washington, United States United States Sang Do Shin Neff Walker Seoul National University College of Medicine, Bloomberg School of Public Health, Johns Hopkins Department of Emergency Medicine, Seoul, University, Baltimore, Maryland, United States Republic of Korea Huihui Wang Richard Skolnik Senior Economist, World Bank Group, Washington DC, Retired, Albuquerque, New Mexico, United States United States Kirk R. Smith Jianxiang Wang University of California, Berkeley, School of Public Institute for Hematology, Chinese Academy of Medical Health, Berkeley, California, United States Sciences, Beijing, China Agnes Soucat Lee Wallis World Health Organization, Geneva, Switzerland University of Cape Town, Cape Town, South Africa Dingle Spence David A. Watkins Hope Institute Hospital, Kingston, Jamaica University of Washington School of Medicine, Seattle, Washington, United States Nicholas Stacey University of the Witwatersrand, Johannesburg, David Wilson South Africa Bill & Melinda Gates Foundation, Seattle, Washington, United States Gretchen Stevens Department of Health Statistics and Information Michael Wilson Services, World Health Organization, Geneva, University of Colorado, Department of Pathology, Switzerland Aurora, Colorado, United States Contributors 379 Nathan Wolfe Gavin Yamey Metabiota, San Francisco, California, Duke Global Health Institute, Durham, North Carolina, United States United States Yangfeng Wu Kun Zhao The George Institute, Beijing, China National Health Development Research Center, China Beijing, China 380 Contributors Advisory Committee to the Editors Anne Mills, Chair Amanda Glassman Professor, London School of Hygiene & Tropical Chief Operating Officer and Senior Fellow, Center for Medicine, London, United Kingdom Global Development, Washington, DC, United States Olusoji Adeyi Glenda Gray Director, Health, Nutrition and Population Global Executive Director, Perinatal HIV Research Unit, Practice, World Bank, Washington, DC, United States Chris Hani Baragwanath Hospital, Johannesburg, Kesetebirhan Admasu South Africa Former Minister of Health, Addis Ababa, Ethiopia Demissie Habte George Alleyne Chair of Board of Trustees, International Clinical Director Emeritus, Pan American Health Organization, Epidemiological Network, Addis Ababa, Ethiopia Washington, DC, United States Ala Alwan Richard Horton Regional Director Emeritus, World Health Organization, Editor, The Lancet, London, United Kingdom Regional Office for the Eastern Mediterranean, Cairo, Edward Kirumira Arab Republic of Egypt Dean, Faculty of Social Sciences, Makerere University, Rifat Atun Kampala, Uganda Professor, Global Health Systems, Harvard T. H. Chan School of Public Health, Boston, Massachusetts, United Peter Lachmann States Professor, University of Cambridge, Cambridge, United Kingdom Zulfiqar Bhutta Chair, Division of Women and Child Health, Aga Khan Lai Meng Looi University Hospital, Karachi, Pakistan Professor, University of Malaya, Kuala Lumpur, Agnes Binagwaho Malaysia Former Minister of Health, Kigali, Rwanda Adel Mahmoud Mark Blecher Senior Molecular Biologist, Princeton University, Senior Health Advisor, South Africa Treasury Princeton, New Jersey, United States Department, Cape Town, South Africa Patricia Garcia Anthony Measham Minister of Health, Lima, Peru World Bank (retired) Roger Glass Carol Medlin Director, Fogarty International Center, National Independent Consultant, Washington, DC, Institutes of Health, Bethesda, Maryland, United States United States 381 Alvaro Moncayo Richard Skolnik Researcher, Universidad de los Andes, Bogotá, Colombia Yale University School of Public Health (retired) Jaime Montoya Stephen Tollman Executive Director, Philippine Council for Health Professor, University of Witwatersrand, Johannesburg, Research and Development, Taguig City, the Philippines South Africa Ole Norheim Professor, Department of Global Public Health and Jürgen Unutzer Primary Care, University of Bergen, Bergen, Norway Professor, Department of Psychiatry, University of Washington, Seattle, Washington, United States Folashade Omokhodion Professor, University College Hospital, Ibadan, Nigeria Damian Walker Deputy Director of Data and Analytics, Bill & Melinda Toby Ord Gates Foundation, Seattle, Washington, United States President, Giving What We Can, Oxford, United Kingdom Ngaire Woods K. Srinath Reddy Director, Global Economic Governance Program, President, Public Health Foundation of India, Oxford University, Oxford, United Kingdom New Delhi, India Sevket Ruacan Nopadol Wora-Urai Dean, Koç University School of Medicine, Istanbul, Turkey Professor, Department of Surgery, Phramongkutklao Hospital, Bangkok, Thailand Jaime Sepúlveda Executive Director, Global Health Sciences, University Kun Zhao of California, San Francisco, San Francisco, California, Researcher, China National Health Development United States Research Center, Beijing, China 382 Advisory Committee to the Editors Reviewers Jishnu Das Zachary Olson Development Research Group, World Bank, Berkeley School of Public Health, University of New Delhi, India California, Berkeley, Berkeley, California, United States Joseph L. Dieleman Institute for Health Metrics and Evaluation, Seattle, Gayatri Palat Washington, United States MNJ Institute of Oncology and Regional Cancer Center, Hyderbad, India Bernard Franck World Education, Inc., Vientiane, Lao People’s Anna Vassall Democratic Republic London School of Hygiene and Tropical Medicine, London, United Kingdom Heike Geduld African Federation for Emergency Medicine, Ron Waldman Cape Town, South Africa Milken Institute School of Public Health, George Washington University, Washington, DC, Peter Heller United States International Monetary Fund, Washington, DC, United States Peter Neumann Center for the Evaluation of Value and Risk in Health at the Institute for Clinical Research and Health Policy Studies, Tufts Medical Center, Boston, Massachusetts, United States 383 Policy Forum Participants The following individuals provided valuable insights to improve this volume’s key findings through participation in the Disease Control Priorities–World Health Organization, Regional Office for the Eastern Mediterranean policy forum on developing a universal health coverage package of high priority interventions. The forum was held in London, United Kingdom, on September 16, 2016, and was organized by Dr. Ala Alwan, Regional Director and member of the DCP3 Advisory Committee to the Editors. Shaikh Dr Mohamed Bin Abdulla Al Khalifa Ferozuddin Feroz Chairman, Supreme Council of Health, Minister of Public Health, Ministry of Public Health, Manama, Bahrain Kabul, Afghanistan Ahmed Mohamed Al-Saidi Assad Hafeez Minister of Health, Muscat, Oman Director General of Health, Ministry of National Health Services, Mohammed Al-Thani Regulations and Coordination, Director, Department of Public Health, Ministry of Islamabad, Pakistan Public Health, Doha, Qatar Ziad Memish Fawzi Amin Professor, College of Medicine, Alfaisal University, Secretary General, Bahrain Red Crescent Society; Riyadh, Saudi Arabia Formerly Director General of Primary Health Care, Ministry of Health, Manama, Bahrain Keshav Desiraju Walid Ammar Former Secretary of Health, Government of India, Director General, Ministry of Public Health, New Delhi, India Beirut, Lebanon Raj Shankar Ghosh Aysha Mubarak Buaneq Deputy Director, Vaccine Delivery and Infectious Undersecretary, Ministry of Health, Manama, Bahrain Disease, Bill & Melinda Gates Foundation, New Delhi, India Rassoul Dinarvand Deputy Minister of Health and President, Iran Food Soonman Kwon and Drug Administration, Ministry of Health, Tehran, Chief, Health Sector Group, Asian Development Bank, Islamic Republic of Iran Manila, the Philippines 385 Index Boxes, figures, maps, notes, and tables are indicated by b, f, m, n, and t following the page number. A mortality rates and, 74, 96–97 Abdul Latif Jameel Poverty Action pandemics and, 326 Lab, 277–78 AIDS. See HIV/AIDS access to health care alcohol emergency care, 247 benefit-cost analysis for interventions, 171 insurance and, 8 fiscal policy and, 360–65, 367–68 mortality rates and, 115 intersectoral policies on, 24, 26, 28, 31, 36 pandemics and, 354 taxation of, 12, 23, 364t, 365 quality of care, 199 Ali, Zipporah, 235 rehabilitation services, 286–87 Allende, Silvia, 235 universal health coverage and, 43–44, 59 Alwan, Ala, 3, 23, 43 accountability Anand, Shuchi, 3 community health platforms and, 272, 278 angina, 126, 133 intersectoral policies and, 37 antibiotics pandemics and, 338 community health platforms and, 268 pathology services and, 226–27 emergency care and, 249, 251 quality of care and, 187, 191, 194–95, pandemics and, 318, 329–31 201, 203–4 quality of care and, 185, 198, 201 acute respiratory illness (ARI), 323 antiviral drugs, 318, 329–31, 334, 337–38 addictive substances, 24–25, 28, 31. See also Argentina alcohol; tobacco use burden of disease in, 133–35 Adeyi, Olusoji, 3, 185 quality of care in, 194 Afghanistan, quality of care in, 194 ARI (acute respiratory illness), 323 Africa. See also specific countries Arreola-Ornelas, Hector, 235 community health platforms in, 267 Atun, Rifat, 3, 43 cost-effectiveness analysis in, 149–50 Australia development assistance in, 307 burden of disease in, 121, 127–28 emergency care in, 251, 262 cost-effectiveness analysis in, 147 mortality rates in, 98, 100–101, 111 development assistance from, 302, 302f pandemics in, 318 life expectancy in, 6 pathology services in, 227 pandemics in, 317, 335 aging population pathology services in, 223, 226–27, 229 benefit-cost analysis and, 181 avian influenza, 315 intersectoral policy priorities and, 31–32 Azerbaijan, mortality rates in, 114 387 B pathology services and, 221 Bangladesh quality of care, 202 burden of disease in, 133–34 Broughton, Edward, 185 fiscal policy in, 362 Brouwer, Elizabeth, 3 pandemics in, 329 Bundy, Donald A. P., 3 pathology services in, 228 burden of disease, 121–43 BCA. See benefit-cost analysis cancer, 122, 126–28, 130, 132–33, 135 Bell, D., 204 cardiovascular disease, 122, 126–27, 130–35 Beltran-Sanchez, H., 98 catastrophic and impoverishing health Bendavid, Eran, 299 expenditures, 122–24, 126, 129–30 benefit-cost analysis (BCA), 10, 167–81, 169–73t defined, 122–23, 122f cancer interventions, 179 future research needs, 138 child health and mortality interventions, indicators, 123, 123–24t 167, 179, 181 measures of, 130–33, 131–35f communicable disease interventions, 168, 172–73 policy recommendations, 137–38 drugs, 171, 173 population-level estimates, 122–26 essential packages of care, 179–80 prevalence, 126–28t, 126–30, 127m, extended cost-effectiveness analysis and, 157 129–30f, 130t life expectancy and, 169, 171–75, 177, 181 relationship between, 123b malaria interventions, 169, 172, 179 universal health coverage and, 124–25b morbidity and, 168–69, 174, 178 chronic diseases and chronic ill health, neglected tropical disease interventions, 169, 171 121–33, 135, 136–39 pandemic interventions, 168, 171 chronic obstructive pulmonary policy interventions, 168, 178, 181 disease (COPD), 126, 128 role of, 167–68 communicable diseases, 126 tuberculosis interventions, 172–73 consequences of illness or injury, 32 universal health coverage, 47 cost-effectiveness of interventions and, 3 vaccines and vaccinations, 173 development assistance and, 305, 308 value per statistical life (VSL) in, distressed financing and, 133, 135f 174–77, 175–78f, 176t families and, 133–35 Bertozzi, Stefano, 3 financial risk protection (FRP) and, 121, 136–37 Bhadelia, Afsan, 235 heart disease, 126, 128 Bhutta, Zulfiqar, 3 HIV/AIDS, 122, 125–26, 128, 134–35 Bill & Melinda Gates Foundation, 178–79, 300–301 influenza pandemic, 322–23, 325, 326, 347–48, Binagwaho, Agnes, 3 352–53, 353t birth control. See family planning kidney diseases, 126, 128 Bishai, David, 267 malaria, 126, 128, 134 Black, Robert, 3 mental illnesses, 126, 128–29 Blecher, Mark, 3 pandemics, 321–23, 322f, 322t, 325, 326, 347–48, Bloom, Barry R., 3 352–53, 353t body mass index (BMI) productivity changes, 132–33, 134f fiscal policy and, 366 treatment discontinuation and, 133 intersectoral policies on, 24 tuberculosis, 126, 134 Brazil universal health coverage and, 46, 49, 121–25, 137 development assistance in, 301 Burkina Faso, mortality rates in, 101, 114 fiscal policy in, 359, 363 Burundi, quality of care in, 194 pandemics in, 335–36 pathology services in, 227 C rehabilitation services in, 287–88 Cambodia breast cancer emergency care in, 250 cost-effectiveness analysis for mortality rates in, 109, 115 interventions, 150–51 pandemics in, 336 mortality rates, 96 pathology services in, 218 388 Index Canada CEA. See cost-effectiveness analysis community health platforms in, 267 CEPI (Coalition for Epidemic Preparedness development assistance from, 302, 302f Innovations), 330, 335 pandemics in, 317–18, 335 Chang, Angela Y., 167 pathology services in, 227 child health and mortality cancer. See also specific types benefit-cost analysis for interventions, benefit-cost analysis for interventions, 179 167, 179, 181 burden of disease in, 122, 126–28, 130, community health platforms and, 273, 275 132–33, 135 cost-effectiveness analysis for cost-effectiveness analysis for interventions, interventions, 149, 153 147, 150–51 development assistance and, 300, 302, development assistance and, 302 304, 305, 307–8 fiscal policy and, 360–61 intersectoral policies on, 27 intersectoral policies on, 25 trends, 69–70, 72, 99–101, 106–7, 109–10, 110t, mortality rates, 70–72, 74, 96, 98, 100 114–17, 117f palliative care and pain control, 240 universal health coverage and, 47, 48, 49, pathology services and, 215–16, 221, 225 51, 53, 55, 56 quality of care, 196–97, 202 Chile universal health coverage and, 51, 53 life expectancy in, 6 Canudas-Romo, V., 98 mortality rates in, 106 capacity China burden of disease and, 121, 123, 126, 129 benefit-cost analysis in, 169, 179–80 community health platforms and, 279–80 burden of disease in, 125, 127, 132–35 development assistance and, 304 development assistance from, 301, 304–5, 307 emergency care and, 252, 254–56, 261 emergency care in, 251 extended cost-effectiveness analysis and, 159 extended cost-effectiveness analysis in, 162 pandemics and, 316–17, 320, 330–31, 338 fiscal policy in, 359 pathology services and, 217, 219–20 household coal use in, 29b rehabilitation services and, 286–87 influenza pandemic in, 355 universal health coverage, 12, 14, 46, 58, 60, 62 mortality rates in, 72, 116–17 cardiovascular disease noncommunicable diseases in, 3 burden of disease, 122, 126–27, 130–35 pandemics in, 318–21, 325, 329, 335–36, 355 cost-effectiveness analysis for pathology services in, 218, 226–28 interventions, 149–52 quality of care in, 190 development assistance and, 302–3 tobacco taxation in, 9–10 extended cost-effectiveness analysis for Chisholm, Dan, 3 interventions, 162 Chongsuvivatwong, V., 132 mortality rates, 74, 98 chronic diseases and chronic ill health. See also palliative care and pain control, 236–37 specific diseases pathology services and, 215 benefit-cost analysis for interventions, 180 quality of care, 193 burden of disease, 121–33, 135, 136–39 catastrophic and impoverishing health expenditures quality of care, 190, 204 burden of disease, 122–24, 126, 129–30 chronic obstructive pulmonary disease (COPD) defined, 122–23, 122f burden of disease, 126, 128 future research needs, 138 emergency care, 261 indicators, 123, 123–24t fiscal policy and, 361 measures of, 130–33, 131–35f intersectoral policies on, 25 policy recommendations, 137–38 mortality rates, 96–98 population-level estimates, 122–26 palliative care and pain control, 236 prevalence, 126–28t, 126–30, 127m, 129–30f, 130t Chu, Annie, 121 relationship between, 123b Cieza, Alarcos, 3, 285 universal health coverage and, 124–25b cigarettes. See tobacco use CCTs. See conditional cash transfer Cleary, James F., 235 Index 389 climate change, 25, 326, 353–54 intersectoral policies, 27, 36 coal, bans on household use, 29b intervention packages, 16–17t, 16–18 cognitive interventions, 290, 292 palliative care and pain control, 236, 241–43, 242t Colombia pandemic interventions, 316, 333–38, emergency care in, 251 334–37f, 353, 355 rehabilitation services in, 288 pathology services, 225, 228–29 communicable diseases. See also pandemics; quality of care and, 186–87, 189, 200–201 specific diseases rehabilitation services, 287–88 benefit-cost analysis for interventions, 168, 172–73 universal health coverage and, 12–14, 13–14t, 43, 45, burden of disease, 126 47–55, 51–55t, 59 cost-effectiveness analysis for interventions, 151 cost-effectiveness analysis (CEA). See also extended development assistance and, 302, 304–5, 307–8 cost-effectiveness analysis emergency care, 252 analysis results, 148–52, 149–52f fiscal policy and, 360–62 benefit-cost analysis and, 161–68, 176, 178, 181 mortality rates, 69, 73–74, 98–99, 101 breast cancer interventions, 150–51 pathology services and, 215, 218, 221 burden of disease and, 3, 137 universal health coverage and, 48, 51, 53, 56 cancer interventions, 147, 150–51 community health platforms, 267–84 cardiovascular disease interventions, 149–52 accountability and, 272, 278 child health and mortality interventions, 149, 153 case studies, 271–79 communicable disease interventions, 151 Haiti, 278–79 development assistance and, 305, 308–9 Indonesia, 273–75, 275f drugs, 150–51, 154 Peru, 276–77 emergency care, 149, 151, 153, 262 Uganda, 277–78 family planning, 154 continuum of functioning for, 272, 272–73t heart failure interventions, 149–51 functions, 269 HIV/AIDS interventions, 149–51, 154 health care facilities and, 269–70 hospitals, 149, 152, 153 historical context, 268–69 influenza pandemic interventions, 335–37 measurement of success in, 270–71, 270–71t intersectoral policies, 35, 36, 37 strengthening, 279–80 malaria interventions, 149–51, 153 conditional cash transfers (CCTs) measurement issues, 154 extended cost-effectiveness analysis and, 160, 162 methodology, 7–8, 147–48, 153 quality of care and, 192–93 neglected tropical disease interventions, universal health coverage and, 59 149, 151, 153 Congo. See Democratic Republic of Congo new technologies, 153 Connor, Stephen, 235 palliative care and pain control, 10–11 contraception. See family planning pandemic interventions, 334–37f, 334–38 cooking fuel subsidies, 366–67 pathology services, 223 COPD. See chronic obstructive pulmonary disease prices and, 153 Copenhagen Consensus, 6 primary care, 150–51 cost. See also benefit-cost analysis; cost-effectiveness quality of care and, 205 analysis; extended cost-effectiveness analysis rehabilitation services, 287–88 benefit-cost analysis and, 167–70, 178–81 syphilis interventions, 149–51 burden of disease and, 9–14, 121–24, tuberculosis interventions, 149–51, 153 128, 130, 133–36 universal health coverage, 44–45, 47–48, community health platforms, 280 58, 60, 147 consequences of illness or injury, 32 vaccines and vaccinations, 148–51, 152, 153–54 development assistance and, 305, 308–10 Creditor Reporting System (CRS), 300–301 emergency care, 250 Cullen, Mark, 3 essential packages of care, 49–55, 51–55t extended cost-effectiveness analysis, 158, 160 D fiscal policy and, 368 DAC (Development Assistance Committee), 300–302 influenza pandemic, 353, 355 DALYs (disability-adjusted life years) 390 Index benefit-cost analysis and, 167, 172–73, 178 palliative care and pain control, 239–40, 242–43 burden of disease and, 126 pandemics and, 324, 330, 332, 334 community health platforms and, 280 pathology services and, 225 cost-effectiveness analysis and, 147–54 quality of care and, 186, 198, 201 development assistance and, 309 universal health coverage and, 50, 59, 60 emergency care and, 250–51, 262 Dua, Tarun, 3 extended cost-effectiveness analysis and, 158 quality of care and, 200, 205 E universal health coverage and, 47–48 Ebola virus, 61, 72, 196, 269, 304, 307–10 value of, 6 ECEA. See extended cost-effectiveness analysis Danforth, Kristen, 3, 23, 43, 235 economic evaluation, 8–11, 10t. See also death registration data, 70, 72, 99–100. See also cost-effectiveness analysis; extended mortality rates cost-effectiveness analysis Debas, Haile T., 3 ECSA (Emergency Care System Assessment), De Lima, Liliana, 235 253–54, 261 delivery platforms. See platforms ECSs. See emergency care systems dementias Egypt intersectoral policies and, 31 development assistance in, 301 mortality rates, 74 intersectoral policies in, 37 palliative care and pain control, 236–37, 239 emergency care, 247–65 quality of care, 196 access to, 247, 248, 249f Democratic Republic of Congo chronic obstructive pulmonary disease (COPD), 261 benefit-cost analysis in, 169 communicable diseases, 252 pandemics in, 324 cost-effectiveness, 149, 151, 153, 262 demographics. See aging population defined, 247–49, 248f Denmark, rehabilitation services in, 288 essential package, 254–61, 255–61t De Silva, Nilanthi, 3 heart failure, 250 development assistance, 299–314 HIV/AIDS, 252, 256, 261 allocation criteria, 304–5 ischemic heart disease, 251 burden of disease and, 305, 308 kidney diseases, 261 case studies, 305–7 malaria, 251–52, 258 China, 307 maternal mortality and, 250–51 PEPFAR, 305–7 morbidity and, 247, 250 changing commitments, 308 neglected tropical diseases (NTDs), 252 effectiveness evaluations, 307–8 obstetric care, 149, 151, 153, 251, 262 goals for, 304 pneumonia, 249–50, 252, 254, 261–62 investment opportunity identification, 309 policy priorities, 262 tracking and monitoring, 300–302, 302f, 303b quality of care and, 204 Development Assistance Committee (DAC), 300–302 study rationale, 249–52, 251f diabetes, mortality rates, 74 Sustainable Development Goals and, 252, 252t diagnosis. See pathology services tuberculosis, 252 diarrheal diseases WHO framework for, 252–54, 254f emergency care, 251 Emergency Care System Assessment (ECSA), mortality rates, 96–98 253–54, 261 distressed financing, 133, 135f emergency care systems (ECSs), 247, Donkor, Peter, 3 252–53, 255, 261–62 Doyle, C., 204 endocrine diseases, 122, 126–32 drugs EPHFs (Essential Public Health Functions), 269 antiviral, 318, 329–31, 334, 337–38 epidemics. See pandemics benefit-cost analysis, 171, 173 EQA (external quality assurance), 226–27 cost-effectiveness analysis, 150–51, 154 essential medicines. See also drugs essential medicines, 59, 60, 239–40, 242–43 palliative care and pain control, 239–40, 242–43 mortality rates and, 100 universal health coverage and, 59, 60 Index 391 essential packages of care, 8t, 43–65 extended cost-effectiveness analysis (ECEA), barriers to intervention uptake, 58 8–10, 157–66 benefit-cost analysis and, 179–80 benefit-cost analysis and, 181 burden of disease and, 122 cardiovascular disease interventions, 162 construction of, 11 conditional cash transfers, 160, 162 costs, 49–55, 51–55t dashboard, 160–61 defined, 7 in DCP series, 161–62t development of, 45 “efficient purchase” of benefits, emergency care, 254–61, 255–61t 159–60, 159–60f financial risk protection, 48 equity benefits, 159 financing, 60 financial risk protection and, 157–60 governance and, 59–60 health benefits, 158, 159f health outcomes and, 56, 57t HIV/AIDS interventions, 157 health workforce and, 60 intersectoral policies, 35, 37 identification of highest-priority package, 45–49 methodology, 158–60 implementation, 56–62, 59t, 61b morbidity and, 158 inclusion criteria, 48–49 nonhealth benefits, 158–59, 159f information and research needs, 60–61, 61b policy decisions and, 162–63, 163f intersectoral policies, 27, 37 policy interventions, 157–61 leadership and, 59–60 poverty reduction benefits, 161, 161f, 163 medical product and technology availability, 60 tuberculosis interventions, 158, 160–62 mortality rates and, 15–16 universal health coverage, 44, 158 palliative care and pain control, 235–36, vaccines and vaccinations, 158 238–43, 239t external quality assurance (EQA), 226–27 pandemics, 316 pathology services, 215, 217, 219–32, F 220–21t, 220b families platforms for, 46b burden of disease and, 133–35 priority-setting institutions, role of, 61–62 national health accounts and, 33b quality improvement, 58–59 palliative care and, 236, 238, 240–43 rehabilitation services, 288–91 pathology services and, 215, 228 value for money, 47 family planning Essential Public Health Functions (EPHFs), 269 community health platforms and, 273–75 essential universal health coverage (EUHC), 11–17, cost-effectiveness analysis, 154 44–45, 54–60. See also universal pathology services and, 219 health coverage quality of care and, 195, 199 Essue, Beverley M., 121 universal health coverage and, 46–47 Ethiopia Fan, Victoria Y., 347 extended cost-effectiveness analysis in, financial risk protection (FRP), 8–11 157, 159–62 burden of disease and, 121, 136–37 mortality rates in, 109, 116–18 essential packages of care, 48 pathology services in, 221 extended cost-effectiveness analysis and, 157–60 quality of care in, 194 intersectoral policies and, 37 universal health coverage in, 55 palliative care and pain control, 236, 243 EUHC. See essential universal health coverage pathology services and, 219 Europe and Central Asia. See also specific countries quality of care and, 204 community health platforms in, 267 universal health coverage and, 43–44, 47–48 development assistance in, 305 financing mortality rates in, 98, 107–9, 114 development assistance and, 302, 304, 309 pandemics in, 317, 325 distressed, 133, 135f pathology services in, 223 essential packages of care, 60 quality of care in, 190 pathology services, 219 evidence-based policy decisions, 8, 9b first-level hospitals. See hospitals 392 Index fiscal policy, 359–68. See also financial risk protection Global Health Estimates (GHE), 26, 57, 69–73, analytic framework, 361–62, 362–63t 100–101, 236 role of, 360–61, 360–61t Global Health Observatory (GHO), 26 subsidies, 364t, 366–67. See also subsidies Global Health 2035 (Jamison et al.), 15 taxation, 364–66, 364t. See also taxation González-Pier, Eduardo, 23 Fitoussi, J. P., 8 governance, 59–60 Flanigan, John, 215 grand convergence (GC), 12, 48–49, 55, 106, 300 Fleming, Kenneth A., 3, 215 Gray, Glenda, 3 fossil fuels, taxation of, 364t, 365–66 Guatemala Framework Convention on Tobacco Control, 37 pandemics in, 336 France quality of care in, 191 development assistance from, 302, 302f Guinea pandemics in, 334 mortality rates in, 114 pathology services in, 227 pandemics in, 318, 323–25, 335, 347 FRP. See financial risk protection Gwyther, Liz, 235 G H Gallivan, Mark, 3, 315 Habte, Demissie, 3 Garcia, Patricia, 3 Haiti Gavi—the Vaccine Alliance, 148–50 community health platforms in, 278–79 Gawande, Atul, 3 fiscal policy in, 362 Gaziano, Thomas, 3 mortality rates in, 109 GBD. See Global Burden of Disease HALE (Healthy Life Expectancy), 31, 305 GC. See grand convergence Halstead, S. B., 105 Geater, A., 132 Hammitt, J. K., 175 Gelband, Hellen, 3 Hamzah, Ednin, 235 gender. See women Haninger, K., 175 Gerbangmas movement (Indonesia), Hanson, K., 167 273–75, 275f head injuries, 34–35b Germany health care facilities. See also hospitals benefit-cost analysis in, 171 burden of disease and, 127–28, 136 development assistance from, 302, 302f emergency care, 247 palliative care and pain control in, 236 quality of care and, 185, 191, 201 pathology services in, 216 health care services. See also platforms; specific services Ghaffar, Abdul, 215, 267 benefit-cost analysis and, 181 Ghana, pathology services in, 230 burden of disease and, 121, 124, 136 GHE. See Global Health Estimates community health platforms, 276–77, 278 Ghebreyesus, Tedros, 7 development assistance and, 306, 307 GHO (Global Health Observatory), 26 emergency care, 248 Giedion, U., 8 extended cost-effectiveness analysis, 158 Glass, Roger, 3 intersectoral policies and, 37 Glassman, Amanda, 3, 8, 43, 62 mortality rates and, 115 Global Burden of Disease (GBD) study. pandemics and, 354 See also burden of disease quality of care and, 186, 204 benefit-cost analysis and, 178 universal health coverage and, 43–48, 52, 54, 58, 60 cost-effectiveness analysis and, 154 health care systems development assistance and, 305 burden of disease and, 121, 136–37 emergency care and, 250 community health platforms, 268, 271 intersectoral policies and, 26 emergency care, 248 mortality rates, 70–73, 100–101 fiscal policy and, 359 rehabilitation services, 285–86 palliative care and pain control, 235, 238, 240–41 Global Fund to Fight AIDS, Tuberculosis, pandemics and, 323, 327, 335 and Malaria, 150, 153 pathology services and, 216, 224, 228 Index 393 quality of care and, 186–87, 190–91, 200 HIV/AIDS rehabilitation services, 287 benefit-cost analysis for interventions, 172–73 health care workers burden of disease, 122, 125–26, 128, 134–35 community health platforms and, 277 community health platforms and, 269 continuing education, 222–24 cost-effectiveness analysis for interventions, essential packages of care and, 60 149–51, 154 palliative care and pain control, 238 development assistance and, 300, 302, 304–6, 308 pandemics and, 323–25, 329 emergency care, 252, 256, 261 pathology services, 218, 222–24 extended cost-effectiveness analysis for Health in All Policies approach, 23 interventions, 157 health outcomes fiscal policy and, 361–63 community health platforms and, 272–73 intersectoral policies on, 27 essential packages of care and, 56, 57t mortality rates, 16, 70–72, 101, 102n1, 106, 115–16 fiscal policy and, 361, 367–68 palliative care and pain control, 236–38 intersectoral policies and, 27, 37 pandemics and, 317–18, 323, 354 quality of care and, 186, 190–93, 199–200, 204–5 pathology services and, 215, 218–19, 225 universal health coverage and, 47–49 quality of care, 196, 198 Health Results Innovation Trust Fund, 193 universal health coverage and, 48–51, Healthy Life Expectancy (HALE), 31, 305 53–54, 56–60, 62 heart disease. See also cardiovascular disease Ho, Jessica, 69 burden of disease, 126, 128 Hofman, Karen, 359 fiscal policy and, 361 Hogan, Dan, 69 mortality rates, 100 Holmes, King K., 3 palliative care and pain control, 236–37 Hong Kong SAR pandemics and, 354 pandemics in, 318, 325, 329 heart failure pathology services in, 227 cost-effectiveness analysis for interventions, Horton, Susan, 3, 43, 50, 147, 167, 215 149–51 hospitals. See also specialized hospitals emergency care, 250 burden of disease and, 122, 127–28, 135–36 palliative care and pain control, 236 cost-effectiveness analysis, 149, 152, 153 quality of care, 197 development assistance and, 299 rehabilitation services, 288 emergency care, 248–49, 253–61, 262 helmet laws, 34–35b intensive care unit (ICU), 190, 249–50, 327, 334 hematology, 216–18, 220, 225, 229, 231 intersectoral policies and, 11 HICs. See high-income countries palliative care and pain control, 239–41 highest-priority package (HPP), 12–17, 44, pandemics and, 323, 327, 329 47–50, 55–60 pathology services and, 215–16, 220, high-income countries (HICs) 222, 228, 230–31 benefit-cost analysis in, 171, 174–77, 179 quality of care and, 185, 190–92, 194–202, 204 burden of disease in, 126–27 rehabilitation services, 285–88, 290–92 cost-effectiveness analysis in, 147 universal health coverage and, 13–14, 45–46, 55 emergency care in, 250 household use of coal, bans on, 29b fiscal policy in, 368 HPP. See highest-priority package intersectoral policies in, 31, 34 HPV vaccination, 148, 150–51, 153 mortality rates in, 70, 114, 116 human capital, 168–70, 192 palliative care and pain control in, 236, 238, 241 Hutton, Guy, 3 pandemics in, 316, 318, 322, 324, 331, 333–37, 352, 354 I pathology services in, 216–17, 223 ICD (International Classification of Diseases), quality of care in, 189, 192 70–71, 235 rehabilitation services in, 288 ICER. See incremental cost-effectiveness ratio universal health coverage in, 12, 43–44 IHD. See ischemic heart disease histopathology, 216–17, 225, 229 IHME. See Institute of Health Metrics and Evaluation 394 Index immunizations. See vaccines and vaccinations Institute of Medicine, 200 impoverishing health expenditures. See catastrophic insurance. See also universal health coverage and impoverishing health expenditures access to health care and, 8 incremental cost-effectiveness ratio (ICER), extended cost-effectiveness analysis and, 159–62 47, 159, 161, 337 pandemics and, 333 incremental costs pathology services and, 228 emergency care, 262 intensive care unit (ICU), 190, 249–50, 327, 334 quality of care, 205 Intergovernmental Panel on Climate Change (IPCC), universal health coverage, 12–14, 50, 55 353–54 India International Association for Trauma Surgery and benefit-cost analysis in, 169–70, 179–81 Intensive Care, 199 burden of disease in, 127–28, 132–35 International Classification of Diseases (ICD), development assistance in, 301, 304 70–71, 235 extended cost-effectiveness analysis in, International Decision Support Initiative, 61–62 158, 160–62 International Monetary Fund, 12 fiscal policy in International Standards for Tuberculosis Care, 186 analytic framework, 361–62, 362–63t intersectoral policies and interventions, 11–12 role of, 360–61, 360–61t accountability and, 37 subsidies, 364t, 366–67 cancer and, 25 taxation, 364–66, 364t consequences of illness or injury addressed by, intersectoral policies in, 26 31–32, 34 Million Death Study, 61b distal determinants of health and, 26–27, 26t mortality rates in, 71–72, 116–17 essential packages and, 27–31, 28–29t, 29–30b, 37 pandemics in, 319, 325, 336, 349 financial risk protection (FRP) and, 37 pathology services in, 218, 226–28, 230–31 health conditions and risk factors amenable to, quality of care in, 186, 190–91, 202 24–27, 25f subsidies in, 364t, 366–67 implementation of, 34–37 taxation in, 364–66, 364t life expectancy and, 26, 31 tobacco use in, 26 priorities for, 23–42 universal health coverage in, 61 subsidy-related strategies, 37 Indonesia Sustainable Development Goals and, 37 community health platforms in, 268, 273–75, 275f taxation-based strategies, 36–37 pandemics in, 336 universal health coverage and, 46 pathology services in, 227 interventions. See also intersectoral policies and infectious diseases. See communicable diseases interventions; specific interventions influenza pandemic, 347–58 barriers to uptake, 58 burden of disease, 322–23, 325, 326, 347–48, benefit-cost analysis and, 167–68, 178. See also 352–53, 353t benefit-cost analysis cost-effectiveness of interventions, 335–37 cost, 16–17t, 16–18 economic burden, 347 cost-effectiveness analysis, 147–48. See also cost- mitigation of, 327, 329, 331, 335 effectiveness analysis mortality rates, 316, 322, 323, 349t, 354 extended cost-effectiveness analysis, 158. See also risk of, 315–17, 348–49, 350–51b extended cost-effectiveness analysis severity of, 348–49, 350–51b nutrition interventions, 152, 170 study methodology, 349–52, 352t pandemics and, 316 injuries rehabilitation services, 288 catastrophic health expenditures and, 126, 128 universal health coverage and, 44, 49 head injuries, 34–35b IPCC (Intergovernmental Panel on Climate Change), mortality rates, 69–70, 73, 98, 101 353–54 rehabilitation services, 286 Iran road traffic, 12 burden of disease in, 128 Institute of Health Metrics and Evaluation (IHME), 70, intersectoral policies in, 34 72, 101, 106, 116, 300–305, 359, 364–65 mortality rates in, 115 Index 395 Iraq, emergency care in, 250 fiscal policy in, 362–63 Ireland, household coal use in, 29b mortality rates in, 115 ischemic heart disease (IHD) tobacco taxation in, 9–10 burden of disease, 126 Lennox, L., 204 emergency care, 251 Lesotho, benefit-cost analysis in, 175–76 intersectoral policies on, 25 Levin, Carol, 3 pandemics and, 354 Liberia, pandemics in, 318, 323–25, 329, 335, 347 LICs. See low-income countries J life expectancy Jamison, Dean T., 3, 23, 43, 105, 157, 158, benefit-cost analysis and, 169, 171–75, 177, 181 167, 235, 347 intersectoral policies and, 26, 31 Jan, Stephen, 121 pandemics and, 351–52 Japan trends in, 6, 98, 98f benefit-cost analysis in, 175–76 liquid propane gas (LPG), 28, 361, 363–64, 366–67 development assistance from, 302, 302f long-term care, 32–33b, 287, 292 fiscal policy in, 362 Looi, Lai Meng, 3, 215 intersectoral policies in, 26 low- and middle-income country (LMICs) mortality rates in, 101 benefit-cost analysis in, 172, 174–77 pandemics in, 317, 319 burden of disease in, 121–22, 126–28, 130, 132–37 tobacco use in, 26 cost-effectiveness analysis in, 147–50, 153–54 Jha, Prabhat, 3, 23, 43 development assistance in, 300, 304–5, 307–10 emergency care in, 250–52 K extended cost-effectiveness analysis in, 157 Kazakhstan, mortality rates in, 109 fiscal policy in, 359 Khanh, Quach Thanh, 235 intersectoral policies in, 23–25, 29, 31–32, 34, 36 kidney diseases mortality rates in, 73, 96, 100–102, 106, burden of disease, 126, 128 109, 114–15, 117–18 emergency care, 261 palliative care and pain control in, 235–38, 240–41 Knaul, Felicia, 3, 121, 235 pandemics in, 316, 318–19, 324, 327, 329–33, 335–38 Kobusingye, Olive, 3 pathology services in, 215–19, 223–30 Korea. See Republic of Korea quality of care in, 186–93, 200–205 Krakauer, Eric, 3, 32, 235 rehabilitation services in, 288 Kruk, Margaret E., 3, 43, 185, 203 universal health coverage in, 43–45, 47–48, Kumar, Suresh, 235 55, 59, 61–62, 64 Kuti, Modupe, 215 low-income countries (LICs) Kwete, Xiaoxiao, 235 benefit-cost analysis in, 172, 175–77, 179 burden of disease in, 121, 128, 130, 135 L cost-effectiveness analysis in, 148–53 Laba, Tracey-Lea, 121 development assistance in, 303–4 Lachmann, Peter, 3 mortality rates in, 70, 98–99, 114, 116–17 Lago, Nestor, 215 pandemics in, 322, 324, 331–33, 336–37, Lancet Commission on Investing in Health, 351–52, 354 6, 44, 49, 55, 106 pathology services in, 219, 222–23, 228 Lao People’s Democratic Republic, mortality universal health coverage in, 12–15, 44, rates in, 109 47–48, 50–51, 55–56, 60–62 Latin America and Caribbean. See also LPG. See liquid propane gas specific countries Lubell, Y., 167 cost-effectiveness analysis in, 150–52 Luyirika, Emmanuel, 235 development assistance in, 308 mortality rates in, 107–9, 114 M pandemics in, 318 Macedonia, former Yugoslav Republic of, mortality Laxminarayan, Ramanan, 3, 158, 359 rates in, 109 Lebanon Madhav, Nita, 3, 315 396 Index Mahanani, Wahyu Retno, 69 intersectoral policies in, 26, 36 Mahmoud, Adel, 3 national health accounts in, 32–33b malaria noncommunicable diseases in, 3 benefit-cost analysis for interventions, 169, palliative care and pain control in, 236, 242, 242t 172, 179 Progresa/Oportunidades in, 192b burden of disease, 126, 128, 134 quality of care in, 192 cost-effectiveness analysis for interventions, rehabilitation services in, 287 149–51, 153 microbiology, 215–17, 220, 223, 225, 229 development assistance and, 302, 304–5, 308 micronutrient interventions, 152 emergency care, 251–52, 258 MICs. See middle-income countries fiscal policy and, 361 Middle East and North Africa. See also specific countries intersectoral policies on, 24 mortality rates in, 107–9, 114 mortality rates, 70–72, 74 pandemics in, 318, 330 pandemics and, 323, 327 middle-income countries (MICs) pathology services and, 215, 218, 220 benefit-cost analysis in, 171–76, 178–81 quality of care, 198 burden of disease in, 123, 125, 127–29, 131 universal health coverage and, 47–48, 50–51, 53, 57 cost-effectiveness analysis in, 147–53 Malawi development assistance in, 303–4 emergency care in, 251, 262 emergency care in, 250–51, 253 mortality rates in, 114–15 fiscal policy in, 368 palliative care and pain control in, 236 intersectoral policies in, 26, 31–32, 34 pathology services in, 230–31 mortality rates in, 70, 96–99, 114–17 Malaysia pandemics in, 328, 336, 347–48, 352 burden of disease in, 128 pathology services in, 218–19, 229–30 extended cost-effectiveness analysis in, 162 rehabilitation services in, 286–88 fiscal policy in, 363 universal health coverage in, 43–44, 46, 49–57, 59 pandemics in, 321 Millennium Development Goals (MDGs) pathology services in, 218, 223, 226–27 community health platforms and, 268, 273 Marks, Elanie, 285 cost-effectiveness analysis and, 147, 149 maternal mortality development assistance and, 302, 304 development assistance and, 307 mortality rates and, 69, 99, 101, 106, 108–9 emergency care and, 250–51 quality of care and, 189 pandemics and, 354 Million Death Study (India), 61b pathology services and, 218 Mills, Anne, 3, 43, 167 trends in, 69, 72, 99–101, 102n1, 106–9, 111, Mills, Jody-Anne, 3, 285 111t, 114–17, 117f Minh, Hoang Van, 121 Mathers, Colin, 69 Mock, Charles N., 3, 23, 247 Mbanya, Jean-Claude, 3 money-metric value of insurance, 159–62 MDGs. See Millennium Development Goals Mongolia Measham, Anthony R., 3 benefit-cost analysis in, 180 Médecins sans Frontières, 153 burden of disease in, 124–25 medications. See drugs universal health coverage in, 124b Medina-Mora, María Elena, 3 Montoya, Jaime, 3 Medlin, Carol, 3 morbidity Memirie, T., 157 benefit-cost analysis for interventions, Mendez, Oscar, 235 168–69, 174, 178 mental illnesses, 126, 128–29 development assistance and, 305, 307 Merriman, Anne, 235 emergency care and, 247, 250 Mexico extended cost-effectiveness analysis and, 158 burden of disease in, 133 fiscal policy and, 367 cost-effectiveness analysis in, 147 intersectoral policies and, 34 fiscal policy in, 363 pandemics and, 315–17, 320–21, 323, 333, 338, 354 influenza pandemic in, 355 quality of care and, 199–200, 202 Index 397 Morocco, fiscal policy in, 362–63 N morphine, 238–40, 242 Naidoo, Mahendra, 215 mortality rates, 69–120 National Clean Energy Fund (NCEF), 366 access to health care and, 115 national health accounts (NHAs), 16, 32–33b benefit-cost analysis and, 177 National Tobacco Control Programme breast cancer, 96 (NTCP), 365 cancer, 70–72, 74, 96, 98, 100 NCEF (National Clean Energy Fund), 366 cardiovascular disease, 74, 98 neglected tropical diseases (NTDs) by cause and country income group, benefit-cost analysis for interventions, 169, 171 73–75, 74–75f cost-effectiveness analysis for interventions, 149, cause-specific trends, 72, 75, 76–95t, 151, 153 96–97f, 97t, 98 development assistance and, 300, 302, 307 child mortality. See child health and mortality emergency care, 252 chronic obstructive pulmonary disease (COPD), pathology services and, 218 96–98 universal health coverage and, 48, 51, 53 communicable diseases, 69, 73–74, 98–99, 101 Nepal community health platforms and, 273 burden of disease in, 128, 134 development assistance and, 304, 307–8 emergency care in, 250 emergency care and, 250, 252 pandemics in, 321 essential packages of care and, 15–16 Netherlands estimates of, 25, 115–16, 352 benefit-cost analysis in, 177 fiscal policy and, 362, 365–66 development assistance from, 302, 302f global trends in, 69–104, 75–95t intersectoral policies in, 31 heart disease, 100 pandemics in, 334 HIV/AIDS, 16, 70–72, 101, 102n1, Nevzorova, Diana, 235 106, 115–16 New Zealand indicators, 106–7, 109, 115–16 life expectancy in, 6 influenza pandemic, 316, 322, 323, pandemics in, 336 348–49, 349t, 351–52, 354, 355 pathology services in, 227 intersectoral policies on, 27 Ng, M., 58 malaria, 70–72, 74 NGOs. See nongovernmental organizations maternal mortality. See maternal mortality Nguyen, Thi Kim Phuong, 121 methodology, 69–73, 71f, 106–7, 107–8f NHAs. See national health accounts noncommunicable diseases, 96, 106–7, 109, 113t, NHAs (national health accounts), 16, 32–33b 114–15, 118–19f Nicaragua pandemics and, 316, 322, 323, 329–30, benefit-cost analysis in, 180 348–49, 349t, 351–52, 354, 355 quality of care in, 205 pneumonia, 98 Nigeria quality of care and, 195, 197, 199 development assistance in, 304 reduction in, 14, 17, 27, 56 mortality rates in, 101, 114 suicide, 30b pandemics in, 331 Sustainable Development Goals and, noncommunicable diseases (NCDs), mortality rates, 98–99, 99f 96, 106–7, 109, 113t, 114–15, 118–19f. See also trends, 69–70, 72, 99–101, 102n1, 106–7, specific diseases 109–10, 114–17 nongovernmental organizations (NGOs), tuberculosis, 56, 70–72, 105–9, 111, 112t, 267, 270–75, 278–79 113–18, 118f Norheim, Ole, 3, 14, 43, 56, 98, 105, 157 universal health coverage and, 14–15, 47, Norway 49, 56–57 development assistance in, 301–2, 302f motorcycle helmet laws, 34–35b pandemics in, 321 Mpanumusingo, Egide, 235 NTCP (National Tobacco Control Mulembakani, Prime, 315 Programme), 365 Myanmar, burden of disease in, 128 NTDs. See neglected tropical diseases 398 Index Ntizimira, Christian, 235 pandemics, 315–46. See also specific diseases Nugent, Rachel, 3, 11, 23, 105, 299 access to health care and, 354 nutrition interventions, 152, 170 accountability and, 338 benefit-cost analysis for interventions, 168, 171 O burden of, 321–23, 322f, 322t obstetric care, 149, 151, 153, 251, 262 consequences of, 323–25 OECD countries cost-effectiveness analysis for interventions, 149 benefit-cost analysis in, 175 development assistance and, 305, 310 development assistance from, 301 economic impacts of, 324–25 official development assistance. See development health impacts of, 323–24 assistance intervention costs and cost-effectiveness, Ogbuoji, Osondu, 105 334–37f, 334–38 Olson, Zachary, 3, 105 investment priorities, 338–39 Omokhodion, Folashade, 3 knowledge gaps, 316 Oppenheim, Ben, 3, 315 mitigation, 316, 326–34, 326b Ord, Toby, 3 pathology services and, 218 Osman, Hibah, 235 patient care and treatment protocols, 330–31 Ottersen, Trygve, 299 preparedness for, 7, 11, 316, 320m, 321t, 326–27, 331–34, 336, 338, 340, 345, 355 P reducing spread of, 329–30 packages. See essential packages of care risk communications and, 328–29 Padian, Nancy, 299 risks, 316, 317–19t, 317–23, 320m, 338, 341, PAHO (Pan American Health Organization), 348, 351–55 269, 280 risk transfer mechanisms, 332–33, 333f palliative care and pain control, 235–46 situational awareness and, 327, 328b access to, 238 social and political impacts of, 325 cancer, 240 universal health coverage and, 51, 53, 61 cardiovascular disease, 236–37 vaccines and, 330b, 331, 331t chronic obstructive pulmonary disease Patel, Vikram, 3 (COPD), 236 pathology services, 8, 11, 215–34 cost, 241–43, 242t accountability and, 226–27 cost-effectiveness analysis, 10–11 accreditation, 225–27, 227b equipment, 240 breast cancer and, 221 essential package for, 235–36, 238–43, 239t cancer and, 215–16, 221, 225 extended cost-effectiveness analysis, 162 cardiovascular disease and, 215 families and, 236, 238, 240–43 challenges in LMICs, 217–18 financial risk protection (FRP) and, 236, 243 communicable diseases and, 215, 218, 221 heart disease, 236–37 continuing education, 222–24 heart failure, 236 costs, 228–29 HIV/AIDS, 236–38 data handling, 225 hospitals and, 239–41 economics of, 228–31t, 228–32 human resources, 241 education, 222–24 intersectoral policies and, 31–32 emerging diseases and, 216, 219, 221, 225 medicines, 239–40, 242–43 emerging technologies, 224–25, 224b need for, 236–38, 237t essential package for, 215, 217, 219–32, platforms for delivery, 235, 240, 242 220–21t, 220b psychological and spiritual counseling, 240 financial risk protection (FRP) and, 219 social supports, 240–41 HIV/AIDS and, 215, 218–19, 225 tuberculosis, 237 leadership and, 222–23 universal health coverage and, 47–48, 54, 59 maternal mortality and, 218 Pan American Health Organization misdiagnosis issues, 202, 202b (PAHO), 269, 280 neglected tropical diseases (NTDs) and, 218 Pandemic Influenza Preparedness (PIP), 336–37 platforms for delivery, 224–25 Index 399 point-of-care testing, 225 pathology services and, 216 quality management, 202, 202b, 225–27, 226b quality of care and, 186, 187–88, 194, reimbursement policies, 227–28 195–98t, 199, 205 services, 216–17, 217b, 217t, 221–22, 227–28 population aging. See aging population staffing inadequacies, 218 Prabhakaran, Dorairaj, 3 Sustainable Development Goals and, 218, 218–19t pregnancy. See maternal mortality; obstetric care training, 222–24 President’s Emergency Plan for AIDS Relief universal health coverage and, 52, 54 (PEPFAR), 305–7 Patton, George C., 3 Preston, S. H., 98 pay for performance (P4P), 193–94, 193b Price, Christopher P., 215 Peabody, John, 3, 185 primary care Peilong, Liu, 299 community health platforms and, 267, PEPFAR (President’s Emergency Plan for 273, 276–77 AIDS Relief), 305–7 cost-effectiveness analysis, 150–51 Perez-Cruz, Pedro, 235 quality of care and, 191 Peru Progresa/Oportunidades (Mexico), 192b community health platforms in, 272, 276–77 mortality rates in, 109 Q pesticides, 30b QALYs (quality-adjusted life years), 10 pharmaceutical drugs. See drugs benefit-cost analysis and, 167 Philippines cost-effectiveness analysis and, 148, 150–52, 154 benefit-cost analysis in, 169 emergency care and, 262 burden of disease in, 124–25 rehabilitation services and, 288 mortality rates in, 114 universal health coverage and, 47–48 quality of care in, 191–93, 202 value of, 6 physical inactivity, 24–26 Qi, Jinyuan, 3, 43, 50, 105 PIP (Pandemic Influenza Preparedness), 336–37 Quality Improvement Demonstration Study (QIDS), platforms 192–93, 193b community health platforms, 267–84. See also quality of care, 185–214 community health platforms access to health care and, 199 defined, 7 accountability and, 187, 191, 194–95, 201, 203–4 development assistance and, 306, 308 affordability, 201 intersectoral policies and, 4–5 assessment challenges, 201–3 palliative care and pain control, 235, 240, 242 breast cancer, 202 pathology services, 224–25 cancer, 196–97, 202 quality of care, 189, 194, 199, 203 cardiovascular disease, 193 rehabilitation services, 289–91 chronic diseases and chronic ill health, 190, 204 universal health coverage and, 13–14, 45–46, 62 conditional cash transfers (CCTs) and, 192–93 pneumonia costs of improving, 204–5 development assistance and, 305 effectiveness of interventions, 201 emergency care, 249–50, 252, 254, 261–62 equity, 201–2 mortality rates, 98 essential package for, 58–59 pandemics and, 316, 322–23 family planning and, 195, 199 quality of care, 196 financial risk protection (FRP) and, 204 policy interventions framework for, 199–201, 199f, 200t benefit-cost analysis, 168, 178, 181 health care facilities and, 185, 191, 201 catastrophic and impoverishing health heart failure, 197 expenditures, 137–38 HIV/AIDS, 196, 198 development assistance and, 307 hospitals and, 185, 190–92, 194–202, 204 extended cost-effectiveness analysis, 157–61 incentives, 192–93 fiscal policy, 359–68. See also fiscal policy infrastructure requirements, 187–94, 195–98t intersectoral policies, 24, 27–28. See also malaria, 198 intersectoral policies and interventions measurement of, 187–90, 188t 400 Index misdiagnosis and, 202, 202b financial risk, 8–11, 121, 124 morbidity and, 199–200, 202 fiscal policy and, 360–62 patient focus, 201 influenza pandemic, 315–17, 348–49, 350–51b perceptions of quality, 202–3 intersectoral policies and, 24–27 performance-based financing, 193–94, 193b, 194f mortality rates and, 115 platforms for delivery, 189, 194, 199, 203 pandemics, 315–17, 317–19t, 317–23, 320m, 338, pneumonia, 196 341, 348–49, 350–51b, 351–55 policy interventions and, 186, 187–88, 194, quality of care and, 191 195–98t, 199, 205 risk reduction primary care and, 191 benefit-cost analysis and, 175, 177 safety and efficacy, 200–201 emergency care and, 252 severe acute respiratory syndrome pathology services and, 219 (SARS), 196, 198 risk transfer mechanisms, 332–33, 333f standards, 190–91 Rodriguez, Natalia M., 235 supervision, 192 rotavirus syphilis, 195 burden of disease, 128 timeliness, 201 cost-effectiveness analysis for interventions, 151–52 training, 191–92 Rottingen, John-Arne, 299 tuberculosis, 196 Ru, Kun, 215 universal health coverage and, 58–59, 201, 203 Ruacan, Sevket, 3 vaccinations and, 192, 196–98 Rubiano, Andrés M., 247 variations in, 186–87 Rubin, Edward, 315 Russian Federation, fiscal policy in, 359 R Rwanda Radbruch, Lukas, 235 emergency care in, 251 Rajagopal, M. R., 235 mortality rates in, 109, 114–15 RBF (results-based financing), 59, 193–94 palliative care and pain control in, 236, 242, 242t rehabilitation services, 285–95 quality of care in, 194 access to, 286–87 cost-effectiveness, 287–88 S demand for, 286–87 Sample Registration System (SRS), 61, 71 essential package, 288–92, 289–91t Sanderson, W. C., 175 heart failure, 288 Sankaranarayanan, Rengaswamy, 3 hospitals and, 285–88, 290–92 SARA (Service Availability and Readiness planning tools, 292 Assessment), 194 platforms for delivery, 289–91 SARS. See severe acute respiratory syndrome policy priorities, 293 Saudi Arabia, development assistance in, 301 universal health coverage and, 43 Sawe, Hendry, 247 renal diseases Saxenian, Helen, 23 burden of disease, 122, 126, 128–32, 134, 136–37 Schäferhoff, Marco, 299 pathology services and, 221 Scherbov, S., 175 Republic of Korea SDGs. See Sustainable Development Goals burden of disease in, 127–28 Sen, A., 8 fiscal policy in, 363 Sepúlveda, Jaime, 3 pandemics in, 318 Service Availability and Readiness Assessment respiratory diseases, 122, 126–30 (SARA), 194 results-based financing (RBF), 59, 193–94 severe acute respiratory syndrome (SARS) Reynolds, Teri, 3, 247, 285 development assistance and, 307 risk factors impact of, 315 burden of disease and, 137 mitigation of, 329–30, 355 community health platforms and, 269 quality of care, 196, 198 conceptual model for interactions risks of, 318 among, 24–25, 25f Sherry, Melissa, 267 Index 401 Shimkhada, Riti, 185 subsidies Shin, Sang Do, 247 benefit-cost analysis and, 172 Sierra Leone fiscal policy and, 360–63, 366–68 mortality rates in, 114 intersectoral policies and, 11–12, 23, 37 pandemics in, 318, 323–25, 335, 347 palliative care and pain control, 242 Singapore pandemics and, 335 benefit-cost analysis in, 177 sugar, 366 pandemics in, 318, 334–36 universal health coverage and, 46–47 Skolnik, Richard, 3 sugar-sweetened beverages (SSBs) Smith, Kirk R., 3 fiscal policy interventions, 361 Smith, P. C., 8 intersectoral policies on, 12, 36 Soucat, Agnes, 43 taxation, 363–64, 364t, 366 source measure units (SMU), 174, 348–52 suicide, 30b South Africa Summan, Amit, 359 burden of disease in, 128 Summers, Lawrence H., 347 cost-effectiveness analysis in, 149 surveillance development assistance in, 301 palliative care and pain control, 239 extended cost-effectiveness analysis in, 162 pandemics, 320, 327–28, 333, 335 fiscal policy in, 359 pathology services and, 216, 218–21, 224 mortality rates in, 114–15 universal health coverage and, 60–62 pathology services in, 223, 226–29 Sustainable Development Goals (SDGs) South Asia, benefit-cost analysis in, 170 on burden of disease, 121, 137 South-East Asia. See also specific countries on community health platforms, 268, 280 development assistance in, 308 on cost-effectiveness analysis, 147 intersectoral policies in, 34 development assistance and, 304 mortality rates in, 109, 114–15 on emergency care, 252 pandemics in, 319 intersectoral policies, 23 pathology services in, 226 on mortality rates, 14, 16–17, 69, Spanish flu pandemic, 317, 322 98–99, 106, 115 specialized hospitals on pathology services, 218–19 emergency care, 255–61 on quality of care, 189 quality of care and, 199 universal health coverage and, 43, 46, 56, 61–62 rehabilitation services, 287 swine flu pandemic, 318, 322 universal health coverage and, 45 syphilis Spence, Dingle, 235 cost-effectiveness analysis for interventions, 149–51 Sri Lanka quality of care, 195 benefit-cost analysis in, 169 burden of disease in, 134 T mortality rates in, 109 Taiwan, influenza pandemic in, 355 suicide rates in, 30b Tanzania SRS (Sample Registration System), 61, 71 burden of disease in, 128, 133–35 SSBs. See sugar-sweetened beverages quality of care in, 189, 191, 201, 204 Stacey, Nicholas, 359 Tapela, Neo, 235 Stevens, Gretchen, 69 taxation Stigliz, J., 8 cost-effectiveness analysis, 148 Stoltenberg, Mark, 235 development assistance and, 301 Sub-Saharan Africa. See also specific countries extended cost-effectiveness analysis, 158, 162 development assistance in, 302, 305, 308 fiscal policy and, 362, 364–65, 367–68 intersectoral policies in, 34 intersectoral policies and, 9–12, 23, 36–37 mortality rates in, 98, 101, 107–9, 114–15 quality of care and, 203 pathology services in, 215, 218, 221, sugar-sweetened beverages, 363–64, 364t, 366 223, 226–28, 230 tobacco, 9–10, 27, 173, 364–65, 364t rehabilitation services in, 288 universal health coverage and, 46 402 Index Temmerman, Marleen, 3 United Arab Emirates, development Thailand assistance from, 301 benefit-cost analysis in, 181 United Kingdom burden of disease in, 129 household coal use in, 29b development assistance in, 301 intersectoral policies in, 26 intersectoral policies in, 35 pandemics in, 317, 324, 334–36 pathology services in, 226–27, 230–31 pathology services in, 215, 218, 226–27, 229, 231 Timor-Leste, rehabilitation services in, 286 rehabilitation services in, 288 tobacco use tobacco use in, 26 benefit-cost analysis for interventions, 173 United Nations Children’s Fund, 106, 267 community health platforms and, 269 United Nations Population Division, 14, 181, 351 cost-effectiveness analysis for interventions, 150 United States extended cost-effectiveness analysis for benefit-cost analysis in, 167, 171, 173–76 interventions, 162 burden of disease in, 127–28 fiscal policy and, 360–65 community health platforms in, 267 intersectoral policies and, 11–12, 23–24, 28, 31, 36 development assistance from, 301, 302, 302f, 305 taxation of, 9–10, 27, 173, 364–65, 364t extended cost-effectiveness analysis in, 157 universal health coverage and, 56 influenza pandemic in, 350–51, 354–55 Tobago, pandemics in, 321 intersectoral policies in, 23, 26, 31 Tollman, Stephen, 3 mortality rates in, 105, 116 Trans-Pacific Partnership Agreement, 227 palliative care and pain control in, 235–36 tuberculosis pandemics in, 315, 317, 319, 334–36, 350–51, 354–55 benefit-cost analysis for interventions, 172–73 pathology services in, 215–16, 218, 226–27, 229–31 burden of disease, 126, 134 quality of care in, 185–86, 193, 201 cost-effectiveness analysis for interventions, tobacco use in, 26 149–51, 153 universal health coverage in, 43 development assistance and, 302, 304 universal health coverage (UHC), 43–65 diagnostic tool subsidies, 367 access to health care and, 43–44, 59 emergency care, 252 barriers to intervention uptake, 58 extended cost-effectiveness analysis for burden of disease and, 46, 49, 121–25, 137 interventions, 158, 160–62 cancer and, 51, 53 fiscal policy and, 361–64, 366–67 catastrophic health expenditures and, 124–25b mortality rates, 56, 70–72, 105–9, 111, child health and mortality and, 47, 48, 49, 112t, 113–18 51, 53, 55, 56 palliative care and pain control, 237 communicable diseases and, 48, 51, 53, 56 pandemics and, 323, 326, 354 conditional cash transfers (CCTs) and, 59 pathology services and, 215, 218, 220, 225 cost-effectiveness analysis, 44–45, 47–48, 58, 60, 147 quality of care, 196 costs and, 12–14, 13–14t, 49–55, 51–55t universal health coverage and, 48–49, 51, development of, 45 53, 56–57, 59 emergency care and, 247, 252, 261 Tunisia, fiscal policy in, 363 essential, 12–16, 12f Turkmenistan, mortality rates in, 114 extended cost-effectiveness analysis and, 44, 158 family planning and, 46–47 U financial risk protection (FRP) and, 43–44, 47–48 Uganda financing, 60 community health platforms in, 277–78 fiscal policy and, 359 palliative care and pain control in, 238 governance and, 59–60 pathology services in, 218 health outcomes and, 56, 57t UHC. See universal health coverage health workforce and, 60 UMICs. See upper-middle-income countries HIV/AIDS and, 48–51, 53–54, 56–60, 62 UNAIDS, 70, 72, 100, 115 hospitals and, 13–14, 45–46, 55 under-five mortality rates. See child health and identification of highest-priority package, 45–49 mortality implementation, 56–62, 59t, 61b Index 403 inclusion criteria, 48–49 value for money, 3, 7–8, 16, 44–49, 60–62. See also information and research needs, 60–61, 61b benefit-cost analysis; cost-effectiveness analysis; intersectoral policies and, 24 extended cost-effectiveness analysis leadership and, 59–60 value per statistical life (VSL), 168, 170–72, 174–78, malaria and, 47–48, 50–51, 53, 57 175–78f, 176t, 181, 348 medical product and technology availability, 60 Verguet, Stéphane, 3, 43, 105, 106, 157, 158, 235 mortality reduction from, 14–16, 15t Vietnam neglected tropical diseases (NTDs) and, 48, 51, 53 intersectoral policies in, 34 palliative care and pain control, 235–36 motorcycle helmet laws in, 34–35b, 35f pandemics and, 51, 53 palliative care and pain control in, 236, 242, 242t pathology services and, 219 pathology services in, 218 platforms for delivery and, 13–14, universal health coverage in, 124b 45–46, 46b, 62 VSL. See value per statistical life priority-setting institutions, role of, 61–62 VSL-to-income ratio (VSLr), 174–78 quality of care and, 58–59, 201, 203 rehabilitation services and, 285 W surveillance and, 60–62 Walker, Damian, 3 tuberculosis and, 48–49, 51, 53, 56–57, 59 Walker, Neff, 3 vaccinations and, 47 Wallis, Lee, 247 value for money, 47 Walsh, J. A., 105 universal public finance (UPF), 157, 159–62 Wang, Huihui, 185 upper-middle-income countries (UMICs) Wang, Jinaxiang, 215 benefit-cost analysis in, 176, 179 Warren, K. S., 105 burden of disease in, 121, 128 Watkins, David A., 3, 11, 23, 43, 50, 56, 235 cost-effectiveness analysis in, 149–52 West Africa. See also specific countries extended cost-effectiveness analysis in, 160 development assistance in, 307, 310 fiscal policy in, 368 influenza pandemic in, 347, 355 mortality rates in, 70, 96, 98–99 pandemics in, 316, 319, 321, 326, 347, 355 palliative care and pain control in, 236 WHO. See World Health Organization pandemics in, 333, 336, 352 Wilson, David, 43 pathology services in, 218 Wilson, Michael, 215 rehabilitation services in, 287 Wolfe, Nathan, 315 universal health coverage in, 62 women U.S. Agency for International Development benefit-cost analysis and, 178 (USAID), 302, 304, 328, 333–35 community health platforms and, 271 development assistance and, 299 V emergency care and, 251 vaccines and vaccinations maternal mortality benefit-cost analysis, 173 development assistance and, 307 burden of disease and, 121, 137 emergency care and, 250–51 community health platforms and, 268, 273, 275 pandemics and, 354 cost-effectiveness analysis, 148–51, 152, 153–54 pathology services and, 218 development assistance and, 300, 308, 310 trends in, 69, 72, 99–101, 102n1, 106–9, 111, 111t, emergency care and, 252, 255, 260 114–17, 117f extended cost-effectiveness analysis, 158 national health accounts and, 33b fiscal policy and, 360–62 quality of care and, 199, 201, 204 HPV, 148, 150–51, 153 suicide rates among, 30b influenza, 355 universal health coverage and, 56 pandemics and, 318, 326, 329–31, 330b, 331t, 333, World Bank 334–35, 337–38, 354 on benefit-cost analysis, 179–81 pathology services and, 219 on community health platforms, 268, 277 quality of care and, 192, 196–98 on cost-effectiveness analysis, 148 universal health coverage and, 47 development assistance and, 306 404 Index on fiscal policy, 365 on pathology services, 225, 227–28 on mortality rates, 116 on quality of care, 199–200, 202–4 on palliative care and pain control, 242 on rehabilitation services, 285–88 on pandemics, 318, 321, 324, 334–35, 347, 349, 352–53 on universal health coverage, 43–44, 47–50, on quality of care, 186, 192–94, 202 55–57, 59–60 on rehabilitation services, 286 World Health Report 2010 (WHO), 8 on universal health coverage, 44, 55, 60 Wu, Yangfeng, 3 World Health Assembly, 43, 235, 241, 315 World Health Organization (WHO) Y on burden of disease, 136–37 Yamey, Gavin, 23, 105 on cost-effectiveness analysis, 148 on emergency care, 247–49, 253–55 Z on fiscal policy, 364–65 Zaire, pandemics in, 324 Global Health Observatory, 318, 323 Zhang, J., 132 on influenza pandemic, 354–55 Zhao, Kun, 3 on intersectoral policies, 26, 34, 36–37 Zika virus, 216, 318, 323 on mortality rates, 69–72, 98, 106, 115–16 Zimbabwe on palliative care and pain control, 238, 240 pandemics in, 336 on pandemics, 315–16, 318, 323–24, 330–31, pathology services in, 228 347–48, 354–55 quality of care in, 194 Index 405 ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, we leverage electronic publishing options and print- on-demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consumption, chemical use, greenhouse gas emissions, and waste. We follow the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC)–certified paper, with nearly all containing 50–100 percent recycled con- tent. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine–free (PCF), or enhanced elemen- tal chlorine–free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/corporateresponsibility.