cn cr C.0 a - ___*_l_ U! I~ as . Iw -- - ------- -- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ U a a c ot- _ ~~ . _ l . _ Cl. mi a - --- ------ a a a= L _D 1I - _ *~~~ 1 _-- ' @1~~ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ K le - sh -| >WI .iLe KX Confronting AIDS A World Bank Policy Research Report A Confronting AIDS Public Priorities in a Global Epidemic Published for the World Bank OXFORD UNIVERSITY PRESS Oxford University Press OXFORD NEW YORK TORONTO DELHI BOMBAY CALCUTTA MADRAS KARACHI KUALA LUMPUR SINGAPORE HONG KONG TOKYO NAIROBI DAR ES SALAAM CAPE TOWN MELBOURNE AUCKLAND and associated companies in BERLIN IBADAN C 1997 The International Bank for Reconstruction and Development / SHE WORLD BANK 1818 H Street, N. W. Washington, D.C 20433 Published by Oxford University Press, Inc. 200 Madison Avenue, New York, N. Y 1001 6 Oxford is a registered trademark of Oxford University Press. All rights reserved. No part ofthis publication may be reproduced. stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior permission of Oxford University Press. Manufactured in the United States ofAmerica Firstprinting October 1997 Cover credits: Mother and child, Curt Carnemark/World Bank; condom box, Population Services International (PSI); Anti-AIDS Club signboard Warren Parker/PSI; clinic worker and patient, UNAIDS/Yoshi Shimizu. The boundaries, colors, denominations, and other information shown on the maps in this vol- ume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. Library of Congress Cataloging-in-Publication Data ConfrontingAIDS public priorities in a global epidemic. p. cm. (A World Bank policy research report) Includes bibliographical references. ISBN 0-19-521117-0 1. AIDS (Disease)-Prevention-Governmentpolicy. I WorldBank. I. Series. RA644.A25C6339 1997 362.1'969792-DC21 97-17434 CIP ( Textprinted on paper that conforms to the American National Standard for Permanence of Paperfor Printed Library Materials, Z39.48-l984 Contents Foreword xiii Introduction xv The Report Team xvii Acknowledgments xix Definitions xxi Summary 1 1. AIDS: A Challenge for Government 13 What Is AIDS and How Is It Spread? 17 The Impact of AIDS on Life Expectancy and Health 22 AIDS and Development 27 The Government Role in Confronting AIDS 38 Social Norms and Politics Make AIDS Challenging 44 Overview of the Book 46 Appendix 1.1 Alternative Estimates of the Current and Future Magnitude of the HIV/AIDS Epidemic 47 2. Strategic Lessons from the Epidemiology of HIV 53 HIV Incidence and Prevalence, and AIDS Mortality 54 Biology and Behavior Affect the Spread of HIV 57 Implications for Public Policy 77 The Level and Distribution of HIV Infection in Developing Countries 86 3. Efficient and Equitable Strategies for Preventing HlV/AIDS 103 Influencing Individual Choices 105 Easing Social Constraints to Safe Behavior 123 Setting Government Priorities in Preventing HIV 132 The National Response 153 v CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 4. Coping with the Impact of AIDS 173 Health Care for the Person with AIDS 174 Difficult Health Policy Choices in a Severe AIDS Epidemic 183 AIDS and Poverty: Who Needs Help? 206 How Governments Can Cope with the Impact of HIV/AIDS on Health Care and Poverty 233 5. Working Together To Confront HIV/AIDS 239 Government, Donors, and NGOs 240 Bilateral and Multilateral Funding and the Stage of the Epidemic 250 Who Will Invest in New Knowledge and Technology? 258 Overcoming Political Impediments to Effective AIDS Policy 272 6. Lessons from the Past, Opportunities for the Future 283 Lessons from Two Decades of Experience 284 The Role of Government 284 Opportunities To Change the Course of the Epidemic 286 Challenges for the International Community 293 Appendixes 295 Appendix A: Selected Evaluations of Interventions To Prevent Transmission of HIV in Developing Countries 295 Appendix B: Selected Studies of the Cost-Effectiveness of Prevention Interventions in Developing Countries 301 Statistical Appendix 305 Selected Bibliography 329 Background Papers 329 Bibliography 330 Boxes 1.1 Pauline: One Woman's Story 16 1.2 The Natural History of HIV/AIDS 19 1.3 AIDS and the Chad-Cameroon Oil Pipeline Project 31 1.4 Looking for the Impact of HIV/AIDS on a Sarnple of African Firms 35 1.5 Orphans and AIDS 37 2.1 The Impact of an Early Peak in Infectivity 62 2.2 HIV Transmission through Medical Injections 66 vi CONTENTS 2.3 STDSIM: Modeling Behavior and Sexually Transmitted Diseases 76 2.4 Are Governments Intervening Early Enough? 80 2.5 Who Is Most Likely To Contract and Spread HIV? 82 2.6 The Multiplier Effect of Reducing HIV Transmission among Sex Workers in Nairobi 83 2.7 Monitoring the Spread of HIV 88 2.8 Declining Seroprevalence in Uganda 92 2.9 The Case for Early Intervention in Madagascar 95 3.1 Best Practices in HIV Prevention and Treatment Online 105 3.2 Behavior Responds to Increased Risks: AIDS Incidence and Increased Condom Use in the United States 107 3.3 Who Knows How Much about Preventing HIV/AIDS? 108 3.4 What Is "Information"? 110 3.5 Preventing HIV among Kinshasa Sex Workers 113 3.6 Harm Reduction among Injecting Drug Users in Nepal 118 3.7 Health Benefits of Regulating Prostitution in Pre-Independence India 121 3.8 What Data on HIV and Other STDs Should Governments Collect? 134 3.9 The Cost-Effectiveness of Prevention among Those with the Highest Risk 145 3.10 Educating Adolescents on HIV/AIDS: A Sound Investment 149 3.11 Thailand's Response 159 3.12 STDs and HIV in the Military 161 3.13 Preventing HIV on the Road to Ho Chi Minh City 167 4.1 The Government Role in Ensuring Clean Blood 187 4.2 Cost of Preventing Secondary HIV Infections through Blood Screening in Uganda 189 4.3 Estimating the Impact of AIDS on the Health Sector 192 4.4 The Effective Price of Care 193 4.5 Is Antiretroviral Therapy an Effective Way To Prevent Sexual Transmission? 202 4.6 Preventing Mother-to-Child Transmission 204 4.7 Three Factors Determine the Household Impact of a Death 209 4.8 Studies of the Household Impact of Adult Death from AIDS and Other Causes 210 4.9 Using Adult Death as a Targeting Criterion for Antipoverty Programs 234 5.1 Government, Private, and Donor Expenditures on AIDS in Five Countries 249 5.2 Helping Calcutta Sex Workers Avoid AIDS 256 5.3 Challenges To Be Overcome in Developing an HIV Vaccine 268 vii CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 5.4 Can Companies Make a Reasonable Return from the AIDS Vaccine? 270 5.5 Someone with AIDS Who Made a Difference 280 6.1 Estimating the Power of Prevention in Three Countries 287 Text Figures 1.1 Estimated Number of Adults with HIV/AIDS, by Region, 1997 14 1.2 Number of New Adult HIV Infections, by Region, 1977-95 15 1.3 The Current Impact of AIDS on Life Expectancy, Six Selected Countries, 1996 23 1.4 Breakdown of Deaths from Infectious Diseases, the Developing World, by Disease Category, 1990 and 2020 25 1.5 Causes of Death from Infectious Diseases among People Ages 15 to 59, the Developing World, 1990 and 2020 26 1.6 HIV/AIDS as a Percentage of the Infectious Disease Burden of Adults, the Developing World, 2020 26 1.7 Relationship of Four Societal Variables with Urban Adult HIV Infections, 72 Developing Countries, circa 1995 28 1.8 Annual Treatment Cost for an AIDS Patient Correlated with GNP per Capita 39 1.9 Number of Adults Living with HIV Infection by Region: by Comparison of Estimates, circa 1995 48 1.10 Current and Projected Future AIDS Death Rate per 1,000 People, by Region, 1990-2020: Comparison of Estimates 49 2.1 Estimated Trends in the Percentage of Adults Infected with HIV, by World Region 54 2.2 HIV Incidence, Prevalence, and AIDS Deaths 55 2.3 Distribution of Men and Women, Ages 15 to 49, with at Least One Casual Sexual Partner, by Number of Nonregular Partners in the Past Year, Rio deJaneiro, Brazil, 1990 67 2.4 Increasing Prevalence of HIV among Sex Workers, Seven Cities in Developing Countries, 1985-95 69 2.5 Rapid Diffusion of HIV among Injecting Drug Users, Asia and Ukraine, Various Years 70 2.6 Annual Incidence of AIDS Cases in Latin America and the Caribbean according to Risk Factor, 1982-95 71 2.7 Probability of a Casual Sex Partner over a 12-Month Period, by Age and Gender 73 2.8 The Impact of Different Baseline Patterns of Sexual Behavior on a Heterosexual HIV/AIDS Epidemic: STDSIM Results 75 2.9 The Impact of Increased Condom Use by Men on Adult HIV Prevalence, Early and Late in the Epidemic 78 2.10 The Impact of Increased Condom Use by Various Subpopulations on Adult HIV Prevalence in the Entire Population 84 viii CONTENTS 2.11 HIV Infection in Urban Sex Workers in Sub-Saharan Africa, Various Years 89 2.12 HIV Seroprevalence among Pregnant Women in Selected Areas of Sub-Saharan Africa, 1985-95 90 2.13 HIV Infection in Africa and the Middle East 91 2.14 HIV Infection in Latin America and the Caribbean 94 2.15 HIV Infection in Asia 97 2.16 HIV Infection in Eastern Europe and Central Asia 99 2.17 Reported Cases of Gonorrhea in Eastern Europe, 1986-94 100 3.1 Socially Marketed Condom Sales in Six Countries, 1991-96 112 3.2 Percentage of Men and Women Using Condoms with a Casual Partner, by Education, Eight Countries 131 3.3 The Impact of Changes in Condom Use and STD Treatment in Four Populations with Different Patterns of Sexual Behavior 142 3.4 Classification of Groups by Riskiness of Their Behavior and Their Accessibility 147 3.5 Resource Availability and Program Coverage 153 3.6 Percentage of People with a Recent Nonregular Partner Who Are Aware That Condoms Prevent HIV Transmission 157 3.7 Coverage of Subpopulations with High-Risk Behavior, Estimates of UNAIDS Country Programme Advisers in 32 Countries 162 3.8 Government Support for Prevention Targeted to Groups with High-Risk Behavior, Estimates of UNAIDS Country Programme Advisers in 32 Countries 163 4.1 Percentage of AIDS Patients with Three Opportunistic Infections, Seven Countries 175 4.2 Impact of AIDS on Utilization of and Mortality at Kenyatta National Hospital, Nairobi, 1988/89 and 1992 195 4.3 Public Share of Health Expenditure in Selected Countries, by Income Group, Various Years, 1990-97 197 4.4 Simulated Impact of a Severe AIDS Epidemic on Health Expenditure, India 198 4.5 Percentage of AIDS-Related and Total Treatment Expenditures Financed by the National Government, Four Selected Countries and Sao Paulo State, Brazil, 1994 200 4.6 Average Medical and Funeral Expenditures, by Gender and Cause of Death, Kagera, Tanzania, 1991-93 211 4.7 Consumption in Kagera, Tanzania, Households by Whether the Household Experienced a Death in the Past Year (Results from Wave I of Kagera Study) 212 4.8 The Time Pattern of the Impact of Adult Illness on Per Capita House- hold Expenditure, Cote d'Ivoire Study 213 4.9 Median Value of Assistance Received among Sample Households Receiving Transfers, by Source, Wave, and Occurrence of Adult Death, 1991-94 219 Ix CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 4.10 Average Cost per Year of Survivor Assistance in 1992 by Government and Nongovernmental Organizations, Kagera, Tanzania 221 4.11 Short-Term Impact of the Death of an Adult Household Member on Food Expenditure and Consumption per Adult Equivalent Member, Kagera, Tanzania, 1991-93 222 4.12 Stunting among Orphaned and Nonorphaned Children under 5, by Household Assets, Kagera, Tanzania 224 4.13 Enrollment Rate for Children Ages 7 to 14, by Orphan Status, Nine Countries 227 4.14 Enrollment Rates by Age, Orphanhood, and Household Assets, Kagera, Tanzania, 1991-93 228 4.15 Poverty in Kagera Region, by District and Adult Mortality Rate 231 5.1 Comparison of Average Annual AIDS Spending by Donors with That of National AIDS Programs, 1991-93 244 5.2 Relationship between the Number of HIV-Infected People in a Country (in Millions) and the Amount of National and International AIDS Expenditures 246 5.3 Relationship between GDP Per Capita and National and International AIDS Expenditure 247 5.4 Donor Funding for HIV/AIDS Interventions in Developing Countries in 1993 by Type of Donor and Stage of the Epidemic 251 5.5 Differing Strengths of Four Types of Nonprofit NGOs 253 5.6 Deaths of Adults from HIV and Other Infectious Diseases in the Established Market Economies, 1990 and Projected to 2020 262 Box Figures 1.5 Trends in Maternal Orphan Rates, Three Hard-hit East African Countries, Various Years 37 2.1 Shape of the Epidemic Curve under Alternative Assumptions about Infectiousness 62 2.4 Lag between the First Reported AIDS Case and Initiation of a National AIDS Control Program, 103 Industrial and Developing Countries 80 2.6 Infections Averted per Year by Raising Condom Use to 80 Percent in Two Populations in Nairobi 83 2.8 HIV Seroprevalence in Rural Masaka, Uganda, 1989 and 1994 92 3.3 Percentage of Adults Who Know That Condoms Are a Means of Protection against HIV Transmission, by Individual Characteristics, Seven Sub-Saharan African Countries 108 3.5 Incidence of HIV-1 and Other STDs among HIV-Negative Sex Workers over Three Years 113 3.1 la Rising Condom Use by Sex Workers and Declining STDs in Thailand, 1988-95 160 x CONTENTS 3.1 lb Declining HIV Prevalence among Young Thai Army Conscripts, 1989-96 160 3.12 HIV Prevalence in the Military 161 4.1 The Cost per Unit of Blood Transfused in Uganda 188 4.3 The Impact of a 5 Percent Infection Rate on the Quantity and Price of Health Care 192 6.1 Projected Impact of Behavioral Interventions in Three Countries 288 Text Tables 1.1 Annual Burden of Infectious Disease and HIV, as Measured by Deaths and Lost DALYs, the Developing World, 1990 and 2020 24 2.1 ProbabilityofHIV-1 InfectionperExposure 59 2.2 Estimated Prevalence and Annual Incidence of Curable STDs among Adults Ages 15-49, by Region 63 2.3 Distribution of Two Samples of Thai Men, by Type of Sexual Network, 1992 74 3.1 Prevention Programs in Cities That Have Kept HIV Infection among Injecting Drug Users below 5 Percent 116 3.2 Percentage of Women Ages 19 to 37 Infected with HIV, by Their Partners' Socioeconomic Status, Kigali, Rwanda 128 3.3 Relation between Education and HIV Status, Men and Women, Mwanza Region, Tanzania, and Rakai District, Uganda 129 3.4 Summary of the Assumptions before and after Interventions, STDSIM Modeling 141 3.5 Distribution of Countries by the Number of Sentinel Surveillance Sites and the Stage of the Epidemic, January 1995 155 3.6 Condom Social Marketing and Government Condom Distribution Programs, by Stage of the Epidemic 165 4.1 Annual Cost per Patient of Palliative Care and Treatment of Opportunisiic Illnesses, Sub-Saharan Africa and Thailand 177 4.2 Annual Cost of Antiretroviral Therapy, Thailand and the United Kingdom or the United States 180 4.3 Deaths per Thousand Adults Caused by a Constant Rate of HIV Infection 184 4.4 Evidence of Possible Crowding Out of HIV-Negative by HIV- Positive Patients, Six Countries, circa 1995 194 4.5 Asset Ownership in Households with and without an Adult Death 217 5.1 Average 1993 Donor-Funded HIV/AIDS Expenditures, by Stage of the Epidemic and Income Level 243 5.2 International AIDS Expenditures through Bilateral and Multilateral Channels by Major Donor Countries in 1993 and Net Immigration in 1992 260 xi CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 5.3 Responses to the AIDS Epidemic in Mexico: Government, NGOs, and Mass Media 275 6.1 Distribution of Developing Country Population by Stage of the Epidemic and Income 289 Appendix Tables (Appendix A) Selected Evaluations of Interventions To Prevent Transmission of HIV in Developing Countries 296 (Appendix B) B.1 Annual Costs per Infection Averted, per Condom, and per Contact for Interventions To Prevent HIV 303 (Statistical Appendix) 1 HIV Infection Rates by Subpopulation 310 2 Classification of Countries by Stage of the Epidemic, with Selected Economic and Policy Variables Affecting the Spread of HIV 318 3 Socially Marketed Condom Sales in Developing Countries, 1991-96 326 Box Tables 1.2 Opportunistic Illnesses Often Diagnosed in HIV-Infected People, Developing Countries 20 1.4 Worker Attrition in Ghana, Kenya, Tanzania, Zambia, and Zimbabwe, Total and by Sickness or Death, 1994 35 3.9 HIV Infections Averted by $1 Million Annual Spending on Prevention, U.S. Estimates 145 4.2 Effectiveness of Blood Transfusion at Averting HIV Infection, Uganda, 1993 189 4.9 Social Safety Net Programs in Which Prime-Age Adult Death Could Be Used as an Additional Targeting Criterion, Five Countries 234 5.1 Per Capita AIDS Expenditures Broken Down by Source of Financing in Four Countries and Sao Paulo State, Brazil 249 xli Foreword A IDS HAS ALREADY TAKEN A TERRIBLE HUMAN TOLL, NOT ,A,only among those who have died but among their families and communities. Short of an affordable cure, this toll is certain to rise. Ninety percent of HIV infections are in developing countries, where resources to confront the epidemic are most scarce. But the course of the epidemic is not carved in stone. This book argues that the global epidemic of HIV/AIDS can be over- come. National governments have unique responsibilities in preventing the further spread of HIV and in mitigating the impact of AIDS. But governments alone cannot overcome the epidemic, nor have they always risen to the task. Nongovernmental organizations and other groups in civil society, including people living with HIV, have played and must continue to play a critical role in shaping government action and in bringing prevention and care to people that governments cannot easily reach. The international community can also do much to support devel- oping countries and regions in financing programs to ensure prevention and improved equity in access to care. It can also support the production and dissemination of information worldwide, and invest in research on prevention approaches, vaccines, and low-cost, effective prophylaxis and treatment that can be used in developing countries. This report is itself an example of the potential benefits of interna- tional cooperation in response to the epidemic. The preparation of this volume by World Bank researchers has benefited greatly from the tech- nical inputs, advice, and financial support provided by the Joint United Nations Programme on HIV/AIDS (UNAIDS) and the European Commission. This research report makes a valuable contribution to the international debate on the role of government in addressing the AIDS epidemic in developing countries. The report's recommendations are those of the authors and do nor necessarily reflect the positions of our respective institutions. The world can overcome HIV Given the necessary information, means, and a supportive community, individuals can and do alter their CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC behavior to reduce the risk that they will contract and spread HIM But there are certain actions that only governments can take, and political commitment has often been lacking. The costs of inaction are poten- tially enormous. Policymakers who demonstrate commitment, by work- ing in creative ways with people most severely affected by HIV/AIDS, have a unique opportunity to contain a global epidemic and save mil- lions of lives. James D. Wolfensohn President The World Bank Joao de Deus Pinheiro Member of the European Commission Peter Piot Executive Director Joint United Nations Programme on AIDS xiv Introduction D EVELOPING COUNTRIES SIMPLY CANNOT IGNORE THE HIV/AIDS epidemic. According to UNAIDS, about 1.5 mil- lion people died from AIDS in 1996. Each day about 8,500 people, including 1,000 children, become newly infected. About 90 per- cent of these infections occur in developing countries, where the disease is likely to exacerbate poverty and inequality. But HIV/AIDS is not the only problem demanding government attention. In the poorest coun- tries especially, confronting AIDS can consume scarce resources that could be used for other pressing needs. How can developing country governments and the international community identify the public prior- ities in confronting this global epidemic? This book provides information and analysis to help policymakers, development specialists, public health experts, and others who shape the public response to HIV/AIDS to design an effective strategy for con- fronting the epidemic. It draws upon three bodies of knowledge: the epi- demiology of HIV; public health insights into disease control; and espe- cially public economics, which focuses on assessing tradeoffs in the allocation of scarce public resources. The report offers persuasive evidence that, for the 2.3 billion people living in parts of the world where the epidemic is still nascent, an early, active government response encouraging safer behavior among those most likely to contract and spread the virus has the potential to avert un- told suffering and save millions of lives. Even where the virus has spread widely in the general population, prevention among those most likely to contract and spread it is still likely to be the most cost-effective way to reduce infection rates. Of course, national governments are not alone in their fight against the epidemic. Bilateral and multilateral donors have provided leadership and major funding for national AIDS prevention programs, especially in the poorer developing countries, as well as for basic research for a vaccine and a cure. Local and international nongovernmental organizations have often assisted and sometimes led the fight against the epidemic. Govern- xv CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC ments can greatly improve the effectiveness of their response by working collaboratively with these actors. But only the government has the means and mandate to provide what economists call public goods. In the case of HIV/AIDS, these include in- formation about the distribution of infection and behaviors that spread it and knowledge of the costs and effectiveness of prevention and miti- gation programs. Similarly, governments have a unique responsibility to reduce the negative externalities from risky behavior, by encouraging safer behavior among those most likely to pass the virus to others. Although sound, these policies can be politically difficult. Indeed, be- cause the spread of HIV involves private behaviors that many people deplore--multiple sexual partners and injecting drug use-govern- ments that attempt to reduce the spread of HIV by these activities may appear to their constituents to condone immoral acts. Governments must make clear that the best way to protect everyone from HIV is to help people who engage in the riskiest behavior to avoid infection. Because resources are scarce, one must think through how best to al- locate them. The consequences of these decisions for particular individ- uals can be enormous. And there are painful dilemmas. In countries where HIV has spread widely, the epidemic will greatly increase the de- mand for health care and the need for poverty assistance. Governments of poor countries face the challenge of responding to the new needs of the AIDS-affected poor while not neglecting the needs of the poor who suffer from other illnesses and other causes of poverty. Drawing on the experience of countries that have faced these dilemmas, the report sug- gests responses that are both humane and affordable. ConfrontingAlDS: Public Priorities in a Global Epidemic is the sixth in a series of Policy Research Reports designed to bring findings of World Bank research on a key development issue to a wide audience. It is a product of the staff of the World Bank; the judgments made in the report do not necessarily reflect the views of the Board of Directors or the governments they represent. Joseph E. Stiglitz Senior Vice President and Chief Economist Development Economics The World Bank xvi The Report Team T HE PRINCIPAL AUTHORS OF THE REPORT WERE MARTHA Ainsworth and Mead Over. Nina Brooks and Samantha Forusz wrote many of the boxes, compiled the statistical appendix, and provided research assistance. Kathleen Mantila provided additional research assistance. Deon Filmer produced the material for box 3.3 and other results based on Demographic and Health Survey data. Tim Brown and Werasit Sittitrai contributed box 3.11. Eduard Bos gener- ated the projections for box 6.1. Julia Dayton and Michael Merson contributed appendix A. Lawrence MacDonald edited the report. The report was produced under the direction of Lyn Squire and Joseph Stiglitz. The editorial-production team was led by Deirdre T. Ruffino, with additional help from Paola Brezny, Joyce Gates, Audrey Heiligman, Brenda Mejia, and Anthony Pordes. Jeffrey Lecksell produced the maps in chapters 1 and 2. Grace 0. Evans provided support in production of the manuscript, with assistance from Thomas Hastings and Jim Shafer. Secretarial support was also provided by Joanne Fleming. xvii Acknowledgments T HIS REPORT BENEFITED FROM CLOSE COLLABORATION, technical review, and financial support from the European Commission HIV/AIDS Programme (EC) and by the Joint United Nations Programme on HIV/AIDS (UNAIDS). We would like to express particular appreciation and gratitude to Drs. Lieve Fransen of the EC and Stefano Bertozzi of UNAIDS for their exceptional per- sonal contributions in sponsoring workshops and review meetings, commissioning background studies, and reviewing the draft report. We are indebted to the European Commission HIV/AIDS Pro- gramme for sponsoring most of the background papers and an authors' workshop in Limelette, Belgium, in June 1996. Comments from the workshop participants were very helpful: Tony Barnett, David Bloom, Marijke Bontinck, Jean-Claude Deheneffe, Dominique Dellicour, Deon Filmer, Michel Garenne, Paul Gertler, Dik Habbema, King Holmes, Roberto lunes, Jose Antonio Izazola, Wattana Janjareon, Emmanuel Jimenez, Tony Kilouda, Tiekoura Kone, Sukontha Kongsin, Michael Kremer, Ajay Mahal, Allechi M'bet, Rekha Menon, Anne Mills, Martina Morris, Phare Mujinja, Amadou Noumbissi, I. 0. Orubuloye, Nicholas Prescott, Pamela Rao, Innocent Semali, Zmarak Shalizi, Donald Shep- ard, Lyn Squire, John Stover, Paula Tibandebage, Inge Van Den Bussche, Peter Way, Marc Wheeler, Alan Whiteside, and Debrework Zewdie. A complete list of the background papers is provided at the end of this report. The important technical contribution of the following UNAIDS ex- perts is gratefully acknowledged: Bai Bagasao, Michel Carael, Renu Chahil-Graf, Suzanne Cherney, Mark Connolly, Sally Cowal, Isabelle de Vincenzi, Jose Esparza, Purnima Mane, Peggy McEvoy, Rob Moodie, Joseph Perriens, Peter Piot, Joseph Saba, Bernhard Schwartlander, Wer- asit Sittitrai, and Country Programme Advisers who responded to the survey highlighted in chapter 3. We are also grateful to UNAIDS for sponsoring a review meeting on the first full draft of the report in xix CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Geneva, and to the following developing country policymakers who pro- vided extensive comments at that time: Akan Akanov (Kazakhstan), Papa Fall (Senegal), Mary Muduuli (Uganda); and Jaime Sepulveda (Mexico). In addition to the individuals mentioned above, many others inside and outside the World Bank provided valuable contributions or com- ments: Peter Aggleton, Sevgi Aral, Natalie Bechu, Seth Berkeley, Dorothy Blake, John Bongaarts, Kenneth Bridbord, Denis Broun, Tim Brown, Richard Bumgarner, Tony Burton, Anne Buve, Julia Dayton, David de Ferranti, Jacqueline Dubow, Richard Feachem, Steven Forsythe, Mark Gersovitz, Ronald Gray, Jacques du Guerny, Salim Habayeb, Jeffrey Hammer, David Heymann, Philip Harvey, Richard Hayes, Estelle James, Dean Jamison, Prabhat Jha, Christine Jones, Arata Kochi, Kees Kostermans, Maureen Lewis, Samuel Lieberman, Bernard Liese, Georges Malempre, Jacques Martin, Raymond Martin, Clyde McCoy, Tom Merrick, Michael Merson, David Metzger, Norman Miller, Susan Mlango, Stephen Moses, Philip Musgrove, Jeffrey O'Mal- ley, Junko Otani, Cheryl Overs, David Paltiel, Lant Pritchett, Hnin Hnin Pyne, Bill Rao, Wendy Roseberry, Lewis Schrager, Thomas Selden, Guy Stallworthy, Karen Stanecki, Daniel Tarantola, Kitty Theurmer, Anne Tinker, Dominique van de Walle, Carina Van Vliet, Maria Wawer, Roger Yeager, and Fernando Zacarias. The opinions and conclusions expressed in this report are nevertheless those of its authors and do not necessarily reflect positions of the World Bank, its member governments, or other collaborating or sponsoring institutions. The financial assistance of the governments of Australia and Switzer- land is gratefully acknowledged. xx Definitions Data Notes H ISTORICAL DATA IN THIS BOOK MAY DIFFER FROM THOSE in other World Bank publications if more reliable data have become available, if a different base year has been used for constant price data, or if countries have been classified differently. The former Zaire is referred to as the Democratic Republic of the Congo (Congo DR), and Hong Kong (China, as of July 1, 1997), is sometimes referred to as Hong Kong. *All dollar ($) amounts are current U.S. dollars unless otherwise specified. * Billion is a thousand million Abbreviations and Acronyms AIDS Acquired immune deficiency syndrome AIDSCAP AIDS Control and Prevention Project CSM Condom social marketing (see Glossary, below: Social mar- keting of condoms) DALY Disability-adjusted life year DHS Demographic and Health Survey DOTS Directly observed treatment short course (for tuberculosis) EC European Commission FSU Former Soviet Union GAPC Global AIDS Policy Coalition GDP Gross domestic product (see Glossary, below) GNP Gross national product (see Glossary, below) GPA Global Programme on AIDS CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC HIV Human immunodeficiency virus IDU Injecting drug user IEC Information, education, and communication MSM Men who have sex with men (see Glossary, below) NEP Needle exchange program NGO Nongovernmental organization (see Glossary, below) OECD Organization for Economic Cooperation and Development ODA Overseas Development Administration (U.K.) 01 Opportunistic infection PAHO Pan American Health Organization PCP Pneumocystis carinii pneumonia PSI Population Services International (private firm) SOMARC Social Marketing for Change (private firm) STD Sexually transmitted disease TB Tuberculosis UNAIDS Joint United Nations Programme on HIV/AIDS UNDP United Nations Development Programme UNESCO United Nations Educational, Scientific, and Cultural Organization UNFPA United Nations Population Fund UNICEF United Nations Children's Fund USAID U. S. Agency for International Development wHO World Health Organization WHO/GPA World Health Organization/Global Programme on AIDS Glossary Adverse selection: the selection into an insurance pool of people likely to have higher claims than others. Assortative sexual mixing: the extent to which people with similar num- bers of sexual partners pair with each other. Asymptomatic: infected by a disease agent but exhibiting no medical symptoms; subclinical. Commercial sex: the selling of sexual services for compensation; prostitution. xxii DEFINITIONS Concurrentpartnership: partnerships that overlap in time. Disassortative sexual mixing: the extent to which people with many sexual partners pair with people with few partners. Discordant couple: a couple in which one partner is infected with HIV and the other is not. Endemic: usually prevalent; persistent at relatively constant levels. Epidemic: a sudden unusual increase in cases that exceeds the number expected on the basis of experience. Epidemic, concentrated: an HIV epidemic in a country in which 5 per- cent or more of individuals in groups with high-risk behavior, but less than 5 percent of women attending urban antenatal clinics, are infected. Epidemic, generalized: an HIV epidemic in a country in which 5 percent or more of women attending urban antenatal clinics are infected; in- fection rates among individuals in groups with high-risk behavior are also likely to exceed 5 percent in countries with a generalized HIV epidemic. Epidemic, nascent: an HIV epidemic in a country in which less than 5 percent of individuals in groups with high-risk behavior are infected. Epidemiology: the study of the distribution and determinants of disease and injury in human populations. Externality: an unpriced side effect on a third party of a transaction between two parties. Gross domestic product: a crude measure of national economic well- being: aggregate expenditure by the residents of a country or final goods and services for consumption, investment, and government services. x0dii CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Gross nationalproduct: an alternative to gross domestic product for mea- suring national economic well-being. Adds to the gross domestic product income obtained by nationals from labor or property outside the country and deducts the income of foreign nationals residing in the country. High-risk behavior: unprotected sexual intercourse (i.e., without a condom) with many partners, or sharing of unsterilized injecting equipment. HIV-positive: having antibodies to HIV Incidence of HIV: the number of new cases of HIV in a given time pe- riod, often expressed as a percentage for a given number of the sus- ceptible population. Low-risk individuals: individuals practicing behavior that puts them- selves and their partners at low risk of HIV infection; depending on the extent to which they mix with high-risk individuals, however, they may nevertheless be at high risk of becoming infected. Men who have sex with men: homosexual, bisexual, and heterosexual men who have sex with other men. Meritgood: a good (or service) whose consumption by the poor is valued by society as a whole. Moral hazard.' the increase in the average loss incurred by people who are insured compared with those who are not; term used by health insur- ance companies to refer to the increased demand for health care exhibited by the insured. Nongovernmental organization: for-profit firms and private nonprofit organizations. Opportunistic illness: an illness that affects people with weak immune systems. Pandemic: an epidemic occurring simultaneously in many countries. XXv DEFINITIONS Prevalence of H1V: the number of people with HIV at a point in time, often expressed as a percentage of the total population. Public good: a good or service having the following two attributes: (1) consumption by one person does not diminish the amount available to others, and (2) excluding people from consuming the good is impossible or costly. Reproductive rate: the average number of susceptible people infected by an infected person over his or her lifetime. Seroprevalence: the prevalence of an infection as detected in blood serum. Sex worker: someone who offers sexual services for money. Social marketing of condoms: programs designed to raise condom use by improving the social acceptability of condoms, making them more widely available through nontraditional outlets and offering them for sale at subsidized prices. Susceptible: vulnerable to becoming infected. Symptomatic: exhibiting sufficient symptoms to require medical treatment. xxv Summary WO DECADES AFTER THE APPEARANCE OF THE human immunodeficiency virus (HIV), an esti- mated 30 million people have contracted the virus, and 6 million have died of acquired immune defi- ciency syndrome (AIDS). About 90 percent of infec- ~~~~tions occur in developing countries, where the dis- ease has already reduced life expectancy, in some cases by more than a decade. HIV is already widespread in many countries in Sub-Saharan Africa and may be on the verge of exploding in other regions. Because most people who develop AIDS are adults in the prime of life, the dis- ease exacts a heavy toll on surviving family members, especially children, and may exacerbate poverty and inequality. Clearly, the human toll of the epidemic is great. But low-income countries face a multitude of pressing human needs. How should developing country governments and the international community respond? In answering this question, Confronting AIDS: Public Priorities in a Global Epidemic draws on three bodies of knowledge: the epidemiology of HIV; public health insights into disease control; and especially public economics, which focuses on assessing tradeoffs in the allocation of scarce public resources. In relying primarily on public economics, we do not intend to deny the validity of other points of view. Much has been written on the epidemic from the perspectives of public health, medical science, and human rights. This Policy Research Report differs by ad- dressing the epidemic in a way that fits more dosely the perspective of decisionmakers outside the health sector who shape national efforts to combat the disease. To this audience, the report argues that AIDS is a large and growing problem and that governments can and should con- front the epidemic by preventing new infections and by mitigating the impact of infections that occur. It finds that some policies will be much more effective than others in reducing the spread of HIV and CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC mitigating its impact, and it provides a framework that helps to distin- guish among activities that can be undertaken by households and the private sector, including nongovernmental organizations (NGOs), those that should be initiated by developing country governments, and those that should be most strongly supported by the international develop- ment community. Although there are clear arguments in favor of government interven- tion to slow the spread of HIV, social norms and politics make AIDS policy uniquely challenging. This is especially true during the early stages of the epidemic, when the advantages of government intervention are greatest but the potential severity of the problem is not yet apparent. The report argues that governments have a mandate to endorse and sub- sidize risk-reducing preventive interventions, especially among those most likely to contract and spread HIV, while protecting them from stigmatization. This report is a strategic document. It has been written to inform and motivate political leaders, policymakers, and development specialists to support the public health community, concerned civil society, and peo- ple living with HIV in confronting the AIDS epidemic. Some readers will already know a great deal about public policy and HIV/AIDS; oth- ers may be considering the disease from a policy perspective for the first time. It is just as relevant for countries in the earliest stage of the epi- demic as it is for those that have suffered the ravages of the disease for more than a decade. Although the report offers examples of programs from many countries, some of which have worked remarkably well, it is not intended as a how-to guide for designing and implementing specific programs. There are many other sources of such information, and sum- marizing them is beyond the scope of the report. Rather, the report of- fers an analytical framework for deciding which government interven- tions should have high priority for addressing the HIV/AIDS epidemic in developing countries and, based on that framework, advocates a broad strategy that can be adapted by countries according to their resources and the stage of their epidemic. Chapter 1 AIDS: A Challenge to Government T HIS CHAPTER PROVIDES BASIC INFORMATION ABOUT THE nature of HIV/AIDS, the extent of the epidemic, and its cur- rent and likely future impact on such measures of well-being as 2 SUMMARY life expectancy, health, and economic growth. Because AIDS strikes adults in their economic prime and, despite recent medical advances, is almost always fatal, the disease reduces average life expectancy (some- times dramatically), increases the demand for medical care, and is likely to exacerbate poverty and inequality. The relationship between eco- nomic development policies and HIV is complex: cross-country data and other evidence indicate that the AIDS epidemic is likely to both affect and be affected by economic development. Nevertheless, policymakers have often been reluctant to intervene. Faced with competing demands for scarce public resources, and aware that HIV/AIDS is spread primarily through private sexual and drug- injecting behavior, governments may conclude that the disease is not a public priority. Drawing on well-accepted principles of the role of gov- ernment, which have been the subject of the discipline of public eco- nomics, the chapter explains why governments must be actively involved in the fight against AIDS. Starting from the view that government has a mandate to advance economic well-being and to promote a fair distribution of society's out- put, the chapter applies public economics to argue that government can- not leave the battle against HIV/AIDS to the private sector. First, in countries that wish to subsidize most of the cost of health care, AIDS will generate enormous government health care expenditures; this alone is sufficient justification for early, effective prevention. Second, whenever a transaction between two parties imposes negative effects, or externali- ties, on a third party, as is the case when a sexual encounter between two people increases the risk of HIV infection to their other partners, public economics argues for government intervention. Third, the provision of information about the state of the epidemic or about the effectiveness of alternative remedies meets the economist's definition of a public good; that is, something that benefits society but that private entrepreneurs have insufficient incentive to produce on their own. Public economics argues that governments can often enhance the welfare of society by ensuring the adequate provision of such services. Fourth, fairness and compassion for the poor warrant a government role in both preventing and mitigating the epidemic. Finally, governments influence social norms and promulgate legislation that affect the rights of both the HIV- infected and the uninfected. Measures that protect the powerless from prejudice, bigotry, and exploitation will simultaneously help to protect everyone from the AIDS epidemic. 3 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Chapter 2 Strategic Lessons from the Epidemiology of HIV IN SOME COUNTRIES, HIV HAS INFECTED ONLY A TINY PERCENT- age of the population and its effects are all but invisible; in others the virus has spread so widely that few families have been spared the tragedy of AIDS illness and death. What accounts for these differ- ences? In reviewing how HIV spreads in populations and the behav- ioral and biological factors behind the epidemic, this chapter identifies important principles for an effective response, based on the epidemiol- ogy of HIV. These provide the foundation for considering government priorities for preventing the spread of HIV (chapter 3). In order for HIV to sustain itself in a population, an infected person must, on average, transmit the virus to at least one other person over his or her lifetime. Both biological and behavioral factors affect the rate of spread of HIV through the population. The key biological factors in- clude the long asymptomatic period of HIV, the risk of infection per contact by different modes of transmission, and cofactors, such as infec- tion with other sexually transmitted diseases (STDs). However, HIV transmission can be slowed dramatically by changes in behavior: reduc- ing the number of sexual and drug-injecting partners, using condoms during sexual intercourse, and using sterilized injecting equipment. Until there is a vaccine or cure affordable to developing countries, the most ef- fective way to arrest the epidemic will be by enabling individuals to re- duce the risky behavior that may lead to their infection and the spread of HIV. The specific measures that can be taken to reduce risky behavior at both the individual and societal levels are discussed in chapter 3. The epidemiology of HIV/AIDS suggests two important objectives for public programs to slow and stop the spread of HIV: Act as soon as possible. Nearly half of the world's population lives in areas where HIV is rare, even among people whose behavior might put them at high risk of infection. By investing in prevention when few peo- ple are infected with HIV, before AIDS becomes a significant health issue, governments can contain the epidemic at relatively low cost. Even in countries where the virus has already spread widely, effective preven- tion now can save the lives of many people who would otherwise have become infected. Prevent infection among those most likely to contract and spread HIV. Not everyone in the population who contracts HIV is equally likely to spread it to others. People with the highest number of partners and the 4 SUMMARY lowest levels of protective behavior (such as use of condoms and of ster- ile injecting equipment) are the most likely to contract and inadvertently to spread HIV Each case of HIV infection directly prevented among people who practice these high-risk behaviors will indirectly prevent many secondary infections in the rest of the population-a kind of "multiplier" effect. Others in the population who practice lower-risk behavior by having few partners, consistently using condoms, or using sterilized injection equipment are unlikely to spread HIV, even if they contract HIV themselves. The likelihood that an individual will contract and spread HIV is determined by the level of the individual's risk behav- ior. Behavioral studies show that observable individual characteristics, such as occupation, age, or sexual orientation, can partially predict risk behavior and therefore can be useful in guiding prevention efforts. How- ever, those with the riskiest behavior vary from country to country and over time. For example, sex workers have large numbers of sexual part- ners and, if they do not use condoms, are among those who are highly likely to contract and inadvertently spread the virus. However, in places where condom use in commercial sex has become the norm, others may be more likely to contract and spread HIV. The chapter concludes with an overview of the level and distribution of HIV in developing countries, by region. In countries with "nascent" epidemics, HIV prevalence is very low, even among people whose be- havior would put them at high risk of contracting it. In countries with "concentrated" epidemics, HIV has risen to high levels among those practicing the riskiest behaviors and is set to spread more widely in the rest of the population. In countries with "generalized" epidemics, HIV prevalence is high even among those whose behavior is unlikely to spread HIV to others. The stage of the epidemic has important implications for government priorities in preventing the spread of HIV; these are dis- cussed in chapter 3. Chapter 3 Efficient and Equitable Strategies for Preventing HIV C AN PUBLIC POLICY AFFECT THE VERY PRIVATE BEHAVIORS that spread HIV? If so, what course of action should govern- ments pursue as a priority to have the largest impact? This chapter addresses these two key issues. 5 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Despite the private nature of the behaviors that spread HIV, govern- ments do have options for influencing decisions among those most likely to contract and spread the virus. Public policy can directly influence in- dividual high-risk behavior, either by lowering the "costs" of safer behav- ior (for example, by subsidizing information of various types, condoms, and access to clean injecting equipment) or by raising the "costs" of be- havior that can spread HIV (for example, by attempting to restrict pros- titution or the use of injected drugs). The chapter highlights examples of successful programs of the first type. Although the second approach is sometimes politically appealing, enforcement actions can exacerbate the epidemic by making it harder to reach those most likely to contract and spread the virus and encourage them to adopt safer behavior. An important complementary approach is to promote behavioral change indirectly through policies that remove social and economic con- straints to adopting safer behavior. One set of activities involves promot- ing social norms conducive to safer behavior, including improving the social acceptability of condoms. A second set aims to improve the status of women, whose lower social and economic status reduces their ability to insist upon sexual fidelity and to negotiate safe sex. These policies include those to expand female education and employment opportuni- ties; to guarantee basic inheritance, property, and child custody rights; and to outlaw and severely punish slavery, rape, wife abuse, and child prostitution. Finally, policies that reduce poverty will ease the economic constraints faced by the poor in paying for essential HIV prevention ser- vices, such as STD treatment and condoms. Many of these actions address fundamental development objectives and have numerous other benefits besides slowing the spread of HIV Their benefits are sometimes difficult to quantify, but they are highly complementary to policies that directly affect the costs and benefits of risky behavior. What prevention strategy should governments pursue to have the maximum impact with limited resources? In keeping with the principles of public economics, governments should either ensure financing for or implement directly those interventions that are essential to stopping the spread of HIV but that private individuals or firms would not have suf- ficient incentive to pay for on their own. As noted in chapter 1, three major areas in which this is likely to be the case are reduction of the neg- ative externalities of risky behavior, provision or regulation of public goods, and protection of the poor from HIV infection. Programs that address these issues will improve the efficiency and equity of government 6 SUMMARY prevention efforts. In addition, following the principles of epidemiology discussed in chapter 2, program effectiveness will be improved if govern- ments act as soon as possible and if they succeed in preventing infection among those most likely to contract and spread HIV. Thus both public economics and epidemiological principles argue strongly for giving pri- ority to measures that prevent infection among those most likely to con- tract and spread HIV The effect of specific program components may be direct or indirect and their impact immediate or long term, but their effectiveness in slowing the epidemic will depend on the extent to which they contribute to this goal. These recommendations are not meant to limit the scope of government involvement if there are ample resources and public will to undertake even more. Rather, the intention is to point out the minimum set of activities that all governments should engage in to improve the efficiency and equity of prevention programs, and a ra- tional order in which to expand activities if resources permit. Governments have many tools for implementing this strategy, such as direct provision of services, subsidies, taxes, and regulatory powers. Meeting any one objective will often require a combination of comple- mentary interventions. To maximize the impact of scarce resources, pub- lic prevention programs should avert as many secondary HIV infections as possible per public dollar spent. Furthermore, priority should be given to interventions that augment (not substitute for) private sector services. HIV prevention programs often have considerable benefits for society beyond those of preventing the epidemic; these benefits and the syner- gies between interventions and policies should be taken into considera- tion in evaluating costs and benefits. Some interventions, such as repro- ductive health and HIV/AIDS education in schools, offer widespread social benefits in addition to benefits for HIV prevention, are inexpen- sive, and are therefore often a sound investment. Programmatic targeting criteria are imperfect, and reaching people at high risk of contracting and spreading HIV can be difficult. The cost-effectiveness of government programs for HIV prevention often can be improved by working with NGOs and those severely affected by the epidemic in the design and im- plementation of programs. This broad prevention strategy based on epidemiology and public economics offers guidance for countries at all stages of the epidemic. For example, both epidemiology and the need to reduce negative externali- ties of high-risk behavior argue for heavily subsidizing safer behavior among those most likely to contract and spread HIV. This action alone 7 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC may be sufficient to dramatically slow the spread of a nascent epidemic. In countries with concentrated and generalized epidemics, preventing HIV among those with the highest chances of contracting and spreading the virus is still essential to slowing the epidemic and is likely to be highly cost-effective. However, in addition, behavioral change among others who practice risky behavior will be necessary to reverse the course of the epidemic. As the epidemic spreads, the cost-effectiveness of prevention among those who practice moderately risky behavior increases. With respect to the equity of HIV prevention programs, in areas where HIV has not yet spread widely, governments can protect the poor best by tak- ing appropriate early action to prevent an epidemic. In countries with generalized epidemics, governments can ensure that the poor have access to the knowledge, skills, and means to prevent HIV While the chapter identifies some basic principles underlying an effi- cient and equitable national strategy for preventing the spread of HIV, it remains for individual countries to identify the specific combination of programs, policies, and interventions to pursue this strategy in a cost- effective way. Program choices are necessarily country-specific because the costs and effectiveness of interventions are likely to vary widely across settings, depending on factors such as the stage of the epidemic, under- lying patterns of sexual and drug-injecting behavior, social and economic constraints on safe behavior, local costs, and implementational capacity. The characteristics and accessibility of those most likely to contract and spread HIV are also highly country-specific. To what extent are governments already pursuing the strategy sug- gested by this chapter? Many developing countries have launched HIV prevention programs, representing a constellation of interventions, but little is known about the extent to which they collectively have reached those at highest risk of contracting and spreading HIV and enabled them to adopt safer behavior. A review of the limited evidence found the following. First, basic data on the patterns of HIV infection and sexual behavior, essential for making sensible decisions about allocating resources among alternative preventive interventions, are deplorably scarce. Many govern- ments, particularly those with nascent or undocumented epidemics, need to expand their collection and analysis of data about HIV infection levels in various groups and about the nature and extent of behavior pat- terns that could spread the virus. This information is essential for estab- lishing an operational definition of those most likely to contract and 8 SUMMARY spread HIV In countries with concentrated or generalized epidemics, governments need to ensure that costs and effects of interventions are more closely tracked to improve the cost-effectiveness of prevention. Second, despite the best efforts to date, programs to change the be- havior of those most likely to contract and spread HIV reach too few of them. Few national programs appear to have systematically assessed the coverage of government and NGO prevention programs--that is, the proportion of people most likely to contract and spread HIV who are reached by prevention interventions. Occupational groups such as the military and police, whose members in many places have more sexual partners on average than the rest of the population, are relatively easy and inexpensive for government to reach. Yet programs to provide mem- bers of these groups with condoms and prevention information are often lacking or inadequate. Finally, the effectiveness of government programs in ensuring access to prevention for the poor has rarely been evaluated. For example, the so- cial marketing of condoms (promoting the sale of subsidized condoms) has been very effective in increasing condom use. However, the extent to which these programs are benefiting the poor, are raising condom use among those with the highest rates of partner change, and are supple- menting rather than crowding out private condom supply has not been established. Taken together, chapters 2 and 3 argue that the effectiveness of gov- ernment HIV prevention programs depends critically on the extent to which they reduce the risk behavior of those most likely to contract and spread HIV. Chapter 3 concludes that the greatest impediments to im- proved effectiveness of government HIV prevention programs are the lack of political will: first, to collect the data on HIV prevalence, risk be- havior, and cost-effectiveness necessary to mount effective programs and, second, to work constructively with those most likely to contract and spread HIV Chapter 4 Coping with the Impact of AIDS W H HILE SOME COUNTRIES STILL HAVE THE OPPORTUNITY to avert a full-scale AIDS epidemic, others already find themselves facing the consequences of widespread HIV 9 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC infection. What can be done that is effective and affordable to help people with AIDS in developing countries? What will be the conse- quences of AIDS morbidity and mortality for health systems and pov- erty? And what can society and governments do to mitigate those impacts? These are the three issues addressed in chapter 4. The first and most basic impact of HIV/AIDS is on those who con- tract the disease. The chapter discusses how medication to relieve symp- toms and treat opportunistic infections can ease suffering and prolong the productive lives of people with HIV, sometimes at low cost. But as the immune system collapses, available treatments become increasingly expensive and their efficacy less certain. Antiretroviral therapy, which has achieved dramatic improvements in the health of some individuals in high-income countries, is currently unaffordable and too demanding of clinical services to offer realistic hope in the near term for the millions of poor people infected in developing countries. An analysis of alternative treatment and care options concludes that community-initiated care provided at home, while often shifting costs from the national taxpayer to the local community, also greatly reduces the cost of care and thereby offers hope of affordably improving the quality of the last years of life of people with AIDS. Second, the epidemic will increase demand for medical care and re- duce its supply at a given quality and price. As the number of people with HIV/AIDS mounts, access to medical care will become more diffi- cult and more expensive for everyone, including people not infected with HIV, and total health expenditure will rise. Governments will likely be pressured to increase their share of health care spending and to pro- vide special subsidies for the treatment of HIV/AIDS. Unfortunately, because of the scarcity of resources and the inability or unwillingness of governments to increase public health spending enough to offset these pressures, either of these policies may exacerbate the impact of the epi- demic on the health sector and may make it more difficult for the ma- jority who are not infected with HIV to obtain care. However, there are things that governments can do. Governments should ensure that HIV- infected patients benefit from the same access to care as other patients with comparable illnesses and a similar ability to pay. Sometimes, be- cause of discrimination, people with HIV are denied treatment or face barriers to care that others do not encounter. In other situations, people with HIV receive subsidized access to advanced therapies while people sick with other severe and difficult-to-treat diseases lack comparable ac- I0 SUMMARY cess to therapies of similar cost. Although patients with HIV-related ill- nesses need and should receive a different mix of services than those with, say, cancer, diabetes, or kidney disease, they should pay the same percentage of their health care costs out of their own pockets as would patients with other diseases. Other measures that governments can and should undertake include providing information about the efficacy of al- ternative treatments for opportunistic illnesses and AIDS, subsidizing the treatment of STDs and infectious opportunistic illnesses, subsidizing the start-up of blood safety and AIDS care programs, and ensuring access to health care for the poorest, regardless of their HIV infection status. The third major impact of the epidemic is on households and, in the aggregate, on the extent and depth of national poverty. Households and extended families cope as best they can with the loss of prime-age adults to AIDS. They reallocate their resources, for example, by withdrawing children from school to help at home, working longer hours, adjusting household membership, or selling household assets, and they draw on their friends and relatives for cash and in-kind assistance. Poorer house- holds, having fewer assets to draw on, have more difficulty coping. Their children may be permanently disadvantaged by worsening malnutrition or withdrawal from school. However, in responding, governments and NGOs should not forget that low-income countries have many poor households that have not experienced an AIDS death but are nonetheless so poor that their children suffer similar disadvantages. At the same time, some households will have enough resources to cope with an adult death without government or NGO assistance. The government's equity objec- tive will thus typically be served more effectively by targeting assistance based on both direct poverty indicators and the presence of AIDS in the household, rather than on either indicator alone. The chapter closes with specific recommendations to ensure that available resources reach the households that most need help by coordinating targeted poverty reduc- tion efforts with programs to mitigate the impact of the epidemic. Chapter 5 Working Together To Confront HIV/AIDS N ATIONAL GOVERNMENTS BEAR THE RESPONSIBILITY FOR protecting their citizens from the spread of the HIV epidemic and for mitigating its worst effects once it has spread. But CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC they are not alone in the effort. Bilateral and multilateral donors have provided both leadership and major funding for national AIDS preven- tion programs, especially in low-income developing countries. Local and international NGOs have stepped forward to help, and sometimes to prod reluctant governments into action. The challenge for national governments is to define their role in the struggle against the epidemic in collaboration with these other actors. This chapter turns from specific national policies to the strategic roles played by various actors in the policy arena. First, it examines the roles that national governments and donors have played in financing AIDS policies within developing countries, arguing that the governments of many low-income countries should confront the epidemic more force- fully, both directly and in collaboration with NGOs. Many types of NGOs are potential and actual contributors to this effort, including for- profit and nonprofit firms, broad-based private charities, and "affinity groups" of those affected by HIV/AIDS. Second, the chapter argues that, despite their substantial contributions to combating the epidemic, bilateral donors and multilateral organizations have invested too little in international public goods, including knowledge about prevention ap- proaches and treatment methods and research on a vaccine that will work in developing countries. Furthermore, both bilateral and multilat- eral donors have a responsibility to coordinate their activities more effec- tively at the country level. Finally, the chapter discusses how public opin- ion and politics shape AIDS policy and how developing country governments can listen to and work with a variety of partners to mini- mize and overcome the obstacles to sound policies for fighting AIDS. Chapter 6 Lessons from the Past, Opportunities for the Future T HE FINAL CHAPTER SUMMARIZES THE MAIN POLICY RECOM- mendations of the report and discusses opportunities for coun- tries to change the course of the epidemic at various stages. For on-line information about the economics of HIV/AIDS, visit http.//www. worldbank. org/aids-econl. I2 CHAPTER 1 AIDS: A Challenge for Government ORE THAN A DECADE AFTER THE HUMAN immunodeficiency virus (HIV) was first identi- fied as the cause of acquired immune deficiency syndrome (AIDS), the disease has been reported in nearly all developing and industrial coun- tries.1 UNAIDS, the United Nations joint pro- gram dedicated to combating the AIDS epidemic, estimates that at the end of 1996 about 23 million people worldwide were infected with HIV and more than 6 million had already died of AIDS. More than 90 per- cent of all adult HIV infections are in developing countries (figure 1.1) About 800,000 children in the developing world are living with HIV; at least 43 percent of all infected adults in developing countries are women (AIDSCAP and others, 1996). In many developing countries the HIV/AIDS epidemic is spreading rapidly. In major cities of Argentina, Brazil, Cambodia, India, and Thai- land, more than 2 percent of pregnant women now carry HIV. These levels are similar to those found ten years ago in such African countries as Zambia and Malawi, where more than one in four pregnant women are now infected. In two African cities, Francistown, Botswana, and Harare, Zimbabwe, 40 percent of women attending antenatal clinics are infected. Figure 1.2 presents UNAIDS estimates of the number of new adult infections by region and over time. While new infections are thought to be leveling off in Sub-Saharan Africa as a whole, in some countries military conflict and civil unrest may be spreading the epi- demic. Meanwhile, the disease is spreading rapidly in Asia. Extrapolation of the trends in figure 1.2 leads some observers to think that Asia may al- ready have surpassed Africa in the number of new infections per year. In Latin American and the Caribbean countries the number of new infec- I3 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 1.1 Estimated Number of Adults with HIWAIDS, by Region, 1997 .NORTH AMERICA WESTERN EUROPE CENTRAL ASIA 750000 510.000 so000 EASTERNASIA, SOUTH ASIA & THE PACIFIC NORTH APRICA A CARUBEAN EMDIAT URImBA 1.111 IlA^SS^SxA&£IA 270,000 200.000 LATIN AMERICA SUB-SAHARAN AFRICA ; 1UJ 300 f*** f0 00,0 10 A£A{AAAAAAKAAAAAAAAlAA^AAAAiA^tAA£AA££ AUSTRAUA& WHIMMI*Si*il* fl11l^£ NEW ZEALAND 13,000 )C- w.1,pxrmon! R.inq .A.HltIG; A;SWA£t£p£HSDt£S£( A tAE 14,0001000 0 Source: UNAIDS data 1997. tions has been steady at about 200,000 per year for several years, while the countries of eastern Europe and central Asia are experiencing the ini- tial stages of rapid spread (not shown). Only in North America and west- ern Europe has the number of new infections declined from its peak in 1986, but even here the future of the epidemic is unclear as it invades lower-income populations whose education and access to health care more closely resemble those of the developing world. AIDS is clearly taking an immense and growing human toll. The dis- ease is catastrophic for the millions of people who become infected, get sick, and, in stark contrast to the recent hopeful news of treatment break- throughs, die. It is also a tragedy for their families, who, in addition to suffering profound emotional loss, may be impoverished as a result of the disease. Because AIDS kills mostly prime-age adults, it increases the num- I4 AIDS: A CHALLENGE FOR GOVERNMENT Figure 1.2 New Adul HIV Infections, by Region, 1977-95 1,600,000 1,400,000 - Sub-Saharan Africa - Asia 1,200,000 1,200-000 Latin America and the Carribean 1,000,000 North America and Western Europe 8000 / * - - - Middle East and North Africa 600,000 600,000// 400,000 200,000 T -- lhe number of new HIV infections has started to decline in Africa, but is 0 osing rapdly in Asia. 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 Year Source: UNAIDS 1997. ber of children who lose one or both parents; some of these orphans suf- fer permanent consequences, due to poor nutrition or withdrawal from school. Numbers cannot begin to capture the suffering caused by the dis- ease. Each infection is a personal tragedy; box 1.1 describes the experi- ence of one of the nearly 30 million people who have contracted HIV. AIDS is not alone in causing human suffering, however. In low- income countries in particular, many urgent problems compete for scarce skills and resources. In the year 2000, malnutrition and childhood diseases that can be prevented or treated much more easily than AIDS are expected to kill 1.8 million children in the developing world; tuber- culosis (TB) is expected to kill more than 2 million people; and malaria, about 740,000. Worldwide, annual deaths from smoking are expected to increase from 3 million in 1990 to 8.4 million in 2020, and nearly all of this annual increase is expected to occur in developing nations (Murray and Lopez 1996).2 And disease is only one of many problems facing governments in improving the welfare of their citizens. About a billion people lack access to clean water, and about 40 percent of women and one-quarter of men in the developing world are illiterate. Throughout the world, inadequate transportation and communications hinder the efforts of billions of people to improve their lives. I5 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 1.1 Pauline: One Woman's Story P'AL'LINF. THE 'iJLN(ANSf,F OF SEVEN CHILDREN 'Alter three months I had enough money and in a farm family in Ghana. was in her early twentics came home. Thai %%as tvo years ago. After a vear a sshen an older cousin promised her w%ork as a wait- hoil grT,w under m! arm." ress in Abidjan, Ctoe d'Ivoire. and oatered to lend Pauline's sister took her to a herbalist who solId her the bus F'are. Because she was unable to make hcr a potion. \XThen chat failed ro rclie,e her worsen- enough moncy mo support hicrself trading fish near ing s'mptoms shc wvent [o a private doctor and %%as her home, she readily accepted. admirted to a ho'pital. %%here she remained for three "WX hen I got there I tound there was no job as a months. The doctors did not tell her that shc had w%aitress. Nis cousin -aid I must work as a pTrostitue ADS. on.l that he 'must nor go with mnen. to pay back the bus tare. I lived in a house %- th ses- 1 never used a condom %ihile I %%as in Abidjan. era] other women doing the s;ame w-ork. Men never aslked F'or them. I never even heard ot "I did the business in the bars. There %%ere so AIDS until I returned home and met t'riends s:ho many other women there I couldii't counti them. had it." she said. Some days I had ibout four or tfie men-the num- Ar the time oF the iner' jew. in l1Wil. Pauline ber depends on %our beaun-. But I didn't likc the %sas 'en' thin. had septic ulcers on her chesrt and work. so I did it until I had enough For m- tfood and shoulders. and suft'ered trom consiant itching on her rent and then I'd stop for a Few days. arms and legs. Shc is one of the six million people -'Ize You didn't pa" Your rent the landlord %%ould w%ho hate so far died oFAIDS. seize your belongings and thro%% you our. I onl' wanted ro sa%c up enough income to come home. Norunr. Himn.pEn VI)Z. Given these many other pressing problems, how much time, effort, and resources should governments devote to fighting AIDS? Views dif- fer widely. Some people consider AIDS to be a late-twentieth-century version of the Black Death, which devastated Europe in the fourteenth century. According to this view, governments should do everything in their power to slow the epidemic. Others think that governments should do little or nothing, either because they think AIDS will not be a major problem in their country or because they think that governments are powerless to change the private behavior that spreads the virus. Most people would probably agree that governments should attempt to do something. But even among those who share this opinion, there are many views about which actions should be regarded as public priorities. This book is addressed to policymakers, development specialists, pub- lic health personnel, and others in a position to influence the public response to HIV/AIDS. It provides an analytical framework for consid- ering how society generally and government in particular should con- i6 AIDS: A CHALLENGE FOR GOVERNMENT front the epidemic. In doing so, it draws upon three bodies of knowl- edge: the epidemiology of HIV; public health insights into disease con- trol; and especially public economics, which focuses on tradeoffs in the allocation of scarce public resources. The report argues that AIDS is a large and growing problem and that governments can and should actively confront the epidemic. It finds that some policies will be much more effective than others, and it distin- guishes between activities that can be undertaken by households and the private sector, including nongovernmental organizations (NGOs); those that should be initiated by developing country governments; and those that should be most strongly supported by donor governments and the international development community. This chapter provides basic information that the remainder of the book will draw upon to analyze government roles in preventing and mit- igating the HIV/AIDS epidemic. Following a short summary of the biology of HIV and how it spreads, we discuss the impact of AIDS on life expectancy and health and compare this with other health threats. We then set AIDS in the context of development, showing that the epi- demic both affects and is affected by many aspects of economic growth. Drawing on this material, we analyze the various rationales for govern- ment involvement in confronting AIDS; this analysis will provide im- portant guidance in identifying public priorities in the global HIV/AIDS epidemic addressed in later chapters. The chapter concludes with a discussion of why social norms and politics can make AIDS pol- icy particularly difficult for governments. What Is AIDS and How Is ft Spread? H IV IS A FATAL, SEXUALLY TRANSMITTED DISEASE (STD). After an initial week or two of flulike symptoms, the disease has no visible effects on the infected person during an asymp- tomatic period, which can be as short as two years or as long as 20. Although the average time without symptoms is about ten years in industrial countries, limited data suggest that it might be as short as five years among the poorest people in the poorest countries (Mulder 1996). Then, in all but a very small proportion of cases, the disease destroys the immune system. This leaves the infected person vulnerable 17 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC to other infectious diseases, which are typically fatal within six to 24 months (box 1.2). As discussed in more detail in chapter 4, recent medical breakthroughs in treating HIV infection in high-income coun- tries, although encouraging, are still very far from offering a technically feasible or affordable cure for the developing world. Like other STDs, HIV is difficult to transmit except by sex or other direct contact with the bodily fluids of an infected person. The major modes of transmission are sexual intercourse, reuse of contaminated syringes by injecting drug users, infection via birth or nursing from mother to child, reuse of needles in medical settings, and transfusions of contaminated blood or blood products. HIV cannot be transmitted by a sneeze, a handshake, or other casual contact.3 About three-quarters of HIV transmission worldwide is through sex; of these sexual transmissions, about three-quarters involve heterosexual intercourse and one-quarter involve sexual relations between men. In developing countries, sex accounts for an even greater proportion of cases. In Sub-Saharan Africa, Asia, and the Caribbean, sexual transmis- sion is overwhelmingly between men and women; less than 1 percent involve homosexual acts. In Latin America and Eastern Europe, how- ever, sex between men still accounted for most sexual transmission as re- cently as the early 1990s (Mann, Tarantola, and Netter 1992). The next most important means of transmission after sexual inter- course is the sharing of unsterilized needles among injecting drug users. Transmission through injecting drug use has been the primary mode of transmission in China and Southeast Asia, except in Thailand, where heterosexual transmission has outpaced transmission by needle sharing. Injecting drug use is also thought to account for about one-quarter to one-third of transmissions in Brazil and Argentina. HIV can spread through a population of injecting drug users extremely rapidly, in some locales infecting the majority within a few months. The importance of mother-to-child transmission varies widely across countries. The major mode of infection among infants can occur in the uterus through contact with the mother's blood at birth or later through breastfeeding (see box 4.6). Since mother-to-child transmission can occur only if the mother is herself infected, it is most common in widespread heterosexual epidemics, such as in Sub-Saharan Africa. By one estimate, 15 to 20 percent of all HIV infections in Africa occur in infants infected by their mothers. Worldwide mother-to-child transmission accounts for about 5 to 10 percent of infections (Quinn, Ruff, and Halsey 1994). i8 -'3C CA ~~ ~~d~~~' rLc,a - , r ' rc , -r -' - aW r9 3X e '- -,cc a- - -. ,, C~~! EV C~ ve - " 7-3- - - :~ 2 z = C I -'r ~ 2V lb C - 0 c g- -~~~~~~ ; ,- *r *r Vtc~ ~j~ ~ n r, r 9J S Z '2 an i- ~~~~~~~~~ rl IL L IL I- tj C 2r n -2 LL , ,, n, ft~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~V z:- r -;5I :3 r =~n'~ .7 I r C - - I~~~eL- - :~ 2'. fl-'Vt,- - C- j~~~~Ci - - C.. nfl ~~~~~~~~~-i c- 0-i"n c~~~~~~_-i :3 ~ t - nVt -.n2 2Vt -V~~~~~i, -E- ; - Z-< ' fl. n c, a-L - c- e3r I,.-c3C a-Lc C--c -2~V - IL a e2 fb ~ ~ ~ ~ ~ V -r Vt !Z 9 .7-~~~ r- ' C Z =' -r- Vt I .- e,, t 2 ta V o a- Vt-t V I Vt f :3~~~~~~~~~~~~~~~~~~~~~~~~~U a- c ~~~~~~~~~~ = Vt L tVt ~~~~~~~~~~~~~~~~~~~ C~ o C ? r W~~~~, r- 0 _ r - e l. - J r C t l I ~ -V - .-I nu, I - - V a-V C - 1V C.- -~~ C -~~ = ~ Vt Vt - ~~~ c n n ~~-' ~ *~~ Vt Vt - -U ~~~ - = ~~~~~~~~~~~ ~ ~~~~ c, ~~~~~~~~~~~ -- 7 c , a - - 4; K, c ' .Vt - C~~~~~~~~~~~~~~~~~~ c-c - a-aA cC ~~ 9o a-. a-n Vt C~~n VtC c ~- Vt ~~c- c, 2 ~ ~ I Z Vt 9~~~~- -.-Ha~~L : 3-C. n t~ l0,~~l -C -J jc~~~~~~~~-.- r.;w--R VtfA2ILf a-Vt 2~~~"~~=9-C2 C3.~~~9 ~~~LV' rL n Z~~~t-c n"C t -I n =c* -a.a i t V a- VC- 2 U n ,R C.s r~~~~~~~~~~~~~~~~~~~~~~~~~~ CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 1.2 Icontinued) mally rare diseases that would not have taken hold at access to medical treatment for opporrunistic illnesses. all had the person becn HIV-negative. Some AIDS- Most research on this question has focused on the associated infections can be treated with conventionial industrial world. Prior to the use of triple-drug thera- antibiotics. particularlr at the earl. stages of clinical pies (see table 4.2). the median time from HR'- I infec- AIDS. As die immune system continues to deterio- [ion to death in industrial countries was around nwelve rate. how-eter. treatment becomes increasingls diffi- vears: the first tw%o stages comprising eight to ten ryears cult and the number and xarient of illnesses increases. and the final-stage. clinicalA IDS comprising about leading to death. Box table I.2 lists the main AIlDS- fourteen to 25 months (KitahaLa and others 19'46). associated illnesses diagnosed in deseloping countries. Mluch less is known abouL the survival rates of The length oFsur- ival after infection depends on HIV\'-infecred people in developing countries. but many factors, including the strain and subrtpe of the both the time from inle&tion to AIDS and the time sirus. the general tarie ot' he person's health. and from rIDS to death are believed to be much shorter. Box Table 1.2 Opportunistic Illnesses Often Diagnosed in HNV-Infected People, Developing Countries NameN Varei Tuberculosi, Because latent TB ' common among HIX-negartie people in deseloping ioun- tries. it is the mo.r common opportunistic infecton there, occurring in -sO to 61 percenr of rhe HIV-infecredl As in people without HI'. TB usually occurs as i lung infection. although the likelihood ot TB intecting ther parts ot rhe bods is higher in the H-i,nfected. Pnreumococcal diseasc This bacterial infection is the must common cause of pneumo3ni in people vwihout HI\ and also causes bacteremia. sinusitis. and mcningitis among the HI\-irifected. P[le:1 i,.-l!hcl .'rk/.cr'.l rz Although almost unknown among people wirh normal immune systems. this small parasite is the most common cau,e ot pneumonia amnong HI\-infected people ourside Africa Toxoplasmosis Pretiously knotsn as a cause of an occasional birth defecr vhhen it intects preg- nant w%omen. in people %. ith AIDS ir is a common cause ofencephalimi. or intec- tion oftche interior oftLhe brain, %hhch causes seizures. coma. and death. Candid;asis CrommonlY knov; n as oral or esophageal thrush. this Frungus infect,or occurs in almost evern person kith HI\ and makes swallowing painful Crpltococcnss A\Jihou'_h almrro unknown in pcople without AIDS. this fungus inrection occurs in about i percent of'AIDS patienis w%orldwkide. usually as meningitis. an intlam- marion ot'Lhe surtace of'the bram. %thich causes se%ere headrache. fever. coma. and death A.IDS-associated cancers Common among upper-income piople in developing countries 1%%ho have access to treatment tor more common opporrunisric Uinessesm. Nore. Co-intkdii, wvih TB an ,one cr more other opponuni;ric infections msv he common in the devtloping isorid OtLher impor rant *-ppc.riunn,rtc infmtiion, 'uch is cCi:VmegsJc.crris CTi and " ,.o... rpm; .;i.n. 'i comples i r.txiC i. do occur in I' elepireg ooun- rric4. but are rarrlh d;ianoej because of! HkL ot [e-oiirce NSea r or.)to t blebund&r;. and tChin 1's;-i. . '.4p.qc-r. P'errien' I0C, 20 AIDS: A CHALLENGE FOR GOVERNMENT Box 1.2 [continued) with a rotal iur ival [imc from infecrion to death of trill has nor been proven [hat HIV infection i, perhaps around we%en s-ears. Asid, from the generally dIh'ars fatal. Rather it ippears that survival after HI\ poorer health and nutrilional s[atus ot many in the intecuon t;llows a bell curne. Just as a tfem people developing world. lack oF treatment for opportunis- progress [o AIDS and die 'enr quickl;. at the other tic infeccions that appear earl% in the course ot AIDS end of the cunre are a te%i who havc been infected is one factor in the shorrer suri' ii rimes. For exam- With HIX' tlr more than a dozen -ears bur are srill ple. people %%irh H1V in developing countries are health%'. Nledical researchers are 'ery interested in mure liklel than their counterparts in rich countries rhese long-cerm nonprogressors as they are called. to succumb to TB. which is more prevalent and les, because theY maY shed some light on the characteris- likelk to be treated in poor countries. In addition, ticS of the immune system that could be boosted. TB has been associated %%th the faster e%olution of tor examplc. by- a vaccine, to protect the a%erage HI\ disease iDe Cock 1 )').1 person againsr HIV infection. The prospects-and In all countries. AIDS is ovcrwhelmingl v fatal. importance-of %kork on a vaccine are discussed in but more than tifteen rears after its emergence. it chapter 5. HIV may also be spread through medical injections. In some of the poorest countries, injections are the preferred delivery system for a vari- ety of medications, and the same syringe may be used on many people in one day without sterilization between injections. However, even in these countries, medical injections with dirty needles are thought to account for less than 5 percent of all HIV infections. Transmission through blood transfusions, once a cause for concern in many countries, has been nearly eliminated in many high- and middle- income countries by routine screening of blood for transfusions. In de- veloping countries, transmission through the blood supply has yet to be eliminated, especially where HIV prevalence rates are high among blood donors and where screening blood for HIV has not yet become routine. In Africa, young children may be given transfusions for treatment of malaria-related anemia, putting them at risk of acquiring HIV But while transmission through transfused blood and other blood products greatly increases the risk of medical care and can rapidly spread HIV among spe- cific populations-for example, among hemophiliacs in industrial coun- tries in the 1980s-HIV transmission through transfusions has never accounted for more than about 10 percent of total HIV infections, even in developing countries. 21 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC The Impact of AIDS on Life Expectancy and Heafth HE MOST OBVIOUS IMPACT OF AIDS IS ON LIFE EXPECTANCY and health. Measuring and predicting these impacts are diffi- cult, not only because of the lack of quality data, but also because the relative size of an impact depends on many factors besides the spread of AIDS, including success in fighting other health prob- lems. Available evidence discussed below suggests that in the most severely affected countries AIDS threatens to reverse a century of progress in the fight against infectious diseases. Elsewhere, it is likely to account for an increased share of the infectious disease burden. Even so, AIDS is only one of many health problems confronting people in developing countries. Indeed, the poorer the country, the more likely it is that other problems-including easily treated problems like malnu- trition and diarrhea-account for a large share of the burden of disease. Life Expecancy Life expectancy is a basic measure of human welfare and of the impact of AIDS. From 1900 to 1990, dramatic progress in the fight against in- fectious disease raised life expectancy from 40 to 64 years in developing countries, narrowing the gap between these countries and industrial countries from 25 to 13 years. AIDS has slowed and in some countries reversed this trend. For example, life expectancy in Burkina Faso, a mere 46 years, is 11 years shorter than it would have been in the absence of AIDS (figure 1.3). Life expectancy in several other hard-hit countries also has been pushed back to levels of more than a decade ago. The im- pact of AIDS on life expectancy in Thailand is less, because its infection rate is less than that of the other countries in the figure. Disability-Adjusted Life Years (DALYs) AIDS accounted for about 1 percent of all deaths worldwide in 1990; this proportion is likely to rise to 2 percent of all deaths in 2020 (Murray and Lopez 1996). However, the proportion of total deaths caused by a disease is an imperfect representation of its burden on society, because it ignores illness and does not distinguish among the deaths of people of different ages. Murray and Lopez (1996) have estimated the cost of dis- eases in terms of disability-adjusted life years, or DALYs. Introduced by 22 AIDS: A CHALLENGE FOR GOVERNMENT Figure 1.3 The Cument Impact of AIDS on Lfe Expectancy, Six Selected Countries, 1996 Burkina Faso 11.3 Cunent life _ expectancy C6te d'lvoire _ 11 Impact of AIDS South Africa 7] Zimbabwe -2 - _2.2 Brazil - -3 Thalland - !MIIIII_ AIDS has already sharply reduced life I I expedan~~~~~~~^|cy in some countis 0 20 40 60 80 Life expectancy (years) Source: U.S. Bureau of the Census, 1996, 1997. the World Development Report 1993 (World Bank 1993c), a DALY in- dudes the disability as well as the mortality effects of disease and uses age weights to discount the importance of infant and elderly deaths. In 1990, poor health resulted in the loss of about 265 DALYs per thousand per- sons per year in developing countries, almost twice the 124 DALYs per thousand per year lost in industrial countries. Since HIV/AIDS deaths entail substantial disability before death and disproportionally strike prime-age adults, HIV/AIDS has a larger impact on health measured as DALYs than when measured as a share of total deaths. However, the dif- ference is not large: Murray and Lopez (1996) project that HIV/AIDS would account for almost 3 percent of all DALYs lost in developing countries in the year 2020, up from 0.8 percent in 1990 (table 1.1).4 One reason that HIVIAIDS does not account for a larger percentage of lost DALYs is that other causes of death in developing countries also en- tail substantial disability and premature death. Further, some of the in- creased impact of HIV/AIDS is offset by the decreasing share of prime- age adults in the population associated with the demographic transition. HW/AIDS as a Share of Infectious Diseases The contribution of HIV/AIDS to the disease burden looms larger when we focus attention on infectious disease. Such a focus is particu- 23 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Table 1.1 Annual Burden of Infectious Disease and HIV, as Measured by Deaths and Lost DALYs, the Developing World, 1990 and 2020 1990 2020 Los, Lost Deuh.s, D.{I ) Dentin DAL VS .4A inln,, burden of lisease (percentage of toted) (peren'age of otal)a [lncriou; diSeamz tl, pcrcencagc 'A rWEa] burden) - 4.2 1t3 I 3. HI\ 'uiaperenra.geo) rr,al burden, '1 0.S 2.0 2.( HlI\ laperceinrageaaotintriLasi burdrn 2.0 3.2 I3.t II Hl plu a pr..-rion (t TB 'as pereenrz_e of enrftcou kurden I 2 8 3.t' --.3 * - Total burdein pcr ! people -I";. 2C). 2 hIifcr[ious burden per I.000 people 311 uli.L) 1.2 2'.i Hl\ tburden per I I 11A, people 1 I 2.1 0 2 a. 1he K'.urih r- ,U ro l he 1ible !4 c-ir ed ts .±ddrng S' perc.cr ofih- 1cs I "h bUl cdC 2-J piT.n[ rhc L i 2i.i TB burden r- [h., number. tor HR. I hbe percenrige. rc chs u.,rhcwr c[im[cc r at crhi porrii.n ot HlN -nevirc'c TB dejrh4 rhir .iould nor hits oc,urrcd had HIl\ p,;TIlc pcople ntc soncrib,rred ri ch, 'prsidof c- B S... biel-rehcin, irco trrn Nlurni and I opc: N 'il larly relevant to our overarching purpose-identifying the appropriate roles for developing country governments in the fight against AIDS- because economic theory, public health teaching, and long-standing practice all affirm that governments should play a significant role in pre- venting the spread of infectious disease. By 2020 infectious diseases, which currently account for about 30 percent of deaths and one-quarter of lost DALYs in developing coun- tries, will have declined to about 14 percent of both measures.5 But the contribution of HIV/AIDS to the infectious disease burden in develop- ing countries is projected to increase sharply, from about 2 percent of deaths and 3 percent of lost DALYs to about 14 percent of deaths and nearly one-fifth of lost DALYs. Moreover, because HIV is an increasingly important factor in the spread of TB, it is estimated that about one out of four TB deaths among HIV-negative people in 2020 would not have occurred in the absence of the HIV epidemic.6 Adding a quarter of TB deaths among HIV-negative people to the deaths directly attributable to HIV/AIDS suggests that HIV/AIDS will be responsible for about one- fifth of all infectious disease deaths in low-income countries in the year 2020.7 In addition, HIV is likely to be responsible for a portion of deaths from several other infectious diseases (figure 1 .4)A8 24 AIDS: A CHALLENGE FOR GOVERNMENT Figure 1.4 Breakdown of Deaths from Infectious Diseases, the Developing World, by Disease Category, 1990 and 2020 (percent) 1990 2020 Dlarrhoal and Diarrheal and childhood childhood HIV 32.9 HIV 23.4 - 13.6 2.0 Othier Other 4/ 2.0 4.6 11.5 TB Mala;ia TB14.6 50 As the epidemic progresses, HIV will Malarlal r TB account for a greatl increased share 6.4 Other 26.7 Of total deaths from infectious dis- - STDs Respiratory easesindevelpingcountries. Respiratory 1.7 25.4 Other 30.8 STDs 1.3 Source: Murray and Lopez 1996. HIV/AIDS as a Major Killer of Prime-Age Adults Because HIV/AIDS is sexually transmitted, AIDS usually strikes prime-age adults-often people who are raising children and are at or near the peak of their income potential. In the absence of AIDS, prime- age adults tend to be less vulnerable to sickness and death than children, adolescents, or old people. Accordingly, AIDS casts an even larger shadow on the health of prime-age adults and the welfare of their depen- dents. In 1990, HIV was already third after TB and non-TB respiratory infections as a cause of adult death in the developing world; by 2020, HIV will be second only to TB as a killer of prime-age adults in develop- ing countries (figure 1.5). Adding one-quarter of TB deaths among FIV- negative prime-age adults makes HIV/AIDS the largest single infectious killer of prime-age adults in the developing world in 2020, responsible for half of all deaths from infectious disease among this important group. The HIV/AIDS share of the adult infectious disease burden varies widely across developing regions. In Africa, where other infectious dis- eases decline less quickly than in other regions and HIV/AIDS infection rates are assumed to be leveling off in many areas, HIV/AIDS will account for about one-third of these deaths (figure 1.6). Because Latin America and the Caribbean countries are projected to make the most 2-5 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 1.5 Causes of Death from Infectious Diseases among People Ages 15 to 59, the Developing World, 1990 and 2020 (percent) 1990 2020 Malaria HIV Malaria Other 6.4 8.6 1.3 23. 5 Other HIV 23.5 t ~~~~~~~~~~~4.4 37.1 /t Respiratory | Between directly causing AIDS deaths X Rsa10.0tr and indirectly facilitating the spread of TB, HW will be responsible for up to half of all adult deaths from infectious Respiratory disease in the year 2020. - 24.7 TB 51.4 Source: Murray and Lopez 1996. Figure 1.6 HIV/AIDS as a Percentage of the Infectious Disease Burden of Adults, the Developing World, 2020 HIV as % of adult deaths from infectious disease 80 73.5 70 60- 50 - 40 38.1 37.1 30 AIDS will account for a larger share of 20 18.2 the infectious disease burden in 10 regions where other infectious dis- eases are less of a problem. o SSA Asia LAC MEC Developing SSA Sub-Saharau Africa LAC Latin America and the Caribbean Countries (total) MENA Middle East and North Africa Source: Murray and Lopez 1996. 26 AIDS: A CHALLENGE FOR GOVERNMENT progress in reducing other infectious diseases, and HIV infection is pre- dicted to continue to rise, HIV will be responsible for almost three- quarters of the infectious disease burden there.9 AIDS and Development A LTHOUGH THE HEALTH IMPACTS OF THE DISEASE ALONE ,A,are ample cause for concern, there are additional reasons why the development community in general and policymakers in particular should be concerned about the HIV/AIDS epidemic. First, widespread poverty and unequal distribution of income that typify underdevelopment appear to stimulate the spread of HIV Second, the accelerated labor migration, rapid urbanization, and cultural modern- ization that often accompany growth also facilitate the spread of HIV. Third, at the household level AIDS deaths exacerbate the poverty and social inequality that are conducive to a larger epidemic, thus creating a vicious circle. Policymakers who understand these links have the oppor- tunity to break this cycle-through policies that are suggested below and analyzed in detail in the rest of the book. Poverty and Gender Inequality Spread AIDS While the determ-inants of an individual's sexual activity are subtle and complex, it is reasonable to expect that at the aggregate level social condi- tions would influence the frequency of risky sexual behavior and hence the size of the epidemic. One hypothesis is that poverty and gender inequality make a society more vulnerable to HIV because a woman who is poor, either absolutely or relative to men, will find it harder to insist that her sex partner abstain from sex with other partners or use a condom or take other steps to protect herself from becoming infected with HIV.'0 Poverty may also make a man more prone to having multiple casual partners, by pre- venting him from attracting a wife or by causing him to leave home in search of work. The idea that poverty and gender inequality exacerbate AIDS is supported by an exploratory analysis of national-level aggregate data on HIV infection rates. Eight epidemiological, social, and economic variables can explain about two-thirds of the variation in cross-country HIV infection rates. Figure 1.7 27 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 1.7 Relationship of Four Societal Vaniabes with Urban Adult HIV Infections, 72 Developing Countries, circa 1995 Urban adult HIV (percent) Urban adult HIV (percent) 35 - 35 * Botswana Zambia Guyana *zimbbwe 20 * Haiti. Z?hibabwe 2 Guyana 5f * M ala R- f**SBotswana 2 HaIt mb onduras * ep of Korea Repaoraora * Mai Mala * PDRJ 5 ThaI0* * *- Thtailand 5 J Lao PDR J. * *Tifaliand * ~~~~~~~Pakta * *.. * 1-~ ~~ . *; > a- .*f* .* 0 *. ~~~~~~~~tARE 4 ARE* China * ' *China 1,100 3,000 8,100 0.3 0.4 0.5 0.6 GNP per capita (1994 dollars) Inequality of income distribution (Gini coefficient) Urban adult HIV (percent) 3S Urban adult HIV (percent) Botswana a Botswana GU n ~ * Zimbabwe 20 - * y"p of Korea * 20 * Rep. of Korea -Halti Zambia Guyana Zimbabe Zambia. Haiti. * Tilnd * 5 4.ao PD0R. 0 5 Honduras 1haia D1-4 102 Honduras Mplfawla po ARE Ar.b Republic of EgtARE China;1n 0 ~ ~ ~ *0 ______________ 0.7 0.8 0.9 1.0 1.1 1.2 1.3 1.4 1.5 0 10 20 30 Ratio ovf urban maies to femaies, age 20-39 Male-femaie literacy gap (percentage points) ARE Arab Republic of Egypt Note: The vertical axis measuring HIV infection has been transformed to a logarithmic scale. Points on a given scatter plot represent the data for 72 individual countries after removing the effects of the other seven variables included in the regression analysis. Income inequality is measured by the Gini coefficient. The methodology and the detailed statistical results are presented in Over (background paper, 1997). Source: Authors' estimates. Poverty, inequaliy of income across shows the associations of four of these variables with the percentage of households, and low status for women urban adults infected with HIV. " The top two panels of the figure demon- all contribute to the spread of HIV. strate that, holding constant the other variables, both low income and un- equal distribution of income are strongly associated with high HIV infec- tion rates. For the average developing country a $2,000 increase in per capita income is associated with a reduction of about 4 percentage points in the HIV infection rate of urban adults. Reducing the index of inequality from 0.5 to 0.4, the difference in inequality between, for example, Hon- duras and Malawi is associated with a reduction in the infection rate by about 3 percentage points.'2 These findings suggest that rapid and fairly distributed economic growth will do much to slow the AIDS epidemic. 13 28 AIDS: A CHALLENGE FOR GOVERNMENT When examining the influence of gender inequality on HIV infec- tion, one must hold constant as much as possible other cultural influ- ences, such as Islam, which may be correlated with gender inequality across countries. The bottom two panels of figure 1.7 show that, after controlling for the percentage of the population that is Muslim (as well as per capita gross national product (GNP), income inequality, and four other societal characteristics), two measures related to gender inequality are associated with higher HIV infection rates. The first of these, the ratio of males to females in urban centers, varies remarkably across coun- tries: some countries have fewer male urban residents than female, and others have 40 percent more males. Other things being equal, one might suppose that commercial sex would be more common in cities where men greatly outnumber women, and that HIV levels would therefore be higher. The evidence of the lower left panel of figure 1.7 is that cities in which men ages 20 to 39 greatly outnumber women do in fact have sig- nificantly higher HIV infection rates. For the average country, increasing the job opportunities for young women so that the ratio of males to fe- males in urban areas falls, for example, from 1.3 to 0.9, would decrease the HIV infection rate by about 4 percentage points. The second measure related to gender inequality included in the analysis is the gap between adult male and female literacy rates. Again, there is great variation across countries, the literacy rate among men being as much as 25 percentage points higher than among women in some countries. When women are much less literate than men, they may be less able to bargain effectively with men and thus be at greater risk in sexual encounters. Furthermore, illiterate women will have diffi- culty finding jobs and thus may depend more on sexual relationships for economic survival, again reducing their bargaining power. The lower right panel of figure 1.7 supports these ideas, suggesting that the average country that reduces the literacy gap between genders by 20 percen- tage points can expect urban HIV infection to be about 4 percentage points lower. The Dynamics of a Growing Economy May Facilitate the Spread of AIDS From the evidence of figure 1.7, a country that improves per capita income and reduces inequality, for example, by implementing invest- ment policies that generate jobs and raise economic growth, will reduce its risk of suffering an AIDS epidemic or help to minimize an epidemic 29 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC already under way. If, in addition, the country acts to close the literacy and urban employment gaps between men and women, HIV would have even more difficulty spreading. Unfortunately, some of the very processes that might achieve these goals can also stimulate the spread of AIDS. And other policies that sometimes accompany growth, without necessarily contributing to it, can likewise worsen the epidemic. An open economy is recognized as a key requirement of rapid growth. Openness primarily refers to the ease with which entrepreneurs can move goods and capital across national borders. Yet a higher degree of openness to trade and financial flows is typically also accompanied by a greater de- gree of openness to movements of people, including immigration. More- over, some studies have suggested that immigration itself contributes to economic growth. This would hardly be surprising, since immigrants are often among the hardest-working and most enterprising people in any country. However, a cross-country regression analysis suggests that coun- tries with larger immigrant populations tend to have larger AIDS epi- demics: other things being equal, a country in which 5 percent of the population is foreign-born can expect to have an infection rate about 2 percentage points higher than a country with no foreign-born. Does this mean that governments should restrict immigration to avoid an AIDS epidemic? No, it does not. Indeed, if immigration is ben- eficial to economic growth, reducing immigration may slow it, a result that, in addition to many other negative effects, may stimulate the spread of AIDS. Attempting to screen out HIV-positive immigrants is unlikely to be very effective, since immigrants are more likely to become infected after they have arrived in a new country, when they are disconnected from the social networks of their former homeland, than before leaving home. Worse, attempts at screening potential immigrants may exacer- bate the epidemic: if people infected with the virus evade screening and arrive illegally, identifying and reaching them with programs to prevent them from infecting others is extremely difficult. Sometimes a specific project promises significant economic benefits but carries with it the danger of worsening the epidemic. An example of such a project-and of an effective response by the governments con- cerned-is the Chad-Cameroon oil pipeline project described in box 1.3. The challenge to governments, donors, and multilateral institutions is to recognize the potential AIDS hazard inherent in such a project and in- corporate into the project design elements that eliminate or at least min- imize and mitigate these problems. Economic development projects that 30 AIDS: A CHALLENGE FOR GOVERNMENT Box 1.3 AIDS and the Chad-Cameroon Oil Pipeline Project lIHE CEHAL -ACAMEK N C)1. PllP LINE Is lVHt FIRST cion, the World Bank. the consortium, and the rwo large-scale infra.trucrure project supported by the go%ernments in%olved have identiFied a package of \World Bank to assess the potential for an adverse menares to avoid exacerbating the HRV/AIDS epi- impact on STDC. including HIV/kIDS. and to in- demic in the project area. Since preliminary esti- corporate pre%enrion efforts in the project design. mates suggest that effective interventions can bc The 30-Year. £3.5 billion project. due to begin implemented for less than $1 million a year. the construction in 1998. insolkei deselopmenr of oil substantial returns on the project are more than suf- fields in southern Chad and construction of' an ficient to justify project execution despite these 1.100 km pipelinc to port facilities on the Atlantic costs. Ulsing baseline data and experience gained coast of Canieroon. A cooperative elfort between elsewhere in Africa. the consortium is developing a the \World Bank, the go%ernments of Chad and layered intervention strateg% that includes: Cameroon. and a consortium of prixate oil conipa- nies. the project promises the rwo countries sub- * monitoring of the STD and HI\' status of' stantial economic benefits,. the wootkforce But the project also involves a potential risk of U vigorous markering of subsidized condoms exacerbating the HI/VAIDS epidemic. During the U intormation. education. communication IIEC) peakl construction period. from Ifl98 to 2001. the U treatment ofclassic ST'Ds project %%ill draw- an estimated 2.Oi)0 construction U interventions ro modictS high-risk behavior %%orkers from Chad 3nd Cameroon. and employ an U coordination with existing government and additional 400 to 600 truckers who %sill travel the NGO programs. particularlY those direcred length of the pipeline. Mlost of the ssorkers vvill be tow3rd sex vorkers. single and unaccompanied males. Those wvorking in Chad will commute from their villages of' origin. To supplemenr the work of the consortium in while those working along the pipeline in Came- this area. the \'World Bank is preparing two techni- roon wvill live in temporary barracks. Some areas cal assistance projects that will help the go%ern- along the propoied pipeline alreaId have extremel) ments of Chad and Cameroon monitor and assess hiRh lev,els oF HI\. a 199i report From an area the health impact of the project. lmplementing adjacent to the ChadiCentral African Republic bor- these programs wvill involve significant challenges, der. and directly on the proposed pipeline route. including the difficulries of reaching the highlyN indicated thar more than halF of the sex ssorkers mobile truckers and the sex workers they Frequent. and one in Ftour truckers ssere already- inrfected v,with the virus - i'O. CIld%%ell and Caldwell 1Q03. pp 8l-tS: Carsv.ell .Alerted to these problems by an environmental and Ho\wells 19!89. pp. -594O: D-me, and Moore 11'96; and assessment performed as part of the project prepara- Innaarubi nd otheri 119)9 do not generate sufficient net economic returns after covering the cost of mitigating negative impacts, including the spread of AIDS, should be rejected as undesirable-even if the gross returns are quite large.14 Sometimes a low-income society beginning rapid growth may face an increased risk of AIDS as a result of a broad shift from conservative social 3I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC norms to more liberal attitudes; these attitudes often include greater individual freedom, especially for women. Lacking an objective measure of social conservatism, the regression used the percentage of the popula- tion that is Muslim as an imperfect proxy for a country's degree of social conservatism. Controlling for all other variables described above, a high degree of social conservatism is associated to a statistically significant de- gree with lower rates of HIV infection. This does not necessarily imply that governments should attempt to instill or maintain social conser- vatism simply to minimize HIV; such broad social values may in any event be very difficult for governments to shape. However, the evidence does suggest the usefulness of an explicit government education policy that would help young people entering a rapidly modernizing society to recognize and avoid risky sexual encounters. A final factor in the regression analysis that is not associated with development but can be readily affected by government policy is the level of militarization. In developing countries, military forces are often based near urban centers and consist predominantly of young, unmar- ried men. Using a variable that measures the number of men in the na- tion's armed forces as a percentage of its urban population, the regression analysis shows that, even after controlling for the ratio of male to female urban residents, countries with more soldiers will have higher infection rates. For the average country, reducing the size of the military from 30 to 12 percent of the urban population will reduce seroprevalence among urban adults by about 4 percentage points. An alternative discussed in chapter 3, which may be more feasible (and is reasonable regardless of the size of the military) is a vigorous HIV prevention program covering everyone in the military. AIDS Has Little Net Macroeconomic Impact Because HIV/AIDS is spreading rapidly and is nearly always fatal, some observers have concluded that it will significantly reduce popula- tion growth and economic growth; a few have suggested that populations in badly hit countries will decline in absolute terms and that there will be an associated collapse in economic output (Anderson and others 1991, Rowley, Anderson, and Ng 1990). However, the available evidence sug- gests that the impact of AIDS on these variables, although varying across countries, will generally be small relative to other factors. Moreover, at a very crude level, declines in population growth due to HIV/AIDS will 32 AIDS: A CHALLENGE FOR GOVERNMENT tend to offset declines in economic growth, so that the net impact on gross domestic product (GDP) growth per capita will generally be small. HIV/AIDS is expected to reduce population growth rates in many countries, but in no country is an absolute decline in population ex- pected. The latest projections suggest that declines in population growth rates from HIV/AIDS mortality will range from about 0.1 percentage point in Thailand to 2.3 percentage points in Botswana, and that the median country's population growth rate will decline by about 1 per- centage point (U.S. Bureau of the Census 1997).15 Over time, such a re- duction in growth would result in a significantly smaller population than would have existed in the absence of AIDS. In Zambia, for example, the population is predicted to be 7 percent smaller in 2005 than it would have been without AIDS. In two countries with very severe AIDS epi- demics, Botswana and Zimbabwe, the projections suggest that by 2010 the population will cease growing. The impact of AIDS on economic growth is a much more complex issue than the impact on population growth. The inadequacy of GDP per capita as a measure of human welfare is striking when changes in per capita GDP are used to measure the impact of AIDS. Other things held constant, the death of people with higher incomes will reduce average income-even though the welfare of those who remain alive has not changed. Conversely, the death of those with lower incomes raises aver- age income, without necessarily improving the lot of any surviving indi- vidual and despite the suffering and economic losses of the families of those who died. Further, increased spending on health care and funerals is included in GDP calculations. As a result, per capita GDP may increase, even though overall well-being has not improved and the in- comes of survivors have been reduced. With these caveats in mind, it is nonetheless possible to estimate the size of the epidemic's impact on individual incomes. This impact will de- pend on characteristics of the country, including the severity of the epi- demic, the efficiency of the labor market, the proportion of the treatment cost of AIDS financed from savings, the distribution of HIV infection by productivity of worker, the time lost from work by the person with AIDS and others as a result of his or her sickness, and the effectiveness of house- hold and community formal and informal insurance mechanisms. Because AIDS kills prime-age adults, many of whom are at the peak of their economic productivity, the shock of AIDS to the labor market is one mechanism through which AIDS might affect growth. However, in 33 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC economies with substantial unemployment, firms should find it easy to replace sick or deceased workers, particularly if they are not key person- nel. Other things being equal, the impact of the AIDS epidemic will be small until the economy begins to grow and is constrained by labor sup- ply rather than by insufficient demand. Box 1.4 provides evidence from a sample of 992 firms that the departures of lower-skilled workers due to AIDS may have had little effect on firm profits in five Sub-Saharan African economies. Another factor that is likely to have a significant effect on the size of the macroeconomic impact of the epidemic is the percentage of AIDS treatment costs financed from savings. Since spending on AIDS treat- ment is likely to reduce the capital available for more productive invest- ment, the higher the proportion of care financed from savings, the larger the reduction in growth resulting from the epidemic. If one takes these factors into consideration, a rough estimate would be that a generalized epidemic, as defined in chapter 2, would reduce per capita GDP by as much as half a percentage point per year.16 The importance of an impact of this size will vary depending on the country's underlying growth rate. In some very poor Sub-Saharan African countries, growth rates in GDP per capita that are already nega- tive may deteriorate even further as a result of the AIDS epidemic. But some countries with severe AIDS epidemics, including Botswana, Thai- land, and Uganda, have been growing rapidly. With per capita growth rates in excess of 5 percent per year, a reduction of 0.5 percent in per capita growth will not be crippling. For these countries, as for many other countries where the epidemic may peak at lower infection levels, the more serious consequences will be the impacts on public health spending and on poverty. Poverty, Inequality, and Orphanhood Although the macroeconomic effects of AIDS are likely to be small in most countries, severely affected countries will experience quite large impacts on their health sector and on the poor. The effect on the health sector will be to increase the price and reduce the availability of health care for everyone, which will tend to hurt the poor most. Furthermore, among the households that suffer an AIDS death, lower-income house- holds will be less able than others to cope with the medical expenses and other impacts, including loss of income. 34 AIDS: A CHALLENGE FOR GOVERNMENT Box 1.4 Looking for the Impact of HIV/AIDS on a Sample of African Firms IN CULNTRI[S W-ITH GLNLRXL.JZED HPV EPIDEMIC>. administcred to 92 ftirms from tour segments of the the mortalinr rare of prime-age ssorking adults %-iil manufacturing sector of five African countries (back- evenrualv rise by a multiple of 2 to 10. depending groam/n11parer. Biggs and Shah 1996l.0 on the baseline mortalint rate in the country and the Box table 1-4 presents the data on HR' infection extent of H1EV infection (rable 9.3). Such incrcases rate in rhe urban population of each of the five should raise the firm's laboir costa by requiring it to countries and the percentage of %sorkers leaving their replace wvorkers more frequently. to spend more on employment in 199q due to sickness and death sickness and death benefits. and, perhaps. to imple- (right-most column). Clearly there is a strong corre- ment AIDS education programs designed to present lation betveen these rwo variables at the country the w%orkers from becoming infec[ed. X'hether these lesel. Zambia. ssith [he highest measured infection changes will hase a measurable effect on firm profits rate. also has the highest rate of' turnover due to depends on sshether they are large relative to the sickness or death. Ghana is at the other extreme on other components r-f labor costs and %% herher labor both variables. costs themielves are large a, a portion of total firm The impaCt of the higher levels of sickness and costs. dearh would be large if the resulting attrition rates Although seseral studies have shown that AIDS are large in relation to overall firm attrition or if ir increases the death rate of w%orkers in specific firmns takes a long tinie to replace the w%orkers. However, it none has compared these death rates with the rates appears that neither is usually the case. The aserage of worker attrition from other causes or has esti- attrition rate from all causes i.s from 8 to 30 times mated the impact of rhe deaths on firm profits larger than that from sickness and death. The time iGiraud 1992, Smith and \Vhiteside 1995, Baggaley to replace deceased w-orkers varies from an average and ochers 19 4, lones 199')-. 'o analsze the ot two weeks for unskilled workers to only- hree impact ofA.IDS deaths within the context of overall ssceks for skilled workers, not enough to signifi- firm perfcormance. a background paper for this study canrlv raise cosLs. The only hint in this labor firce analyzed data on wvorker attrition due to sickness and death that were collected as part of a surv-es tiBx ruontinues oeni thle I;o/lrbning,q pgqe. Box Table 1.4 Worker Attrition in Ghana, Kenya, Tanzania, Zambia, and Zimbabwe, Total and by Sickness or Death, 1994 Percentage of uvarkers learingfirm Irrban Hjjjl Totatl in Smnple DLfe to Due to Coutinn- prenalence Finrm 1Wnorkrr a/i CrflLes sic-b,ei or dentl/ Zambia 24 - 19-t I- .82 21) s 25 Zimbabwe 2(0.5 1) 5'i.2 10 'ii 1.2 Kenya .I 21 -t I 12i!' . 0.'9 Tanzania 16.1 19-T I-t.61 1 ('13 t ) .6 Ghana 2.2 18 Fi616- I I 6 iJ 3 Fotal 092 11;. 1 3 1.9 1 .1 L.. Secprct.caence diai reer tIc -rzzk 'r .uall% actoe 3dl11ri a r-puriedJ i i S Rure3, Cd th (C-n;u.s idmabas,, Ot'her. lrhcr diLi from RP'ED [linel Rurn-e- de'crbed in BLeig and S'ab , bad .:dacv. I 'i'(!l A prcldinarL t-r,ion -ri' tL' iible appear; in Nation.uI Rew.sirch (L un,il i p 25T p 35 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 1.4 (continued) data that AUDS might be cosily to firms is that they to reduce the firm's value-added per wvorker by a sta- took S times longer on average to find a replacement risEicallY significant. but small. amount i b iaroird for a dcceased professional than they did for a skilled paper. Bigs and Shah I S'.96 wvorker. However. exen 24 4seeks does not seem long Of course, these results are far from definitive. to search for a skilled professional. First, they are specific to Africa and to economies The ultimate question is whether labor force that w%ere doing rather poorly: for any sick w-orker sickness and death sisibly reduce firm profirs. in an many replacements wvere a%ailable. Second. e%en a output-constrained firm. hiring more ssorkers random sample otf 9)2 Firms is a small sample in increases output. Hotviz%er. a firm that is suffering w%hich to study- adult niortaliry. especi311y if the from a drop in dema3nd for irs product can raise events of most interest are the deaths of the most profits (or drcrease lossesi bV releasing w%orkers. If high]!' skilled wvorkers-the profession31s-of w%hom some of dhe firms in rhe sample are experiencing an there are few, in ans' firm Nevertheless, until more increase in demand and others a decrease. it Would definitive studies are performed. the evidence sug- be impossiblc to disentangle these two elfects. and gescs that the impact of UIDS sickness ind deaths is the estimated impact of AIDS deaths would be not a major de[erminant of the economic perfor- meaningless. mance of the average firm in dev eloping countries. One solution to this problem %%ould be to -- assume That the departure of %sorkers due ro sickness I Tl; c.:.r. . crc tbvd pr.; .lC rnci Aorkrng cod vriwLne. aind tc'.t.It Ilid iarrn',nt Thc tii, .icre cm lcctcd bi and death is beyonnd the control of the firm. "%herea.s r.rd; priect, dc.-gncd i, astute thattemiud bc repret-rcrr- the departure of either. presumabl' healths. %torkers ire ci tLhe 'ecior, ro v.hivh .1et aterc drmAn Thi .itcsion- n tire i%.as de~ine b% ec'norr'ts. 't3r' ican.. und inimnaemrnen is partly decided bs [he firm. Linder these assump- n in orer ro e about tir vJu!c 1 lam,uC.I srtlsi. !1 vT_r 1! {Ztjyn Jl!!l .h .3 ; .-.t' ,rMn,.;; I rions the deparrure of a worker due to sickness and A\ric. Sherily be;fore it itenr to rEh fldi. qiei,norL ;'ere addcd death is estimated by instrumental variable methods on norier jrur,tori We argue in chapter 4 that, because low-income households are more adversely affected by an AIDS death than other households, a severe epi- demic will tend to worsen poverty and increase inequality. One important way in which AIDS is likely to exacerbate poverty and inequality-and indeed, one of the most tragic effects of the epi- demic-is the increase in the number of children who lose one or both parents. To be sure, AIDS is not the only cause of orphanhood: in some countries other causes of prime-age adult death may orphan many more children than AIDS. Nonetheless, as AIDS mortality rises, growing numbers of children will be orphaned by the disease; the result- ing impact on orphanhood rates in three hard-hit countries is shown in box 1.5. The impact of an adult death on surviving children is discussed at length in chapter 4. Here it is sufficient to note that, even if we set aside 36 AIDS: A CHALLENGE FOR GOVERNMENT Box 1.5 Orphans and AIDS TIHE IIMPACT (U-F \ S LiRE E AIDS EPIDEMIlC ON Box Figure 1.5 Trends in Matnal Orphan Rates, maternal orphan rates can be secn in census data Three Hard-hit East African Countries, Various Years over the past 20 tear, from three East African coun- tries (box figure l.A). In the absence of AIDS, grad- 'ia 5| 6j 5 ual improvements in maternal health over the past c 5 rmo decades nou[d hase reduced maternal orphan 4 3.1 rates. Instead. we see that in Kenya the maternal _ 33 2.8 - orphan rate has remained nearly constant. In 2 1.91 1.8 2.23 Tanzania the maternal orphan ra[e fell between the LI f * I)-Os and late 1980s but then climbed rapidly to 0 BI I S i s_o almost 3 percent in the 1990s. Finally, maternal 1969 1993 1978 1988 1994 196919911995 orphan rates in Liganda show steady increases since Kenya Tanzanla Uganda 1969. a trend that can probably be attributed to a *.. ;;A n-d orph.r rr rei ci'i tf,cre .nclule ch'Ide?r th, m1 combination ot AIDS and civil war. Because AIDS ro-parxnrorpbnln tends to be geographicallY clustered, maternal --.., dnv.it il TanrlA-i] ..- S. andr Lardaj I i..iar' bad.n orphan rates are ev en higher in areas hard hit by the i jfl!sifli j;,, Ind Lnd r;, i E-aI Jg 'i it 'icm DHS dau epidemic. Across Fifteen %illages in Rkali District of Tinzania IU J-S 3id Liv,ndar 11i; n re rr -r;, .n!±. dmo. cpin;id Ldb, Uiganda. for exanmple. the maternal orphan rate in Hunter oid .\ itr,r inrrh.corinp. 1990 was o.6 percent. double that for the rest of the country IlKonde-Lule and others 199- i. Losing a parent can hase profound consequences for any child. and rhese are likely to be worse in with HIV at or around the time of birth. Also, poor households. Governments and NGOs trying AIDS orphans are more likely to become rs'o- to mitigate the impact of AIDS should be careful to parcnt orphans bccause HIV is transmitted sexually. consider overall needs and avoid creating programs For example. in a population-based surveY of rural that favor AIDS orphans over other orphans who areas of Miasaka Distric. Llganida. 10 percent of all may be equally or even more needy. children under 15 had lost one or both parents -et a consideration of the impact of the epi- iKamali and others 1992). FifTeen percent of che deniic must also recognize [hat AIDS orphans often survi ing parents o single-parenlt orphanis w%crc face uniquely seser problems. Ver- young orphans infected with HI\-. three times the parental infec- whose mothers are infected or die of AIDS have tion rate among nonorphans. Finally, AIDS higher mortaliry rates than other orphans because orphans may suffer social stigma from having lost roughly one-third of them are themselves infected their parents to a sexually tran5mirted disease. the unmeasurable grief and psychological pain experienced by children who lose a parent, the measurable declines in nutritional status and reductions in schooling can cause profound and lasting damage to a child. These effects, which are likely to be greatest in the poorest fami- lies, can greatly reduce an individual's ability to aquire the skills and knowledge needed to escape poverty. 37 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC The Government Role in Confronting AIDS G IVEN THE IMMENSE IMPACT OF HIV/AIDS ON LIFE EXPEC- tancy and health, and the possibility that AIDS will exacerbate problems of poverty and inequality, the need for governments to confront the epidemic appears straightforward. Indeed, for many people the human suffering caused by the epidemic is reason enough for governments to be involved. However, there are other important reasons for government involvement, some of which are not so readily evident. Analysis of these rationales for government involvement is a necessary foundation for considering how governments should confront HIV/AIDS. The Impact of AIDS on Public Heafth Spending One economic rationale for government involvement in HIV preven- tion is very straightforward: prevention is much cheaper than treatment and avoids the sickness and death that are the final outcome of the dis- ease. This argument is particularly important in the many low-income countries where governments remain committed to publicly financed curative health care. In such countries, the high cost of AIDS treatment starkly reveals the scarcity of resources. Figure 1.8 illustrates the difficult tradeoff governments face. In the figure, each country is represented by a point indicating, on the vertical axis, the estimated total cost (public and private) per year of AIDS treat- ment and, on the horizontal axis, the national GNP per capita. Not sur- prisingly, the amount spent on treatment rises sharply with GNP per capita. The upper of the two regression lines fits those points well and suggests that in the average country the annual treatment cost of AIDS is about 2.7 times GNP per capita. The second line in the figure (esti- mated from other data) demonstrates that for less than this amount the average developing country could finance a year of primary education for ten students. And this is just one of the many alternative productive uses to which these financial resources could be put. As the number of AIDS cases and treatment costs mount, it becomes painfully evident that AIDS treatment is consuming public resources that could have been used for other human needs. Yet governments may find it very difficult to restrict financing for AIDS treatment without at the same time reassessing their commitment to publicly financed health 38 AIDS: A CHALLENGE FOR GOVERNMENT Figure 1.8 Annual Treatnent Cost for an AIDS Patient Conrelated with GNP per Capita (dollars) Annual cost $1,000,000 Japan $100,000 Annual cost of treating G Aust Puerto one AIDcae CstaR P United States $1000 ~ ~ ~~Cng DR K e n y * 'Barbaclos $10,000Chl Tanzan aThailand ISpending on AIDS treatment increases Itond uraly \ with GNP; on average, treating an $100 AIDS patient for one year costs about India Annual cost of educating 10 the same as educating ten prinnary $10 primary school students school students for one year. $100 $1,000 $10,000 $100,000 GNP per capita Note: The trendline for AIDS is: Annual cost = 2.7 X (GNP per capita)0 95. a. Formerly Zaire. Source: Annual AIDS treatment costs are from Mann and Tarantola (1996) and Ainsworth and Over (1 994a,b). Annual cost of educating ten primary school students are the authors' cal- culations based on data from 34 countries in Lockheed and others (1991). care. Indeed, in many countries there are political pressures to subsidize AIDS treatment at a higher level than other health care services, and these pressures are likely to grow as the number of people with HIV in- creases. For all these reasons, a government that wishes to continue sub- sidizing health care should initiate aggressive prevention efforts as early in the epidemic as possible. Even governments that are attempting to reduce subsidies for curative care would do well to invest in HIV pre- vention, since political pressures for subsidized care may be very difficult to resist. The Public Economics Rationale for Governments to Fight HIV/AIDS Let us assume, for the sake of argument, that a government does not subsidize curative health care and that it is able to resist all pressures to do so. Given that HIV/AIDS is primarily transmitted by sex, would there still be a rationale for government intervention to reduce its spread? The answer from a public economics perspective is yes. To under- stand why, it is useful to first consider the public economics rationale for 39 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC government intervention against other contagious diseases, such as TB. If all markets worked perfectly, governments would not need to be in- volved in the fight against these diseases. Instead, each person at risk of infection would pay an appropriate share of the cost of reducing his or her risk. In reality, of course, there is no mechanism other than govern- ment through which individuals can pay this amount. Since a person in- fected with TB is likely to consider only the benefits to himself when de- ciding whether to pay for treatment, without government intervention people infected with TB would be cured less often than everyone else would prefer. Economists call the benefits of treatment that are not cap- tured by the person paying for the treatment external benefits and the negative impacts on others of not getting treatment external costs. These externalities, if large, are a strong economic justification for government intervention. A related problem can be best understood in the case of a vector- borne disease, such as malaria. Even if people know that draining a stag- nant pool where the anopheles mosquito breeds will greatly reduce their chances of catching malaria, people may not voluntarily pay the costs of drainage, since everyone benefits regardless of who pays. Thus, each per- son may hope to benefit from the actions taken by others. Elimination of stagnant water is an example of what economists call a publicgood Be- cause individuals hope to gain from what others have paid for, a public good may not be produced at all unless government taxes everyone in order to finance its production. In giving advice on how governments should spend scarce public re- sources, economists look for evidence of large externalities or public goods. Where these exist, markets are said to have failed, and public in- tervention to address the market failure is warranted. In the case of TB, malaria, and other diseases that may strike people regardless of their individual behavior, the public economics arguments for government intervention due to market failure are clear. At first consideration, it may appear that externalities and public goods are not a significant concern in the case of STDs, including HIV. Since most STD transmission occurs as a result of a voluntary act be- tween two people, each can weigh his or her risk and proceed only if the benefits of doing so outweigh the risks. If both parties agree to have unprotected sex even though they may catch an STD, why should the government intrude in these private decisions? The problem, of course, is that the decision of these two people has consequences for many 40 AIDS: A CHALLENGE FOR GOVERNMENT others, endangering marital sex and procreation as well as more casual sexual relationships. Ideally the couple should take the interests of others into account when they decide whether or not to engage in unprotected intercourse. However, even if they agree to use a condom or otherwise reduce their risk of infection, they cannot demonstrate to potential future partners that they have behaved prudently. In public economics terms, there are external benefits associated with refraining from risky sex. Since an individual cannot capture these benefits, he or she will be less prudent than would otherwise be the case.17 The result: higher STD infection rates and a greater risk of infection for everyone who is sexually active, even those who are monogamous, since most people can- not be sure that their partner is also monogamous. In such a situation, government intervention is justified if it can increase the incentives for the most sexually active individuals to practice safer sex (or for injecting drug users to adopt safer injecting behavior) to the point that their deci- sions more closely reflect consideration for the social consequences of risky behavior. The arguments above for government intervention to prevent the spread of STDs apply even more strongly to HIV/AIDS. In addition to being sexually transmitted, HIV/AIDS has two characteristics that worsen the market failures associated with the disease-and suggest that governments should be particularly concerned about HIV prevention. The most obvious is that AIDS cannot be cured and is almost always fatal. Since adult deaths impose costs on other family members and the rest of society, as argued above and shown in chapter 4, they are an argument for goverment intervention. In addition, we have seen that HIV makes people vulnerable to other infectious diseases, including TB. Since individuals have little control over their exposure to TB, and since people who have HIV and TB can spread TB even to people who are HIV-negative, the link between HIV and TB further strengthens the arguments for a government role in controlling HIV While this link suggests that HIV/AIDS should receive special atten- tion, the close epidemiological relationship between HIV and other STDs, discusssed in chapter 2, means that in practice any effective strat- egy against HIV will almost certainly involve stepped-up prevention ef- forts against other STDs, and vice versa. Since the monitoring problem described above applies equally to all STDs, governments would have a role in STD control even in the absence of HIV. Because HIV multiplies manyfold the external costs associated with a case of gonorrhea or geni- 41 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC tal ulcer disease, the presence of HIV strengthens the argument for gov- ernment intervention to control the spread of all STDs. The Government Role in Generating Information The above argument for government intervention assumes that peo- ple already know of the risks of HIV, or have the means to find out what they need to know. However, this is often not the case. Thus there is yet another compelling rationale for a government role in confronting the epidemic: the provision of information to enable individuals to decide whether or not to change their behavior to reduce the chance of infec- tion. In some countries, HIV/AIDS has now been present for two decades, long enough for most people to know that it adds a mortal risk to sexual liaisons; yet surveys show that an alarmingly large proportion of people in some countries still do not know how to protect themselves. In other societies, the disease is still new, an invisible danger spreading through an unsuspecting citizenry that, because of HIV's two-to-twenty- year asymptomatic period, has yet to be jolted by a sudden upsurge in AIDS deaths. In either society, only the government has the incentives and the capacity to generate the information that permits people to take the first steps toward self-protection. Information about the state of the epidemic and how to avoid infec- tion is a true public good. As in the case of malaria eradication, each individual who benefits from new information subtracts nothing from its value to others. Although it is possible to restrict access to informa- tion, for example, by selling it in magazines available only to subscribers, valuable information has a tendency to spread beyond those to whom it is sold. Therefore, private firms have less incentive to produce and sell information and will produce less of it than is socially desirable. This is particularly true of information from epidemiological monitoring of infection rates in various groups in society. In contrast to information generated by military surveillance, which is often particularly valuable when kept secret, the value of public health surveillance lies in announc- ing the results, so that people are aware of the disease in their midst and can take steps to protect themselves by reducing risky behavior. The public role in the production of new information reaches beyond monitoring the epidemic to include various types of research that enable a more effective response. In all countries, governments will require country-specific information about how to identify and reach people at 4z AIDS: A CHALLENGE FOR GOVERNMENT highest risk of becoming infected and spreading HIV to others. Research that improves the effectiveness of interventions throughout the country has important public good qualities, and therefore deserves government support. Some information relevant to prevention efforts, including bio- medical insights into the virus, is an international public good. Chapter 5 argues that the production of such information, especially research on a vaccine suitable for developing countries, deserves broad support from the international community. AIDS and Human Rights HIV/AIDS has created new human rights concerns and cast harsh new light on long-standing problems. Thus, the universally recognized obligation of governments to protect people from harm at the hands of others is a compelling reason for governments to play an important role in society's response to HIV Because people can be infected with HIV and transmit it to others for years before they become ill, the disease defines and creates a new minor- ity group in society. Government responses to the difficult task of bal- ancing the interests of the infected against the interests of others have var- ied widely. In Cuba, for example, HIV-infected people are confined to protect others from the risk of infection (Leiner 1994). At the other ex- treme, courts in the United States have upheld the right of an individual not to reveal his HIV infection, even to the extent of prohibiting govern- ment authorities from informing a woman of her husband's HIV infec- tion after his death (Burr 1997). Some prevention strategies manage to avoid tradeoffs between the rights of the infected and the rights of the uninfected, yielding benefits to both; we present evidence of the success of such approaches in chapter 3. More difficult are choices that arise in allocating public health care expenditures and in determining the extent and type of assistance to provide to surviving household members, issues we discuss in chapter 4. In all these instances, governments will unavoid- ably be involved in shaping social and legal attitudes toward the rights of the HIV-infected and the rights of those who are not infected. The human rights rationale for a government role related to HIV/AIDS is dramatic and unambiguous in instances where sexual rela- tions are compelled by force. Here the government's overarching respon- sibility in protecting individuals from physical harm and forced exploi- tation is reinforced by the public interest in preventing the spread of 43 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC HIV. The universal obligation of governments to prevent rape and invol- untary sexual servitude has been recognized in international human rights accords for decades. Although human rights advocates and pro- tectors of traditional mores may argue whether the arranged marriage of a 14-year-old girl should be condemned or protected, all would agree to vigorously enforce prohibitions against rape and the sale of individuals, often adolescents, to houses of prostitution. Always abhorrent, rape and involuntary sexual servitude have become even more reprehensible in an age when its victims may be involuntarily exposed to risk of HIV infec- tion. Governments that have been lax in prosecuting rape and forced prostitution must recognize that in the age of HIV/AIDS these crimes have become even more heinous than before. Social Nonns and Politics Make AIDS Challenging D ESPITE THE COMPELLING ARGUMENTS FOR GOVERNMENTS to confront AIDS, social norms and politics make designing and implementing effective AIDS policies uniquely challeng- ing. The specific problems and their solutions will vary across coun- tries. Nonetheless, four types of issues have commonly arisen: * denial that HIV/AIDS may be a problem * reluctance to help people who practice risky behavior to avoid infection * preference for moralistic responses * pressure to spend on treatment, at the expense of prevention. Denial is typically seen at the earliest stages of the epidemic, when the long asymptomatic period of the disease makes its effects nearly invis- ible. An extreme form of denial includes unwillingness to acknowledge that extramarital sex and illicit drug use exist in the society. Some offi- cials in societies with conservative social mores may be genuinely un- aware of the extent of extramarital sex or illicit drug use; the more con- servative the society, the greater the likelihood that such activities are clandestine. More typically, officials may be aware of such activities but lack information to evaluate their relevance to the AIDS threat. In such a situation, officials concerned about the possible negative reaction of their constituents may be reluctant to initiate the frank, public discus- 44 AIDS: A CHALLENGE FOR GOVERNMENT sion that could provide the basis for formulation and implementation of an effective HIV prevention program. In some instances, officials may acknowledge that HIV/AIDS poses a threat to society but nonetheless may be reluctant to advocate HIV pre- vention programs that focus directly on the people most likely to con- tract and spread the disease: sex workers, injecting drug users, male homosexuals and bisexuals with many partners, and others with high rates of sexual partner change. Although such approaches are most cost- effective-as we show in chapter 3-they may be impeded by two types of forces. On one hand, politicians and policymakers responding to the vast majority of their constituents, who do not engage in high-risk be- havior, may feel little pressure to focus prevention programs on those most likely to contract and spread HIV This is because few constituents are likely to understand the links between infection rates in people prac- ticing high-risk behavior and their own risk of infection. On the other hand, to the extent that people engaged in high-risk activities have polit- ical influence, they or their advocates may oppose prevention efforts focused on themselves, out of concern that such programs will fuel dis- crimination. Given the lack of demand for prevention programs for those practicing high-risk behavior from the majority of their con- stituents, and resistance from those who would be the primary benefi- ciaries of such programs, officials may find it easier to initiate a general public information campaign, even if this is less likely to reach those most apt to contract and spread the disease. Even if politicians and policymakers get past denial and the reluctance to target prevention interventions to those with the riskiest behavior, some interventions may well have broader social support than others. In many societies, encouraging abstinence from extramarital sexual rela- tions or from injecting drug use would be widely viewed as morally cor- rect, while providing free condoms for sex workers and their clients and offering clean needles to injecting drug users would be regarded by many people as facilitating immoral activity. Chapter 3 discusses why attempt- ing to proscribe potentially risky behavior and to encourage more so- cially acceptable, low-risk behavior, while sometimes politically expedi- ent, may have the unintended effect of exacerbating the spread of HIV. Societies and their governments should be aware of these costs when choosing how to confront the epidemic. A final political obstacle to an effective government response arises only after people start getting sick and dying from AIDS. At this point, people who are infected with HIV and their families may be strongly 45 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC motivated to lobby the government for subsidized treatment and care. We discuss the government response to the increased need for treatment and care in detail in chapter 4. Here it is sufficient to note that to the extent that such spending drains resources from effective HIV preven- tion, it will lead to more infections, sickness, and deaths. Overview of the Book T HIS CHAPTER HAS PROVIDED BASIC INFORMATION ABOUT HIV that the remainder of the book will draw upon in analyz- ing how society in general, and governments in particular, can identify the public priorities in confronting the global epidemic of HIV/AIDS. We next analyze the epidemiology of HIV to identify some key principles that are essential to an effective response. This analysis concludes that acting as early as possible to prevent infections among people most likely to catch and spread the virus-people who have unprotected intercourse with many partners and people who share nee- dles to inject drugs-would avert the largest number of secondary infec- tions, not only among similar individuals but also in the general popula- tion (chapter 2). Are such measures possible? Which approaches are most cost-effective? How can governments improve upon their current efforts? Examining the experience of countries in confronting HIV/ AIDS, we find that helping people most likely to spread HIV to protect themselves and others can indeed work and be extremely cost-effective. However, we also find that many governments have yet to implement programs with sufficient coverage of those most likely to contract and spread HIV or have failed to support these programs with broader social interventions, and so are missing valuable opportunities to prevent the spread of the epidemic (chapter 3). What steps can government take to mitigate the impact of AIDS on infected individuals, the health sector, and surviving household mem- bers? Even when resources are very scarce, there are humane and afford- able actions that governments can take to help people cope. However, these efforts should not draw resources from prevention, nor should government assistance be provided simply because of an AIDS diagnosis. Instead governments should integrate AIDS mitigation efforts with existing health care reform and poverty programs in ways that ensure that government assistance reaches those who need it most (chapter 4). 46 AIDS: A CHALLENGE FOR GOVERNMENT We next consider the strategic roles played by developing country governments, NGOs, bilateral donors, and multilateral institutions in financing and implementing AIDS policies within developing countries, and suggest ways in which these efforts could be improved. This discus- sion of partnerships concludes with an analysis of how public opinion and politics shape AIDS policy, and how developing country govern- ments can work with other players to confront the epidemic (chapter 5). The book concludes with a summary of the main policy recommenda- tions from the report for countries at various stages of the epidemic (chapter 6). Appendix 1.1 Alternative Estimates of the Current and Future Magnitude of the HIV/AIDS Epidemic E XTENSIVE SURVEILLANCE IN SOME COUNTRIES, COMBINED with spotty ad hoc surveys in others, allows an estimate of the extent of HIV infection for all countries in the world. Although more information is available about HIV infection than for any other important disease, the data in many countries are sparse or unrepresen- tative. Differences in judgment lead different experts to differing esti- mates of national infection, which aggregate to large differences in the estimated total number of HIV-infected people in the world. Since uncertainty also exists regarding the future trends in infections by country, differences among experts in the estimated current number of infections can translate to even larger differences in projections for the future. The number of people in the world who were living with HIV infec- tion in 1995 was variously estimated to be between 13 million (Murray and Lopez 1996) and 20 million (Global AIDS Policy Coalition, or GAPC). Two other estimates (UNAIDS 1996b and Bongaarts 1996) agree on the intermediate figure of 17 million. Figure 1.9 breaks down the global total by region for each of these four sources. The figure makes clear that the four estimates are in rough agreement about the number of HIV-infected people in Latin America, the Middle East, North America, and Europe. For Africa and Asia, however, major discrepancies exist, greater than those that would be expected from the one-year difference in estimates. In Africa, UNAIDS and the GAPC agree on estimates that are 40 percent larger than those of Murray and 47 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 1.9 Number of Adults Living with HIV Infection, by Region: Compaison of Estimates, circa 1995 (millions) Number of adult HIV infections 12 10 -0 0i UNAIDS O- GAPC 8 - Murray & Lopez 6 Bongaarts 6 4 2 0. _IF Africa Asia Latin Middle North Europe America East America Region Source: The four projections are by UNAIDS (1996a) for 12/95, the Global AIDS Policy coalition (Mann and Tarantola 1996) for 1/96, Murray and Lopez (1996) for 12/94 and Bongaarts (1996) for 12/94. Lopez and Bongaarts. In Asia, the GAPC is alone in predicting more than twice as many infections as any of the other three sources. Much of the disparity in Asia is due to uncertainty surrounding the number of infections in India, which is almost exclusively derived from a few scattered surveys of urban high-risk populations. Extrapolating HIV infection from these small urban samples to a country of over 850 million is highly problematic. As Asia's second-largest and most popu- lous nation, India has the potential to dominate the future course of the Asian epidemic and heavily influence infection estimates for Asia as a whole. Figure 1.10 presents the estimates from three of the same sources for the future course of the epidemic, as reflected in the mortality rate from AIDS in five regions.18 The disparity among HIV infection estimates extends to AIDS death rate estimates. While India was the main source of disagreement in estimates of current HIV infection, the countries of Eastern Europe and Central Asia (EECA) present an even greater chal- lenge to those who attempt to predict the future course of the epidemic. Although Murray and Lopez and GAPC both project a vanishing AIDS death rate in EECA through the year 2020 (figure 1. 10), more recent in- formation suggests a larger and potentially explosive epidemic. HIV in- fection has spread with extraordinary speed among intravenous drug users in the former Yugoslav Republic of Macedonia, Poland, and 48 AIDS: A CHALLENGE FOR GOVERNMENT Figure 1.10 Current and Projected Future AIDS Death Rate per 1,000 People, by Region, 1990-2020: Compailson of Estimates Asia Latin America and Caribbean Sub-Saharan Africa AIDS death rate (per 1,000) AIDS death rate (per 1,000) AIDS death rate (per 1,000) 0.8 0.8 -- 0.6 - GAPC 0.6 - 0.4, 0.4 -Bongaarts 25-GP / Bngaarts GAP C 0.2 -a 0.2 - Murray & Lopez 0 Murray& Lopez 2 1990 2000 2010 2020 1990 2000 2010 2020 1.5 Eastem Europe and Central Asia Established market economies Bongaarts AIDS death rate (per 1,000) AIDS death rate (per 1,000) 0.8 0.8 - 0.6 - 0.6 - 0.4 - 0.4 - Murray & Lopez 0.5 Murray & Lopez 0.2 Murray & Lopez 0.2 - GAPC 0- GAPC Mu 0 _iongaarts 1990 2000 2010 2020 1990 2000 2010 2020 1990 2000 2010 2020 Source: The three sets of projections are by Murray and Lopez (1996), Bongaarts (1996) and the Global AIDS Policy Coalition (Mann and Tarantola 1996). Since the Global AIDS Policy Coalition (GAPC) does not explicitly project an adult death rate, the projection for 2005 used here is derived from their estimate of the number of new HIV infections in 1995 (table 1.5) by using Bongaarts' (1996) rule of thumb that AIDS deaths in any given year will approximately equal the number of people infected with HIV during the year ten years before. Ukraine. For example, the percentage of HIV-infected injecting drug users in Nikolayev, a Ukrainian city on the Black Sea, rose from 1.7 per- cent in January 1995 to 56.5 percent eleven mnonths later (AIDSCAP and others 1996). Furthermore, dramatic increases in sexually transmit- ted diseases in the region suggest an increasing vulnerability to HIV infection. Projections for EECA that do not take into account these recent outbreaks are likely to underestimate the severity of the HIV epi- demic in those countries. The most dramatic differences in the projected size of the AIDS epi- demic are in Africa and Asia, where experts nevertheless agree that the impact will be large enough to measurably affect the growth and struc- ture of the population. Just as experts differ in their projections of the future number of deaths from AIDS, so they differ on the epidemic's impact on popula- 49 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC tion levels and growth rates. Although negative population growth is not projected for any African country, both life expectancy and the depen- dency ratio will be adversely affected.19 Differences in the projected im- pact of AIDS are sensitive to factors such as: the estimate of the infection rate in the base year, projection of future infection rates, the length of the asymptomatic period, the prenatal transmission rate, methodology, length of time from diagnosis of AIDS to death, the age and sex distri- bution of AIDS deaths, and the start year of the epidemic. For a detailed discussion of differing projections for Africa, see Stover (backgroundpaper, 1996). Notes 1. Garrett (1994) describes the first appearance of oped countries, by either measure (Bobadilla and others AIDS cases in several countries and the subsequent inves- 1993). tigation leading to the 1984 discovery that HIV is the cause of AIDS. 6. As of the end of 1993, an estimated 4.2 percent of worldwide TB infections were thought to be attributable 2. Appendix 1.1 to this chapter compares the Murray to the AIDS epidemic, with the share expected to rise to and Lopez projections, relied on in the text, with the 13.8 percent by the end of the century (Dolin, Raviglione, other, higher projections for the future course of the AIDS and Kochi 1993). In developing countries with severe epidemic. HIV epidemics, the share is even larger. For example, 39 percent of adult TB infections in Abidjan, C6te d'Ivoire, 3. HIV can be isolated from the saliva of an infected can be attributed to HIV (De Cock 1993). In Africa, 19.5 person. Although there are a handful of cases of trans- percent of all TB deaths in 1990 were attributable to HIV, mission through oral sex, there are no confirmed cases of and this was projected to increase to 29 percent by 2000 transmission via saliva alone. (Dolin, Raviglione, and Kochi 1993). Murray and Lopez 4. Murray and Lopez (1996) are unique in estimating (1996) exclude all HIV-positive people from their count current and future mortality disaggregated by disease. of TB deaths, even if the individual suffered from TB at Their estimates of the current and future death rates from the time of death. HIV/AIDS by region are smaller than those of Bongaarts (1996), especially in Africa, where Murray and Lopez 7. As this book was being finalized in the spring of estimate half the death rate in 2020 that Bongaarts esti- 1997, the World Health Organization (WHO) an- mates in 2005. Mann and Tarantola (1996) present much nounced that the new "directly observed treatment strat- higher estimates than Bongaarts. See appendix 1.1 at the egy" for TB (the DOTS approach) has been so effective end of this chapter for a comparison of various estimates. that the global number of TB cases is projected to remain flat rather than to increase. A revision of the Murray and 5. For comparison, infectious disease currently ac- Lopez projections, which takes into account this new de- counts for about 6 percent of the disease burden in devel- velopment, would decrease the number of TB deaths, in- 50 AIDS: A CHALLENGE FOR GOVERNMENT cluding TB deaths among HIV-negative people, attribut- urban residents at low risk is estimated to increase at the able to HfV. However, because the overall infectious dis- rate of 2.7 percentage points per year. ease burden would also decline, the relative importance of projected future deaths due directly to HIV/AIDS would 14. The European Commission has sponsored the de- increase. The same applies to projections for adult deaths velopment of a "toolkit" to assist planners in evaluating caused by infectious diseases, discussed below. the potential links between their projects and the HIV/ AIDS epidemic and incorporating those links into project 8. A similar figure constructed from DALYs would design (European Commission 1997). yield roughly the same conclusions. 15. In 1993 and 1994, the United Nations, the World 9. Because Murray and Lopez (1996) had not seen the Bank, and the U.S. Bureau of the Census released country- very recent data on STDs and HIV in Eastern European specific demographic projections for Sub-Saharan Africa, and Central Asian countries described in chapter 2 of this which for the first time reflected the impact of the AIDS book, they projected zero adult deaths from AIDS in epidemic on population growth. The United Nations and those countries in the year 2020. the U.S. Bureau of the Census updated their estimates in 1996. Stover (backgroundpaper, 1997) analyzes the sources 10. Individual-level data do not always show a nega- of the discrepancies among these alternative projections. tive relationship between individual or household income and HIV infection. Chapter 3 discusses the contradictory 16. Estimates of the macroeconomic impact of AIDS individual-level studies of this topic and possible ways to include Over (1992); Kambou, Devarajan and Over reconcile them with the aggregate finding reported here. (1992); Cuddington (1993); and Bloom and Mahal (1997). Ainsworth and Over (1994) review the literature. 11. Each panel of figure 1.7 presents the relationship between one of the societal variables and HIV infection 17. This argument is made in detail in Kremer (back- after purging the effects of the other seven explanatory groundpapers, 1996 a,b) and in Over (1997). The same variables. The figures were constructed by the avplot com- arguments apply when needle sharing is the mode of mand in the 1997 STATA software package. See Over transmission. (backgroundpaper, 1997) for details and further results. 18. Murray and Lopez (1996) present projections to 12. The index ofpoverty used in the upper-right panel the year 2020 based on the background paper by Low- of figure 1.7 is called the Gini coefficient and is defined on Beer and Berkeley (1996). Bongaarts (1996) presents pro- a scale from 0 to 1: 0 represents the perfectly equal distri- jections by region for the years 1995 and 2005. Since bution in which each person has exactly the same income Mann and Tarantola (1996) do not present their projec- and 1 represents the other extreme, absolute inequality tions in table form, the death rates by region were com- such that all income is received by one person. puted from their tables of new AIDS cases, January 1, 1995, through December 31, 1995 (to get the death rate 13. Since the dependent variable in these regressions is for 1995), and new cases of HIV infection, January 1, transformed into a "logit" as described in Over (back- 1995, through December 31, 1995 (to get the death rate ground paper, 1997), changes in the independent variables from AIDS in 2005). are associated with specific changes in these logits, which correspond to changes in the infection rate. All the results 19. Asian countries with lower birth rates than African control for the age of the epidemic, which is statistically nations are at risk of experiencing negative population significant. In the average country, prevalence among growth. 5' I CHAPTER 2 Strategic Lessons from the Epidemiology of HIV IV IS PRESENT IN ALMOST ALL COUNTRIES, BUT H | the speed with which the virus has spread has varied tremendously. In some countries HIV has infected only a tiny percentage of the population so far and its effects are all but invisible; in others, the virus has spread so widely that few families have been spared the tragedy of AIDS illness and death. Figure 2.1 shows the trend in the percentage of adults infected in various regions of the world. Only some of the differences across and within regions can be ex- plained by differences in the timing of the introduction of the virus. Consider the following: * As of 1996, infection rates were still rising in all developing regions. In contrast, infection rates appear to have stabilized in North America and Western Europe at relatively low levels, even though the virus was introduced at nearly the same time as in Africa and Latin America. * HIV has only recently been introduced in Eastern Europe and the former Soviet Union (not shown) but, as noted in chapter 1, the number of new infections is rising exponentially. * In Thailand and parts of India, HIV infection among sex workers has climbed precipitously; however, it has so far remained low among sex workers in Indonesia and the Philippines. * In Yunnan Province, China, and Manipur State, India, more than two-thirds of injecting drug users are infected, but in nearby Nepal the infection rate among injecting drug users has remained very low. 53 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 2.1 Estimated Trends in * The level of infection among pregnant women in urban areas of the Percentage of Adults Infected the Democratic Republic of the Congo (formerly Zaire) has lev- with HIV, by World Region eled off at 4 to 5 percent, but in Botswana and Zimbabwe rates are Percent Infected with HIV six times as high and still climbing. 2.5_ Africa * Infection levels are declining among Thai military recruits and young adults in Uganda. 2 What explains the different trajectories of the epidemic and what are the implications for policies to prevent HIV? In reviewing how 1. 5 /HIV spreads in populations and the behavioral and biological factors behind the epidemic, this chapter identifies some important epidemi- 1 ological principles that provide a foundation for the discussion in chapter 3 of government policies to prevent HIV. In the first part of the chapter, we review the determinants of the spread of HIV in pop- 0.5 L. America ulations. HIV does not strike individuals at random: both biology and N Amedca - - individual behavior affect its spread. Most of the variation in trajecto- 2 W. EUlfiff ries of the HIV/AIDS epidemic across world regions can be explained 0 by differences in behavior across societies and across groups within so- 1980 1985 1990 1995 cieties, which are in turn influenced by many of the economic and cul- Source: Bongaarts 1996, figure 2. Used tural factors described in chapter 1. In the absence of a cure or a vac- by permission cine, the key to arresting the spread of HIV is changing behavior. The epidemiology of HIV/AIDS suggests two important objectives for public programs to slow the spread of HIV, discussed in the second HIV is most widespread in Africa and part of the chapter: governments should act as soon as possible and, ir- rising in all developing regions. respective of the stage of the epidemic, they should ensure prevention of infection among those with the riskiest behavior, who are most likely to contract and spread HIV. The chapter concludes with an overview of the level and distribution of HIV in developing countries. The extent to which HIV has saturated subpopulations with high-risk behavior and spread outward to those with low-risk behavior has im- portant implications for government priorities in preventing HIV, dis- cussed in chapter 3. HIV Incidence and Prevalence, and AIDS Mortality T HE RATE OF SPREAD OF HIV AND CURRENT LEVELS OF infection are measured by incidence and prevalence. 54 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV * The incidence of HIV is the number of new cases, that is, the num- ber of people who become infected during a specified period of time, usually over a twelve-month interval. * The prevalence of HIV is the number of people currently infected with HIV at a given point in time. Because there is no cure for HIV/AIDS, HIV prevalence reflects the cumulative numbers of infections from the past and the mortality rate of those infected. Incidence and prevalence of HIV and AIDS are often expressed as a rate-for example, in terms of the number of infections per 1,000 adults. Figure 2.2 shows the relationship between HIV incidence, HIV prevalence, and AIDS mortality in a simulated epidemic for a typical Sub-Saharan African country. At the beginning of the epidemic, HIV prevalence grows rapidly and AIDS mortality is not yet evident because of the long asympromatic period of most of those infected. Years later, when the first cases of AIDS appear, large numbers of people are already infected with HIV. Incidence may still be climbing but growth in prevalence may slow because of rising HIV/AIDS mortality or satura- tion of the population. As long as incidence exceeds mortality, the prevalence of HIV will continue to rise. Prevalence will peak in the year in which incidence exactly equals the rising mortality rate. Whether prevalence then levels off, declines, or resumes climbing toward a Figure 2.2 HIV Incidence, Prevalence, and AIDS Deaths Infections per 1,000 adults 30 By the iie people start dying of AIDS, HIV prevalence may already be very 25_ Prevalenc high. 20 15 - _ .... 10 _ _ _ Incidence 5 AIDS deaths 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Years since the start of the epMiemic Source: Based on 1995 WHO data. 55 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC new peak will depend on whether the number of new infections- incidence-is equal to, less than, or greater than the number of deaths of people with HIV/AIDS. In the absence of a cure, the key to reducing future HIV prevalence is by preventing new cases-that is, lowering incidence. Stable or declining prevalence does not necessarily signal the end of the epidemic. Eventually, HIV prevalence will level off in all popula- tions; in some it will stabilize at a high level, and in others at a low level. The factors affecting the height of the plateau in HIV prevalence are discussed below. However, a plateau simply indicates that there is an equilibrium in which the number of new infections exactly offsetr AIDS mortality. In populations where prevalence is declining, mortal- ity is occurring at a faster rate than new infections. The number of new infections may still be quite high, coexisting with high mortality. The relationship between HIV incidence and prevalence and the lag in the appearance of AIDS cases have important implications for public policy: * Early intervention is critical to prevent an AIDS epidemic that can persistfor decades. Only a fraction of those infected with HIV are showing symptoms of AIDS at any given point in time. By the time that AIDS morbidity becomes a significant health issue, HIV may have spread widely in the population, making preven- tion efforts very difficult. Countries with few reported AIDS cases should not be complacent about launching prevention cam- paigns. While figure 2.2 alone is sufficient to make the case for early intervention, there are other compelling reasons that we shall return to later in this chapter. * The full impact of infection levels on mortality is delayed. Even if all new HIV infections could be prevented, in the absence of a cure, AIDS deaths would continue for years because of the popu- lation already infected and the long asymptomatic period between HIV infection and AIDS. Countries where HIV prevalence is high are only beginning to experience the profound mortality im- pact of the epidemic, which will last for decades even with the best prevention efforts. These consequences are laid out in chapter 4 and reinforce the case for intervening as soon as possible to pre- vent HIV STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Biology and Behavior Affect the Spread of HIV N OT ALL INFECTIOUS AGENTS INTRODUCED INTO A POPULA- tion will be self-sustaining. If each infected person transmits the infection, on average, to less than one other person over his or her lifetime, then the infection will eventually disappear; if to more than one other person, then the infection will expand. The repro- ductive rate of a sexually transmitted disease is the average number of susceptible people infected by an infected person over his or her life- time (May and Anderson 1987, Thomas and Tucker 1996).1 If each person infected with a disease transmits it to exactly one other person, then the reproductive rate is 1. In populations in which HIV has a reproductive rate of less than 1, the epidemic will not be self-sustain- ing. Thus, the greater the reproductive rate of HIV, the more rapidly the epidemic will spread. What factors, then, determine the HIV reproductive rate in various populations? We have seen in chapter 1 that the most common mode of transmission of HIV is through sexual contact. Three main factors have a large influence on the reproductive rate of all sexually transmitted dis- eases (STDs), including HIV: * the amount of time a person remains infectious * the risk of transmission per sexual contact * the rate of acquisition of new partners.2 These factors are similar for transmission through contaminated in- jecting equipment, except that the risk of transmission per contact refers to the risk per injection, and the number of partners refers to number of people with whom injecting equipment is shared. The broad points in the following discussion therefore apply to transmission through sharing of contaminated needles, as well as through sexual contact. Each of these three factors is in turn influenced by the biology of the virus and by individual behavior. Biology plays an important role in the amount of time a person remains infectious and in the risk of transmis- sion per contact. But individual behavior also has a strong influence on the risk of transmission per contact-for example, through decisions on condom use, disinfecting shared needles, and seeking treatment for other STDs. And individual behavior has a direct relation with the rate 57 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC of partner acquisition. Until medical science discovers either a cure or a vaccine, the most important avenue for reducing the spread of HIV will continue to be changing individual behavior. The Duration of Infectiousness The lack of a cure and the long duration of infectiousness are the main characteristics that distinguish HIV from most other STDs. The long duration of HIV infectiousness increases the likelihood that an in- fected individual will pass the infection to others. Further, because a per- son with HIV typically remains asymptomatic for years, an infected individual and his or her sexual (or injecting) partners are often unaware of the risk of transmission. Thus, the long duration of asymptomatic HIV infection potentially puts many more partners at risk than is the case for other STDs. The impact of recently developed drugs that extend the lives of peo- ple with HIV may lengthen the infectious period. However, if these drugs significantly reduce the viral load, they might reduce the risk of infection per contact. Be that as it may, unless both dramatic medical advances and significant reductions in the costs occur, these new drugs are unlikely to have a significant impact on the duration of infectious- ness in developing countries, since few developing countries have the financial or human resources to provide them. That leaves two primary mechanisms for prevention-reductions in the risk of infection per con- tact and reductions in the acquisition of new partners. The Risk of Infection per Contact The average risk of infection with HIV per sexual exposure is much smaller than that for other sexually transmitted diseases; however, be- cause of the long period of infectiousness and numerous cofactors that enhance HIV transmission, the chance that an HIV-positive person who does not take precautions will eventually infect others can be quite high. The most extensive studies of the risk of HIV transmission per expo- sure have been conducted in industrial countries. Because of generally superior health levels and the ready availability of treatment for other STDs, the average risk of HIV infection per sexual contact in industrial countries is quite small (table 2.1). For example, the average chance that an infected male will sexually transmit HIV to an uninfected female 58 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Table 2.1 Probability of HIV-1 Infection per Exposurea Aloth ofrrairsnzii;ion In lctious per 100 epoewurei Nfale-ro-tcmale. inpro[cued %aginal sex 0.I -0.2 Female-mo-male. unprortcred vaginal sexh 0.033-0.1 Niale-ro-male. unprortccrd anai sex 0.5-3.0 Needle stick° 0.3 M1other -io-child trisnnss ion 13-48 Expcsiure to coonraminattd blood producs 90-100 a. In the aktence ot otairor,. such as other STD- and v inm3tn. L nkncrowune" o.cr the mncubruion period. b Cakula[cd : 1 I t3ro I. cima- rthe ra[e or male ro temile vur-rwi Dabs ind other; 1'V13 DeG(rurtula mnd owhers l'S' Dunn and oEhurs i)q42. Eur,yein S Croud I K')2 Hal erLko and BlrJes I'PC; MI .rio and de \ incene P-I I',l.,r. Sh-bo;k-. arnd lew.el 199 Tokar' and OEhers 1993 partner by unprotected vaginal sex is estimated at between 1 and 2 per 1,000 exposures. The risk of transmission from an infected female to an uninfected male partner through unprotected vaginal sex is one-third to one-half as great (Haverkos and Battjes 1992).3 Thus, women are be- lieved to have a somewhat greater probability of becoming infected from an infected male partner than the reverse. Anal sex carries the highest risk, especially for the receptive partner. The risk of transmission in un- protected anal intercourse, based on a study of men, is estimated to be between 5 and 30 per 1,000 exposures for the receptive partner. How- ever, all of these figures very likely underestimate-perhaps severely- the average transmission probability per sexual act. They are generally based on studies of transmission within discordant couples-couples in which one partner is HIV-positive and the other is HIV-negative. Cou- ples that are discordant for a very brief period are not captured in these samples; thus, the most infectious individuals are likely to be excluded. These studies also fail to capture couples in which neither partner tests positive for HIV, but one recently has become infected. Below we review evidence that this may be the period of highest infectiousness. If true, then studies of discordant couples are measuring HIV transmission dur- ing a less infectious period (Mastro and de Vincenzi 1996). A transmission rate perpartnership, not taking into account the length of partnership, might be a more realistic measure of the risks of sexual transmission of HIV within relationships.4 A review of studies of per- partner transmission rates among heterosexuals in the United States and 59 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Western Europe found an average transmission probability of about 23 percent from men to women, and of about half that rate (12 percent) from women to men (Mastro and de Vincenzi 1996). Yet even these rates are probably lower than those faced in developing countries because many people in developing countries are infected with other STDs that enhance HIV transmission, an issue we discuss below. The "per contact" risk of HIV transmission with a commercial or casual partner in developing countries is thus likely to be substantially higher than the figures in table 2.1. Notwithstanding these considera- tions, the average infectiousness of HIV is believed to be substantially less than that for other STDs. In the case of gonorrhea, for example, the probability that an infected woman will transmit the disease to an unin- fected male partner during intercourse is 20 to 30 percent per exposure, while the probability that an infected male will transmit the disease to his female partner is 50 to 70 percent (Hethcote and Yorke 1984). Although sexual intercourse is the primary means of HIV transmis- sion in both developed and developing countries, other modes of trans- mission carry a higher probability of spreading the infection. The chance that a mother will transmit the virus to her infant is variously estimated at 13 to 48 percent. Transmission probabilities through sharing contam- inated injecting equipment between infected and uninfected injecting drug users are variable, depending on the specific injecting practices and which equipment is shared. The probability of transmission through an accidental needle stick in a medical setting, when the needle has been exposed to HIV-infected blood, is only about 1 in 250, or 0.3 percent. The transmission rate for transfusion of contaminated blood is nearly 100 percent. The risk of infection per contact is not a constant; it can be influenced by a variety of factors, some of which may tend to exacerbate the epi- demic. We discuss the most important of these below. Risk may be highest soon after infection. Recent studies suggest that infectivity can vary dramatically according to the stage of HIV in- fection. The two peaks of infectivity are thought to coincide with the periods of highest viral load-the first and highest within the first few months of infection (before the production of antibodies to the virus) and the second, which is thought to be lower, at the very end of the asymptomatic period, as the body loses its batde with HIV (Pinkerton and Abramson 1996). Studies of homosexual men suggest that an indi- vidual faces a 10 to 30 percent chance of becoming infected during a sin- 6o STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV gle act of unprotected receptive anal intercourse if his partner is at the early, acute stage of the infection Uacquez and others 1994). In the mid- dle stage, the likelihood of infection drops to between 0.01 and 0.1 per- cent, but at the end stage it rises again to between 0.1 and 1.0 percent. The variation in probability of heterosexual transmission in early and late periods of infection has not been estimated but could have impor- tant implications for the size of the epidemic (box 2.1). Greater risk of transmission immediately after infection may be one reason that the epidemic has taken off so rapidly in some developing countries. In Thailand, the average female-to-male risk of sexual trans- mission was estimated to be 3 to 6 infections per 100 exposures-much higher than the rates in table 2.1-perhaps because more people in Thailand were in the earliest, most infectious stage of the disease (Mas- tro and others 1994).5 In addition, the likelihood of HIV transmission also differs by the type of the virus. HIV- 1 is more easily transmitted and has a shorter incubation period than HIV-2 (De Cock and Brun-Vezinet 1996). HIV-1 has many subtypes with specific geographic distributions. However, there is no condusive epidemiological evidence to date that any of these subtypes are more or less infective than others (Anderson and others 1996, Expert Group 1997). Untreated STDs raise the risk of HIV infection per sexual exposure. STDs are far more common in developing countries than in industrial countries (table 2.2). Studies in both industrial and develop- ing countries have found that people with current or past STDs are 2 to 9 times more likely to be infected with HI .6 However, because HIV and other STDs are both highly correlated with risky sexual behavior- high rates of partner change in particular-it is difficult to disentangle the extent to which conventional STDs actually enhance the transmis- sion of HIV. Nonetheless, there are compelling biological reasons for believing that untreated ulcerative STDs such as herpes, syphilis, and chancroid greatly increase the risk of HIV transmission per exposure: the lesions caused by these diseases provide a ready portal for transmission of HIV, whether they are on the HIV-infected or the uninfected partner. En- hanced HIV transmission in the presence of nonulcerative STDs such as gonorrhea, chlamydia, or trichomoniasis is also biologically plausible but the epidemiological evidence to support it has been weaker, mainly for methodological reasons (Laga and others 1993). For example, a recent study in Malawi found that the amount of HIV virus in the semen of 6i CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 2.1 The Impact of an Early Peak in Infectivity NIEDIC.l RESEA RCHERS .ARE ,TILL LINSiL ABOUT infeciousness e%enruaJlv conmerge to the same HRl Lhe precise patu:rn oF infecrivin otFHI\ over its long pre%alence level. Ho%%e er. if the virus is more infec- inCLibition period. However. Ro. rnderson t 19961 live earlY in the incubation period. many more peo- has denionstrated that if H[V' is che most infecTious ple are cumulartvely intect ed. If people are most earl\ on, as some medical researchers suspect. HRV infectious immediately after becoming infecced and in;idence %vill climb niore rapidly and reach a higher before they develop antibodies to the \irus. the% peak prc%ilence than if it is equallv intectious earl ixould cest negatike for HI\ precisely %,hen rhey are and late in rhc incubation period. or i it iv more mo,t infiecrious. HIV could spread ern rapidls intectious lare in the iniubation period In Ander- a-mong those with high rates of partner change dur- son'> himulation. epidemics %%ith all three parterns of in, [his brief. highl% infcctiou, period Box Figure 2.1 Shape of the Epidemic Curve under Alternative Assumptions about Infectiousness Incidence of AIDS (per quarter-year) (thousandsl 11 10 Pattern of infectiousness over the Incubation 8 period 6 L 4 2 0 0 20 40 Time In years sm. ,^.e. ARLpud tic, in ,aider5en I co 1;w8. rigur -, Re i, nEcd t pCr11L151-sl..1 ItO%ford i n-,tra.17 P5r.-.. HIV-positive men with urethritis was eight times higher than in a con- trol group of HIV-positive men without it, and these concentrations di- minished significantly when the urethritis was treated with antibiotics (Cohen and others 1997). Women are more likely than men to have STD infections without any apparent symptoms, and therefore many infections go untreated. Half of all women with gonorrhea, for example, 6z STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Table 2.2 Estimated Prevalence and Annual Incidence of Curable STDs among Adults 15-49, by Region Prevalhece Poemilenca hence Inecieencepo Region1 (millions) per 1.000 ( lmillionslI 1.000per year Is ri.'9 r ;i, .':' Ir Nsorth Americi82 I-i 91 Australaia 11 A ;' Cl \'N;t ern Europc 10 -c I16 -- we>:'rlo,rw,g .r'aur!rie, Sub-Saharan .Africi 5 108 65 25i ho-th hAsia 121i 158 1 % 166 Latin America inrd Caribbean ' 4 36 ( 1-j Eastern Europe and Central Asia 1 2 I 11 2 Northern kfrica and Middle East 6 i 1li 60 East A6ia and Pacific 16 162 28 for ~~~~~~~~~~~~~~~~~ ~~~250 8 3.33 1/3 ..er1. in.:ludc' .-xphdIi .-nmorrhea. chhinad.a. richornon.i;i Inrernat,c'nal diffcrence' in STD pr.- alnce and in iddnce reflet diHcr nces in , urfl bch. ior inrd in heilth cire-c-c.inre chd. ..1-r V: H \'HC-P.GPA 19iS have no symptoms, compared with only 5 percent of men (Hethcote and Yorke 1984). Thus, if nonulcerative STDs do facilitate HIV transmis- sion, they are likely to differentially raise the transmission probabilities to and from women, since a higher percentage of women are likely to have asymptomatic STD infections that go untreated. A recent review found that, in eleven African countries, from 5 to 17 percent of pregnant women tested positive for syphilis; in Jamaica the rate was 5 percent, and in Haiti more than 10 percent (Van Dam, Dallabetta, and Piot 1997). Whatever the precise nature of the link between HIV and other STDs, there is evidence that treating symptomatic STDs reduces HIV transmis- sion. In the early 1990s a randomized controlled trial in rural areas of Mwanza region, Tanzania, found that treatment of symptomatic classic STDs lowered the incidence of HIV among adults by more than 40 per- cent (Grosskurth and others 1995a). The extent to which this result can be generalized to other countries is likely to depend on many country- specific factors, including the underlying prevalence of HIV and STDs, the types of STDs that are prevalent, the quality of treatment services, and existing levels of STD treatment prior to the intervention. At the outset of the Mwanza study, HIV prevalence among adults 15 to 54 was already high-4 percent (Grosskurth and others 1995b). Simulations of the HIV 63 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC epidemic in rural Uganda indicate that the proportion of HIV infections for which STDs were a cofactor was highest early in that epidemic (Robinson and others 1997). This suggests that the effectiveness of STD treatment in slowing HIV incidence in Mwanza might have been even greater had it occurred much earlier in the epidemic. Male circumcision may be a factor. Some researchers have found a correlation between HIV infection and lack of circumcision among men and believe that this may account in part for the rapid spread of HIV in Sub-Saharan Africa.7 Ethnographic studies suggest that men are least likely to be circumcised in central, eastern, and southern Africa, along a north-south swath through the Rift Valley (Bongaarts and others 1989). This also happens to be the area with the highest rates of HIV infection in urban areas. In 1989, in five countries where more than three-quarters of men were not circumcised, the urban prevalence of HIV was roughly 16 percent. In contrast, the average level of urban HIV infection was only 1 percent in 20 other countries where more than 90 percent of men were estimated to be circumcised. One reason why uncircumcised men could be at higher risk of con- tracting HIV and passing it to others is that they are at higher risk of developing ulcerative STDs, particularly chancroid. Poorer genital hy- giene among uncircumcised men may also play a role, particularly in low-income and unsanitary settings. A study in Kenya found that, even among men without chancroid, uncircumcised men were more likely to sero-convert (29 percent) than those who were circumcised (2.5 percent) (Plummer and others 1991). However, the amount of increased risk of HIV infection from lack of circumcision alone has not been established and whether or not such a risk exists is still debated. This is because cir- cumcision is highly correlated with many other factors besides chan- croid. In particular, ethnicity and religion are strong determinants of whether or not men are circumcised. It is therefore difficult to disentan- gle the effect of male circumcision (or lack of it) from that of other cul- tural norms that affect sexual behavior. Clearly, even if male circumcision is protective against acquiring and spreading HIV, it is not sufficient to prevent infection. High proportions of men in West Africa are circumcised, yet HIV has nonetheless spread rapidly there. More than three-quarters of U.S.-born men are circum- cised, but that has not prevented a sexually transmitted HIV epidemic in the United States (Laumann, Masi, and Zuckerman 1997). Conversely, in Western Europe and South America, circumcision is uncommon, yet 64 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV the HIV epidemics in those areas have not reached the scale of the one in eastern and central Africa (de Vincenzi and Mertens 1994). Behavior affects the probability of transmission. Although the basic transmission probabilities for HIV per exposure are founded in the biology of the virus, fortunately they can be substantially reduced through behavioral change. Using latex condoms and obtaining treat- ment for conventional STDs can lower the probability of transmission through sexual contact. Sterilization of injection equipment can dramat- ically reduce transmission among injecting drug users and among pa- tients in medical facilities. And mother-to-child transmission can be reduced through both medical treatment and behavioral changes. The prospects for changing behavior to reduce HIV transmission are dis- cussed in chapter 3. The Rate of Partner Change While transmission probabilities have an important influence on the reproductive rate of HIV, the rate of sexual partner change probably accounts for the greatest differences in the rate across groups and coun- tries. Similarly, the rate at which injecting drug users change partners with whom they share unsterilized injecting equipment strongly influ- ences the HIV reproductive rate among them. Finally, in medical set- tings, the rate of reuse of unsterilized injecting equipment for multiple patients is analytically equivalent to the rate of partner change (box 2.2). In all three situations, the higher the rate of partner change is, the greater is the likelihood that the virus will pass from infected to unin- fected people. In contrast, although the probability of becoming infected per expo- sure is higher for people receiving transfusions of infected blood and for the children of HIV-positive mothers, these groups are unlikely to infect many others. The rate of partner change among transfusion recipients, for example, is quite low, on average. Because the HIV reproductive rate for these modes of transmission is probably less than 1, if the virus were spread only by transfusion or from mother to child, the epidemic would most likely not be sustained. In the absence of condom use or steriliza- tion of shared injecting equipment, rapid rates of partner change sustain the epidemic. Both the average rate of partner change in a population and the vari- ation of the rate across individuals have an impact on the spread of HIV 65 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 2.2 HIV Transmission thmugh Medical Injections IN ALkNY PrARTS OF THE WORLD. MEDiCAL IRCY- cis. syphilis, malaria, and polio through unsterile ritioners or pacients prefer injections over oral med- injection equipment has been documented in devel- icarion-often because of patiens' or medical practi- oping councries. and unsterile injection equipment tioners' perceptions that injections are more effective is a major cause of abscess iNCvart 1993i. Patients in than oral medication. Swudies in countries as diverse countries hard-hit bx AIDS are often well aware of as India, Kenya, Nigeria, Llganda, and Vietnam have the riskls of H[\' cransmission by medical injection; confirmed rhe popularirv of injections among pa- suney-s in rwo communities in rural Llganda found tients IBirungi and Whyre 1993. Reeler 1990). They thar 63-83 percent of households kept their own are often favored by both public and private practi- needles and syringes at home to atoid havingy to rioners. including traditional healers, pharmacists. share w%ith others in medical situations (Birunpi. and professional injectors. As a result, behavioral stir- Asiimwe. and \W'hvie 109-4i. se%s conducted in 1989-90 found that from one- The number or share of HIV infections transmit- third to more than half of adults in eight of nine ted through unsterile medical injections is nor countries received medical iniections in the past year known and, given the frequcnicy of injections in (Ferry 1995). Armong those receiving injections. che many couniries. mighr be dicficult to idocumnen aserage was between three and six per year. 1The age distribution oF HI\' cases and [he cor-rcla- In the resource-scarce settings of developing tion w-ith other kno%%n risk factors strong,ly suggest countries, injecting equipment-including suppos- that sexual cransmission is nevertheless the main edly disposable equipment-may be reused on mode of transmission in most countries. Howevcr, mans patients without proper sterilization. A study in countries where HIV infection is w-idespread and of three dispensaries in Burkina Faso, for es-ample. wvhere sterilization practices are poor. the risk of t'ound that for everv 1,000 injections, the dispen- HRN' transmission by medical injection i., rcal. This saries used from 14 to 250 syringes and from -( to risk can be minimized by impro%ing sterilization T0O needles IWyatt 19931. If such equipment is not practices and by minimrzing the use of injections. properly sterilized between uses. HIV and other Because of the public and professional popularity blood-borne diseases can be transmitted bersteen of injections in many countries. minimizing their patients. in the same %ay that they are transmitted unnecessary use may require substantial public among injecting drug users. Transmission of hepari- education. in populations. Other factors being constant, the higher the average rate of partner change is, the higher is the reproductive rate of HIV. How- ever, in a population in which a few people have very high rates of part- ner change and many people have very low rates, HIV and other STDs will spread more quickly than if the same average number of partners were distributed more equally across the entire population (Anderson and May 1988, Over and Piot 1993). Surveys of sexual behavior suggest that there is in fact quite significant variation in the rate of partner change across subgroups in a single pop- 66 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV ulation.8 To take one example, figure 2.3 shows the distribution of men and women 15 to 49 in Rio de Janeiro, Brazil, who had at least one non- regular sexual partner in the previous twelve months, according to the total number of nonregular partners. There are two "peaks" in the distri- bution of people by their number of nonregular partners-one large peak among those with no nonregular partners or only a few, and an- other small peak among those with very many partners. Roughly half of men (56 percent) and 90 percent of women reported no nonregular part- ner, meaning that they either did not have any sexual partner or that they had sex only with their spouse or with some other regular partner. Those who did have nonregular partners usually acknowledged only a few. For example, roughly 12 percent of men and 6 percent of women reported Surveys of sexual behavior show that most people have no casual partners only one nonregular partner in the past twelve months. On the other or only a few, but a few people have hand, a small percentage of men-nearly 2 percent-reported having 20 very many partners. Figure 2.3 Distribution of Men and Women, Ages 15 to 49, with at Least One Casual Sexual Partner, by Number of Nonregular Partners in the Past Year, Rio de Janeiro, Brazil, 1990 Men Women (percent) (percent) 7 ±0 5- e ~~~~~~~~~~~~~~~~~4- 3- 4 _ _ 2- 2- 0 . . 11. . ., ,, 1 1 2 3 4 5 6 7 8 9 10 1±12 13 14 15±e6 17 18920+ ± 2 3 4 5 6 7 8 9 10 i 12 13 14 15 16 17 I8 1920+ Number of nonregular partners Number of nonregular partners Note: 56 percent had no nonregular partners. Note: 90 percent had no nonregular partners. Source: Backgroundpaper, Deheneffe, Carael, and Noumbissi 1996. 67 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC or more partners in the same period. The "two-peaked" distribution of people according to their nonregular sexual partners is typical of those found from sexual behavior surveys in other countries (background paper, Deheneffe, Carael, and Noumbissi 1996) . The variation, or heterogeneity, in sexual behavior is even more strik- ing in Thailand, where, in 1990, 28 percent of men 15 to 49 had a non- regular partner in the past year and almost 4 percent had 20 or more partners, while only 2 percent of women reported any nonregular part- ner.9 Sampling methods for these surveys are typically less successful in capturing sex workers, who are usually a small percentage of the popula- tion but an important component in the second peak of the bimodal dis- tribution. High rates of partner change in a very small subgroup may be sufficient to sustain an STD or HIV epidemic that can gradually spread to the rest of the population. Mixing patterns. The path of the epidemic within the overall popu- lation depends on the degree and pattern of mixing among people with high-risk behavior, and the mixing between people with high-risk be- havior and people with low-risk behavior. By "high-risk behavior," we mean unprotected sexual intercourse with multiple partners or sharing of unsterilized injecting equipment. People with high-risk behavior are very likely to become infected and to unknowingly pass HIV to others. People with low-risk behavior-who have few partners, who consistently use condoms, who do not inject drugs, or (if they do) do not share in- jecting equipment-are less likely to pass HIV to others. However, they are nevertheless at risk of becoming infected through transfusion of con- taminated blood or sexual mixing with people who practice high-risk be- havior. And young children are at risk of becoming infected from their mothers perinatally but are highly unlikely to spread HIV. In a sexually-transmitted HIV epidemic, the speed at which HIV spreads from people with a large number of partners to those with very few partners depends on the extent of mixing between people with dif- ferent levels of sexual activity. If people with large numbers of partners have intercourse only with others who are similarly active (known as assortative sexual mixing), then HIV will tend to rise rapidly within those groups but only very slowly and to a limited extent in the rest of the pop- ulation. As a result, the epidemic will achieve lower peak levels of infec- tion in the entire population than if those with large numbers of partners also have sex with those who have fewer partners (known as random or disassortative mixing) (Anderson 1996; Anderson, Gupta, and Ng 1990). 68 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Mixing patterns explain why HIV does not spread through a population at a uniform rate. Rather, it spreads in a series of smaller epidemics that race through overlapping subpopulations whose behavior puts them at various degrees of risk, then outward to those with less-risky behavior with whom they mix. Sex workers whose clients do not use condoms, injecting drug users who share unsterilized injecting equipment, and others with very high rates of partner change are typically the first to be infected in an HIV epidemic. HIV prevalence in these groups can rise very rapidly. Figure 2.4 shows rapid increases in HIV prevalence among sex workers in sev- eral cities in developing countries. Some of the differences in the rate of increase across cities can be attributed to differences in the timing of the introduction of the virus. However, other factors also appear to be at work. HIV prevalence among sex workers in Santo Domingo, Domini- can Republic, has risen more slowly; this is believed to be due to very high rates of condom use among brothel-based sex workers (Peggy McEvoy, personal communication). HIV tends to move even more rapidly among injecting drug users who share injecting equipment than among sex workers because the risk of transmission per contact is much greater. In countries where those who inject drugs commonly share injecting equipment, HIV can infect Figure 2.4 Increasing Prevalence of HIV among Sex Workers, Seven Cities in Developing Countries, 1985-95 HIV seroprevalence (%) 100 HIV prevalence has reached high levels and continues to climb among Nairobi, Kenya female sex workers in many develop- 80 -. ing countries Abidlan. Cole diholie 60 - Pune. Indla 40 - Phnom Penh. Cambodia Bangkok. Thalland 20 Mandalay, Santo Domingo. * Mvanmnar 0 , Dominlcn Republic I I 1985 1987 1989 1991 1993 1995 Source: U.S. Bureau of the Census (database), 1997. 69 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC the majority of users in a matter of months, as has been the case in parts of Asia and in Ukraine (figure 2.5). Figure 2.6 shows the incidence of AIDS in various population groups in six areas of Latin America. In Brazil, the first wave of the epidemic was among men who have sex with men.10 This was followed a few years later by an epidemic among injecting drug users, most of whom are also men. Later, the disease spread to sex workers and the female partners of bisexual men and injecting drug users. The timing and pattern of the epidemic waves can be very different, even within a single region. In the Andean Area, Mexico, and the South- ern Cone, the epidemic first struck men who have sex with men, as it did in Brazil. In the Caribbean and the Central American Isthmus, hetero- sexual transmission quickly outpaced transmission in other groups. In Thailand, HIV spread first among men who have sex with men and in- jecting drug users, and then spread among sex workers and their clients. Researchers have since determined that the epidemics among injecting drug users and sex workers in Thailand were largely independent, spreading two distinct variants of the virus (Ou and others 1993). Figure 2.5 Rapid Diffusion of HW among Injecting Drug Users, Asia and Ukraine, Various Years HIV prevalence (%) 80 1989 so 1990 1990 ---_9199 Once HW is introduced among inject- 20 ing drug users who share needles, HIV 1989 prevalence can go from practically zero to near saturation levels in just a _3 1987 1988 1995 few months. 0 _ _I | _ |, * r * n Yunnan Manipur Myanmar Bangkok, Chiang Ral, Nikolayev, Province, State, Thailland Thalland Ukraine China India City/Country Source: Stimson 1996; for Ukraine, UNAIDS 1996d. 70 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Figure 2.6 Annual Incidence of AIDS Cases in Latin America and the Caribbean according to Risk Factor, 1982-95 Cases Andean Area Cases Brazil 1t,00c 5,000 Homo/bisexual Homo b*lexua 800 - - 4.000 600 - -- - 3.000--- - 400 2-- -- -- 2.000 - - Heirersexual ~ . / / 200 - Ote 1.000 - e - Olher 0 I 0 83 84 85 86 87 88 89 90 91 92 93 94 95 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 Cases Caribbean Cases Central American Isthmus 1,000 1,000 Soo lo 800Heterosexual/ 600 Soo 400 Heterosexua 400 7 ~~~~~~~~~~~~~~~~~H o mlo biseso al 200 200 - Hom o bi.e-uai 20l/r Other 8182 83 84 85 86 87 88 89 90 91 92 93 94 95 82 83 84 85 86 87 89 90 91 92 93 94 95 96 Cases Mexico Cases Southern Cone Homo/bisexual ,500 400 Homc. b&.?ua.a | IDU 1,000 ~~~~~~~~~~~~300 t 000 ~~~~~~~Heterosexua30 / 200 vit /200HIV moves through sub- Soo / / populations with different ,/'-v ... 8|°°X! 00 heterr,e %ua degrees of risky behavior / Other 2 olrer in a series of overlapping o I . I I I I I _ I epidemics. 82 83 84 85 86 87 88 89 90 91 92 93 0 2 83 84 88 86 87 88 89 90 91 92 93 94 IDU Injecting drug users. Andean Area: Bolivia, Colombia, Ecuador, Peru, Venezuela. Southern Cone: Argentina, Chile, Paraguay, Uruguay. Source: PAHO data, 1996. 71 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Subpopulations that practice high-risk behavior are constantly in flux (Weniger and Berkley 1996). Over time, some individuals abandon high-risk behaviors or die, while others initiate high-risk behavior, adding to the high-risk subpopulation. High-risk behavior usually varies over an individual's life cycle. Sexual activity is often highest among young, unmarried adults, particularly men. As men and women marry and age, levels of casual sex typically decline. The age profile of the per- centage of men and women engaging in sex with a nonregular partner in figure 2.7 demonstrates this well, although it also reflects to some extent temporal changes in social norms. As we shall see in chapter 3, changing socioeconomic factors can also induce people to adopt or abandon risky behavior. The dynamic nature of the subpopulation practicing high-risk behavior at any point in time prevents their HIV-infection rate from reaching 100 percent. If there is very little mixing between people with different degrees of risky behavior, the overall epidemic may have multiple peaks. Incidence may rise and fall several times as first one group and then another be- comes nearly saturated by the virus. Thus, a sustained decline in inci- dence in a specific group of people with risky behavior does not neces- sarily signal the end of the epidemic in the entire population (Anderson 1996; Anderson, Gupta, and Ng 1990). Concurrent partnerships. Partnerships that overlap in time are con- current partnerships. Examples of concurrent partnerships include: part- nerships between married men or women and commercial or casual sex- ual partners; people engaging in long-term relationships with more than one casual partner; and polygyny, the practice of having more than one wife. In two populations in which individuals have the same average number of partners in a given period, HIV and other STDs will spread more rapidly in the population in which partnerships are concurrent than in the population in which partnerships occur sequentially (background paper, Morris 1996). This is because when partnerships are sequential, the virus cannot spread to a new susceptible person until the dissolution of one relationship and the start of another. In a concurrent partnership, it can infect more susceptible people in a shorter period of time. Recent research has focused on the role of "bridge populations" in the spread of HIV. These are partnerships that link people in groups that otherwise might not mix, such as partnerships between people with high-risk behavior and those with very low-risk behavior (background paper, Morris 1996; Morris and others 1996). For example, men who 72 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Figure 2.7 Probability of a Casual Sex Partner over a 12-Month Period, by Age and Gender Men Probability of casual sex (%) 70 Rio de Janeiro (Brazil) 60 50 - - _ 40~ Tanzania - -. 30 Thailand 20 10 Central African Republic 0 Sri Lanka, - I I l 0 15-19 20-24 25-29 30-34 35-39 40-44 Age Group Women Probability of casual sex (%) 25 Kenya 20 Rio de Janeiro (Brazil) Is _ Tanzania 10 - Levels of sexual activity vary over an Central Afrian Republic individual's life cycle. Sexual activity Central Afrca Republic --is often highest among young, 0 0 2 , , , unmarried adults. 0 15-19 20-24 25-29 30-34 35-39 40-44 Age Group Note: These results control for schooling, urban residence (where applicable), and occupation. Source: Backgroundpaper, Deheneffe, Carael, and Noumbissi 1996. have unprotected intercourse with sex workers and have either a wife or steady girlfriend may transmit HIV to monogamous women who would not otherwise be at risk. The extent of such mixing in Thailand was re- cently captured by a survey of sexual behavior among low-income men and long-haul truckers in three provinces (table 2.3). The bridge popu- lation comprises a very large percentage of both groups-about 17 per- cent of low-income men and 25 percent of the truckers. Sexually active 73 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Table 2.3 Distuibution of Two Samples of Thai Men, by Type of Sexual Network, 1992 I oai-inanjue m1.7n Tri,rrk s ,Venerork rypt 55 IV i, A An I, /r,ik, k.2-,_ 81' I 2_ ' i No parrntr 16 .1) 1 I Sii \\'i[e ink', 0~ -i-I 4'.4 Ohchir unk C..~ 63 i-ikc . o[htr ' - 'S 41 "e\ z .orI:tr u-nIv 21 'I 1'' ;4' 16l. / 2i I if * >.\. %%orker -.' li. lOh.r , ie %c.rkc r ktr -r.3 1- WiFe c.mrher + sex wc:rker '.1 21 5 IS ori 'F / Or 1 to 9.'2 ' 0Il I 3, - . Nurnbf.*t.i rn~n . Reercrn- pari(d , in rhe pa.r l! rrn.:-h. injecting drug users are another potential bridge population. In Manipur state, India, within two years of the first reported HIV case among drug injectors, 6 percent of the noninjecting sexual partners of injecting drug users were infected (Sarkar and others 1993). The impact of heterogeneity, mixing, and concurrency: A simula- tion. The combined effect of heterogeneity in sexual behavior, mixing between groups of different behaviors, and concurrent partnerships can have a profound impact on the potential course of the HIV/AIDS epi- demic, absent any intervention. Figure 2.8 shows simulations of hetero- sexual HIV epidemics in four imaginary populations with different un- derlying patterns of sexual behavior (background paper, Van Vliet and others 1997). The four simulated populations are identical in every re- spect, except in their patterns of sexual behavior, as follows: * Commercial and casual sex: In this population, some men engage in commercial and casual sexual relationships before and after mar- riage; some women have relations with casual partners before mar- riage but all women are monogamous after marriage. The com- mercial and casual partnerships can be concurrent with each other and with marriage. Sex workers comprise about a one-fifth of 1 percent of all women. 74 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Figure 2.8 The Impact of Different Baseline Patterns of Sexual Behavior on a Heterosexual HW/AIDS Epidemic: STDSIM Results Adult HIV prevalence (%) 35 30 Commercial and casual sex 20 ____ ______ Casu~~~~al sex 10 _ /_Simulaions show that HIV spreads Commercial sex most rapidly in populations with 5 _ < _ __ Serial monogamy _ _ concurrent commercial, casual, and marital sex, and last rapidly if there 0 _ I_I_,_,_I_,_,_I_I_I_I_I , , I are no concurrent partnerships. 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 29 30 Years since the beginning of the epidemic Source: Backgroundpaper, Van Vliet and others 1997. * Commercial sex: Some men engage in commercial sex before and after marriage with a small group of female sex workers comprising only one-quarter of 1 percent of the female population. Neither men nor women have casual sexual relationships before or after marriage. Commercial and marital relationships can be concurrent. * Casual sex: Some men and women have casual sexual relationships before and after marriage. There are no sex workers in this popula- tion, but casual partnerships can be concurrent with each other and with marriage. * Serial monogamy: Some men have commercial and casual sexual re- lationships before marriage and some women have casual relation- ships before marriage. After marriage, neither men nor women have extramarital relationships. All relationships before marriage are serial; this is the only one of the four populations with no con- current relationships. Sex workers are 0.05 percent of the female population. Using the STDSIM simulation model described in box 2.3, it is pos- sible to see how the course of the epidemic would differ in these four 75 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 2.3 STDSIM: Modeling Behavior and Sexually Transmifted Diseases THE S i i L.\ Tl I-Ns 1N Pii -HAFTF R F NT ['4 ti1:innhips bcr%ecen men and ",omen arc explicily chaptcr I c:re conduc[cd by researchers at the modeled. %iith parameters tfor [he freqUen.a ot inter- Eraimui LUniv%rsirs . Rotterdam. [he Nerherland4. course and the duration of each rlariomnship. \Vhen using STDSNIN1. a computer model he; developed a imulated person becomes infected. the program to estimate [he rinsnmission of and impact of prc- considers that his or her pariner nia; also become -cnre In icer'entio-n4 on HIV and four classiic intered S [[rs-_onorrhea. chlam%dia. siphiIi. 3nd chan- Foir example. [he litc history of a man and a roid Vm-kg an 'ut paera. Yin Vh cc and other' 'oniman %%ho enter into a stead; relationship might Iqcfl~ The STD)SINI model can describe the spread evo!ve as folldoN,. Before [he nmarriaige. beth have of HIV and other S lDLs in populations irh diffcr had o)ther sexual relationships After the rriarriagc. ent demographic characteristics. se:ual beha% or. the manl continues ca,ual rclationships vwith other health care sertings. and feasible intrsenctons. scomen and be--mes intfcted w%ith an STD. s;hich Each STDSINI simulation show-s the cumulatite he transmnits to hi,s %%ifr. The man seeks treatment oucconie of inceractions bertween a large number of for his STD bur quite ioon hc is rcintected b% his hNpothetical indiidlulls. Each indiiduaI has a life wife, w%ho is unav.are that she is infected. This historv anrd specitc characrerisuics. so,me ot "hich process might repeat icself until b,th pirrncrs sek- renain constant khile orhers changce. In rhih model, treatment for their STDs. The STDSINI nmodel is indi%iduals srart and end sexual relationships. eon- described in detail in Van der Ploeg and o)thcrs tract STEs. and. e:cept fi-or HRV are cured. Rela- i, I 09-). imaginary populations with no specific intervention or behavior change.'1 Baseline levels of condom use are assumed to be 5 percent among casual partners and 20 percent among sex workers. HIV is intro- duced in the population in year zero. The results in figure 2.8 show the trends in HIV prevalence for the entire population, including members with high- and low-risk behavior. The first curve, at the top of the figure, shows the path of the epidemic in the population in which people have concurrent relationships and commercial and casual sex. Thirty years into the epidemic, HIV begins to show signs of leveling off, but at a very high level-30 percent. In the population in which there is only commercial sex and marital sex, HIV prevalence peaks eight years after the virus is introduced at about 13 per- cent, then declines to an equilibrium prevalence of about 8 percent 20 years into the epidemic. Prevalence declines even in the absence of any behavior change because the people with the riskiest behavior, infected at the outset, begin to die. Most new cases of HIV infection at that point occur among people who have recently adopted high-risk behavior. In 76 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV the population in which there is casual but no commercial sex, the epidemic progresses more slowly, reaching a prevalence rate of only about 3 percent eight years after the start of the epidemic. However, prevalence continues to climb, reaches 30 percent, and is still increasing 30 years after HIV is introduced. Finally, in the population that practices serial monogamy with commercial, casual, and marital partners but no concurrent partnerships ("serial monogamy," the lowest curve), HIV prevalence rises more quickly than in the population with only casual sex, but levels off at about 9 percent. Thus, even without behavior change to prevent HIV, the course of an HIV epidemic can be quite dif- ferent across populations, depending on the heterogeneity of behavior, the extent of mixing, and the degree of concurrency in partnerships. Implications for Public Policy T HE PRECEDING DESCRIPTION OF HOW HIV/AIDS SPREADS through populations, particularly the role that variations in indi- vidual behavior play in determining the course of the epidemic, have important implications for government policies to prevent HIV. Act as Soon as Possible Governments that invest in effective prevention before HIV/AIDS becomes evident can avert suffering, save lives, and avoid potentially massive expenditures on AIDS treatment and care. To demonstrate the importance of changing behavior as early as possible in the epidemic, fig- ure 2.9 shows the impact of reducing HIV transmission by a moderate increase in condom use in the hypothetical population in which HIV is spread by both commercial and casual sex.12 The top line shows the course of the epidemic without any intervention or spontaneous behav- ior change; the lowest line shows the impact of raising condom use from 5 percent to 20 percent among men with commercial and casual sexual partners, three years after the start of the epidemic; the middle line shows the impact of the same increase in condom use 15 years after the start of the epidemic. An intervention that succeeded in raising condom use rel- atively early in the epidemic would reduce peak HIV prevalence from 30 percent to 20 percent. Increased condom use that occurs later prevents HIV prevalence from rising, leaving it constant at 22 percent. By the end 77 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 2.9 The Impact of Increased Condom Use by Men on Adult HIV Prevalence, Early and Late in the Epidemic 30 Adult HIV prevalence (%) Baseline 25 Increased condom use. year 15 20 7 15 Increased condom use. yeat 3 10 5 Early interventions to change behavior will reduce peak prevalence and save many lives. 0 1 , I , I , , , I , I , , 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Years since the beginning of the epidemic Note: Baseline is the STDSIM model in which HIV is spread by commercial and casual sex and 5 percent of men use condoms. In the simulation, condom use rises to 20 percent. Source: Background paper, Van Vliet and others 1997. of the 30-year simulation, the earlier intervention has prevented more than twice as many infections and three times as many deaths as the later intervention (backgroundpaper, Van Vliet and others 1997). Note, how- ever, that the level of behavior change modeled in year 3 of the epidemic, when prevalence is already high (5 percent), is not early enough to pre- vent substantial infection of the population. There are several reasons why early intervention to change high-risk behavior is preferable to later action. Early in an epidemic, HIV spreads exponentially. Because very few people are infected, the probability is greater that unprotected sex or needle sharing involving an infected per- son and a random partner will result in a new infection. Further, if the viral load is highest in the first months of infection, as scientists suspect, early in the epidemic a very large proportion of those infected will be highly infectious. Later in the epidemic, unprotected sex between an infected person and a random partner is relatively less likely to result in transmission-both because of the lower average infectiousness and be- 78 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV cause there is a greater likelihood that a random partner is already in- fected.13 Another reason for acting early is that interventions cannot be put in place instandy; a period of trial and error may be necessary to dis- cover which approaches work best in particular settings. Finally, as will be shown in chapter 4, from a budgetary perspective it is far less costly to prevent HIV infection than to treat people with AIDS. As we shall see later in this chapter, governments representing nearly half of the population of developing countries-2.3 billion people-are in areas with nascent HIV/AIDS epidemics and are still in a position to act early and decisively to prevent a widespread epidemic. Among those countries are China, certain states of India, Indonesia, the Philippines, Eastern Europe, North Africa, and the Middle East. Yet even in coun- tries where HIV is already prevalent among those with risky behavior, there are still opportunities for prompt intervention to prevent the spread of HIV into the next wave of people susceptible to infection. Despite the compelling case for acting early, governments have often been slow to respond. This is understandable in those countries where HIV struck first, since very little was known in the 1970s and early 1980s about how HIV was transmitted and how its spread could have been prevented. In particular, the existence and implications of the long asymptomatic period of HIV infection were not well understood. Coun- tries where the epidemic hit later have had the opportunity to benefit from increased knowledge about the disease, and governments are be- coming quicker to launch national AIDS prevention programs, no doubt due in part to the efforts of international programs like the Global Programme on AIDS and UNAIDS (box 2.4). Yet even today, some policymakers dismiss the significance of a relatively small number of re- ported AIDS cases, not recognizing that such a situation is precisely when prevention is needed. As recently as 1994, a senior health official in one populous developing country chided reporters that "to make a few thousand reported AIDS cases into an epidemic is just absurd." Tragically, such attitudes persist, despite the now widely understood fact that reported AIDS cases are but the tiny, visible portion of a deadly, mostly invisible, and potentially explosive HIV epidemic. Officials who ignore the signs of a pending epidemic miss an opportunity to mount an early, less costly, and highly effective response. There are many reasons why policymakers may be slow to confront HIV/AIDS. In virtually every country, the HIV/AIDS epidemic has been preceded by a period of denial. AIDS is perceived to be a problem 79 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 2.4 Are Governments Intervening Early Enough? THll I-it-ST w_;)UNT R[iFs Hii Bi\ THE HI\ kiD)s tL I Box Frgure 2.4 Lag between the First Reported denic in the la[e Iel-os and early Si.'it: were AIDSCaseand Initiationofa NationalAIDS caught hr sorpri>e: the 'iral source AIDS, the Control Progam, 103 Industrial and Developing alg .1svymptromatic period of H IV infecion. and Countries h aes char HI' is sprcad "%crc nor %cll under- tod(.d. Go%crrement, in these countries *%ere n.-ot Lag In years abhle- tol launch prevention progranmi untli %ear; aFter thL introduction of H[\. in man% .Mrictrn co.un[ric' HIV had already ipread to the general population. 4 and the numnbcr. of AI[)S cases " er, already ri.ing 1 at alarming rates. (Councries %%hcre HIX "as in[ro- 3 - duced later ha. e had an oppor[unrn tto learn trom thi, tr3gic experience and to act early- to pre%ent a 1.5 ' idelpread epidenlic Haae rhe% done o: I The :,ns'ser appear, [o be 'rc:. Box tigure '..-i 0.1 1990-94 .lhowus [he imount of time berx%een the first repnrred o _- . AIDS case aid thie rtart of a go'ernnient narironl 1980-8 1985-86 1987-89 A.IDS progranm for 103 countries tcr w%hrch such -1 YearofthefirstreportedAIDScase data are aadlble. In nearl, a third otf rhc countries. the firsi case e. AIDS l.as reported betore 19S -2 -1.1 For this group. %%hrch includ.e, most of the indus- \u ,; . t4;c-Ijial, b.t.iJ -,i. d,i1 r..m LuNAiDS trial countries. nationial AIDF control program, i.-unrr Suppcrr I inN an Iri c,ird - itifli aI 5) il wcere launched an average of almost the years after rite firit reported case. More thant a tlhird i.. per- Iccn of the countries reported their first AUDS case in 19 S5-S6z: tihesc countries launchcd AIDS pre- Qhich helped countries to de%elop narional AIDS '-enrion projgram an a%erage of IS month, later programs and action plan%. Fully ol) percent of the Among countrie, that had their first reported AIDS countries in ho-e figure -.i launched their national Lase in I 9C00-i4. national AIDS programs %%cre AIDS program in I'iS6-.iS. launched an average of one 'ear before the first Atrhough the launching of a national AIDS pro- repForted AIDS case. Viernarn. for example. gram is a w%elcome indication that poliie-smake-rs rec- launched its program in 1 i1t% three years before og-nize the importianc of prevention. organizing and clieir first reported AdDS case. implemen[ing an effective program takes time. In fact. the improved response cime cf country Further. an early re>ponse does r neces'arrlv fore- AIDS pre%rntion program' w%as largely associated shados the quaiinr ot the response. the coverage ot %i[th thc mobil7iaiorn ot the internatronal commu- progranms. or rhu dcgree ot political commitment. njrr in the mid- l'1is to launch international anti which in large part deterinile a progranm's efltctkie- AiDS prograns. This movement ssa> spearheaded nes,. Chapter 3 -,\ill examine the a'arlable cross- bv \WHO and irt Global Programme on Al DS. countr) dara on the quality of the national response. 8o STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV of foreigners, an imported problem confined to tourists or to people who have lived abroad. Injecting drug use and unprotected sex with multiple partners are said to be problems only in other countries, not in one's own country. The lack of reliable data on HIV prevalence and risk behavior makes the potential for an epidemic easy to ignore. Even when officials and politicians understand that the conditions for an HIV epidemic exist in their country, they may be reluctant to raise the issue publicly or to initiate programs to prevent HIV among people engaged in high-risk behavior, who are often stigmatized. Or, they may believe that the lim- ited resources available would be better spent fighting more common health problems, such as tuberculosis and malaria. Finally, officials and business leaders have sometimes attempted to conceal information about the extent of the epidemic because they fear it will discourage tourism or investment. Although some of these obstacles cannot be dismissed lightly, failure to overcome them and act early to prevent the spread of HIV can be ex- tremely costly. In rich and poor countries alike, denial and other social and political barriers have delayed interventions that might have signifi- cantly reduced the epidemic, saving money and, most important, lives. The Multiplier Effect of Changing the Highest-Risk Behavior We have seen that an HIV epidemic can only persist if the reproduc- tive rate of HIV is greater than 1 in at least some subgroup of the popu- lation, and that the shape of the epidemic is largely influenced by the de- gree of mixing between people with different degrees of risky behavior. Other things being equal, those with the highest rates of partner change are most likely to contract and spread HIV (box 2.5). It follows that pre- venting HIV infection in someone with a high rate of partner change- sexual or injecting-will indirectly avert many more future infections than preventing infection in a person who practices low-risk behavior and is thus less likely to infect others. Therefore, for HIV, as for other sexually transmitted diseases, the most efficient strategy for reducing the spread of the disease is to prevent transmission among those with the highest rates of partner change (Hethcote and Yorke 1984, Over and Piot 1993). Prevention of infection among those with the highest rates of partner change has a "multiplier" effect in terms of preventing many more subsequent, secondary infections, most of them among individu- als with low-risk behavior (box 2.6). QT CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 2.5 Who is Most Likely To Contract and Spread HIV? NOT EVERYONE IN THE POPULATION IS EQUALLY ulation, over time. In general, over the course of a likely to become infected with HIV and to transmit year, most people have no casual sexual partners or it to others. People with large numbers of sexual and at most one or two; their reproductive rate of HIV needle-sharing partners who do not take precautions will tend to be quite lo Some pople . however, to prevent HIV infection by using condoms or clean l1ac se i ral partners and a somc%% hat higher repro- injecting equipment are very likely to become in- ductive rate of HIV, and a much smaller number of fected. These unsafe practices then . III create many people have very many partners and a very high opportunities for HIV to spread through the popu- reproductive rate of HIV, much e,rearer thin one. lation. On the other hand, people who have no sex- TD .- !. y portion of this distribution-that i'.. Pc pIr' ual or injecting partners, or very few partners, or who have the greatest numbers ofpartners amd v,i ',, r who always take precautions to avoid contracting use condoms or clean injecting equipment-are those and spreading HIV, are very unlikely to become most likely to contract and spread HIV Ailhou_h infected or to infect others. these individuals are a very small peicrntage ol th. The distribution of individuals according to the total population, enabling their o adopt safer likelihood that they will contract and spread HIV behavior, thereby protecting themscki ind oihcrs. varies across populations and, within any given pop- is essential to curbing the epidemic. How much more effective is preventing transmission among people with the highest rates of partner change? The top panel of figure 2.10 compares the impact of slowing transmission by raising condom use in three different subpopulations in a heterosexual epidemic: women in steady relationships, men engaging in commercial and casual sex, and sex workers. The top line shows the baseline increase in prevalence in the absence of any change in condom use and is identical to the baseline epi- demic in figure 2.9. The second line shows the impact o(instantaneously raising condom use from zero to 20 percent among women in steady relationships. Because these women do not have a high rate of partner change, higher condom use has almost no effect until HIV prevalence reaches about 15 percent, and only a small effect thereafter. The third line shows the impact of instantaneously increasing the rate of condom use from 5 to 20 percent among the 40 percent of men who have a com- mercial or casual partner annually. 14 This intervention, like the previous one, has very little effect for several years, until prevalence reaches about 13 percent. The fourth and lowest line shows the impact of increasing from 20 to 90 percent the rate of condom use among sex workers, a mere 0.20 percent of the adult female population.15 Raising condom use 8z STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Figure 2.6 Annual Incidence of AIDS Cases in Latin America and the Caribbean according to Risk Factor, 1982-95 Cases Andean Area Cases Brazil 1t,00c 5,000 Homo/bisexual Homo b*lexua 800 - - 4.000 600 - -- - 3.000--- - 400 2-- -- -- 2.000 - - Heirersexual ~ . / / 200 - Ote 1.000 - e - Olher 0 I 0 83 84 85 86 87 88 89 90 91 92 93 94 95 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 Cases Caribbean Cases Central American Isthmus 1,000 1,000 Soo lo 800Heterosexual/ 600 Soo 400 Heterosexua 400 7 ~~~~~~~~~~~~~~~~~H o mlo biseso al 200 200 - Hom o bi.e-uai 20l/r Other 8182 83 84 85 86 87 88 89 90 91 92 93 94 95 82 83 84 85 86 87 89 90 91 92 93 94 95 96 Cases Mexico Cases Southern Cone Homo/bisexual ,500 400 Homc. b&.?ua.a | IDU 1,000 ~~~~~~~~~~~~300 t 000 ~~~~~~~Heterosexua30 / 200 vit /200HIV moves through sub- Soo / / populations with different ,/'-v ... 8|°°X! 00 heterr,e %ua degrees of risky behavior / Other 2 olrer in a series of overlapping o I . I I I I I _ I epidemics. 82 83 84 85 86 87 88 89 90 91 92 93 0 2 83 84 88 86 87 88 89 90 91 92 93 94 IDU Injecting drug users. Andean Area: Bolivia, Colombia, Ecuador, Peru, Venezuela. Southern Cone: Argentina, Chile, Paraguay, Uruguay. Source: PAHO data, 1996. 71 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 2.10 The Impact of Increased Condom Use by Vanious Subpopulations on Adult HW Prevalence in the Entire Population Baseline Steady relationship: 20% condom use Casual sex: 20% condom use ... Commercial sex: 90% condom use: | Condom interventions among those Adult HIV prevalence (%) Behavior change in year 3 with the highest rates of partner 30 change are highly effective in stopping 25 the increase in HW early in this simu- lated heterosexual epidemic in which 20 _ - HIV is spread by commercial and casualsex. 15 - ' - 10 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Years since the beginning of the epidemic Increased condom use among those Adult HIV prevalence (%) Behavior change In year 15 with the riskiest behavior is also highly effective in a later-stage epi- 25 demic, but behavior change among other groups of the population will be 20 - -- - necessary to lower HW prevalence more rapidly. 15 10 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Years since the beginning of the epidemic Note: Baseline is the model in which HIV is spread by commercial and casual sex, with no behavior change. Source: Background paper, Van Vliet and others 1997. among the small number of sex workers and their clients has a much larger impact on prevalence, both because they have a higher rate of part- ner change and because we have assumed higher condom use. If 90 per- cent of sex workers maintained consistent condom use, prevalence would rise to about 14 percent after 30 years, while if condom use 84 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV among men with commercial and casual partners can be maintained at 20 percent-a change in behavior involving many times as many people in a given year-prevalence would stabilize at about 20 percent.16 These benefits of preventing transmission among those with the larg- est number of partners during the early stages of the epidemic have been well documented (Garnett and Anderson 1995; Stover and Way 1995; backgroundpaper, Van Vliet and others 1997). However, as HIV spreads, an increasing number of new cases occur among people who are not themselves engaging in risky behavior, but who nevertheless become in- fected by their partners. In fact, in advanced epidemics the majority of those infected may practice low-risk behavior. 'When the overwhelming number of new cases is among people with low-risk behavior, does it still make sense to try to prevent HIV transmission among the relatively small segment of the population that continues to engage in the highest- risk behavior? The answer is yes. In the bottom panel of figure 2.10, we show a sim- ulation of the impact of raising condom use in the same three subpopu- lations shown in the top panel, but applied much later-fifteen years after the start of the epidemic. The long delay and higher prevalence rates notwithstanding, high condom use among those with the highest rate of partner change-sex workers, in this simulation-is still more effective in lowering prevalence in the whole population than is moder- ate condom use among men engaged in commercial and casual sex, who have fewer partners. Either of these strategies is far more effective than raising condom use among women in steady relationships. One can imagine other sexual behavior regimes where condom use among sex workers might make less of an absolute difference than condom use in casual relationships, or where these levels of condom use in any one of the subpopulations or even in all of them combined are not sufficient to reverse the course of the epidemic; some of these other scenarios will be discussed in chapter 3. Even so, interventions that prevent transmission among those with the highest rates ofpartner change willprevent more sec- ondary cases per primary case averted than interventions that change only the behavior of the population that practices low-risk behavior. It follows that, if interventions that reduce risky behavior among those with the highest rates of partner change are not much more expensive per primary case averted than programs for low-risk populations, then they can be very cost-effective. CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC The Level and Distribution of HIV Infection in Developing Countries DE EVELOPING COUNTRIES FIND THEMSELVES AT DIFFERENT stages of the HIV/AIDS epidemic due largely to differences in the timing of the introduction of HIV and in sexual and drug-injecting behavior. In this section, we classify countries according to the level and distribution of HIV infection, which, as we shall see, has a significant influence on the cost-effectiveness of preventive inter- ventions (chapter 3), and on the scope of interventions to mitigate the impact of the epidemic (chapter 4). Despite the obvious value of a country typology, low-quality and inadequate data and our still-incomplete knowledge of the disease itself make any evaluation of a country's status very tentative. Because collect- ing incidence data is very costly, HIV monitoring systems collect data on prevalence. However, prevalence data are rarely collected for representa- tive samples of the population. Our typology is therefore necessarily based on prevalence among frequently monitored groups with presumed high-risk behavior-sex workers, injecting drug users, homosexual and bisexual men, STD patients, and the military-and one frequently monitored group assumed to be at lower risk-pregnant women attend- ing antenatal clinics. Most of these groups present significant sampling problems. For sex workers, injecting drug users, and homosexual and bisexual men, it is often impossible to identify a representative sample. Even the prevalence data for pregnant women, which may be systemati- cally collected, are not usually nationally representative but are confined to women in urban areas who attend certain clinics. Further, because women only become pregnant if they are sexually active and most preg- nant women are from younger age groups, women at antenatal clinics cannot be a proxy for the general population. Thus, much of the infor- mation about HIV prevalence comes from ad hoc samples, and in some cases the samples are very small. Notwithstanding these and additional problems that we discuss below, by using available data from research studies and epidemiological surveillance, countries can be classified according to two broad criteria: first, the extent of HIV infection among groups of people often found to engage in high-risk behavior,17 and, second, whether the infection has spread to populations assumed to practice lower-risk behavior. The ty- pology includes three stages of the HIV/AIDS epidemic: 86 STRATEGIC LESSONS FROMf THE EPIDEMIOLOGY OF HIV * Nascent: HIV prevalence is less than 5 percent in all known sub- populations presumed to practice high-risk behavior for which information is available. * Concentrated: HIV prevalence has surpassed 5 percent in one or more subpopulations presumed to practice high-risk behavior, but prevalence among women attending urban antenatal clinics is still less than 5 percent. * Generalized: HIV has spread far beyond the original subpopula- tions with high-risk behavior, which are now heavily infected. Prevalence among women attending urban antenatal clinics is 5 percent or more. Due to a lack of data, this typology does not address several important factors. In particular, it does not distinguish between countries based on incidence, the rate of new infections. As we have seen, because HIV can- not be cured and lasts many years, prevalence can rise even if incidence is declining. Prevalence data do not reveal whether the number of new in- fections is increasing, declining, or level, either in specific subpopulations or among the entire population."8 Moreover, prevalence data are available only for a few specific subpopulations. Prevalence can stabilize in one or more of these subpopulations even as it spreads rapidly through others that are not monitored. These shortcomings highlight the importance to governments of collecting additional data on HIV incidence and preva- lence so that policymakers can formulate an effective response (box 2.7). Even with much better data on incidence and prevalence, we would still lack sufficient behavioral information to confidently predict the course of the epidemic in a specific country. Surveys of sexual behavior by the World Health Organization's Global Programme on AIDS in the late 1980s and early 1990s were among the first attempts to measure behavioral risk factors for HIV infection in developing countries (Cle- land and Ferry 1995). WHO also studied drug-injecting behavior and the risk of HIV infection in thirteen cities in industrial and developing countries in 1989 (WHO, Program on Substance Abuse 1994). These and more recent studies have increased our knowledge about risk factors in developing countries. Even so, almost two decades into the epidemic only a few geographical areas and a handful of developing countries have been covered. Policymakers in most countries simply do not know how many people engage in commercial sex, casual sex, or injecting drug use; how frequently they do so; or the extent to which they take actions to re- 87 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 2.7 Monitoring the Spread of HIV HIV/AIDS PROGRANM PLANNERS NEED) TIMNELY HRN and STD prevalence in the general population dclta on rrends in HR' prevalence among certain sub- are expensive and often do not caprure enough peo- populations to design. implement. and moni- ple ssirh high-risk behavior to detect trends. Fur- tor the impact of' HIX'/AIDS inrtrsentions. The thermore. for ethical reasons. HIV lesting in these W;orld Health Organization and UNAID) recom- larger surveys canl oniv be done with the informed mend that countries conduct periodic -surves of consent of participants. This wvill bias the results if HR' pre%alence among populations likely- to engage people who choose not to participare are more or in high-ri,k behavior. such as parients ar STD clinics less likely to engage in risky behavior than those who and people enrolled at drug treatment centers do participate. At any rate. HIV is unlikel y to show IAIDSCAP and others 19(96. Chin 1990. Saro up in a major way- in the gencral population until 1996). Residual blood collected for othcr tests per- fairl- late in the epidemic. Important trends can be formed at these Facililies is tested for HIV after derected by inonitoting sentinel populations. %sirth remoxing all information that would allow individu- the understanding that these groups are not repre- als to be identified, to nonitor levels and trends in sentatite of the general population and the results HIV prevalence vshile maintaining the confidential- are not adequate to project the number of current it-Y of test results. This is called the " uniniked, anony- and furure AIDS cases. Sentinel sur-eillance should mous" method of HIV sunrcillance. In countries be augmented bs- an AIDS case-reporting system. where HIX' prevalence has reached significant Ievels vw hich %%ill hclp in estimating the start dare of the in populations with high-risk behavior, surveillance epidemic and the impact on mortality and the should be eNxtended to populations thar practice low- health system. HfV and AJD)S epidemiologicap sur- as vmnoecmoen fteesnicalSr risk behavior, such as women attending antcnaral veillance is only one component of che essential clinics. intormnation that government has a key role and Monitoring HIX prevalence of thcce "entinel" comparative advantage in prosviding; other npes oft populations-oftcn rcferred to as "sentinel surveil- inforniation are discussed in chapter 3. lance'-is pret'erred because large-scale surveYs of duce their risk of contracting HIV. Basic information on the levels of condom use and partner change among the general population is also unavailable. Without this information, it is simply not possible to accu- rately predict the course of the epidemic. For these reasons, we cannot predict with certainty in which low- prevalence countries HIV infection will take off or at what level preva- lence will stabilize. Some countries where HIV is currently regarded as a minor problem may turn out to resemble eastern Africa, where the virus spread rapidly through groups with high-risk behavior and widely into the general population. In other countries HIV may infiltrate groups with high-risk behavior but never evolve into a generalized epidemic, even without intervention. Or, knowledge of HIV may cause people to adopt less risky behavior, with or without government urging. Countries 88 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV with a high incidence and prevalence of STDs other than HIV are likely to be particularly vulnerable to a large and rapidly spreading HIV epi- demic, because STDs and HIV are spread by the same behavior and STDs enhance HIV transmission. However, since most other STDs can be cured, countries with low STD prevalence may still have behavior patterns that are conducive to the rapid spread of HIV. Lacking the information to predict the course of the epidemic, and given the terribly high human and financial costs of HIV/AIDS, it would be prudent to assume the worst and act aggressively to minimize the epidemic as early as possible. The remainder of this chapter uses this typology to out- line the state of the epidemic as of mid-1996 in four developing regions. A list of countries and estimates of prevalence in different subpopulations are in table 1 of the statistical appendix to this report. Africa In seventeen Africa countries, more Roughly 90 percent of all HIV transmission in Sub-Saharan Africa is than 20 percent of female sex workers by heterosexual sex. HIV has spread rapidly among people with high- in cities are infected with HIV. Figure 2.11 HIV Infection in Urban Sex Workers in Sub-Saharan Africa, Various Years HIV prevalence (%J 100 80 40 _ 60 401 i'_ -b s _ . 20 _. 1 M a _ 0 i p - i-a - a,,- ~~. M 0 -J m en 'e Country a. HIV-1 and/or HIV-2. b. Formerly Zaire. Source: U.S. Bureau of the Census (database), 1997. 89 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 2.12 HW Seroprevalence among Pregnant Women in Seected Areas of Sub-Saharan Aftica, 1985-95 HIV seroprevalence (%) 35 30 Kampala, Uganda 25 20 _/ Batet i's _ J __ / Gaborone. otswana * Dar er sSalaam. Tanzania Infection rates among women attend- ing antenatal clinics have grown ±0 -- __ -- - '- rapidly to high levels in some coun- KInshaaa. Congo DR . bies, stabilzed at lower levs in 5 others, and appear to be declining in South Atica (national) . ' Kampala, Uganda. 0 , , , , . . - 1985 1987 1989 1991 1993 1995 a. Formerly Zaire. Source: U.S. Bureau of the Census (database), 1997. risk behavior and widely among those assumed to be at lower risk. Preva- lence among urban sex workers exceeds 20 percent in seventeen coun- tries, and is 50 percent or more in nine countries (figure 2.11). Infection rates among women attending antenatal clinics have grown rapidly to high levels in some areas, have stabilized at lower levels in others, and appear to be declining in Kampala, Uganda (figure 2.12). HIV has in- fected more than 5 percent of women attending urban antenatal clinics in nineteen countries, and in six countries more than 20 percent are infected. An estimated two-thirds of all new cases of mother-to-child transmission worldwide occur in Sub-Saharan Africa (UNAIDS 1 996d). The countries with generalized epidemics include most in eastern, southern, and central Africa, plus C6te d'Ivoire, Benin, Burkina Faso, and Guinea-Bissau in West Africa (figure 2.13). There is often consider- able geographic variation in infection levels within countries. Nigeria, which has more than 100 million people and is the region's most popu- lous country, has areas at all three stages of the epidemic. In more than half of Nigerias states the epidemic is concentrated. HIV has spread most widely in Lagos, along the west coast, and in Delta, Plateau, Borno, and Jigawa states, located to the east and northeast. However, in three 90 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Figure 2.13 HIV Infection in Africa and the Middle East 4 P asm43-( | I_ L .: AIIXD Rep E _ Egyp ; l'a- K>~. | al.| d Nascent Mozo'lu [~Unknown Stage of Epidemic - Intemnational Boundariies ltw-,.VS~y #.MopDwenUn' od.WoddRok. :Les0tho - - o b-,h m bX,ndonn JULY 1997 states-Edo, Niger, and Oyo-the epidemic is still nascent with low prevalence levels, even among subpopulations with high-risk behavior. HIV was detected early in the Democratic Republic of the Congo (for- merly Zaire), but in contrast to many eastern and southern African countries, prevalence has stabilized at less than 5 percent on average in urban antenatal clinics (Piot 1994; statistical appendix, table 1). In Uganda, one of the hardest-hit countries in Africa, HIV prevalence among young people has declined (box 2.8). Most Sub-Saharan African countries face the dual challenge of lower- ing HIV prevalence-which can happen only over many years-and of coping with the impact of existing high prevalence on the health system and society. Their domestic budgetary resources to accomplish this are quite limited. Countries with nascent epidemics in Sub-Saharan Africa- Cape Verde, Madagascar, Mauritania, Mauritius, and Somalia-have a unique opportunity to intervene early and aggressively to pre-empt a 9'1 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 2.8 Declining Seroprevalence in Uganda RFCEN I S FUDiEB FROM L'C.A9NT1A SHO5X .\ DECi INk HR1 prev-alence is also declining 3mong pregnant in HI\' seropre%alence rates, particularly among women actending antenaral clinics in LUgand.3 In the Young adult,. Among adults in N:iasaka district. main referral hospital in Kampal. Liganda's capital LUganda. overall HI\' seropresalence declincd by less and largest city. HI\ seroprevalerce among preg- thani I percentage point between 1989) and 199-i. nant women fcll 'rom 28 percent to 16 percent be- from 8.2 percent to .76 percent Niulder and o[hers tween 1989 and 1993 iBagenda and others 199') 1995i. How%e%er. rhe decline was greater among the All age groups under age 3 experienced a declne in young-from 3..4 to 1.0 percent among miles and prevalence. but the decline %,as greatest for those from 9.9 tro 73 percent among females age-s 13 to under 19. Similar declines %scre observed among 24 ibcx igure 2.8'. The greatcsr declines 'wkre seen preenant w%omen 3a anrenatal clinics in other urban in males ages 20 to 24 and females ages 13 3o 1r . areacs o' oficnda berween 1991 and 199a iA4iinsc- Yet HIV seropre%alence increased among males and Okiror and others It )t females ov-er 25 vear, oId In neighboring Rakai Di)s- T'he iocrall decline in HIV prevalence among the trict. HIV pretalence in adults 1i to i9 declincd r'so rural populations of adults in Nlasa.ka and Rakaa from 23.-i in 19 9iJ to 20.9 percent in 1992 5Scr- districts can be accounted lor almost completely by ssadda aid AilRErs 1995). O. diC aglili. tIe s1u,ge[ - Si g1Lm IiI l-liv . siL 11[I o C ha ag i IlC ic dei d group showed the largest decline: flor agcs 13 to 2c S erwadda and others 1'N11i. Hosseser. according to presalencedeclined from 1 3 rto 12.6 percent. nationnide Lar'csv ccnducted in 198l! and 1995. Box Figure 2.8 HIV Seroprevalence in Rural Masaka, Uganda, 1989 and 1994 Seroprevalence (I.) 20 15 0 13-24 25-34 35-44 45+ 13-24 25-34 35 44 4S+ Men's age group Women's age group M.luldtj ind other' I 9Z STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Box 2.8 (confinued) [here have been important changes in sexual be- mental organizations lNGOs) and the Ugandan ha%ior thar could explain the apparent decline in government have focused on reducing the Frequency incidence of HRV among young adults observed in of sexual partner change. disrributing and promot- antenaral clinics. particularly in urban areas lAslimwe- ing condoms, and controlling sexually transmitted Okiror and others 199-. Stoneburner and Carballo infecrions (Niulder and others 19951. However. 1 997). The percentage of young aduLts age 15 lo 19 knowledge of the routes of HIV is widespread in who have ever had sexual intercourse declined From Uganda. and many young people have personal 69 to 4s percent among men and from -4 to 54 experience of a loved one dying from AIDS. In the percent among women. Condom use ha- risen sub- absence of a carefully designed evaluation. as- wsttianll among aUl age groups. and dhe percentLge sessing the contribution of specific interventions u-ith a casual partner has declined, particularly is extremelv difficult. Given the importance of the among the young. issue to other countries with generalized epidemics, Is the decline in prevalence in young adults the international support for such an evaluation is result of policy! In[erventions by both nongovern- warranted. full-scale epidemic. High STD prevalence in Madagascar makes it vul- nerable to the rapid spread of HIV (box 2.9). The epidemic in most of North Africa and the Middle East is also nascent, although there is evidence of rapidly climbing HIV infection among injecting drug users in Bahrain and Egypt, as has been the case in Asia. A large number of North African and Middle Eastern countries could not be classified be- cause of lack of data. Latin America and the Caribbean More than half of the countries in Latin America and the Caribbean have concentrated epidemics (figure 2.14). These include the most pop- ulous countries in the region-Brazil and Mexico. Six countries have nascent epidemics, two (Guyana and Haiti) have generalized epidemics, and two (Bolivia and Panama) have insufficient information to be dassified. Injecting drug use and sex between men have played a major role in transmission in many countries in Latin America. Roughly one-quarter of all HIV infections in Brazil (24 percent, 1992) and a third in Argen- tina (39 percent, 1991) have been attributed to transmission through injecting drug use, which is an important source of transmission in Uruguay as well (Bastos 1995, Libonatti and others 1993). The epi- 93 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 2.14 HIV Infection in Latin America and the Caribbean ..1:.:. I Unknown.StageofEpideT, o n Th :.i - / _ -- - 1 ...pt- d-h b-kri-~~~~~~~~~~~~~~~~~~~~~~~~~~~~~-* ks-4 i C _Generalized | | Concentrate, M a u sif n . Nascent w th | | Unknown Stage of Epidegh: a og -In ternationol Boundaries \, Thisonpwsproducetdy l hewMrps seix a Ues of Th - I n h-i s The b ndone,, color,, deno ounobons dbonyodrer, -l -. Ihoon on this Cieadort0f imnplya onAnrcpsIiof shesp easm s= Groupg onyhjedrostt on the oal sotaos of ony serrizo . - a v mondonemenn or occoptonen of such houndoros. demic is well established among homosexual and bisexual men in Argen- tina, Brazil, Colombia, Mexico, and Peru and has infected significant numbers of sex workers in Argentina, Brazil, the Dominican Republic, Guyana, Honduras, Jamaica, and Trinidad and Tobago. Although the data are spotty the relatively high prevalence of HIV among injecting drug users, homosexual and bisexual men, and sex workers in Latin America suggests that in many of these countries the virus is poised to spread to the low-risk sexual partners of people who engage in high-risk behavior. In the Caribbean and parts of Central America, HIV is spread mostly through heterosexual transmission. Male and female cases are roughly equal in Haiti; the epidemic has spread broadly to 8 percent of pregnant 94 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Box 2.9 The Case For Early Intervention in Madagascar AS A COUNTRY 'WITH RFLATI\ ELY FEW (C-ASES OF all women aged 15 to 4i9 surveyed in 1992 by H[I' infection. M-Iadagascar is in a good position to the Miadagascar Demographic and Health Survey. intervene earls to stop the rapid spread of HR'. only 0.5 percent used condoms (Refeno and others Fewer than half a percent of STED patients and sex 1994). workers were infected with HIX' in 1992. but the The government of NMadagascar launched its potential exists for an cxplosiv-e HNI' epidemic in national AIDS control progranm with the encourage- Mladagascar because of the high prevalence of other ment of and financial support from WHO/GPA in STDs (Behets and others 1996). As of 1995. almost 1988. Lhree 'vears before iLs first reported AIDS case. one-third of sex workers and about one in eight The countrn already had a system of STD surveil- pregnant women in urban areas w%ere infected wvith lancc prior to the introduc[ion of HTI/AIDS. con- syphilis. According to one projection model, hy sisting of fifteen STD trearment clinics in maior 2015 HIV seropre%alence among Malagdsy adults rowns. However, rhe system is poorly equipped and could be either 3 percent. using the course ot needs major revamping. To address the gaps in Thailand's epidemic as a model, or 15 percent. kno%%ledge about HI% prevalence. the government using tht course of Kenya's epidemic as a model, conducted and financed, throueh a W'orld Bank Certain behavioral risk factors common to STD loan, its first survey of STD/HI\V prevalence and and HIM transmission are common in Mfadagascar. risky behavior among high-risk groups in 199--95. In 1992, about one in ten men reported recent sex- As we shall see in chapter 3. strengrthening epidemi- ual contact with a prostirute IAndriamahenina 1995I. ological surveillance, raising levels of awareness of Stable concurrent partnerships outside oft marriage H[\'/AIDS. and vigorously promoting condom use W-cre reported by I I percenr of pregnant wiomen. and STD trearnienr among individuals with the nearly one-third of STD patients and one-quarter riskiest behavior should be the priority for of prostitutes. Coondom use is low in Miadagascar. Madagascar. given the levels and distribution of Only one-third of sex workers reported always using infection that presail Fortunately, condoms will condoms wirh stable extramarital partncrs and only soon have a higher profile in the counrns, as LSAID 5 to 8 percent of STD patients reported always has funded a new condom social marketing program using condoms i Andriamahenina 1 )9S . Among in major cities as of th. end of 1996. women, and there is significant mother-to-child transmission. More than 70 percent of AIDS cases in the Dominican Republic is attributed to heterosexual transmission; the ratio of male-to-female cases now stands at 2 to 1 and is declining (ONUSIDA 1997). HIV prevalence among pregnant women in that country has risen to a national average of 2.8 percent, and in some areas has reached 8 percent. Following a sim- ilar path, 1 percent of pregnant women in Honduras are also infected with HIV. In Guyana, which is in South America but faces on the Caribbean, nearly 7 percent of women attending antenatal clinics were infected, as of 1992. 95 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Asia In most Asian countries for which there is information, the epidemic has reached a concentrated stage either nationwide or at least in some states or provinces (figure 2.15). This includes regions of the world's two most populous countries, China and India, most of Indochina, and Malaysia. In the remaining Asian countries for which there is informa- tion, the epidemic is nascent; infection among those presumed to prac- tice high-risk behavior is less than 5 percent. Patterns of infection in east, south, and southeast Asia have been greatly influenced by the proximity of many countries to the "Golden Triangle" of heroin production, located at the border between Lao PDR, Myanmar, and Thailand, and to its distribution routes (background paper, Riehman 1996). HIV infection was first detected among those who inject drugs in Bangkok in 1987; during the next year it spread rapidly among injecting drug users in the Thai capital (Stimson 1994). The pattern was quickly repeated among injecting drug users in north- ern Thailand and along the border areas between southern Thailand and northern Malaysia. In 1989, HIV infection was identified in Myanmar, Yunnan Province in China, and in Manipur State in India. HIV was de- tected among injecting drug users in Singapore in 1990. Injecting drug use has been the main transmission mode in China, where the most highly infected province, Yunnan, is adjacent to interna- tional drug routes. Male injecting drug users in Yunnan account for 78 percent of HIV infections in China (Zheng 1996). In other Chinese provinces, infection rates are thought to be low, even among those who practice high-risk behavior (Yu and others 1996). Economic reforms that have helped to reduce the number of people in poverty in China by more than half since the late 1970s have also resulted in large increases in in- ternal migration that could generate conditions conducive to the spread of HIV Studies have estimated that nearly 100 million people, roughly one in twelve people in China, have moved either temporarily or perma- nently from their registered residence (Nolan 1993, Peng 1994). Much of the movement involves migration within provinces, but an estimated 20 million migrants have moved from poor areas of western China to eastern provinces (Nolan 1993). Most migrants are young, single, and male, but many women have also migrated; some have reportedly be- come involved in prostitution. STDs, which were all but eliminated in China in the 1960s, are rising rapidly (Cohen and others 1996, Kang 96 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV Figure 2.15 HIV Infection in Asia SJ / 9 , -s>, ' ' | - - ' ~~~~~~~~~~~~~~Generalized e.4f-i Al 5/ ; _ ,}< | | concentrated Nascent [ /< .jr Unknovwn Stage of Epidemic Province/State Boundaries - ) - \ > tI iF;-- International Boundaries .. - ;.-.1:: i.; , Thi m p was prduced by tIe Map De.mn, unit uof The W,dd nomk Bnd n Group, oey ftdq-ennt-a se Iugol sfuhs nf ansy un-i*quy, or-y 2O endorsement or onneptonce 0f sn boonduries.a.' . 1995). Early preventive interventions for migrants and sex workers in areas receiving migrants could reduce the likelihood of an epidemic of HIV and other STDs among these mobile groups. Among the nations of South Asia, the epidemic is believed to be spreading most rapidly in India and Pakistan. In India, HIV is widespread among injecting drug users in the northeastern states of Manipur and Mi- zoram and is spreading to their sexual partners; prevalence in antenatal clinics in Manipur has reached 2 percent. HIV is well established among sex workers and STD patients in much of southern India, induding pop- ulous Maharashtra and Tamil Nadu states Gain, John, and Keusch 1994). In the city of Mumbai (formerly Bombay), HIV prevalence among preg- nant women has reached 1 to 2 percent. In Pakistan, the infection rate among injecting drug users in Lahore was 12 percent; as of 1995, HIV infection among women attending antenatal clinics was still extremely low. Transmission by those who inject drugs also may be a factor near a 97 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC second major heroin-producing area, the "Golden Crescent," where Pak- istan's Northwest Frontier meets the Badakhshan area of Afghanistan and the Baluchistan area of Iran (backgroundpaper, Riehman 1996). How- ever, there are no recent data on HIV prevalence among drug users or other groups in these areas. In Nepal, prevalence so far remains very low among injecting drug users in Katmandu, pardy because of interventions discussed in the next chapter. Bangladesh's HIV epidemic is still nascent, but, without behavior change, HIV could spread quickly among a popu- lation of brothel-based sex workers and their clients. In most of southeast Asia, with the significant exceptions of Indone- sia, Lao PDR, the Philippines, and Papua New Guinea, the HIV epi- demic is at the concentrated stage. Injecting drug use has played a cen- tral role in the launching of HIV, often in conjunction with commercial sex, but heterosexual transmission is now the predominant mode of transmission. HIV is firmly established among injecting drug users and sex workers in Cambodia, Myanmar, and Thailand, and 1 to 3 percent of pregnant women are HIV-positive in those countries. In Thailand, HIV prevalence peaked at 4 percent among military conscripts in 1993, but has recently been declining following a national campaign to reduce sexual transmission of HIV through greater condom use and a reduction in commercial sex. In Cambodia, however, infection levels in the mili- tary have reached nearly 7 percent. In Malaysia and Vietnam, more than three-quarters of HIV infections are attributed to transmission through injecting drug use (Hien 1995, Kin 1995). Yet sexual transmission in Malaysia is clearly on the rise; nearly 40 percent of HIV/AIDS cases seen at the University of Malaya Medical Center since 1986 were thought to be due to heterosexual transmission (Ismail 1996). In contrast, although HIV has been detected sporadically for some time among sex workers in the Philippines and Indonesia, it has not spread rapidly, even within that group; as of mid-1996 these two populous countries remained at the nascent stage Jalal and others 1994, Tan and Dayrit 1994). Eastern Europe and the Former Soviet Union The rapid social change and economic dislocation that has accompa- nied the collapse of socialism in Eastern Europe and the former Soviet Union (FSU) have created a situation in which the potential for an HIV epidemic looms large. The available data on HIV prevalence suggest that most countries in the region are still in the nascent stage (figure 2.16). 98 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV However, reliable information on HIV prevalence by subpopulation is scarce in all but a handful of countries-almost two-thirds of all of the countries in this region cannot be classified based on the available information. Ukraine has a concentrated HIV/AIDS epidemic, based on high HIV prevalence among injecting drug users; between January and August 1995, HIV prevalence among those who inject drugs rose to 13.0 per- cent, from just 1.4 percent. Just five months later, more than half of the injecting drug users in the Ukrainian city of Nikolayev were infected (UNAIDS 1996d). A survey of new injecting users in Poland in 1995 Figure 2.16 HIV Infection in Eastern Europe and Central Asia Generalized Concentrated Nascent Unknown Stage of Epidemic ;. m 'op w os pdd by Intemational Boundaries ihe Mop Deign Unit f Th. W.'I ooc by I I r-: on t - ; - -, . Gnu 'un-o.,' fJ¢~~~~~~~~~~~~~~~~~~~~~~~~~~~~~I .e-.zh Fu- f; ?' E C Z\ YurDslcFiaYR EA-rmfi, _ 46e./ Mlot.ljie&n% Trkey fljTu -r6ist- To g 99 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 2.17 Reported Cases of Gononhea in Eastem Europe, 1986-94 Number of cases (thousands) 250 Estonia 200 1SO Russian Federation Lalsla , 100 Lithuania The rapid increase in STD cases in 50 Eastern Europe signals heightened vulnerability to an HIV/AIDS epidemic. 0 l l 1986 1987 1988 1989 1990 1991 1992 1993 Source:WHOGPA 1995. found HIV prevalence of 4.7 percent; a few years earlier, among longer- term injectors in the city of Warsaw the rate was 45 percent (WHO/EC Collaborating Centre 1996b). Given that those who inject drugs often travel to neighboring countries, it is reasonable to expect a similarly rapid take-off among injecting drug users in the Russian Federation and Belarus (Bourdeaux 1996). In Romania, HIV initially spread primarily among children; over 90 percent of AIDS cases in 1990 were among children under 13 years of age. It was erroneously believed that transfusing blood among children would provide important nutrients and boost the immune system (Hersh and others 1991). Instead, it spread HIV among them. The prac- tice has since been abandoned. A key signal of the potential for an HIV epidemic in this region is the dramatic increase in STDs experienced by most countries since the col- lapse of the Soviet Union. The number of gonorrhea cases has nearly doubled between 1990 and 1994 in four Eastern European countries (figure 2.17). In Ukraine, the number of syphilis cases increased more than tenfold between 1991 and 1995 (AIDSCAP and others 1996). Regardless of the stage of the epidemic, there are compelling reasons for governments to find ways to encourage people who practice risky be- havior to adopt safer practices as soon as possible. Fortunately, in most 100 STRATEGIC LESSONS FROM THE EPIDEMIOLOGY OF HIV areas of the developing world it is not too late to avert a generalized epi- demic. Half of the population of developing countries-2.3 billion peo- ple-live in areas where the HIV/AIDS epidemic is still nascent. An- other third of the population of developing countries lives in areas where the epidemic is already concentrated but has yet to become generalized. In all these areas, action now to help people with the highest-risk behav- ior protect themselves and others from HIV infection can save millions of lives and avoid massive future expenditures on AIDS treatment and care. Even in areas where the epidemic is already generalized, action to prevent infection among those most likely to contract and spread HIV can still make a substantial difference. What steps can be taken to help people who engage in high-risk sex- ual activity or injecting drug use to protect themselves and others from HIV infection? The next chapter discusses two broad complementary approaches: altering the perceived costs and benefits of individual choices, and changing the social environment that shapes and constrains these choices. Notes 1. Thomas and Tucker (1996) point out that the use female transmission rates as high as 140 percent (Haver- of the word "rate" is not technically correct, since it is in kos and Battjes 1992). fact a number, not a rate per unit of time. The reproduc- tive rate was first applied to HIV transmission by May and 4. Transmission rates per partnership are not much af- Anderson (1987). Previous authors had referred to this as fected by the length of the partnership or number of ex- the "infectee number" (Hethcote 1976, Nold 1978). posures. This is because in a relatively long-term partner- ship, either the HIV-positive partner is truly infectious 2. The number of exposures per partner and the type and transmission will occur relatively quickly, or the HIV- of sexual act also affect the spread of HIV through the positive partner is relatively uninfectious and transmission population, but empirical evidence suggests that the rate will not occur despite many contacts over a long period. of partner change is far more important (Anderson, Gupta, and Ng 1990). 5. High infection rates with other STDs may also have contributed to high transmission rates in the study. 3. However, a review of sixteen studies that compared female-to-male and male-to-female transmission found 6. For example, Hook and others 1992, Laga and female-to-male rates as low as 5 percent and male-to- others 1993, Lazzarin and others 1991, Mastro and oth- I0I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC ers 1994, Plummer and others 1991, Quinn and others 13. While this "pre-emption" effect slows the epi- 1990. demic's rate of spread at the population level as the epi- demic matures, the probability that an uninfected individ- 7. For example, Bongaarts and others 1989; Caldwell ual will encounter at random an infected partner has gready and Caldwell 1996; Conant 1995; de Vincenzi and increased because a large share of the population is infected. Mertens 1994; Moses and others 1990, 1995; Simonsen and others 1988. 14. In this model, 40 percent of men have a casual partner or a sex worker in a year. However, the individu- 8. Like all data on sexual activity, such surveys are sub- als are not the same over time. It is assumed that men who ject to a variety of errors. Since they ask about private use condoms with casual partners also use them with sex behavior, misreporting may skew results. In many cultures workers. men may exaggerate the number of their sexual partners, while women may do the opposite. People with high rates 15. We choose to model a higher level of condom use of partner change may not be able to recall their number among sex workers because it has been shown to be more of partners accurately, and sex workers may not be cap- feasible than similarly high levels of condom use among tured by surveys using standard sampling techniques. other subpopulations. Even if we assume a significant degree of error, however, th vaito in th rat of' pate chnewti'oua 16. Note that in this hypothetical population with tionv .istiking, concurrent and casual sex, behavior change in any one of tions is striking. these groups is not sufficient to reduce HIV prevalence to 9By13the percent of men 15-49 with nonregu- zero. That would require simultaneous increases in con- 9. By 1993, ~~~~~~~dom use in more than one group. lar sexual partners in the past twelve months had declined s to 15 percent in Thailand (Thongrhai and Guest 1995). 17. We recognize that the characteristics of certain subpopulations, like military service or sexual orientation, 10. Men who have sex with men include self-identi- are imperfect predictors of risky behavior. While members fled homosexual and bisexual men as well as heterosexual of these subpopulations may practice riskier behavior on men who have sexual relations with other men. average, in some countries they practice low-risk behavior, as evidenced by lower rates of partner change, high rates of 11. The impact of behavior change on the epidemic condom use, or limited sharing of injecting equipment. will be introduced later in this chapter and in chapter 3. 18. An exception would be adolescents, who can be 12. The baseline is the same as the top line in figure presumed to be uninfected as they enter adulthood. HIV 2.8 for an epidemic fueled by commercial and casual sex, prevalence among adolescents at a point in time is likely to with concurrent partnerships. reflect recent infection. 102 CHAPTER 3 Efficient and Equitable Strategies for Preventing HIV/AIDS W _ _ _ HILE RESEARCHERS CONTINUE THE search for a cure or a vaccine that is afford- able in developing countries, the greatest hope for combating HIV/AIDS in the fore- seeable future is that of helping people to choose safer behavior so that they will be less likely to contract and spread HIV But can government affect the very private and intimate behaviors that spread HIVIAIDS? And if it can, what are the priority actions that governments should take to prevent the epidemic, so as to maximize the impact of scarce resources? This chapter identifies government priorities in the prevention of HIM/AIDS. We find that public policy can affect the behaviors that spread HIV. There are effective interventions, and governments have many ways to influence private behavior. In the first two sections of the chapter, we focus on two complementary approaches. The first aims to influence individual choices directly within the existing economic and social context by changing the costs and benefits of various types of be- havior, making safer behavior a more attractive choice. However, indi- vidual behavior is shaped and often constrained by the economic and social context; some individuals' choices are very constrained. A second, complementary approach is to change the economic and social condi- tions that make it difficult or impossible for some people to protect themselves from HIV. The benefits of this approach typically extend far beyond HIV prevention. It is therefore often difficult to assess the cost- effectiveness of measures to affect the economic and social environment, 103 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC but they are likely to improve the efficacy of direct interventions to individuals. Which activities are most justified for government action and, among them, which should be priorities? There are three activities in which gov- ernments have an indispensable role in ensuring the efficiency and eq- uity of prevention programs: providing public goods related to preven- tion, such as the collection and dissemination of information about the epidemic; reducing the negative externalities of risky behavior by pro- moting safer behavior among people who are most likely to contract and spread the virus; and promoting equity by ensuring that the most desti- tute are not denied access to the means to protect themselves from HIV. These activities will reduce the spread of HIV most quickly and will ben- efit everyone in society, including those with low-risk behavior and the poor. Private individuals would not invest in them in sufficient quantity to slow the epidemic. In addition to satisfying these criteria of public economics, government-sponsored preventive activities should be cost- effective. We review some important factors that influence the cost- effectiveness of government actions and propose a set of public priorities for preventing HIV according to the stage of the HIV/AIDS epidemic. In the final section of the chapter, we review the available evidence on the extent to which national governments have succeeded in following the broad prevention strategy outlined in this chapter. We find that while governments, in collaboration with private actors, have imple- mented many worthwhile programs, some of which have had demon- strable impact, there is a need for renewed effort to implement the activ- ities that are likely to be most cost-effective from the government perspective in preventing the epidemic. Perhaps the major impediment to more effective, efficient, and equitable prevention is a lack of political commitment, first, to collect and disseminate the data on HIV preva- lence, risk behavior, the effectiveness of programs, and their costs, and, second, to work constructively with those at highest risk of contracting and spreading HIV to prevent infection. These are important issues that national governments, international donors, and nongovernmental or- ganizations need to act on urgently; they are discussed at greater length in chapter 5. Before we begin, a caveat: this chapter reviews evidence on the impact of policies and suggests priority actions for governments. However, it is not intended as a blueprint for implementing these programs in the field, or as a primer on "best practices" for specific interventions or pro- 104 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.1 Best Practices in HIV Prevention and Treatment Online UjAIFT i! D I' IN' I -Tu -DAT1 INRk L\ITIuN ON T I'uŽu/.ur A Nhort [cchnical o, er' Le%S of best practices in Hl\ pre%ention. treatment, and the topic For managers of HI\ AIDS pro- .AIDS impact online on irs wNeb,ire. ssw%s.unaids.org. grims and prlIcts maiing the main Information w.ill he a%ailable on more than -i top- probilems in'olked Js %ell da the hbe4 prac- ics, among them communint mobilization. nialc and tice responses. Female condoms,. counseling and tescing. HIV edu- U Brit pecrti'r aM r.m Exampls ot 'bet cation. epidemiolocgy. human rights. STDs. and practice on the topic in 'pecit countries. HIV vaccines The collection ot materials on each * trot 'norm; ulon: A selection of clide, or ropic %NIII normallY s he e componentv overheads and tal;king points For pre'c nta- *I N-lIS pm o, ....... IR.c'iir A short adscocav do c- tions ument for lournalisrh and communit, leaders * Adt 1,,zrcr.d i, A niaximum of tcn reports. arti- that cites the key faet5 and figurts. addreswes cles. bookss. conmpact cli5c. or v%ideo that rep- myt%hs andr misconceptions. and sets out %%avs resent the most up-to-dare. aurhcritatiI of dealing %.. ith dhi toFl. thinking on the topic. grams. This huge task is beyond the scope of this volume and the exper- tise of its authors. For this, UNAIDS is assembling an extensive collec- tion of resources to which practitioners may refer (box 3. 1).1 This chap- ter also focuses on changing behavior to prevent HIV Programs to ensure a safe blood supply, while preventing some infections, will not be sufficient to prevent an HIV epidemic spread mainly by sexual or drug- injecting behavior. The same is true for medical interventions to prevent transmission of HIV from mothers to their children. Blood safety and treatment to prevent mother-to-child transmission are discussed in the context of the impact of HIV on the health sector in chapter 4. Influencing Individual Choices M O OST THINGS IN LIFE ENTAIL SOME RISK, YET PEOPLE WILL- ingly take risks when they perceive that the benefits of an action outweigh the costs. For example, drivers speed and pedestrians dart across busy streets, despite the increased risk of injury or death. People start smoking although they know it might lead to lung cancer and heart disease. Sometimes risk enhances pleasure. I05 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Mountain climbers scale the Himalayas, their thrill perhaps intensified by the danger they face. All of these decisions reflect individual prefer- ences and an assessment of costs, benefits, and risks (Philipson and Posner 1993). Sex and injected drugs offer very intense if short-lived pleasure. Do individuals also weigh costs, benefits, and risks when deciding whether and how to engage in these activities? Fortunately for efforts to slow the epidemic, the answer is yes. A sub- stantial body of economic research, much of it in developing countries, has shown that actual and perceived costs and benefits, some of which can be affected by government policies, significantly influence private decisions about marriage, childbearing, and contraceptive use.2 It is therefore reasonable to assume that sexual behavior that spreads HIV can also be influenced by public policy. Similarly, studies have shown that under the right conditions drug users change their injecting behavior to reduce the likelihood of being infected with HIV. This section discusses four types of policies that can alter the perceived "costs" and benefits of various activities so that individuals will avoid behavior that is very likely to spread HIV: providing various types of information, reducing the costs of using condoms, reducing the costs of uging sterile injecting equipment, and attempting to raise the costs of engaging in commercial sex and injecting drug use. 3 Knowledge of HIV Reduces Risky Behavior-But Not Enough Knowledge about the levels of HIV infection in the population, how HIV is transmitted, and how to avoid contracting it can induce some people who engage in high-risk behavior to adopt safer sexual and in- jecting practices or to refrain from casual or commercial sex and inject- ing drug use altogether. In Thailand, for example, the public revelation in 1989 that 44 percent of sex workers in the northern city of Chiang Mai were infected with HIV is believed to have contributed to the grow- ing use of condoms in commercial sex, even before large-scale govern- ment condom programs began (Porapakkham and others 1996). Simi- larly, there is evidence that condom use rose in the United States in the 1980s, in part independently of the impact of prevention programs (box 3.2 ). Moreover, there is considerable evidence that people engaged in high- risk activities are more motivated to learn about HIV than others, since they are more likely to become infected and suffer the consequences. io6 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.2 Behavior Responds to Increased Risks: AIDS Incidence and Increased Condom Use in the United States .\sTULDY 1N THE i 1 itAils -\S sHUWN THAT bY cstablishing ,tronger AID, prevention pro- oung American adulti incre.mcd condorn uLe in grams. disent3angling; che impact of wooing per the 1I'kS0s in response to the groi%ing perceited capita eumulative AIDS cases and sttne-run pre- rh.,ks of HI\. net oftthe impact of state HI\ pre- vent ion programs on condom use n ould be sention programs t Ahirus. Hotz. and Philipsorn errcrmel dilficult. In fact, the study did not find a 1111l.O This 'ugtc t[, hat there %%ill be s:ome spon- itrong correlation beteen the extent ot the XL)DS taneous bchasioral rcisponsc to the rising costs of epidemic and the strength of a sttce § HI\ preven- unprote,ted t ex inposcd h! rhe AID)> epidemic. non efforts. To. t-urther isolate the effect of rhe per In 1'%-. the first of the 5i\ ycar, s co'ered b, the capita cumulati'e AIDS cases on condom use. [he studs. there had been relatively- fe% AI)S cases. anid study controlled for such factors as age. gender. there "crc fe" duffercen nesto corndom use across cen- race and erhnicit-. marital status. urbani residence. sus rceions of the Lnited States. As ti-he cumulative educaction. and parents Income. as %\ell ai for state- number of .AiD case, climbed differentially. so did l specific control' that proxied for AIDS program 'tlndomn u'c: thc laigci tlic plc P a[Pitad 1tin.ulat 1'C c\.pcnd!turC'. number of AIDS cases "as. the lArger %%as the Th sitrong correlation beten per capita cumu- increase in condom use. A.cro.s all 'rateC for the ix- larie AIDS cases and condom use persisted. even year period. condom ase amiLoIng 2-- to 21-v-car-olds after controlling h!r these variables. Furthermore, doubled fronm 8 percent to inorte thn 16 percent: s'th the passage oF time the responsi'eness ot con- tor African Americans in this age groLup condom use dom use to per capita eumulati'. AIDS eases almost tripled. from - to 19 percent. The respon- increased. The authors estimated that 32 to 60 per- siseness ot condom use to per capita cumulati%ve tenl orf he actual increase in condom use could be A_IDS cases was greatest among sexually aci',. single attribueld ro increases in the eumulatike number of men and single men in urban areas. the n"o groups AIDS cases-an obtjecive measure of heightened in the sample mo r likely ro be at high-risk of HI\ rnik-as opposed to other state- and indiv-idual- intecrion. Condom use among, married men. "ho level factors. presumably had a much loker risk of acquiring HW''. did not '3an according to the per capita cumu- 'T h 1atie AIDS cases in the stare.- .i"- u it Se r pnd;ni- t, r th 1lc,-t-w1 a- c Eihe INJIconil L,)ngudindl hurnt if 'wutuh \\Vhat caused these chanres in beha ior? IF the N LS\ -LN. F. uh beg!:lan In !" oh inrer.-mi:. t I 2.0(i per- states witih the largest AIDS epidemics responded fIL ir.-.m Ehib-C!, birth .uh.- Surveys conducted by the Global Programme on AIDS in 1989-90 found that respondents who had engaged in risky behavior knew more about modes of transmission and the severity of AIDS than those who had never engaged in risky behavior (Ingham 1995). Similarly, Demo- graphic and Health Surveys (DHS) in seven African countries found that individuals with characteristics correlated with higher-risk sexual 107 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC behavior were more likely than others to be aware that condoms pre- vent transmission of HIV (box 3.3). A study in Imphal, the capital of Manipur State in northeastern India, found that injecting drug users, more than 80 percent of whom were infected with HIV, were far more Box 3.3 Who Knows How Much about Preventing HIV/AIDS? VXTE WOULD EXP\'F-T- THAT PFoPlE Wi-XlOE BEIL-\\1R- risk ot HIX infection. Their higher levels of school- puts them at higher risk of acquiring or spreading ing also probably contribuced to greater awareness. HI\ would hate gTreaer incentive to find Out hott However. know;led2c of Hl\ prevention among HR' is transmitted anid host it can be prevented. If individuals likelk tc be at highnr risk i' scill '%ell this were indeed the case. penple w-ho are more likely ihort of 100 percent. -T-his suggests that addirional to engage in high-risk behavior nould cend [o kno%s efforts are needed to provide basic information more about HRV prevention than those " hose heha%- about prevention to thow wuho are most likelY to ior doc2 not put [hem at risk, contract and spread [he sirus. Data from seven countries in Sub-Siharan Aftrica supporT this hYpothesis. Box Figure 3.3 shos.;s the Box Figure 3.3 Percentage of Adults Who Know percentage of men and women with various back- That Condoms Are a Means of Protction against ground characteristics w-ho knew% that HEY transmis- HIV Transmission, by Individual Characteristics, sion could be prevented by condonis. The diaa are Seven Sub-Saharan African Counties from a pooled sample of DHS data on adults trom Burkina Faso. the Central African Republic. Core 1w iporcent) d1lvoire. Senegal. Tanzania. Liganda. and Zimbab%te. Women E Those gToups with the lowest level of knots ledge I -- - - ------- about the protective effect of condoms llar left in 80 -.-.......-.-. ...... the figurei w%ere married men and women in their late -Os who lived in rural areas, had lirtlt or no schooling. and no nonregular sexual partners in a 60 - . -... recent time period. The -y were also the least likely u to...-.. nced this knokledge: HIX infection lesels are loiter in rural areas than in urban areas in Sub-Saharan 40 - Africa. and monogamous couples would not nced -- - -- to use condomr to avoid infection. Lack of educa- tion mighr also be responsible for lost lesels of 20 . . ... kniomledge: DHS dati sho%t a srron-. correlarion . benteen education level and knowledge of condons I - I as a means of presenting HIl lbakgrad /itrdk,. Uirb, ijrhan unn,r- tjrb.r,.'- 1C1t791 ~~~~~~~~~~~~~~~r...rjl rn-,-d t-cd Mt-r- Ire mied nee- a4ce FilnilCr I1cf-) . C,,f; d! -m, r.lr &,idr f F ; Irr Ar [he ocher extreme are the r%i-o bars on [he ite C -I ;c- right of the figure. w, hich correspond to the groups L' p, rr- iar Ar. pr"- r1r An of men and wonmen with rhe highest levels of knowl- ., id r;n. . -:.1i prjt edge about condoms as a means of protecting .; PUIFr, . . against HIV. Thev w-ere v-ounger. single. lived in P.-oled [ d1rW dAC r Bulrit na F3,c che Cenisil urban areas. and recently had a nonregular sexual Africir Reptblic. (lS d 1.oirc Scncg;t. Tan7anii. Uganda. mod partner. all facr.rs that ssould also put them at higher ,mbib-,e Io8 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS knowledgeable about HIV transmission than was a comparison group of college students (Sarkar and others 1996). While some of those with high-risk behavior may have become knowledgeable as a result of targeted prevention programs, the fact that they are more knowledgeable than others and yet persist in risky behav- ior underscores the point that simply increasing awareness of HIV will not change high-risk behavior enough to end the epidemic. Indeed, if this were the case, the now-widespread knowledge that HIV is trans- mitted through unprotected sexual intercourse and shared needles would have marked the beginning of the end of the epidemic. Researchers have identified a number of interrelated factors that affect the way in which people use knowledge to assess the costs and benefits of risky behavior and internalize them. These factors include the extent to which: they understand how HIV infection would affect them person- ally; they perceive their own behavior to be risky; and they have the skills necessary to negotiate safer behavior with partners and to resist social pressures. Interventions that address these issues can result in substan- tially more behavior change than knowledge alone (Choi and Coates 1994; Holtgrave and others 1995; Oakley, Fullerton, and Holland 1995). These interventions can take many forms, from public informa- tion campaigns using mass media to training and education programs conducted face-to-face (box 3.4). Some of these approaches are likely to be more effective than others. For example, a lecture coupled with skills training that included role playing, psychodrama, and group discussions was more effective in raising condom use among homosexual men in the United States than was a small group lecture alone (Valdiserri and others 1989). Approaches that change behavior by improving the information base are enhanced by other policies, discussed below, that lower the costs of safer behavior. Nevertheless, many knowledgeable people who have the skills to re- duce their risk and have internalized the dangers will continue to engage in risky behavior. Given the devastating personal cost of contracting HIV, why isn't the danger of being infected sufficient to cause people to forgo all risky activities? One reason is that the costs of reducing risk are clear and immediate while the benefits are uncertain and distant. Whether people are willing to incur these up-front costs to reduce the risk of future sickness and premature death from AIDS depends on their assessment of the probability of infection from a specific act and the dis- count rate applied to future years of healthy life. Assessing the probabil- lO9 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 3.4 What Is "Infonnation"? AT THE. OPERATIONAl. I FEL OF HR' PREVENTION U education. such a- reproductive healih and programs, information sometimes refers to thc first Hl-VlVAIDS education in schools oft three components in "information. education. and * counseling to help people asse;s their owvn risks communication' acriviries. or IEC. Billboards, pam- and take; appropriate actioni. phlets, and public service messages on radio and tele- vision are examples of this narrow- definirion of in- Some waYs of pro-iding intormation are more formacion. However, this report uses int;rmation in effective than others in helping people to adopt the much broader sense commonly understood by safer behavior. Ideniti'ing the way-s in which infor- economists and mans others to include all types of mation can be pros idcd most effectiv-els at the lovs - know%ledge, regardless of how.v ir is acquired or shared. esr possible cosr is an imporrant operational ques- Thus, in this report, the phrase "providing informa- ion, but is bevond the scope of this report. \While tion' encompasses such diverse activ6ities aS: inFormation, broadly defined and efficiently pro- sided. can almnost certainl]y change the behavior of * making available basic knouledge. such as the sonie people to an extent. it is unlikelY ri-. he stffi- facts abour HIl' transmission and hosV to pro- cient to changc behavior enough to stop the tect oneself HIV"AILDS epidemic. especially among those most * training in skills and motivation, such as how likely to contract and 'pread the virus. to negoriate condom use or to sterilize inject- ing equipment ity of infection is not easy, since most people do not know the HIV in- fection status of their partners or even the extent of their partners' cur- rent or past risky behavior. It is hardly surprising that, for people in the thrall of passion, addiction, or difficult economic circumstances, the im- mediate and certain costs of reducing their risk of contracting HIV will sometimes appear greater than the uncertain future benefits. Moreover, even if people making decisions about their private behavior take into full account the potential costs to themselves, they are likely to greatly undervalue the costs of their behavior to society in terms of secondary infections that unintentionally may result. For these reasons, although better information on the risks of HIV will induce people engaging in high-risk behavior to reduce their risk somewhat, they will still engage in riskier behavior than society would prefer in the interest of curbing the AIDS epidemic.4 Additional mea- sures will be necessary to change the costs and benefits of risky behavior and its alternatives in ways that encourage safer behavior. IIO EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Lowering the Costs of Condom Use Condoms are highly effective in preventing HIV transmission, both directly and by reducing the transmission of other STDs (Pinkerton and Abramson 1997). However, even people who are fully aware of the risks of HIV and of the protective benefits of condoms may not use them. The costs of condom use include not only the price of the condom but also the potential inconvenience and embarrassment of obtaining and using one, and, for some people, reduced sexual pleasure. Policies that lower these costs, by lowering the price of condoms, improving their availability, and increasing their social acceptability, would be expected to increase condom use and reduce HIV transmission. Condom social marketing programs aim to achieve all of these objec- tives: they sell condoms at low, heavily subsidized prices; they make con- doms readily available by selling them at nontraditional outlets, such as pharmacies, drug stores, truck stops, bars, and hotels; and they increase the social acceptability of condoms with advertising campaigns and grass roots activities, such as street theater, that show condom use as normal, healthy, and even fun. These programs are often targeted at low-income households, where the price of condoms is more likely to inhibit de- mand. Following the introduction of condom social marketing pro- grams, sales have increased dramatically (figure 3.1). In many of the countries shown in figure 3. 1, condoms were virtually unavailable or un- known prior to the program. Most condom social marketing programs are subsidized by international donors and many, such as the program in India, receive government subsidies. Some governments have also low- ered condom prices by reducing import tariffs and sales taxes on con- doms and latex rubber. The potential impact of condom social marketing coupled with re- ductions in taxes and tariffs can be seen in Brazil. Before the social mar- keting campaign began, condoms cost between $0.75 and $1 each, and market volume was stagnant at about 45 million pieces; a tariff on im- ported condoms kept prices high and sales low. In 1991, subsidized Prudence brand condoms, priced at about $0.20 each, were launched with an intensive information campaign, and the tariff on imported con- doms was gradually reduced. By 1995, total market volume had more than tripled to an estimated 168 million condoms (Clemente and others 1996). Both the domestic suppliers and the socially marketed condoms experienced sales growth, and Prudence condoms claimed about 11 per- cent of the greatly expanded market.5 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 3.1 Socially Marketed Condom Sales in Six Countnes, 1991-96 Sales (millions) 35 30 Vietnam/ 25 20 15 10 Condom swoial maring progranm raise condom sales by offering 5 --'-;/-*- condonms at subsidized prices, by Ethiopia increasing their availbility, and by Cote d'ivolre ..... - Uganda popularizing their use. 0 I l l I I 1991 1992 1993 1994 1995 1996 Source: Statistical appendix, table 3. Some governments, NGOs, and condom social marketing programs also have launched condom promotion programs focused specifically on commercial sex. The social marketing program in Cambodia, for ex- ample, includes special activities to reach sex workers and their dients at bars and hotels. Some of these programs distribute condoms to sex workers for free. Several have succeeded in raising condom use among sex workers and have shown a clear impact on the incidence of HIV Box 3.5 describes the success of one such program involving sex workers in the Congo DR (formerly Zaire); similar success has been achieved in a program involving prostitutes in Nairobi, Kenya (Ngugi and others 1988; see box 2.6). Thailand launched a nationwide campaign that in- cluded condom distribution in brothels and a mass advertising campaign promoting condom use in commercial sex, with the goal of achieving 100 percent condom use in commercial sex. Condom use has risen dra- EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.5 Preventing HIV among Kinshasa Sex Workers .\N 'TD I RE'. ENTTi:)N Pfl CWR.\¶ IN 1[NlSF[.\A,. gonorrhea. trichomoniasis. and genital ulcer dis- Congo DR I formerlk ZaireI. that offered MIG HI\'- eame-which probably contributed to the reduction negati%e sex Unorkers free condom.. STD tests. treat- in HEV transmiswion. \Women w,-ho used condoms ment. counselin, arnd group discussions about pre- conistentis and artended the clinic regularlk had %cntion succeedecl in increasing condom use and 1o%%er HIV -incdcnce rhan those whose condom use reducing the inciden;cc of STDs. including HI\ and viiii to the clinic were ls frequent. I Lagaand othiers ItN-i . Tsso Important lessons troll the success of this Prior to the progra_m. only 11 percent of the sex program are the complementarirv of condom and %%orker, used condoms. and those wkho did used sTD treatment interenuions with sex *sorkers and them occa'ionalls. Three monrh into rhe progrirn. the need to risc rhr willlngness of clients to use con- more than half of the %%omen reported that the% doms to cnsure success. \\hile increased condom use %sere using condoms consistenily-that is. thev %cere %sa- probably most directly responsible for the reduc- engaging in not more than one unprotected act per tion in HlI anid S FD incidence and a less costly Week. By the end of chc three-year project. onsis- component of the project. the STD treatment com- rtnt condom use had riscn to more thin t%o-thirdls ponent was also important for ensuring the health of of the participating women. The main obstacle cited thc sex workers and maintaiing their participation for not achic% ing ZOO percentl condimna use %s3a in the prolect. Not taklr into account in the e%alua- refusal bv mnale clients to use them. cion is the facr [hat midway th-irough the project. the In the courie of the project. the incidlcic of social miirketing progranm for A'Wdence condoms was Hl\- I dropped fromn II to 1.-4 pcr 1i) stoman- launched in Kinshasa. and likely had an import3nt sears of obserxnaion (box figure i.M;i. There 'las abco impact on the ssillingness of clients to agree to con- a decline in the incidecrcc of three trcrable s I-[Ds- donm use I NMaric Laga. personal communication l. Box Figure 3.5 Incidence of HIV-1 and Other STDs among HNV-Negative Sex Workers over Three Years HIV incideice iale pei 100 peson-yea's SrD Incidence rate per 1o0 personmentihs U ____ 6-monthly HIV I ncidence _ 14 13 - rates - 13 12_- Gonoceccal infection - 12 ChiamVdial infection LI 6 Trichomniasis 4 7 10 1 ... Genctai cicer disease -1 9 9 7~~~~~~~~~~~~~~~~~~~ 6 5 ~~~~~~~~~~~~~~~~~~~~5 4 ON1.001% 4 3 3 2 2 3 6 9 ±12 ±15 18 2.1 24 27 30 33 36 Months since start of the study L.h.. i.1, d ,;dwrm [,") I Ih 1 K It- [ -ri [led I'. Perm!.%k. 113 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC matically, to over 90 percent of commercial sex acts, and HIV prevalence has declined among several groups with high rates of partner change (Nelson and others 1996, Rojanapithayakorn and Hanenberg 1996). Policies to increase condom use among people who have unprotected sex with many partners are a potentially powerful means of reducing the size of the HIV epidemic and thereby of providing significant benefits to the entire society. These programs, however, can be politically contro- versial. Providing subsidized condoms to people with many sex partners may be viewed by the public as condoning immoral behavior. Better un- derstanding of the potential benefits of such programs-and the costs to everyone of failing to adopt them-is an important first step in increas- ing their acceptance by society generally and by constituencies that might otherwise oppose them. Lowering the Costs of Safe Injecting Behavior As we have seen in chapter 2, once HIV is introduced to a population of injecting drug users, the risk of infection to those who share injecting equipment is extremely high-higher and more immediate than for any other group engaging in high-risk behavior. Since HIV greatly increases the personal cost of injecting with shared equipment, we would expect to see substantial behavioral change, provided that there are low-cost ways for injecting drug users to avoid becoming infected. Not all injecting drug users are addicted, but for those who are, quit- ting is rarely a low-cost option.6 It can be extremely painful, can take a long time, and is often not successful: within a year or two of complet- ing drug treatment, 70 to 80 percent of people treated for heroin use typically resume injecting (Golz 1993, McCoy and others 1997).7 This is not to suggest that treatment programs are without value, even from the limited perspective of confronting HIV/AIDS. Often they are the only available channel for reaching injecting drug users with HIV pre- vention programs. And although treatment often does not result in permanent cessation of drug use, it can sometimes result in less risky behavior-lower injecting rates and less needle sharing-among those who resume injecting (Blix and Gronbladh 1988; Metzger 1997; Rezza, Oliva, and Sasse 1988). Overall, however, there is little evidence to sug- gest that treatment programs alone will be an effective means of con- taining the HIV/AIDS epidemic among injecting drug users.8 Thus, although providing information on the risks of HIV transmission and "14 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS lowering the costs of drug treatment programs may induce some inject- ing drug users to stop injecting altogether and others to adopt safer prac- tices, many are likely to continue risky injecting behavior unless the costs of safer injecting behavior-primarily the costs of obtaining sterile equipment or bleach-are sufficiently low. Unfortunately, in many countries safer injecting behavior can be very costly to the injecting drug user. Research suggests that difficulty in ob- taining sterile injecting equipment is the most important reason that in- jecting drug users share equipment (National Research Council 1989, Vlahov 1997). Problems in obtaining sterile equipment are likely to be especially severe in developing countries, where low incomes and general scarcity of sterile injecting equipment and even bleach leave many in- jecting drug users with no access to dean equipment. In Myanmar, for example, many addicts go to professional injectors who sell the drugs, injecting one person after another with a single needle attached to an eyedropper or plastic tubing (Oppenheimer 1995). Injecting drug users in Ho Chi Minh City, Vietnam, reportedly obtain injections at "shoot- ing booths" ("Ho Chi Minh City. . ." 1996). A survey of one town in Manipur State in northeastern India found that virtually all who inject (97 percent) share equipment: in most cases, eyedroppers attached to sy- ringes (Sarkar and others 1996). Often, even if a sterile syringe can be obtained at a price that an injecting drug user is able to pay, possessing one can lead to arrest. In such instances, an injecting drug user must weigh the risk of HIV infection from sharing equipment against the risks of prison if caught with a syringe. One inexpensive way for governments to lower the cost of safer in- jecting behavior is to remove legal barriers to the purchase and posses- sion of sterile injecting equipment. When armed with knowledge of the risks and provided with legal, easy access to sterile equipment, injecting drug users in many settings have rapidly adopted safer behavior. After needles were made available from pharmacies without a prescription in 1992 in the U.S. state of Connecticut, the percentage of injecting drug users who shared needles declined from 71 to 15 percent within three years (Span 1996). In Bangkok, Thailand, over 90 percent of injecting drug users surveyed reported changing their behavior to reduce their risk of HIV infection; 80 percent said they were obtaining sterile injecting equipment from pharmacies instead of sharing equipment with others (Choopanya and others 1991).9 HIV prevalence among injecting drug users in Bangkok, which had been climbing rapidly, stabilized at around I"5 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 40 percent, about half the level that was reached in neighboring coun- tries where syringes and information about HIV were less readily avail- able (Weniger and others 1991). In some locales, governments or NGOs have moved sooner and more actively to minimize the harm from injecting drugs in the face of the HIV epidemic. Such "harm reduction" programs have often held sero- prevalence among injecting drug users at remarkably low levels. In the five cities in table 3.1, harm reduction programs have maintained HIV prevalence among injecting drug users at less than 5 percent for at least five years, while in neighboring cities infection rates among injecting drug users soared to 50 percent or more (Ch'ien 1994; Des Jartais and others 1995; Lee, Lim, and Lee 1993; Poshyachinda 1993; Wong, Lee, and Lim 1993). Each of the cities cited began programs early, before HIV had widely infected injecting drug users. The programs included information on safe injecting, needle exchange programs (NEPs), bleach programs, and referral for drug treatment. Besides containing the epi- demic, these programs also raised the demand for drug treatment in all five cities. However, among all of the cities, only Sydney, Australia, launched a massive expansion of drug treatment programs. Table 3.1 Prevention Programs in Cities That Have Kept HIV Infection among Injecting Dnug Users below 5 Percent Evien.shle /Dt 's Lega/ 1,0/u ornr- reporting pitrubdme of E.xpanded Hll change in Progpaill Begai A'eedle Iniecwtmg B/emah Cor,ll,unt,lr dnrg counvelulg b&haiior lite vren I.PVChaqge eqhippneptr dibtribition onrre.uh treannen,r & te4iJlg q ,(%i , i.:'cl!.nd I Lund * i s%dc n i IAusLraliai Tic,,nia 3 ,U) S I T~rc.nco JI CicY n.da,I 3IDU11 Jriqw1 EIe, Ihrl-_n;ld Othtr4 1 rI uiS * and IUV EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Needle exchange programs, by providing new sterile injecting equip- ment in exchange for used syringes, minimize needle sharing and remove contaminated needles from circulation. Many NEPs also supply free condoms to help prevent the spread of HIV to the sexual partners of injecting drug users, as well as education about reducing high-risk behavior and referral for drug treatment.10 NEPs have been widely adopted in Australia, New Zealand, and many Western European coun- tries with great success in reducing transmission of HIV as well as other blood-borne infections, like hepatitis B and hepatitis C. A recent study compared trends in HIV prevalence among injecting users in a sample of 81 cities in North America, Europe, and Asia and the South Pacific (Hurley, Jolley, and Kaldor 1997). In the 52 cities without NEPs, HIV prevalence among injecting drug users increased by 5.9 percent per year, on average, while it declined by an almost equal percent annually (5.8 percent) in the 29 cities that had NEPs. Nonetheless, political opposi- tion to needle exchange programs remains a potent force. In the United States, these programs are rare and distribution of syringes is often ille- gal. A recent study estimated that a national NEP in the United States would have prevented from 4,000 to 10,000 HIV infections between 1987 and 1995 among drug users, their sexual partners, and children; if such a program had been implemented in 1996, it could still have pre- vented from 5,000 to 11,000 infections over the next five years (Lurie and Drucker 1997). Can harm reduction strategies also work in low-income countries? The answer appears to be yes. Although there is less experience with harm reduction strategies in developing countries, the remarkable suc- cess of a needle exchange program in Nepal suggests that such programs should be attempted more widely. Launched in 1992, the program has helped to hold HIV prevalence among Katmandu's 1,500 injecting drug users at less than 2 percent, even as HIV has soared in other Asian coun- tries along the drug trade routes (box 3.6). Where resources for needle exchange are lacking, bleach distribution may offer a less expensive alternative. Bleach is not only cheaper, it is also less controversial than needle exchanges, and, when used correctly, highly effective in killing HIV on infected equipment (Siegel, Weinstein, and Fineberg 1991). Furthermore, improved availability of bleach is virtually the only option for harm reduction among injecting drug users in prisons, since prison authorities are unlikely to distribute needles because they can be used as weapons. However, bleach is not always I17 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 3.6 Harm Reduction among Injecting Drug Users in Nepal NEP1-r. NIOED EALL rl, T FE1 RE,NT [[IF h[RE.AD OF t0ll 21 percent. and the number Ot times equipment HIV amo:ng injiccint drug u%crs. w-ith dramatic re- %%a, shared decreased 21' perccnt. NMosr telling is the sulIt. In 1]%30 a nonegorArnmerital organization called t'ic that HI' prev-alence anmon in1ecring drug user, thb Lifesaiing and Litegiving Socien- teon it-, pro- in Katmandu has renmilned lc, le,is thin 2 percent %ide edLucation. condoms. bleachi. needic e\chanoc. eten asL it has soared in nearb%y countries iNlaharjan and prinnir. health care to about o Of Ckarmandl. and orhirs 1hTS W. \\heEh-l HR prv-.alerIce remains 1. 011i inie,uing drug users. Thu group has c.lIabo- lo%% in the Vcar~ to LUme "'ill depend on continucd rated 'uth rhl\e linirics oF Home Affairs and Hel[lh harm rCduLcionr fforts anong inlecing drui guers. and 1 enforcemntm a2encies. w hile c.ffcring injc- I ill and other, E 19 e tintated that the total ing drug users confidcntial and nuniudgrnemnta] as%.L- cost, hor the program alt'er one %ear of operation [ance i Peak. Nlaharlan, and ( rclrs 19941. wvere -i333. and (hat the avcraoe cost per cient Among the drug users participating in the pro- contact 'vas p2 ' per onrict. since he v%irus oiten gram, the mian frequeny uft injecring ICll from 2-i prcad Ctrjni injecting drug uscrs to hc-ir partners iniections per v eek. ,horrlv before the Start of the and their children. this e\pcndirure represents an program. to I - iniections per xv :ck in 11i9-. Eh k ineitnment not onl% in protecring the intecEtng druig number of unsiaf injecions dropped hY halt' thc users thcmselves, but alsc in pre%enting a much number ofpcople v, ith uehom cquipment '.s shared -; ider epidemik. readily available. For example, before 1991, bleach was all but unknown in Manipur State in India (Sarkar and others 1996); a bleach distribu- tion program launched that year in the city of Churachandpur increased the percentage of injecting drug users using bleach to sterilize syringes from 31 to 72 percent. Bleach programs produce the greatest gains in life expectancy per uninfected injector when they are implemented early, while levels of infection among injecting drug users are still 2 percent or less (Siegel, Weinstein, and Fineberg 1991). Programs that reduce the cost of safer injecting procedures often arouse fears that they will encourage people to initiate drug use or dis- courage addicts from seeking treatment. If efforts to reduce the cost of safer injection behavior did encourage illicit drug use, this would have to be weighed against the benefits of reduced HIV transmission. Fortu- nately, there is substantial evidence indicating that this is not the case (National Research Council 1989). Two evaluations of NEPs in six in- dustrial countries failed to find any evidence that the programs raised the number of injecting drug users or that they increased the number of im- properly discarded needles (Lurie and others 1993; Normand, VIahov, and Moses 1995; U.S. GAO 1993).1l II8 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Raising the Cost of Risky Behavior: The Uncertain Effect of Enforcement We have seen that lowering the costs of safer sex and safer injecting be- havior can encourage less-risky behavior and reduce the transmission of HIV But both prostitution and use of illegal and addictive drugs impose substantial negative externalities on society in terms of sexually transmit- ted and blood-borne diseases, crime, the costs of law enforcement, and in- carceration. It is not surprising, therefore, that programs that attempt to promote safer behavior among sex workers, injecting drug users, and other people who engage in high-risk behavior may be politically unpop- ular, especially if they are perceived to be condoning prostitution or drug addiction. An alternative approach that often has broad popular appeal is to discourage such activities through punitive measures and stepped up enforcement of existing laws. From the standpoint of economic theory, such attempts to raise the cost of potentially risky behavior could dis- courage it, provided that costs can be raised sufficiently. However, because sexual activity and injecting drug use are private activities, enforcing pro- hibitions is cosdy and difficult in practice (Minon and Zwiebel 1995). Moreover, as we discuss below, such attempts may have unintended con- sequences that may exacerbate the epidemic. Raising the costs of commercial sex. Attempts to curtail commer- cial sex have only rarely been effective. Indeed, numerous studies have shown that prohibition and punishment of commercial sex causes sex workers to shift their place of business and to change the way they solicit clients to avoid prosecution. Singapore, for example, attempted to eradicate prostitution by dosing "red-light" districts in commercial areas; brothels soon appeared in residential areas (Ong 1993). Similar attempts in the Philippines have driven sex workers underground (Brown and Xenos 1994). These efforts rearrange the problems associ- ated with commercial sex, but do not eliminate them. Worse, people who elude law enforcement and continue to engage in commercial sex are likely to become more difficult to reach with public health interven- tions to encourage safer behavior. Enforcement problems aside, punishing commercial sex is not very efficient in preventing HIV, since the virus is transmitted not by com- mercial sex per se but by unprotected intercourse with multiple partners, regardless of whether one party pays the other. Commercial sex can be either high-risk or low-risk, depending on the activity and whether or II9 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC not a condom is used. Thus, even if commercial sex could be eliminated by these measures, HIV could persist and spread through casual sex net- works, although perhaps less rapidly. Moreover, it is reasonable to expect that the more effective is the curtailment of commercial sex, the more casual sex networks will expand. These problems notwithstanding, it is possible for prohibition and punishment of commercial sex to slow the epidemic, provided that high- risk commercial sex transactions are reduced sufficiently and that non- commercial high-risk sex activities do not increase by an offsetting amount. China is one of the few countries in history that seems to have dramatically reduced prostitution and STDs in this manner, for about 20 years beginning in the early 1950s (Cohen and others 1996). How- ever, this was achieved not as an end in itself, but in the context of mas- sive social and political revolution and extensive government control of individual freedoms and the economy. The economic controls proved to be incompatible with growth and have since been relaxed; the internal migration that has ensued has created conditions conducive to the reap- pearance of prostitution, casual sex, and the spread of STDs. Few, if any, countries are likely to be willing or able to impose the extensive social controls and bear the high costs experienced by China in the 1950s simply to control HIV. Short of such extensive efforts, at- tempts to eradicate commercial sex will cause some sex workers to find other types of work and will discourage some clients. Those who persist, however, will become more clandestine and thus more difficult to reach with information about HIV and policies that encourage condom use (background paper, Ahlburg and Jensen 1996). An alternative to prohibiting and punishing commercial sex is to le- galize and regulate it. Although this approach sometimes arouses consid- erable political opposition, it can make sex workers and their clients eas- ier to monitor and to reach with information, condoms, and STD treatment. For example, in legal brothels in Australia condoms are used almost universally and STD rates are low (Feachem 1995). The public health advantages of regulated prostitution can also be seen in the his- torical experience of pre-independence India (box 3.7). But regulated commercial sex faces the problems one would expect in any attempt to create a monopoly where there is a ready alternative sup- ply. The likely result is a legal, high-priced market, presumably with lower transmission of HIM and an unregulated, low-priced market, which authorities cannot easily monitor and where transmission of HIV 120 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.7 Health Benefits of Regulating Prostitution in Pre-independence India REGULAXTED PROSTITUTION OFFERS NANY OPPOR. diers. mandating periodic medical inspection of tunities to monitor the health of prostitutes and prostitutes for infection. The act worked w,-ell in their clients. to treat disease, and to prevent infec- low%ering the transmission of STDs). Howe%er. pub- dions. Howeever, chese measures are often polilicall lic opposition to these measures in Britain and India difficult to implement. Historical experience stith wkeakened enforcement. As a result, hospital idmis- regulated prostirution in India under British rule sions for STDs rose among the troops, pea ng in illustrates both the public health benefits of regu- the 13890s. laced prostirution and the pow%er of public opposi- A second Cantonnients Act passed in 1899 tion to undermine such eflbrrs. ga%e the millitar torces more authority in curbing The Contagious Diseases and Cantonment Act the spread of disease. Once prostitution could of India was enacted in 186-i to control the spread again be regulated. hospital admissions for STDs of STDs among [he occupying British forces among the British soldiers fell dramatically-from IFarAwell 1989i. The act regulated first class" Indian 536 per thousand in 1895 to onls 0f per thou- prostitutes and brothels frequented by British so1- sand in 1909. will presunably be higher (backgroundpaper, Ahlburg and Jensen 1996). In Singapore, for example, where brothels have been regulated since the government abandoned efforts to stamp out prostitution, a significant informal commercial sex sector persists. Moreover, mandatory health screening of prostitutes, even in the regulated brothels, is not always en- forceable (Ong 1993). A similar pattern can be seen in Australia, where the 1986 Prostitution Regulation Act reduced the number of brothels in Melbourne by 65 percent. Result: the price of sex in brothels rose, while the number of streetwalkers and escorts increased (Hatty 1993). In summary, prohibiting and punishing commercial sex is unlikely to be an effective approach to reducing HIV transmission. Although fewer people would presumably engage in risky behavior, those who do so de- spite the threat of prosecution will be more difficult to reach with public health interventions. Depending on the balance between these two out- comes, draconian measures to curtail commercial sex could actually worsen an HIV epidemic. The net impact of legalization and regulation is also difficult to predict and will depend on how much regulation raises the price of sex in the legalized sector and the extent to which programs are successful in encouraging safer sex in the illegal, unregulated sector. Raising the costs of drug use. Arguments for and against raising the costs of drug use through prohibition and punishment are broadly I2I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC similar to those for prohibiting commercial sex. Again, it is important to note that neither drug use per se nor injecting per se spreads HIV; rather, it is the sharing of unsterilized equipment that does so. Many who use illicit drugs do not inject, and often those who inject do not share needles. So efforts to eliminate illegal drug use, while they might be rationalized on other grounds, are potentially a very inefficient and costly way of reducing the unsafe injecting behavior that spreads HIV Nonetheless, the already substantial political appeal of drug enforce- ment efforts is likely to be strengthened by the close association between use of injected drugs and HIV. It is therefore useful to consider the likely impact on HIV of the two main strategies for eradicating illegal drugs: restricting supply, by attempting to halt the drug trade, and reducing demand, by punishing drug users or forcing them into treatment. The most politically popular way to attempt to reduce drug use is to reduce the availability of drugs. However, drug interdiction may simply rearrange the problem or make it worsc. For example: * Addicts may switch to other substances. In India, when the gov- ernment tried to restrict the heroin trade, the price of heroin rose and addicts switched to synthetic opiates; injecting behavior was unchanged (Pal and others 1990). * Users may shift from smoking to injecting, which requires a smaller dose to produce euphoria but greatly raises the risk of HIV For ex- ample, efforts to control opium smoking in Bangkok, Calcutta, and other areas in India were followed by an increase in heroin in- jection (Des Jarlais and others 1992; Sarkar and others 1993). * The drug trade may shift to other areas where people not previ- ously exposed to drugs may begin injecting. For example, as a re- sult of efforts to halt the drug trade in other regions, West Africa has emerged as an important transit point for cocaine from South America and heroin from Southeast Asia bound for Europe and North America. Similarly, stepped-up enforcement in Nigeria shifted the drug trade to Cote d'Ivoire, Zambia, and Zimbabwe (Stimson 1993). If restricting the supply of injectable drugs does not effectively reduce risky injecting behavior and may actually increase it, what about at- tempting to reduce demand? Because most injecting drug users are chemically dependent, prohibition and threats of punishment are noto- 122 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS riously ineffective in reducing their demand for drugs. A survey of 450 injecting drug users in Manipur State in India, where addicts are impris- oned, found that only 2 percent regarded the threat of imprisonment as a reason to stop injecting drugs; half of the addicts surveyed had previ- ously been in prison (Sarkar and others 1993). And far from reducing HIV transmission, imprisonment may have the opposite effect. Unable to obtain syringes, prisoners who inject drugs frequently resort to shared, improvised equipment, such as ballpoint pens, which would be very dif- ficult to sterilize, even if bleach were available. Mandatory drug treat- ment is likely to be even less successful in ending drug use than volun- tary treatment, since patients entering such programs presumably have very little desire to change their behavior. In summary, efforts to raise the cost of injecting drugs through drug interdiction or the punishment of injecting drug users may increase rather than decrease risky injection behavior. Although the data on the impact of such efforts on HIV incidence are fragmentary, the available evidence suggests that harm reduction programs, including information about HIV, sterile injection equipment or bleach kits, and referral for voluntary treatment programs will be more effective and less costly in reducing risky injection behavior than drug interdiction or incarceration of addicts. This is particularly true given the strong evidence, discussed above, that injecting drug users do respond to this information by re- ducing risky injection behavior that spreads HIV Easing Social Constraints to Safe Behavior A SECOND APPROACH TO REDUCING THE SPREAD OF HIV AIMS to change the social and economic factors that shape-and sometimes constrain-individual choices about risky behavior. Measures pursued by this approach have many other benefits besides reducing the HIV epidemic and they are already on the agenda of most developing country governments. The benefits are sometimes more dif- ficult to quantify because of their broad impact, which extends far beyond HIV prevention. However, these measures are highly comple- mentary to policies that directly affect the costs and benefits of risky behavior. Included in this approach are measures that alter social norms, raise the status of women, and reduce poverty. 123 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Alteing Social Norms Some social norms discourage behaviors that transmit HIV, while others may be conducive to high-risk behavior or may discourage peo- ple from adopting safer behavior. HIV is likely to spread more widely where multiple, concurrent partnerships are the norm. In urban areas of parts of Sub-Saharan Africa, for example, traditional polygyny has evolved into many forms of formal and informal marriages and consen- sual unions, often concurrent and long term. The resulting sexual net- works are highly conducive to the rapid spread of HIV (Caldwell, Cald- well, and Orubuloye 1989; National Research Council 1996). Social norms and peer pressure that encourage men to use the services of pros- titutes or that venerate men with many female "conquests," while plac- ing a high value on female chastity create the conditions for an explosive HIV epidemic. Recent surveys reveal dramatic differences in the premarital sexual experiences of unmarried youth in different countries, even within the same region. The difference between what is acceptable for women and for men is also striking. In Rio de Janeiro (Brazil), for example, 61 per- cent of never-married men 15 to 19 had sexual intercourse in the past 12 months, compared with only 9 percent of never-married women (Carael 1995). In Manila (the Philippines) and Thailand, 15 and 29 percent, re- spectively, of men 15 to 19 had sexual intercourse in the 12 months be- fore the survey, but for women the share was 0 to 1 percent. There is also substantial diversity across countries in Sub-Saharan Africa; for example, in the Central African Republic, C6te d'Ivoire, Guinea-Bissau, and Kenya, young men and women are both very likely to have had sexual intercourse, while in Burundi and Togo, the proportion of young men and women who have had sexual intercourse is very small. The challenge for policymakers in countries with social patterns that are very conducive to the spread of HIV is to encourage safer behavior, without stigmatizing those who engage in unsafe behavior in ways that make them more dif- ficult to reach with public health interventions. Norms on marriage and childbearing can also affect the spread and prevention of HIV The tradition of a man selecting a bride who is five or ten years younger than himself spreads HIV from one generation to the next (backgroundpaper, Morris 1996; Ssengonzi and others 1995). In eastern Africa, the traditional custom of levirate marriage was once a major contributor to the spread of HIV. According to this custom, a 124 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS woman whose husband has died must marry his brother or, at a mini- mum, have sexual relations with the brother. In part because of the risk of contracting HIV, the practice is now in decline, but still widespread. Throughout Sub-Saharan Africa, mothers improve the survival of young children by breastfeeding as long as two years after birth. However, in some societies there is a taboo against marital sex while the mother is breastfeeding, leaving husbands to seek sexual gratification elsewhere. Fi- nally, efforts to encourage condom use among married couples for HIV prevention will be especially difficult in societies where people want large families and a woman's social status and economic well-being are heavily dependent on the number of children she has. This is largely the case in Sub-Saharan Africa, where the benefits of increased condom use among married couples in suppressing the HIV/AIDS epidemic would be great- est (Bankole and Westoff 1995). While sexual conservatism may be one of the best protections against HIV at the societal level, adherence to norms is never complete. An HIV epidemic can nevertheless occur, and conservatism can result in stigma of those who become infected or who are in social groups associated with high-risk behavior, making it more difficult to support safer behavior. Sometimes religious and political leaders stigmatize condom use as im- moral, erecting additional social barriers and costs to safer and more responsible behavior. Thus, even though broadly based sexual conser- vatism may be helpful, treating HIV and the behavior that spreads it as a moral issue rather than a public health issue can hinder efforts to con- tain the epidemic. Improving the Status of Women In most societies, the lower social and economic status of women re- duces their ability to insist upon male sexual fidelity and to negotiate safe sex. These problems can be particularly acute in societies where women's inheritance rights, property and child custody rights in divorce, and even the right to own land and other property are limited. In some instances, a wife's mere suggestion that her husband use a condom can provoke physical abuse. Even where the situation is not this stark, women's lower literacy, lower incomes, and low economic independence relative to men's give them less access to prevention information, fewer resources to pur- chase condoms and treat conventional STDs, and less ability to leave a relationship that puts them at risk of contracting HIM For all these rea- I25 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC sons, many more women than men face situations in which they are unable to choose behavior that would protect them from HIV Women who sell sexual services often face particularly constrained choices. Unemployment, divorce, desertion, and the breakdown of the extended family are among the factors that can lead women to offer sex for money (Plange 1990). One-half of sex workers interviewed in Cal- cutta cited extreme poverty as the reason for entering sex work, and 22 percent cited "family disturbances" (Chakraborty and others 1994). Often one of the few jobs available to unsupported, single women with limited education, prostitution can be quite lucrative compared with available alternatives. In The Gambia, for example, sex workers earn three times as much per day as women in informal sector jobs and as much as senior civil servants (Pickering and Wilkins 1993). Low-priced sex workers in Bali, Indonesia, earn more per week than the average civil servant does in a month (Wirawan, Fajans, and Ford 1993). In metro- politan Bangkok and a northeastern province of Thailand, the average take-home pay of sex workers is more than twice that of the average woman of the same age in other jobs (Bloom and others forthcoming). Controlling for their age and education, female sex workers earned over 50 percent more than they would have in other jobs for which they were qualified. Women may also engage in commercial sex to fulfill family obliga- tions. Throughout the developing world, poor families seek to insure themselves against economic risks by diversifying economic activities and through networks of family members spread over broad geographic areas. Adult children migrate to urban areas, abroad, or to specific areas to seek lucrative jobs that will allow them to save and remit earnings to their families. This motivation, coupled with the high returns for prosti- tution, explains much of the supply of sex workers, particularly in Asia (Archavanitkul and Guest 1994; Wawer and others 1996a). Fewer than one in ten sex workers in Bali, Indonesia, are from Bali (Wirawan, Fajans, and Ford 1993). Nepali women make up half of the population of prostitutes in Mumbai (Bombay) brothels (Human Rights Watch/Asia 1995). In Indonesia, Nepal, and Thailand, migration net- works between particular villages that supply young women and specific areas in urban commercial sex districts are well established (Archavan- tikul and Guest 1994; Human Rights Watch/Asia 1995; Jones, Sulistyaningsih, and Hull 1994). Ending legal restrictions on women's rights, encouraging social equal- ity, and raising economic opportunities for women not only make it eas- I26 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS ier for women to avoid HIV; they are also important in fostering devel- opment. Policies that specifically help women include increasing girls' school enrollment; guaranteeing equal employment opportunities; out- lawing and severely punishing slavery, rape, wife abuse, and child prosti- tution; and guaranteeing inheritance, property, and child custody rights. A growing economy is also an important ingredient for increased eco- nomic opportunities for women. Of course, improving the status of women will also open new opportunities to choose risky behavior; a few women may do so. However, it is reasonable to expect that the vast majority of women, given expanded choices, would seize the opportu- nity to choose safe behavior and avoid HIV infection. Reducing Poverty Poverty and low socioeconomic status also constrain the decisions people make about risky behavior. Those with low incomes, for example, may not be able to afford to treat STDs or to buy condoms. Poor fami- lies may see commercial sex as a lucrative occupation for young and poorly educated daughters. People with less education may have less access to information about the dangers of high-risk behavior or be less able to understand prevention messages. This explains why those most likely to contract STDs and other infectious diseases within a society are the poor and uneducated. And it is supported by the cross-country analysis in chapter 1, which showed that developing countries with higher incomes have lower levels of HIV infection. However, at the individual level within countries, the probability of HIV infection is often greater among men and women with higher in- comes and schooling. Most of the evidence comes from studies con- ducted in eastern and central Africa in the late 1980s and early 1990s. For example, in a study of female outpatients and pediatric patients in Kigali, capital of Rwanda, women whose main partners had higher edu- cation and income were more likely to be infected than those whose partners had low education and income; infection rates showed similar patterns according to the partners' occupations (table 3.2). Among women attending family planning clinics in Dar es Salaam, Tanzania, those whose partners had more than twelve years of schooling were five times more likely to be infected than those whose partners had no school- ing (Msamanga and others 1996). In Malawi, HIV seroprevalence was lowest for pregnant women whose partners had no education (5 percent) but rose to 16 percent among women whose partners had more than 127 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Table 3.2 Percentage of Women Ages 19 to 37 Infected with HIV, by Their Partners' Socioeconomic Status, Kigali, Rwanda Hril'- posilitre Partner, '(duraalens,7scpr~ I;catitl r. I | Ncine 22 >I - NII :lt Farming C ierVAnr .a,f r S.Trpk .iLc I S... , )JI,nando.chor - 1)! seven years of education (Dallabetta and others 1993). In rural Rakai District, Uganda, household heads with any education were more likely to be infected than those with no schooling (background paper, Menon and others 1996b). In Kagera Region of Tanzania, the probability of dying from AIDS was higher for women with primary schooling or sec- ondary schooling, compared with that of women with no schooling (backgroundpaper, Ainsworth and Semali 1997). The gap in infection rates between the educated and uneducated in the early 1990s was greater in rural areas than in urban areas of Eastern Africa, for both men and women. In the town of Mwanza, Tanzania, for exam- ple, women were more likely than men to be infected, but there was no difference in HIV prevalence among men or women according to their schooling (table 3.3). However, in rural areas of Mwanza and in Rakai District, Uganda, there were sometimes marked differentials in prevalence by education, which were larger among women than among men.12 Finally, higher-income adults in Central and Eastern Africa were more likely to be infected than those with lower incomes. Workers with higher incomes were more likely to be infected in two businesses in Kin- shasa, Congo DR (formerly Zaire), in the late 1980s (Ryder and others 1990). The workers in the (higher-paying) bank had higher HIV infec- tion rates than those in the (lower-paying) textile mill and, within each I28 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Table 3.3 Relation between Education and HIV Status, Men and Women, Mwanza Region, Tanzania, and Rakai District, Uganda HI'prei'ale;lce (pertewr Studi sie Level of srhooling Men WI'omen Nlwanza. Tanzanii turbanil Few%er than .cari 9.( 15.3 4 %vars or more S3 I ^.3 ,%l%%anza. Tan!ani trurab Fenwr thanr - years 3.0 i % cars or more d. '.1 RA-a] Ditrrici. Uganda (rural I Nonc - 13. Primanr i -.6 Sc.:ndnl I -14. 41.1- ', ,rc,,. Barc.ng, and orh.r; I911'2. k_ros>.LLi h in d orher, IN`1b. S Sr.. idda i md [h.r, I12. firm, managerial workers had higher HIV prevalence than manual work- ers. In Rakai District, Uganda, heads of household with higher-quality dwellings were half again as likely to be infected as those without, con- trolling for age, gender, marital status, education, and occupation (back- groundpaper, Menon and others 1996b). Why would adults with higher socioeconomic status have higher HIV infection rates? First, men with higher education and income will find it easier to attract and support additional commercial and casual sexual partners. For example, analysis of data from the WHO/GPA sex- ual behavior surveys found that in five African sites, as well as in Thai- land, Manila (the Philippines), and Rio de Janeiro (Brazil), the more ed- ucation a man has, the more likely he is to have had a casual, nonregular sexual partner (background paper, Deheneffe, Carael and Noumbissi 1996).13 A second reason is that men and women with more education and higher incomes are likely to travel more and thus have more oppor- tunities for a variety of sexual contacts. Do these results imply that for HIV, unlike other infectious diseases, including other STDs, poverty reduction and rising education may ac- tually increase the spread of HIV? This would seem to contradict the finding in chapter 1 that HIV infection rates are lower in countries with higher income and literacy. This seeming discrepancy between findings at the individual and in- ternational levels can be explained by two factors. First, at the time many of these people becarne infected, in the early- and mid-i 980s, awareness and knowledge of HIV prevention were low. Thus, the protective ad- 129 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC vantages that higher education and income would normally provide-a greater ability to learn about HIV prevention and more resources to pur- chase condoms or take other steps to avoid infection-did not come into play. Second, lacking this knowledge and failing to take protective mea- sures, those with higher income with their greater number of partners were more exposed to HIV. Since HIV prevalence is cumulative, over a long enough period this would result in higher HIV prevalence among those with higher education and income than among the poor, with fewer partners. If these explanations are correct, then as knowledge of how to avoid HIV infection becomes available, people with more education and higher incomes would be in a better position to learn about it and avoid infection. As a result, HIV incidence should decline more rapidly among those who are better-off, eventually reversing the positive relation be- tween income and prevalence found in African studies. The limited available evidence suggests that this is in fact occurring. In Brazil, for example, about three-quarters of those newly diagnosed with AIDS through 1985 and for whom educational data were available had a secondary or university education. By 1994 only about one-third of those with newly diagnosed AIDS had that much education (Parker 1996). In urban areas of Butare, Rwanda, HIV incidence was higher among women in low-income households (Bulterys and others 1994). This result is consistent with trends in industrial countries; in the United States, for example, new infections are more likely to occur among those with low socioeconomic status (Cowan, Brundage, and Pomerantz 1994; Krueger and others 1990). Further, DHS data confirm that in all of the developing countries studied, the more education men and women had, the greater was the likelihood that they were using condoms (figure 3.2).14 In another study, Thai men with the highest permanent income and assets were more likely than other men to consistently use condoms with sex workers (Morris and others 1996). Among 21-year- old Thai military conscripts in the early 1 990s, HIV incidence was lower for conscripts with more education (Carr and others 1994).15 Several studies of sex workers have found that women with higher incomes were more likely to use condoms and had lower levels of HIV infection. In three cities in Sao Paulo State, Brazil, for example, sex workers who charged higher prices had fewer clients, were more likely to have always used condoms in the past year, were less likely to have in- jected drugs, and consequently were less likely to be infected with HIV 130 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Figure 3.2 Percentage of Men and Women Using Condoms with a Casual Partner, by Education, Eight Countries Men's condom use (percent) 70 Burkina Faso 60 - r-abe. 50 .t..t....-- lal African RepublicKea 40 *. nzania 0 e.. 10 . -' '----. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . None 1-3 4-6 7-10 11+ Years of education Women's condom use (percent) 60 50-Tanzania_ zir. 30 ~ /~~~~~~- -- ---'- I -- - ---- --- ---- / . . Central African Republic 20--.... 1 0 - .. - . . .. : . .. . . .. . : . .. . . . .. . . .. . . 10;-;^ . . - Men and women with more schooling K , Kenva ,..- are more likely to use condoms for o1 71 1 l casual sex. None 1-3 4-6 7-10 II+ Years of education Note: This figure shows the probability of condom use, holding constant age, urban residence, occupation, and assets. The reference period varied. See note 14 at the end of the chapter. Source: Backgroundpaper, Filmer 1997, from DHS data. 131 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC and other STDs than were low-priced sex workers (Lurie and others 1995). In conclusion, despite the links found between higher socioeconomic status and HIV infection in some areas, government policies to raise in- comes and schooling and reduce poverty should lower the economic constraints to engaging in safer behavior and reduce the incidence of HIV over the long term. These policies are often pursued because they have far-reaching social benefits; they are also highly complementary to short-term interventions to encourage safer sexual and injecting behav- ior and will likely enhance the impact of these interventions. Setting Government Priorities in Preventing HIV L OWERING THE COSTS OF SAFER SEXUAL AND DRUG-INJECT- ing behavior for those who are most likely to contract and spread HIV can reduce risky behavior, and this in turn will have a powerful impact on the course of the epidemic. Given the many pos- sible ways of attempting to achieve these ends, what are the highest pri- ority programs from the perspective of government spending? This section proposes a broad prevention strategy for governments at all stages of the epidemic to maximize the impact of limited government resources in curbing the spread of HIV. In keeping with the principles of public economics, governments should finance or directly implement interventions that are essential to stopping the spread of HIV but that private individuals or firms will not have sufficient incentive to finance on their own-namely, provision of public goods, reduction of the neg- ative externalities of behavior that spreads HIV, and protection of the poor from HIV infection. Programs that address these issues will im- prove the efficiency and equity of government prevention efforts. In addition, following the principles of epidemiology from chapter 2, pro- gram effectiveness will be improved if governments act as soon as possi- ble and if they succeed in preventing infection among those most likely to contract and spread HIV. These recommendations are not meant to limit the scope of government involvement if there are resources and public will to undertake even more. Rather, the intention is to identify the minimum set of activities that all governments should be engaged in to improve the efficiency and equity of prevention programs, and a rational order in which to expand them. 132 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS This strategy is a broad one, based on underlying principles of pub- lic economics, epidemiology, and cost-effectiveness. It remains for indi- vidual countries to identify the specific combination of programs, poli- cies, and interventions to pursue this strategy in a cost-effective way. We discuss a number of factors that sometimes uniquely affect the cost- effectiveness of public activities to prevent HIV However, even within these guiding principles, the cost-effectiveness of HIV prevention pro- grams is likely to vary considerably according to the setting. Program- matic choices are necessarily country-specific because the costs and effec- tiveness of interventions, as well as the characteristics and accessibility of those most likely to contract and spread HIV, vary widely across settings. Public Economics and Govemment Prioities Government policies can potentially address three types of market failures that occur with respect to the prevention of HIV. The first prob- lem involves the underprovision of public good-specifically, the lack of incentives for the private sector to collect and disseminate information crucial to the prevention of the epidemic. The second involves the nega- tive externalities of high-risk behavior: people deciding whether or not to take steps to protect themselves from HIV are likely to consider the costs to themselves of becoming infected but may not consider the cost of sec- ondary HIV infections that may result if they become infected. The third type of market failure involves equity: very poor people are less able to protect themselves against HIV than others. Providing public goods. The collection and production of infor- mation about the prevention and control of HIV is almost entirely a public good, since it is impossible for a private agent to capture the resulting benefits. Essential information includes data on the levels and trends of HIV and STD infections, the prevalence of high-risk behav- ior, and the costs and benefits of prevention programs (box 3.8). Everyone benefits from such information, but there is insufficient incentive for it to be produced privately in sufficient quantity. Reducing the negative externalities of behavior that spreads HIV. If people engaged in risky behavior were fully informed of the dangers of HIV infection and how it could be prevented, if they had the means to prevent it, and if they were the only ones to suffer conse- quences if they became infected, then the argument for government intervention to prevent their infection would be weak. However, unprotected intercourse with multiple partners and unsafe injecting I33 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 3.8 What Data on HIV and Other STDs Should Governments Collect? THE TYPFS OF iNFORMLATION DES.CRIBtED BEL0 \( countr'? Are the- prevalent in ihe general are crucial if the g-overnment is to design and population or conFined to distinct. identifi- implement et ectiee measures ro reduce high-risk able groups? These questions can be ansu'ered behavior. rhrough representarive surveYs of [he general * Leve'el; ageiU rud' ot Hf mH OV/urr wcVr TiT ;T-c- population and smaller surveys of behavior . lntc:rmarian on the leekls and trends oa in groups [hought to be at high risk. Gather- HIV 3nd other STD intfecrions in the general ing such information can bc sern difficul: se,- population, in subpopularions xsth high- and ual behavior sureys are often controaersial. Ic.n-rislk beha'ior. and in specific geographical and suneys of both sexual and drug-use areas is necessar to monitor the spread of the behas ior face logisEical and accuracy prob- epidemn anld the impact oft public policies. lems The information i, nonetheless crucial I_his informarion is typicallk garheretd through ro estimaringr the future course of the epi- unlinked anonrmv.ous tesling of blood samples demic. [he number of individuals at high risk. froti group;s c mndii:duals-w%omen attending and the most effective approach to reducing antenatal cinmcs. blood donors. STE partinrs. riskv behasior. and in jecring druig users in treactentn pro- U LThe-3 Cos t' J i.t-i7'eI/ 0' .rou ai7 r'e. 'iii/½f'[ 0: gram-as part of an epidenmiological survcdl- Hll- :':1-lt-r Much intointia(in ha. becn lance s-sstnl ,m iDSCAP and others 1tN06. collected oi the inpact ofprogranms on kno%s I- Chin 1+111. Sato I i'ioi. Specialized sunre edge and behavior. but regrertabl% few% srudies can track trends in HIX' ptrsalence among have docuinented rht impact of these behav- subpopuladtions that engagc in high-risk behav- iors on the incidcnce of HIV: (ester srill hatc ior In countricss heer the H[\ -pzdemic iS autnipted rto esrimate Asr-ciectiseness. £- nascent, ls.l and trends in the prevalence of ce ,men ts ofsueffecuvcncsc in ternis of the other STD, can bc incd;catie of patterns ol impact on HIt incidenc, rather thain on be- se\ual lchh ior that still spread HIV. havioral change alone. arc likely to. highlight * P t, e o;llur{'i'*:e/.'-.: l hadi a.:'/ H'e i- - rhe crucial importanc, o treducin risk% behav- te?;-IL" 0(1r'l r ".ie .t ri. Informauirr!on an or among those most likel tuo contract and theprc%alence and nature of ris-k heha lor is spread HI\. Such tudies can be helpful not indispcnsable ht'r estimating the potential sizc onlk n shaping more effkctive policies. but also and rate of sprcad oi an HlV epidemic. and in overcoming political objctions t. suhbidiz- deciding s hen. tvhere. and hots to intervnle. ing safer behavi or among people most likely to \What arc the hieh-risk behat ors in a gisen transmi HIV\ to others. practices raise the risk of infection for everyone, even for those who do not engage in high-risk behavior. As shown at the beginning of this chapter, people who practice risky behavior are likely to take some steps to reduce their risk of contracting HIV because of the enormous personal costs if they become infected. However, because they react primarily to the increased risk to them- 134 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS selves, their response will not reflect the full costs-that is, the negative externalities-of unintentionally spreading HIV to the rest of society. Many people who do not practice high-risk behavior would be willing to pay those who do practice it to adopt safer behavior in order to contain the epidemic and thus lower everyone's risk of contracting HIV. Gov- ernment is the only mechanism through which this can happen. Thus, there is a strong justification for government to subsidize safer behavior among people who engage in high-risk behavior in order to protect everyone-either by preventing primary infections among those who engage in high-risk behavior or, when such individuals are already in- fected, by preventing the infection from spreading to others. The principles of public economics and epidemiology are therefore in agreement that governments should give high priority to the prevention of infection among people most likely to contract and spread HIV. The extent of negative externalities of risky behavior can be measured by the number of secondary infections that a person would generate if infected with HIV By this definition, the riskier an individual's behavior, the more secondary infections he or she is likely to generate, and thus the greater the negative externalities of his or her behavior. The reproductive rate of HIV among such individuals is likely to be much greater than 1, sustaining the epidemic. On the other hand, people who practice low- risk behavior-such as abstinence from sex or injecting drug use, sexual monogamy, consistent condom use, or injecting with sterile equip- ment-will generate few, if any secondary infections. They are vulnera- ble to the negative externalities generated by others' risky behavior. Some readers may ask: Shouldn't government devote at least as many resources to preventing the spread of HIV among people who do not engage in high-risk behavior but who may nonetheless contract the virus? After all, such individuals usually constitute the vast majority of the pop- ulation. In countries where government has a broad mandate to provide health services to everyone, and where resources are adequate, expanding government-financed preventive interventions to the low-risk population should not be ruled out. However, even in these cases, governments should first ensure that programs achieve adequate coverage of those most likely to contract and spread HIV, starting from those with the highest- risk behavior, since this is the most efficient way to protect everyone. There are situations, however, in which interventions for the entire population are essential to the success of programs aimed at reaching those most likely to contract and spread HIV. For example, although '35 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC information that helps people with low-risk behavior to further reduce their risk will have very little impact on the overall epidemic, they need to understand how HIV is transmitted and-more important-how it is not transmitted. Inaccurate information or even accurate information presented in ways that create feelings of fear and vulnerability can incite discrimination against those who are at high risk of contracting and spreading HIV and people who are HIV-positive or living with AIDS (Allard 1989). This is not only unwarranted and unjust, but it may also compromise the ability to launch the prevention programs most likely to slow the spread of HIV, as well as efforts to mitigate the epidemic dis- cussed in chapter 4. Information that fuels fear of HIV-infected people has been a serious problem in nearly all countries where there has been extensive early publicity about the epidemic. To avoid this problem, gov- ernments should make certain that the general population understands that HIV cannot be caught from a handshake or other casual contact, and take steps to reduce stigma and protect those at risk of HIV from discrimination. Helping poor people to avoid HIV. Government's first step should be to lower the costs of safe behavior for the poor by improving the functioning of markets, for example, by eliminating tariffs on condoms and restrictions on condom advertising. If prevention is still too costly for the poor, subsidies may also be warranted. In countries where par- ticular groups face significant barriers to acquiring information- because of illiteracy, different languages, or lack of access to newspa- pers, radio, or television-subsidized information for the disadvantaged groups will improve their access to prevention. Besides subsidized information about HIV, the most important preventive measures are those that make it easier for poor people to obtain condoms, treatment for STDs, and access to safe blood. However, unless there is substantial overlap between the poor and those with high-risk behavior, subsidized HIV interventions for the poor will address equity but will not be suffi- cient to contain the epidemic, particularly in its early stages. Cost-Effectiveness from a Government Perspective Governments, like all entities that must operate within a fixed budget, seek "value for money" in deciding between alternative expenditures. Cost-effectiveness analysis is a tool for deciding among alternative courses of action when resources are scarce. It seeks to determine how to I36 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS achieve the maximum effect within a given budget or, viewed differently, how to achieve a desired effect at the lowest possible cost. For HIV pre- vention and other health interventions, the effects or benefits per dollar spent can be calculated without regard to who pays for the interven- tion or who benefits. However, some health interventions that are cost- effective by this measure would be undertaken and financed by individ- uals themselves even without government involvement. (For example, many individuals will seek counseling, use condoms, or use sterile in- jecting equipment to avoid contracting and spreading HIV, whether or not these services are subsidized by the government.) Measures for which individuals would have paid on their own are not a priority for a gov- ernment's limited funds. Rather, governments should first spend their limited resources on interventions that are cost-effective in the conven- tional sense and that would not occur in the absence of government in- volvement-that is, cost-effective interventions that provide public goods and reduce the negative externalities of high-risk behavior. This strategy will ultimately prevent the largest number of subsequent infec- tions among all subgroups in the population, including the majority who do not engage in high-risk behavior. Some governments have the mandate to intervene broadly in the health sector and to assume responsibility for providing curative and pre- ventive health care for all citizens, irrespective of whether government spending is efficiency-enhancing. Would this mandate change the prior- ities discussed above? No, it would not. Unless the costs of changing behavior among those most likely to contract and spread HIV are extra- ordinarily high, promoting behavior change among them is likely to be the most cost-effective in terms of preventing infection in the general population. Several factors that affect the cost-effectiveness of alternative publicly financed HIV/AIDS prevention programs do not necessarily enter into cost-effectiveness calculations for private individuals. There are also a number of issues specific to the cost-effectiveness of HIV interventions that need to be taken into account. Public benefits include averted secondary infections. In calculat- ing the benefits from prevention programs, it is important to consider not only the benefits received by the person directly affected, but also the number of secondary infections that are averted. Secondary infec- tions measure, in effect, the extent to which negative externalities are being addressed by programs. Failure to include secondary infections I37 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC among the benefits of public health programs will result in serious underestimates of the benefits of prevention in groups that are very likely to contract and spread HIV, since the number of people in these groups tends to be small relative to the total population but the num- ber of secondary infections per person tends to be large. Public prevention should augmnent, not replace, private preven- tion efforts. Public programs will be more cost-effective if they can raise additional demand for risk-reducing behavior without "crowding out" private sources of inputs. For example, in societies where condoms are already widely available at low cost, subsidized condom projects that are not specifically directed toward those most likely to contract and spread HIV may simply shift the financing of existing condom sales from private pockets to public ones, with no impact on the epi- demic. Whether the resulting redistribution enhances equity depends on the income distribution of condom purchasers relative to the gen- eral population. As we saw earlier in this chapter, high-income sex workers and men with higher education are more likely to use con- doms for casual and commercial sex. Thus, untargeted condom subsi- dies are likely to help those who are relatively better off rather than enhance equity, and are likely to substitute for condoms that were already being purchased privately rather than generate additional demand. The same argument holds for the potential substitutability of public for private services for STD treatment and safe blood. The effec- tiveness of public programs should therefore be measured as the differ- ence between outcomes with and without the program. HIV/AIDS prevention has other external effects. Many HIV/ AIDS preventive interventions have additional positive effects that might be overlooked if cost-effectiveness calculations considered only the impact on HIV/AIDS. For example, condom programs also prevent other STDs, which are spread in the same ways as HIV and generate similar negative externalities. HIV education in the schools and encour- agement of condom use among adolescents who engage in sex will complement efforts to postpone sexual activity, reduce teenage preg- nancy and abortion, and as a result improve school completion rates among girls. Harm reduction programs among injecting drug users generate demand for drug rehabilitation programs and reduce transmis- sion of hepatitis B and C and other blood-borne diseases. Although these benefits are difficult to quantify, they should not be overlooked. Involving other actors can improve cost-effectiveness. While gov- ernments must support prevention efforts among those who are most 138 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS likely to contract and spread HIV, governments are not necessarily the most effective agent to design and implement such programs. Enlisting and subsidizing NGOs to help in the design and implementation of interventions can greatly improve the cost-effectiveness of public pro- grams, particularly if the NGOs are staffed by or representative of peo- ple at high risk of contracting and spreading HIV The cost-effectiveness of all interventions can also be improved through other government measures, such as relaxing legal prohibitions and minimizing the stigma that people in high-risk subpopulations often face, so that NGOs can work more effectively. Public-private cooperation in confronting AIDS is discussed at greater length in chapter 5. Which Interventions Are Cost-Effective? Only a few HIV/AIDS interventions have been rigorously evaluated with respect to their impact on the incidence and prevalence of HIV; among those evaluated, interventions targeted to those who practice high-risk behavior tend to be more effective (Aral and Peterman 1996; Choi and Coates 1994; National Research Council 1996; Oakley, Fullerton, and Holland 1995). Appendix A of this report presents the results of 22 of the more rigorous evaluations of HIV/AIDS interven- tions that have taken place in developing countries. Unfortunately, in- formation on the costs of these interventions is usually unavailable; their cost-effectiveness has rarely been evaluated. The impact of alternative prevention strategies in four epi- demics. The effectiveness of alternative interventions will be strongly influenced by the nature of the intervention itself and by the hetero- geneity of the behavior that is fueling the epidemic. To illustrate this point, Van Vliet and others (1997) have simulated the impact of increased condom use and increased treatment of curable STDs (chlamydia, gonorrhea, and syphilis), on a heterosexual HIV epidemic in the four hypothetical populations described in chapter 2 using the STDSIM simulation package.16 The simulations show the impact of increased condom use and STD treatment in various groups in each of the four populations fifteen years after the start of the epidemic. (As seen in chapter 2, in each population earlier intervention among those most likely to contract and spread HIV would be more effective than the later behavior change discussed here.) The simulations show the impact of increased condom use in three groups of people with different rates of partner change-female sex 139 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC workers, men with casual or commercial partners, and women in stable relationships. These groups are commonly the focus, respectively, of out- reach programs to sex workers, socially marketed condom programs, and reproductive health services. In these simulations, the term " sex worker" refers to women with the highest rate of partner change-i 0 new part- ners per week, or more than 500 partners per year. In the real world, of course, some women who have very high rates of partner change do not regard themselves as "sex workers" and they may contact their male partners in a variety of settings. There are other men and women in these imaginary populations with large numbers of partners, but fewer than 500 per year. The impact of different interventions in all of these groups can be simulated, individually and simultaneously. However, for expositional purposes, we present simulated interventions with only three groups. In the baseline scenario, before any intervention, we assume that only 20 percent of sex workers and 5 percent of men having sex with ca- sual or commercial partners use condoms consistently, that is, in every act of intercourse. We also assume that none of the women in steady re- lationships are using condoms.17 The simulations show the impact of instantaneously raising consistent condom use among sex workers to 90 percent, and among the other two groups to 20 percent. 18 These levels were selected because the authors believed they are realistically achiev- able in some countries. In other countries, it may be possible to ex- ceed the levels of condom use simulated here. Among those who use condoms, the failure rate through breakage or misuse is assumed to be 5 percent. The simulations also show the impact of increased STD treatment on HIV prevalence. The baseline scenario assumes that 25 percent of all STD cases that produce symptoms are effectively treated and that there is no specific screening or treatment program for sex workers. The sim- ulations show the impact of increasing the share of symptomatic STDs treated to 75 percent in the general population and, in a separate sce- nario, the impact of implementing a monthly screening and treatment program that covers 90 percent of sex workers. In the latter intervention, it is assumed that 5 percent of sex workers are not cured. The assump- tions behind the baseline scenario and the five condom and STD inter- ventions are summarized in table 3.4. The simulated impact of increased condom use and STD treatment on the HIV prevalence of adults in the four hypothetical populations is I40 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Table 3.4 Summary of the Assumptions before and after Interventions, STDSIM Modeling (percent) .Afier Ass5umption Btselinie interentio,, L .'Flg a(ll"dVm,1 ~vtokimr Sex %orkcr' 2") 90 Nien a %h casud or .omnercial parinvr5 20 Wonmen ages 15-5( in , able relanion4hips i SYmptemark5i> rtDrle'l '-c, Se.- evorkenr l; u-ith.w vildi n Pit,l=:e h lng v IrWtP)h fit tfr S TLi' 90 Sb'e'{r. A1a ; ire/j'.q.w. Ylin \.Ia[ :mnd oLher! 1-1- shown in figure 3.3. Despite the different underlying patterns of sexual behavior, the impact of specific interventions shows some striking con- sistencies across populations: * Achieving 90 percent condom use among sex workers results in a dramatic drop in HIV prevalence in all three populations where there is commercial sex (a, b, d), even though sex workers comprise only a very small share of each population (0.25 percent or fewer women). STD screening and treatment for sex workers is far less powerful. * Increased STD treatment among the general population is less effective than raising condom use among those with many part- ners. This is not surprising, since people with high-risk behavior generate a disproportionate number of STD cases, and condoms prevent transmission of both HIV and STDs. STD treatment among the general population and condom use by women in sta- ble relationships have the largest impact in the populations with concurrent casual sex (b, c). * The impact of greater condom use among women in stable rela- tionships is very small, and in the serial monogamy population (d) it has almost no impact on the epidemic. In the population where the epidemic is driven by commercial sex (a), condom use by monogamous women slightly accelerates a decline in HIV preva- lence, while in the other two populations it merely slows the growth of a still-expanding epidemic. I41 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 3.3 The Impact of Changes in Condom Use and STh Treabnent in Four Populations with Different Patterns of Sexual Behavior (a) Commercial sex only (b) Commercial and casual sex Adult HIV prevalence (%) Adult HIV prevalence (%) - ne ~~ ~ ~~20 30 - Baseline30_ 90% STD treatment, sex workers 25 75% symptomatic STD 15 treatment, general 20 population . . 90% condom use, z o 5s sex workers 20% condom use, men in commercial and I0 casual sex s 20% condom use, * 5 women in stable relationships o .... . . ...... o 0 5 10 15 20 25 30 0 5 10 15 20 25 30 Years since the beginning of the epidemic Years since the beginning of the epidemic (c) Casual sex only (d) Serial monogamy Adult HIV prevalence (%) Adult HIV prevalence (%) 30 20 25 ____ 20/ -5 ___ __ 20_/_ Raising condom use among sex is lo workers to 90 percent was the most eHfective way to reduce lo ___- HIV prevalence in three of the / four simulated populations; in 5 the remaining simulated popula- tion there is no commercial sex. .......... 0 5 10 1 5 20 25 30 0 5 10 15 20 25 30 Years since the beglnning of the epidemic Years since the beginning of the epidemic Source: Backgroundpaper, Van Vliet and others 1997. Looking at the ranking of interventions within specific populations, we see that: * In the epidemic fueled by commercial sex alone (a), all of the in- terventions produce an absolute decline in HIV prevalence; in the epidemic fueled solely by casual sex (c), none of the simulated in- terventions is sufficient to cause an absolute decline. 142 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS * In the population with concurrent commercial and casual sex (b), prevalence declines in response to 90 percent condom use among sex workers; 20 percent condom use by men in casual sex keeps prevalence from rising. Other interventions in this population merely slow the expanding epidemic. * In the population practicing serial monogamy (d), increased con- dom use by sex workers is the only intervention that results in an absolute decline in HIV prevalence. In reality, no intervention is an "either/or" proposition. There is al- ways spillover between interventions intended for people with different rates of partner change, so any particular intervention is likely to change behavior in more than one group of people but to differing degrees. Fur- ther, combined interventions to multiple groups will have greater impact than single interventions; for example, working only with sex workers to raise condom use will be less effective than simultaneously working with both sex workers and their clients. However, these simulations show that the greatest share of the impact will be achieved through the interven- tions that succeed in changing the behavior of those with the highest rates of partner change. The broad conclusion that can be drawn from these simulations is that, although the overall pattern of sexual behavior in the population does affect the impact of interventions, prevention of infection among those with the highest rates of partner change has a large effect irrespec- tive of the underlying patterns of sexual behavior in the population. As- suming that it is not a great deal more expensive to increase rates of con- dom use among those with high rates of partner change-sex workers or others-focusing condom subsidies and promotion efforts on changing their behavior is likely to be highly cost-effective. Studies of the cost-effectiveness of HIV interventions in develop- ing countries are rare and not transferable. Fewer than half a dozen studies have documented the costs and effects of preventive interven- tions in developing countries (Beal, Bontinck, and Fransen 1992; Gilson and others 1996; Moses and others 1991). An overview of the results of several cost-effectiveness studies in developing countries is in appendix B of this report. Most evaluation studies measure impact by changes in in- termediate behaviors that are believed to affect risk-such as the increase in condom use, or knowledge about HIV prevention, or the number of people receiving sterile syringes. The number of HIV infections averted '43 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC is then extrapolated on the basis of assumptions about the relationship between the behavior and HIV incidence. However, the lack of good data on sexual behavior and on the relation between sexual behavior and incidence makes these benefits very difficult to assess. Virtually no stud- ies, except those based on simulations, have measured the impact of in- terventions on secondary cases where they are thought to exist (Over and Piot 1996; Stover and Way 1995; backgroundpaper, Van Vliet and oth- ers 1997). None to our knowledge have taken into account the external benefits of interventions or the issue of complementarities between interventions. While cost-effectiveness studies can be very useful in deciding among alternative interventions in a given setting and stage of the epidemic, their conclusions are usually not easily transferable to other settings (backgroundpaper, Mills and Watts 1996). For example, an evaluation of the effectiveness of enhanced treatment of symptomatic STDs in reduc- ing HIV incidence in rural Mwanza Region of Tanzania found that the intervention lowered HIV incidence by 42 percent at a cost of roughly $10 per person treated, or $234 per primary HIV infection averted (Gilson and others 1996; Richard Hayes, personal communication).19 However, treatment costs clearly could be much higher in a middle- income country, and effectiveness may have been quite different in an area with lower HIV prevalence than the Tanzanian study site (4 percent of adults were infected).20 Moreover, without estimates of the costs and impact of alternative interventions in the same area, we cannot say whether a particular intervention is more or less cost-effective than other interventions in reducing HIV transmission. Ideally, we would like to know the costs and effects of alternative interventions implemented in the same setting, but this has rarely been done (box 3.9). Interventions focused on those most likely to contract and spread HIV should be more cost-effective from the public perspective, because preventing infection in a person with risky behavior prevents many sec- ondary infections among individuals with whom they mix-some of whom practice high-risk behavior and some of whom practice lower-risk behavior. In fact, the extent to which ongoing programs affect those who practice high-risk behavior is often unknown. For example, condom so- cial marketing programs improve the access of the poor to condoms, but it is still not known to what extent these condoms are used by those in the highest-risk groups. Information on the level, distribution, and type of high-risk behavior; the number of people involved; and their charac- I44 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.9 The Cost-Effectiveness of Prevention among Those with the Highest Risk HOX \ 1-',N'I Hl I - NFE TI-NS t-C0LD BF \ F RTEED artending i ST) cltnics in Calikornni lovs risk'. \Ven h spending an addittonal .%1 million per y,ar omn los risk includes mostl of the general populatiorn. presenrion in, gruups is!th different Iexcl5 of risk of includtng %iomen of childbearing age in -1 of the HI\ infection? The answers generated in a recent eountr\ 50) states. srud. in the Uinited Statce demonstrate the high The results of the exerctic depend on the ac- co,[-e[tectisene,s ot focusing pri-vention expendirLire sumpliorin %isth respect to the cos;t and impact ut on chosc moq likely to Lontract and sprtad HI\. as differenr intcrcnti-oni. ThK author inittall assumed well as the addi[ional benefits from carlk intersen- th3t prevention costs tor onc individual in an gis en Eion tKahn 199o'. group s%ould be $201) per .ear then t:amined the The study detincd tour risk groups in terms c.f cnsitik iN of Ehc results to thi, assunmpcion. This [he level ot Mteadv -scate HIV pres.alnce they- s'ocd $200-per-year figure 1. L' based on a urticY of the attain ishithout intervention-high risk is(0 percent annual c)[ ot o riou pre'ention inter'-Lcntons tOr prevalencel: medium risk 1 IS percen[t: lowJ risk high- and low-risk groups in the LUnited Stts.' In (I percenti and very loss risk m.I perc1ntv. teadv- terms ot the impact of inkrentons. the auhour state HIV pres alence is detiecd aS thc point[ it assumed that intcr%cnuons lov.cr risk bx i) perc;nE. which the nunmber of ne'% infection, e nxactl matches which he believes is a consera[ivc estmrare. the nuniber oF peo)ple cxiting the group through Box [able 3.L shis rhe numtber -f HI\ infec- dearh or eliminating the risk Factor Itor example. tion, preOenred. gi%en these assumniptons. b% spend- stopping injecting drug use). Examples of ihese ing SI million on each of the se era] groups: either groups in Lhe Linictd States include %oung homosec- iitlh late inteFention iafter srtad; -state prevalnse ual nien in San Francisco i high riski. injecting drug has been reachedi or isith eirls interF-ntion (beFore usrc-s in 5an Francisco Imediumn risk and %vomen Ibox .-oor "U civ to/lon ic p Box Table 3.9 HIV Infections Averted by SI Million Annual Spending on Prevention, U.S. Estimates Baseli/flu Hfi - 5qewr 20-veat Rskg-ouip prevalence - mo) horzon lhorizoni Hop-1. rj; . 1i1C1d Lare Si) 164 i1x I Pre-'tcadv 'tate a) 91 M'eady 'ra[e I V .d' Pr-c-scadv 'rare 3 I 112 Stead% sEjEc 4 2 2 Pre-iicadl% ac stt ' 2 - I ;1 j,'r,t',: S'eddy s,[tar il I 1 ie I) , K r, ,. Ii4hn5 145 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 3.9 (continued) steady-state prevalence has been reached . The averted for the low-risk steadv stare rises EC, only IS impact of each intervention is shown tbr a five-Year and 93. respectivels.. Even it prev,ention s substan- rime horizon and for a 20-vear time horizon. One tiallv more successful at changing behavior in [he million dollars prevents the most cases if focused on low-risk groups. che higher etifcti%eness of pre%en- [he highest-risk group in the stead)' state. However. [ion in high-risk groups remains. rhe bcenefits of intervening early wvith this group only \WUhile mncrventions in high-risk groupw are more become evident in the 20-year time horizon. etfective, they are also potentially more costl-. These tigures anderstate the impact of preven- However. the study ntimt[ed [hat interventions in tion in high-risk groups. however. becau%e sec- the low-risk groups isteady state) .tould have to ondary infections prevented in the par[ners or chil- be roughl one-fortieth to one two-hundredth drenl of people in high-risk groups are not included. ( 1/-I4-l2i_00) the cost of' an intervent[ion in high- The number of infections averted in low-risk groups risk groups (s[eady statel to prevent the same num- w%ill not be much affected by this omission. but ber of infecions averied by intenening in high-risk among high-risk groups the total of averied infec- groups in the steadY irate. In other %cords. in the rions might be seseral times higher, depending on high-risk steady state. interccntions in the low,--risk the group and the degree of sexual mixing w-ith population would have to cost $1 to 55 per person lower-risk groups. per year. compared with $200 per person per year in The result-thar prevention in high-risk groutcps the high-risk population. to prevent an equikalent is most cost-effective-is robust to large changes in numnber of HIV infections for a $1 million outlay. the assumptions about the effectiseness of'the inter- vention. If programs reduce risky beha% ior bs 5( 'The programs and covr per purson pr cu inluded annual percent instead of 10 percent. the number of infec- rering and counicling i-il.-I Im. h!ecbh d;rrib.uron and ou- tions averted in the high-risk steady state rises to rca]- i$6Jii 3-se,is,n grnup conindng or itSLs 15v. fle,die - ~~~~c'.ch engc A -nI-S0i' pe-i '.crl3hp- fur high-r-isk Cas me,n 830 for the Five-year simulation and 3.750 for the S -kJ '; Son c;un;e1in Kr n-'% -P ri-, womrien I $?'i1 20-s-ear simulation. while the number of infecrion. t,.or.-ourling t;m ncdiumr-r.Lk gii MenIi -1-o teristics is a basic public good and will enhance efforts to improve cost- effectiveness by helping to improve the targeting of programs. Cost-effectiveness and the accessibility of target populations. Although it is highly desirable to focus public interventions on those who are most likely to contract and spread HIV, identifying and reach- ing these individuals can be difficult, especially where legal sanctions and social stigma cause these people to want to avoid being discovered. The costs of reaching those most likely to contract and spread the virus can have a significant impact on the cost-effectiveness of interventions. Figure 3.4 shows a stylized classification of groups of people accord- ing to the extent to which they practice high-risk behavior and their pre- 146 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Figure 3.4 Classification of Groups by Riskiness of Their Behavior and Their Accessibility Hgher-risk behavior Sex workers in brothels Streetwalkers IDUs in treatment programs Out-of-treatment IDUs Prisoners Homosexual/bisexual men Military, police, sailors Street children STD patients Truck drivers, bar workers More difcu Eas er access acces Government employees Elderly in rural areas Employees of large firms Women attending antenatal clinics Children in school Lwr-risk behavior IDU Injecting drug user Source: Adapted from Adler and others 1996, figure 8. Used by permission. sumed accessibility. Of course, the extent to which members of these Prevention programs should focus on identifiable groups practice risky behavior varies considerably across- people most likely to contract and spread HIV; some of these individuals settings and with the effectiveness of prior prevention efforts. Thus, such are easy to identify and reach, others a figure would need to be modified according to the situation in a spe- less so. cific country, on the basis of the results of HIV and behavioral monitor- ing systems. In the upper-right quadrant are groups with higher-risk behavior that are relatively easy for government agencies and collaborating prevention 147 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC partners to reach. The benefits of behavior change in these individuals are relatively high, especially given the large number of secondary infec- tions generated, while the costs of locating them will be relatively low, enhancing cost-effectiveness. In the upper left quadrant are groups who also practice high-risk behavior but who are less easy to reach. In these instances the benefits of behavior change will again be great but the costs of locating and working with these individuals may be high, reducing the net benefits. In the lower-right quadrant are people who, on average, are presumed to practice lower-risk behavior but to whom access is easy. The benefits of intervening in these groups may not be great, but the costs of reaching them may be very low. Inexpensive interventions for these groups may still be cost-effective relative to some alternatives (box 3. 10). The lower left quadrant includes people who are very unlikely to contract and spread HIV and are very difficult and costly to reach; of the four types of groups, this is the lowest priority for public-sector HIV pre- vention efforts. Note that the accessibility of most groups can be im- proved through government actions to reduce stigma, decriminalize be- havior, and educate the public on the nontransmissability of HIV by casual contact and the benefits of working with these groups. Of course, the "groups" identified in figure 3.4 are not homogeneous with respect to their risky behavior. Since individuals with high-risk be- havior cannot be easily identified, programs need to focus interventions on people with characteristics that are highly correlated with risky be- havior. However, some sex workers consistently use condoms, and some government employees who have many partners do not. Intervening to change the behavior of people with specific characteristics like age, sex, occupation, or geographic area is not a perfect way to reach those with high-risk behavior. Some members of these groups will be exposed to in- terventions even though they practice low-risk behavior. Moreover, oth- ers with high-risk behaviors who don't belong to any of these groups will be missed. The lack of perfect criteria for focusing interventions on those with the highest-risk behaviors is one source of leakage of program re- sources. This reduces the cost-effectiveness of interventions if resources go to people with lower-risk behavior. On the other hand, leakage may improve cost-effectiveness if resources go to people with even riskier be- havior than those targeted. Surveys of sexual behavior, such as those con- ducted by WHO/GPA and DHS, can help to overcome these problems by establishing the characteristics and geographic location of those who have unprotected sex and high rates of partner change. Unless programs 148 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.10 Educating Adolescents on HIV/AIDS: A Sound Investment IN COLIN TrRIE, V HERE SEXUA.I Att 1\ IT BE(--INS Such programs arc ormctimes unpopular wxith Jl an earls- age and young people hate high rates of parents \%ho w%orn' that infrrmation on reproductive partner change, promoting safer beha% or among heal[h. STDs. and contraception might cause [heir adolescents is clearly inporEant to slowing the children to bccomc sexualv- actime ar an earlier age. spread of HIV. There are many possible inicr- Research has sho\sn chat this is not rhc case. Re. eis 'entions to address risk bchavior among adolok- ot school-bawcd programs ha%e 6tund that partici- cent,. both in and out of ;chool. Ho\%eter, c%en in pating youth havi- not beguin sexual activiN earlier societies w%here ,e\ual aeciwiry doe; no! generally [Gluck and Rosenrhal 1l9tE, KirbY and others 1110-i. begin until alter young people have completed LANAI DS 199 ). Moreover. a retie\t of school- their schooling. reproductiie health education in based progranis in the Linited States ciLund [hat pro- hth school svstem-\%hich includes intormati.in on gramr,s that included sc\ual health education and the benefits ot postponing sewual a.civ it, as well AIDS pre\cn[ion not only delayed the- start ot sexual as how- to prevent pregnancy. STDs. and HI\ acrti I'. but reduced the number of partners and for those %sho do not abstain-is a potentially raised contraceptive usc among those stho became powerful intervention. Besides prescnting HI\ 'ex\uallv atri\e iGluck and Rosenthal 19NS. among students who mighr otherwise cdopt risky- G ien rhe other broad social benefirs and the rel- behavior. these programs ha' e many other bene- an'\ely lo\% cost of addgi- HI\VAIDDS education to fits. They prevent 5T Ds and associated iniferil- e'istinc program,. HIXVIIDS education i likelv to irt. and they prexent unwa-nted pregnancy,. which be a g-ood in%estmenr in preventing Hi's The oVer- may- lead to abortion or to girls' dropping out of tshelming majority of AIDS program managers school. MIore broadly. reproductive health educa- "ho responded to the A
a '' f; ^ -X ^ s ^ p -i s -i ' 0 {ii ^ CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC cents as the subpopulation most likely to be the recipient of an interven- tion, financed either by government or the private sector (figure 3.7). All countries had at least one program focused on youth, even though in many of the countries it is not clear to what extent adolescents engage in risky behavior. About nine out of ten countries reported a public or pri- vate program focused on sex workers, while about seven out of ten coun- tries had a program focused on injecting drug users; a slightly lower pro- portion of countries had programs that focus on the military and on men who have sex with men. However, respondents estimated that these pro- grams on average covered only about one-half of the relevant group with high-risk behavior. Coverage was highest for adolescents and the military, and lowest for men who have sex with men and injecting drug users. In most of the countries assessed by The UNAIDS Country Programme Advisers also reported that gov- UNAIDS Country Programme Advisers, ernments were least likely to finance and most likely to impede preven- prevention programs did not reach the majorit of people most likely to con- tion programs targeted to men who have sex with men and injecting tract and spread HIV. drug users (figure 3.8). Although six out of ten governments funded Figure 3.7 Coverage of Subpopulations with High-Risk Behavior, Estimates of UNAIDS Country Programme Advisers in 32 Countries Countries (% of total) 70 No program 60-0-10% - 50 .[................. ...................................................... - 11-50% 30 . ..... . . .50% - 20 ....... ....-.- ....... .. .. . % ------ 1. . . . . . . . - - - - 30 lb . r-} 20 * -. __= 0 M5M IDU Military Clients of SW Sex workers Adolescents (n17) (n= 18) (n= 20) (n 32) (n =32) (n 24) n Number of countries responding. MSM Men who have sex with men. IDU Injecting drug users. SW Sex workers. Sosmre: Authors' calculations, based on Country Programme Adviser survey results. i6z EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Figure 3.8 Government Support for Prevention Targeted to Groups with High-Risk Behavior, Estimates of UNAIDS Countfy Progamme Advisers in 32 Countries Countries (% of total) 60 60 1 ; Funds- 50 1. . . . . . . . . . . . . . . . . . . . . . . . . . . . . Encourages * 40 - Impedes 30 * .**. 20 ... .. - - . . . 1 .. .. -. E . 20 _ . . .... ... ...... _ _.......... . ...... ......~-- s 10 ...... - .. ..... ......- __ L El L~ 1- I I MSM IDU Military Clients of SW Sex workers Adolescents (n= 17) (n=8S) (n= 20) (n= 32) (n=32 (n= 24) n Number of countries responding. MSM Men who have sex with men. IDU Injecting drug users. SW Sex workers. Source: Authors' calculations, based on Country Programme Adviser survey results. prevention programs for adolescents, only about one-third did so for the Most of the countries assessed did military and for sex workers. Two advisers indicated that the government not fund prevention pmgrams focused on those most likely to contract and in their country promoted prevention for the general population of het- spread HIV; they were most likely to erosexuals, but neither encouraged nor discouraged programs for those impede programs targeted to injecting most likely to contract and spread the virus. drug users and men who have sex . . . . ~~~~~~~~~~~~~with men. To summarize, while some prevention programs have attempted to encourage safer behavior among those most likely to contract and spread HIV, coverage is generally low. The fact that governments may have logistic and political difficulties in reaching groups such as sex workers and injecting drug users, while understandable, does not detract from the urgent need to assure the fullest possible coverage of these groups. Often these obstacles can be overcome through government funding and support of NGOs. Moreover, in many countries coverage is low even among "captive" populations, such as the military; where this is the case, governments have an opportunity to inexpensively reach these groups with information and other prevention interventions. Effective interven- tions with broad coverage of those with high-risk behavior will go a long I63 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC way toward preventing infection among others engaging in high-risk activities and among the lower-risk population. Improving the Equiy of Prevention Programs: Expanding Condom Use The effectiveness of government programs in ensuring access to pre- vention for the poor has rarely been evaluated. However, improving equity of access to condoms is one of the major objectives of condom so- cial marketing programs and free government distribution of condoms. Have they improved equity? Condom availability and use in general have expanded considerably, partly as a spontaneous response to HIV and partly as a result of social marketing and other public, private, and donor-sponsored programs. As of 1996, 60 developing countries had functioning condom social mar- keting programs, although not all were on a national scale. This was twice the number in 1991.25 Many of these programs are supported by international donors through three major contractors-DKT Interna- tional, Population Services International (PSI), and Social Marketing for Change (SOMARC); others, for example, in Botswana, India, South Africa, and some Latin American countries, are also subsidized by na- tional governments. In some countries, such as Indonesia, condom brands launched through social marketing have been taken up by for- profit distributors. Social marketing aside, nearly three-quarters of the 70 countries that responded to condom distribution questions in the AIDS in the World I survey provide condoms through a national AIDS control program (Mann and Tarantola 1996). The likelihood of having a condom social marketing program is more strongly related to the increased spread of HIV/AIDS than is government condom distribution (table 3.6). This is partly because government condom distribution includes distribution through government family planning clinics and health services. Finally, in many countries, such as Brazil, Thailand, and Vietnam, unsubsidized commercial sales have risen. However, the extent to which these programs disproportionately help poor people to obtain condoms is not clear. As we have seen earlier in this chapter, in most countries, people with higher incomes and educa- tion are more likely to use condoms. Providing subsidized condoms to low-risk individuals who would have purchased them at market prices would neither improve equity nor reduce the epidemic. Likewise, while I64 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Table 3.6 Condom Social Marketing and Government Condom Distribution Programs, by Stage of the Epidemic Perientoige otcoun rries w'rth Stage CSMtf Conidomz of'rhe progran1s, dki5nbution bi epidenic 1.9.96 ,V,1CP, 1992 Na,cent 31 -I Concenir3ccd e9 GenerahWzed 0)I 1(Il Unknown 1 3 Tor.a, i o 4') Number of counUriev 123 CSNI Corndm ,ociul maiketing NACP NaikniJ AID.S conLrol prc.grarm So -rr S.taris cal ippendix. [abk ' condom use has risen in both the subsidized and commercial market, it is still not clear the extent to which subsidized programs have squeezed out private sales. This is likely to be an important issue in older condom social marketing programs, after their initial effect on popularizing con- doms and generating greater demand has worn off. A second way in which these programs promote equity is by encour- aging condom use among those most likely to contract and spread HIV, forestalling or slowing the epidemic before it reaches the poor. Unfortu- nately, relatively little is known about the extent to which condom social marketing programs are used by those with the highest rates of partner change-which is the key to their effectiveness in slowing the epidemic. Surveys of sexually active adults confirm that people are far more likely to use a condom for sex with a casual or extramarital partner than a steady partner or a spouse (Agha 1997, Coleman and others 1996, Lowenthal and others 1995, Tchupo and others 1996). But they do not show whether these programs reach people with the highest rates of part- ner change. Do these programs lower the costs of condom use suffi- ciently to bring about high-use rates among sex workers, soldiers, truck drivers, and other people with many partners? By selling through non- traditional outlets like bars and hotels, condom social marketing pro- grams are probably much more likely to reach people with risky behav- ior than are conventional programs that distribute condoms through health clinics. If the majority of individuals with the highest rates of i65 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC partner change are reached through these programs, there may be sub- stantial cost-saving if the programs involve a lower subsidy per condom than free distribution. Furthermore, such programs may avoid the polit- ical controversy and possible stigmatization that may arise with more tar- geted programs. Additional research on the sexual behavior and the economic status of those who use subsidized condoms, and the extent to which those with the highest rates of partner change use condoms from these programs, will help greatly to demonstrate and improve their cost-effectiveness.26 That said, many countries still lack vigorous condom programs that specifically prevent HIV and other STDs. Many condom social market- ing programs, for example, in Bangladesh, Colombia, Costa Rica, Pak- istan, and Sri Lanka, as well as recently launched Chinese programs in Yunnan Province and Shanghai, are oriented primarily toward family planning, with little if any marketing for STD and HIV prevention (DKT International 1997; Kang 1995; "Signs of Change. . ." 1996; "Sri Lankan Condom Sales . . ." 1996). Even in some African countries with concentrated or generalized HIV/AIDS epidemics-for example, Mali, Niger, and Senegal-family planning and reproductive health are the main themes of the programs. Depending on the country, such themes may be less controversial than HIV and STD prevention. However, they may also fail to reach those with the highest rates of partner change. For example, sex workers and young, sexually active men do not frequent health or family planning clinics. Furthermore, women who need con- doms for STD prevention may be reluctant to obtain them from com- munity health or family planning clinics, even if free of charge, because of inconvenience, an unreliable supply, or a desire for anonymity. These problems can be overcome if condoms are promoted specifically for HIV and STD prevention and if they can be readily and cheaply obtained in nontraditional oudets readily accessible to people in situations that tend to be conducive to casual and commercial sex. Such locations include pharmacies, kiosks in red-light districts, bars, nightclubs, hotels, truck stops, and military bases (box 3.13). In Peru, socially marketed condoms are sold in three-quarters of the pharmacies and in strategically placed vending machines (Futures Group International 1 995a). One way for governments to stimulate demand for condoms for disease prevention is to end restrictions on condom advertising. But even when there are no legal barriers, open promotion of condom use can be ex- tremely controversial if it is seen as encouraging promiscuity. Messages i66 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Box 3.13 Preventing HIV on the Road to Ho Chi Minh Ciy \ IETNANIFsE rRI v K DRIJ\ERS HA\E A SAN INGt undcrstand wvhat the disease is and how% to avoid it. Never. ever. evcr hit a child along rhe road because, Hoiiever. it was onlv after [)KTF International, a wvell, he might be Yours. It's no iecret that w%ith all U.S.-based condom social marketing company be- those days and nights on the road and all that cime gan promoting the sale of Totsr and OA'conndoms awav from home. truckers seek out amorous compans. in Vietnam chat trucker% had ready access to reliable Ho Duc Cu is marter-of-fact abour thr issue. t is protection. near sunset. and he is sitting at a noodle shop at the Cu Finishes his tea 3nd w%alks over to his Russian- Goods TransporTation Company truck stop on the made truck. Inside the driser-side door is a pouch outskirts of Hanoi. drinksing tea t'rom a ceramic bowl containir, a handful of OA'condoms. I'm gone and getting ready to haul 10 tons of tractor equip- from my wife and kids 26 dasvs out oF everv mornth.' ment south to Ho Chi Mlinh Cit- (formerly SaugonL Cu says %%irh a tfiinc grin. He adds that not onlv are "It's a three-and-a-half day drive from here to O0'condoms dependable. but vou can buy rheni at Saigon.' Cu sass. For a lot of drivers, that nicans locations along most roads throughout \'ietnarn. rwo to three wcnmen along the iay V" The Vietnamese gomernmenc has N%idely% publi- cized the risks of Hl\/AIDSL so most truck drivers v;, --. DtKT lrnrnAEsonal I n.d 1 l' -d 1i, perni,.4erl. must convey useful information and at the same time be directed to the appropriate populations so as to avoid offending influential leaders and segments of the public. Religious leaders in particular may have strong negative reactions to condom promotion if they are not informed about the benefits of condom use or if they are confronted with messages they find offensive. In Uganda, religious sensitivities led to an unofficial ban on promoting condoms on television and radio from 1991 to 1995 (Buwembo 1995). In the Philippines, with a nascent epidemic, opposition by the Catholic Church to artificial contraception extends to condoms for prevention of HIV and STDs (SOMARC 1996). In Niger, conservative religious groups defaced billboards advertising the SOMARC-sponsored social marketing of condoms (Futures Group International 1995b). Sponsors of social marketing of condoms and other birth control methods have nevertheless found ways of generating support, even among critics. For example, they work closely with religious leaders, po- tential critics, and local spokespeople before launching condom promo- tion campaigns to explain the many advantages of condom use (preven- tion of HIV and other STDs; infertility arising from STDs; unwanted pregnancy, abortion, and withdrawal of pregnant teenagers from sec- ondary school; and the promotion of child spacing, which reduces child I67 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC and maternal mortality). They test promotional messages with their intended audience and with potential critics to avoid giving offense and maintain a low profile until a basis for success is firmly established. Con- doms marketed under names like Trust, Protector, OK, and Couples' Choice encourage the view that condom use is safe, modern, and socially responsible, irrespective of whether they are used for family planning or disease prevention. More explicit messages about condoms will some- times be more readily accepted among the people who need condoms the most. This chapter has provided evidence that people will adopt safer be- havior, particularly people at high risk of contracting and spreading HIV, and that governments have many ways, direct and indirect, to influence individual behavior. It has identified prevention activities in which gov- ernments have a unique role, since private individuals will not finance them sufficiently, and it has outlined important considerations in deter- mining the cost-effectiveness of public spending on HIV/AIDS preven- tion. The chapter highlighted two areas in which most governments can greatly improve the effectiveness of their efforts to prevent HIV, given sufficient political commitment. The first is to increase the amount and quality of information collected concerning the nature and extent of risky sexual and injecting behavior in the population, trends in the inci- dence and prevalence of HIV, and the costs and effects of alternative pre- ventive interventions in the local context. The second is to use this in- formation to ensure that prevention programs result in safer behavior among the subpopulations that are most likely to contract and spread HIV and to ensure access to prevention among the poor. Neither of these issues is easy to resolve; however, both are easier to tackle than the very difficult decisions that are thrust upon governments in countries with widespread epidemics. These are the topics of our next chapter. i68 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS Notes 1. Other useful references include Adler and others 7. A 620-place drug rehabilitation center in Kun- (1996); European Commission (1997); Dallabetta, Laga, ming, the capital of Yunnan Province in China, features and Lamptey (1996); Gerard and others (1995); Lamptey a three-month program, primarily for injecting heroin and Piot (1990); and Nicoll and others (1996). users, that encourages their complete rehabilitation and support from family members (McCoy and others 1997). 2. See, for example, the work of Becker (1981) and Among the patients are those who were arrested and reviews of the literature by Birdsall (1988) and Strauss and many who voluntarily enroll. The price charged to pa- Thomas (1995). tients is $120 for those who are mandated treatment and $220 for those who voluntarily enroll. Families are re- 3. "Costs" here are not limited to the monetary costs portedly willing to pay this price, which includes all ther- of treatment or prevention. Costs of becoming infected apy, drugs, room, and board, finding it cheaper than sup- include suffering and premature death and the stigma and porting the drug habit of their relatives. However, among discrimination sometimes suffered by people with AIDS patients followed after treatment, 80 percent return to in- and their families. The costs of engaging in safer behavior jecting within two years. include, for example, any social stigma associated with purchasing condoms or obtaining treatment for STDs, as 8. The same general arguments-high costs and high well as the time, inconvenience, embarrassment, or mon- relapse rates-apply to programs that rely on methadone, welltary csthe t , oinconveniengcthem. embarrassment, oron a synthetic drug that, taken orally, removes the craving for etary costs of obtaining them. heroin without inducing euphoria. Moreover, because 4Iofbehavior methadone is only effective against heroin addiction, it 4. Information programs encouraging safer bealr does nor substitute for other injected drugs. should not be expected to have much of an effect on the behavior of the low-risk general population, since these in- 9. This change in behavior occurred at a time when lividuals may correctly conclude that they face relatively Thaiand had neither needle exchange nor methadone little risk. This explains the lack of relation found in many treatment programs. studies between knowledge of the risks of HIV (which in some hard-hit countries approaches 100 percent) and be- 10. However, the programs are much less successful at havior change in the general population (Sepulveda 1992, promoting condom use than at modifying risky injecting for example). behavior (Normand, Vlahov, and Moses 1995). Once drug injectors are infected, preventing the spread of HIV 5. Since 1996 the import tax and sales tax have been to others through sex is extremely difficult; thus, early im- reintroduced. Nearly one-third of the cost of running the plementation of harm reduction strategies among inject- condom social marketing program goes to pay the sales ing drug users is critical to preventing the spread of HIV. and import taxes (backgroundpaper, Pyne 1997). 11. The evaluations were undertaken in Australia, 6. An estimated one-third of the 750,000 heroin Canada, the Netherlands, Sweden, the United Kingdom, users in the United States, for example, are considered to and the United States. be occasional users who are not addicted (National Re- search Council 1989). However, the addictiveness of 12. Both the study by Serwadda and others (1992) in drugs depends on their purity. In Yunnan Province, Rakai and Barongo and others (1992) in Mwanza note that China, located adjacent to the Golden Triangle of opium education is no longer significant in multivariate production in Southeast Asia, injected heroin is more than regressions. However, other intermediate behavioral vari- 80 percent pure, most likely making it far more addictive ables are typically included as explanatory variables in these than in the United States and more difficult to stop studies, masking the effect of education (which may be a (McCoy and others 1997). determinant of all of them) and leading to bias in the esti- I69 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC mates. The educational differences in rural Mwanza are sta- amous in populations (b) and (c). However, their rate of tistically significant for women and men, and they remain partner change is very low. Although condoms are used significant in multivariate regressions (although whether for contraception in many developing countries, they are endogenous regressors were included is unknown). usually not the preferred method of family planning for married couples; DHS surveys conducted in the 1990s 13. This result from the GPA sexual behavior surveys found that condom use among married couples ranged is net of the impact of age and occupation. Several other from 0 to 3 percent (Curtis and Neitzel 1996). studies had similar findings. For example, greater school- ing is associated with a higher probability of casual part- 18. In the population with an epidemic fueled by nerships among men in DHS data from Burkina Faso, the commercial sex, the increase in condom use among men Central African Republic, C6te d'lvoire, and Uganda with casual and commercial partners from 5 to 20 percent (backgroundpaper, Filmer 1997). In rural areas of Kenya, represents 20 percent consistent condom use with sex Tanzania, and Zimbabwe, educated women are more workers only. For the simulations of condom use, it is as- likely to engage in casual sex than are uneducated women, sumed that if either partner wants to use a condom, a con- but in urban areas the relation is reversed. In C6te dom will be used. d'lvoire, men and women from wealthier households (with a car and good housing) were more likely to have 19. The cost-effectiveness might have been substan- casual partners. tially higher had the authors included estimates of the 14. The absolute levels of condom use in figure 3.2 are number of secondary infections averted. not comparable across countries, since the reference pe- 20. Mils and others (1993) found that STD treat- rnod for the DHS question about casual partners and con- ment costs for similar interventions in Mozambique and dom use was as short as one month (in the Central African South Africa also amounted to roughly $10 per episode of Republic and Zimbabwe) and as long as one year (in Haiti STD treated. and Tanzania). 15. Prevalence among those with 0-6 years of school- 21. The survey of the managers of national AIDS con- ing was 1.46 percent, among those with 7-9 years was trol programs in 187 countries was conducted between De- 1.06 percent, and among those with more than 9 years of cember 1993 and June 1994. Of these, 118 responses were schooling, 0.65 percent. These prevalence rates were mea- received, for a response rate of 75 percent. However, the sured per 100 person-years of observation (Carr and oth- quality of the responses varied from "complete and detailed" ers 1994). Since the WHO/GPA data on sexual behavior (about one-quarter of the responses) to "sparse and general" from about the same time period showed that men with (half of the responses, which received individual follow-up). higher income and education were more likely to have For more information on the survey methodology, see commercial or casual partners, it is likely that condom use Mann and Tarantola (1996), box 30.1, pp. 315-17. was already on the rise among Thai men before the brunt of the HIV/AIDS epidemic hit. 22. For these countries there were no data whatsoever on groups presumed to have high rates of partner change, 16. The underlying sexual behaviors in these four data were from very small samples (fewer than 100 peo- populations are summarized here as: (a) commercial sex; ple), or data were too old (from 1990 or earlier). (b) commercial and casual sex; (c) casual sex only; and (d) serial monogamy. The first three populations allow some 23. Questionnaires were sent to 120 countries; the re- concurrent partnerships, but the last does not. Also, both sponse rate was 42 percent. The low response rate and the (b) and (d) have commercial and casual sex. For more de- high participation of industrial countries means that the tail, refer back to chapter 2. results cited here are not representative of developing countries but are nevertheless true for the 50 countries 17. Women in stable relationships are monogamous that participated in the survey. Countries that responded in populations (a) and (d), but are not necessarily monog- included 15 in Africa, 8 in Latin America, 6 in Asia and 170 EFFICIENT AND EQUITABLE STRATEGIES FOR PREVENTING HIV/AIDS the Pacific, 12 members of NATO, and 9 European coun- Lesotho, Madagascar, Myanmar, the Russian Federation, tries not in NATO. Senegal, and Uzbekistan. 24. Only the 43 countries with UNAIDS Country 26. The costs of condom social marketing programs Programme Advisers were surveyed. Responses were re- are better documented than is their impact on HIV trans- ceived from 26 advisers representing 32 countries, for a re- mission or the extent to which they are used by the poor. sponse rate of 70 and 74 percent, respectively. Among the The cost per condom sold over a five- to six- year period 32 countries, 15 were from Africa, 7 from Asia, and 5 each through 1995 in eighteen Sub-Saharan African social were from Eastern Europe and Latin America and the marketing programs was $0.19 (1995 dollars), including Caribbean. The countries are: Barbados, Belarus, Benin, the cost of the commodity and overhead (Guy Stallwor- Bulgaria, Burkina Faso, Cambodia, China, Congo DR thy, PSI, personal communication). The net costs ranged (formerly Zaire), C6te d'Ivoire, Cuba, Dominican Re- from $0.08 to $0.20, depending in part on whether the public, Eritrea, Ethiopia, Ghana, Haiti, Indonesia, Ka- project was new, which raised costs. Cost recovery to the zakhstan, Kenya, Lao PDR, Moldova, Mozambique, Paki- program is only about $0.01 per condom. A review of stan, Philippines, Rwanda, Senegal, South Africa, Togo, CSM programs in ten countries (Bolivia, Congo DR [for- Uganda, Ukraine, Venezuela, Vietnam, and Zambia. merly Zaire], C6te d'Ivoire, the Dominican Republic, Ecuador, Ghana, Indonesia, Mexico, Morocco, and Zim- 25. Condom social marketing began in eleven coun- babwe) found that net costs ranged from $0.02 to $0.30 tries in 1996: Albania, Chad, China (Yunnan Province per condom sold, including the value of donated condoms and Shanghai), Republic of Congo, Guinea-Bissau, (Mills and others 1993). I71 CHAPTER 4 Coping with the Impact of AIDS WT T HILE SOME COUNTRIES STILL HAVE THE opportunity to avert a full-scale AIDS epi- demic by acting early to change the behav- ior of those at highest risk, others already have large numbers of infected people v v across many groups in the population. Chapter 1 presented evidence of the terrible impact of HIV/AIDS on individual welfare, in terms of human suffering and losses in life ex- pectancy. What can be done to mitigate the impact of the AIDS epi- demic on people and society? There are many impacts of the AIDS epidemic that cannot be quantified-for example, the emotional pain experienced by infected individuals and their families and the psycho- logical damage wrought on surviving family members. These impacts are very important, but how to respond to them is beyond our expertise and best left to others. This chapter considers the economic aspects of three types of impacts-on infected individuals, on the health sector generally, and on surviving household members-and the ways in which government policies can help people to cope, given the many other pressing demands for scarce public resources.1 The first part of the chapter shows that there are affordable, effective, and humane ways for governments in low-income countries to help ease the suffering of individuals infected with HIV However, both govern- ments and individuals in the poorest countries should be wary of fund- ing expensive treatments with uncertain benefits. The second part of the chapter suggests how governments can cope with the increased demand for and scarce supply of health care brought on by the AIDS epidemic in ways that are effective and compassionate, as well as fair and affordable. The third part proposes a strategy for developing countries to address the '73 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC needs of poor families hit by the AIDS epidemic in the context of other poverty programs. The chapter concludes with a summary of the policy recommendations for governments attempting to cope with the impact of HIV/AIDS on health care and poverty. Health Care for the Person with AIDS W H HAT IS THE HEALTH IMPACT OF HIV/AIDS ON AN infected individual over the course of the disease? Are there )VVV,effective, affordable treatments for people with AIDS in low-income countries? To answer these questions, this part of the chap- ter reviews the many illnesses that often afflict people with HIV/AIDS, the available treatments, and their cost. It distinguishes between three types of care: relief of symptoms, such as headache, pain, diarrhea, and shortness of breath, which is sometimes called palliative care; preven- tion and treatment of opportunistic illnesses (Ols); and antiretroviral (ARV) treatments, which attempt to combat HIV itself. Next it pre- sents the amounts that developing countries are actually spending to care for people with HIV/AIDS. While this amount is often large rela- tive to a country's GNP per capita, it is usually too little to buy all the drugs needed to treat opportunistic illnesses, much less to pay for anti- retroviral therapy. The section closes with a review of programs to assist with the home care of people with HIV/AIDS. The discussion finds that although treatment of HIV itself is difficult and extremely expensive, some of the symptoms and opportunistic ill- nesses typically suffered by people with AIDS can be treated simply and at low cost. Some infectious diseases associated with HIM especially tuberculosis, are somewhat more expensive to treat, but because they are infectious there are sound reasons for governments to subsidize treat- ment of any infected individual who would not otherwise get treated, regardless of the individual's HIV status. Palliative Care and Treatment of Opportunistic Illnesses The pattern of opportunistic illnesses differs from country to country, depending on which diseases are prevalent, and the quality and amount of treatment available. The natural history of HIV illness and several of I74 COPING WITH THE IMPACT OF AIDS the most important opportunistic illnesses are defined in box 1.2. Figure 4.1 presents the proportion of AIDS patients who suffer from each of three Ols-tuberculosis, cryptococcosis, and Pneumocystis carinii pneu- monia (PCP)-in six developing countries and the United States. Tuber- culosis is most common in the three poorest countries, the Congo DR (formerly Zaire), India, and C6te d'lvoire, becoming less common as per capita income rises. At the other end of the income gradient, PCP is most common in the United States, and is also common in the middle- income developing countries, Brazil, Mexico, and Thailand, but is rarely reported in the three lower-income countries. Cryptococcosis, a generic name for a group of fungal diseases that includes cryptococcal meningi- tis, shows no consistent pattern by income level, but infects at least 5 percent of people with HIV in all six countries. Among these three dis- eases, and indeed among all Ols, tuberculosis spreads most readily from people with HIV to others. As we discussed in chapter 1, tuberculosis greatly exacerbates the health impact of HIV in many developing coun- tries, particularly in Africa and India, where it is the most common op- portunistic infection. Figure 4.1 Percentage of AIDS Patients with Three Opportunistic Infections, Seven Countries 70 68 64 Ia P. carinii pneumonia - 60 - -- 54 3 Tuberculosis 50 _ _ _ | Cryptococcosis 41 40 - - _ 30 31 30 29 2 22 20 19 17 The pattern of opportunistic illnesses 13 differs across countries, depending on 10 5 4 5 9 5 7 which diseases are prevalent and the 2 - 1 3 3 quality and amunt of treabtent 0o available. Congo DRO India Cote Thailand Mexico Brazil United d'lIvolre States Note: Since only three of the 20 or more Ols are included, and since a patient may suffer from many OIs before death, percentages for a given country need not total 100 percent. a. Formerly Zaire. Source: Background paper, Perriens 1996; Kaplan and others 1996. I75 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Because of the variation in symptoms and in opportunistic illnesses, the cost and number of health care episodes for an HIV-infected person vary widely. Table 4.1 presents rough estimates for the average costs of pharmaceutical and inpatient palliative care of symptoms, prevention of tuberculosis and PCP, and curative care of the more common oppor- tunistic illnesses. Estimated lifetime cost per patient for this care ranges from $300 to $1,000, depending on which drugs are used and the cost per day of inpatient care. How effective are these treatments? In the early stages of HIV illness, palliative treatment can inexpensively relieve some of the pain, discom- fort, and incontinence that otherwise rob people of the ability to enjoy life and contribute to their family and their community. Without symp- tomatic treatment, dehydration that results from diarrhea and nausea can kill in a few days. Fever and headache can be disabling for days or weeks. As shown in the top panel of table 4. 1, drugs for palliative care are quite cheap. Hence, all but the poorest HIV-infected patients and their families are likely to be willing and able to buy these drugs, provided they are available. The sad truth is that these drugs are often not avail- able, an issue we discuss below. Moving down the table, we see that the opportunistic illnesses that commonly arise early in the course of AIDS can also be treated quite inexpensively. Treatment for thrush, toxoplasmosis, and pneumonia/ septicemia can buy one to four years of life at an additional drug cost of $30 to $150-all but the very poor would probably be willing and able to pay for these treatments. Rarer opportunistic illnesses like the fungal diseases tend to occur later in the course of the HIV infection and are more difficult and expensive to treat. For example, in the United States the average life expectancy after diagnosis with cryptococcal meningitis, the most com- mon of the cryptococcosis diseases, is 320 days, while in the Congo DR, perhaps because of later diagnosis, this drops to 180 days, even with ex- pensive state-of-the-art drugs (backgroundpaper, Perriens 1996). Since a patient in the Congo DR might survive 30 days without treatment, such drugs would extend life by about 150 days for about $870. In Thailand earlier diagnosis would result in treatment extending life by perhaps 330 days for $1,740. Many patients in these two countries might decide against buying these drugs, even if they have the money to do so. In the final stage of AIDS, the immune system is so weak that a vari- ety of infections spread throughout the body, leading to death. At this point, morphine to assuage extreme pain and the sensation of suffoca- 176 COPING WITH THE IMPACT OF AIDS Table 4.1 Annual Cost per Patient of Palliative Care and Treatment of Opportunistic Illnesses, Sub-Saharan Africa and Thailand (1996 dollars) Diagnosed epiiodes per Average cost per 100 patient-years Cost per episode9 patient-year Sub-Saha ran Su;b-Saharai Sub-Saharan Symptom or illness Afria Thail.anb Africa Thailand Africa Thailand Palliative care> e Diarrhea 63 13.00 8.19 Scaling skin rash 15 1.50 0.23 Itching skin rash 52 2.00 1.04 Cough 120 1.40 1.68 Fever 105 0.60 0.63 Headache 52 0.25 0.13 Pain, mili 52 1.12 0.58 Pain, severe 17 14.00 2.38 Nausea 75 1.75 1.31 Shortness of breath 43 6.50 2.80 Subtotal. 594 594 18 96 1&96 Treatmeent of inexpen''ve OL Tulberculosisd 47.5 40 37.00 261.88 17.58 10i .75 P rarini pneumonia 3 20 8.00 207.76 0.24 41.55 Toxoplasmosis 0 2 8.00 207.76 - 4.16 Oral thrush 77 77 2 00 2.48 1.54 1.91 £sophageal thrush 14 14 10.00 4.96 1.40 0.69 Pneumonia/scpticemia 20 20 60.00 25.38 12.00 5.08 Subtotal 161.5 173 32 76 15& 1- Treatnzent ofeexpensive Ols Cryptococcosis 5 25 870.70 1,741.40 43.54 435.35 Herpes simplex virus 5 18 140.00 46.80 7.00 8.42 Peniclliosis 0 9 1,852.50 697.40 - 62.77 Other Ols in, ludting: Cvtomn.ga1o virus Mycobacteriumiavium/complex 19.5 19 717.88 717.88 139.99 136.40 Subtotal 29 5 71 190.52 642.94. Inpatient daysf 3.000 3.000 7.25 22.44 217.50 673.34 Outpatient visitsf 1,200 1,200 2.50 13.60 .30.00 163.20 Gramd totals per case Palliatnve plus inexpensive Ols 29922 1,013.65 Palliative plus all OIs 489 74 1.656.59 a Costs per episode arc estimiated in Perriens (backyroundpaper. 1996) b Frequency of various symptoms and Ols tor lThailand are from Perriens (backgroumd pape,, 1996), or Kaplan and others (1996) or, where both givc a value, an averagc of the two c The frequency and creatment cosrs of symptoms listed under palliative care are assumed to be the same in Sub-Saharan Africa and Thailand d In Zambia. expeiience sug,gests that preventing tuberculosis in a group of HIV-infected pactents costs approximately the same as treat- ing Those in the grouip who gei the disease (Foster, Codrey-Faussetr. and Porcer 1997). e. Since palliative care involves generic drugs, these costs should be similar in all countrnes able to buy esscntial drugs in bulk through inrernanonal tenders. f. Paticnts in Thailand pay about 30 percent of these costs for drugs. inpatient days, and ourpatient visrs. The cost per patient-day in Sub-Saharan Africa is tak-en from Chela and others (1994) 177 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC tion provides relief to the dying patient, and this in turn helps to ease the distress of the patient's family. If purchased in bulk at international generic prices, enough morphine to ease the last two weeks of life would cost less than $4. But because of international controls on morphine dis- tribution, this essential drug is rarely legally available in poor countries at any price. The foregoing discussion has shown that many of the symptoms and opportunistic illnesses that occur in the early stages of AIDS can be effectively treated at low cost. Unfortunately, the low-cost generic forms of the needed drugs are often unavailable; even when they are available, people often lack information about their efficacy. Thus, many people pay much more than the $10 to $20 cited in the table for palliative treat- ments, while achieving no additional benefit. Governments can address these problems by facilitating the availability of generic drugs needed for palliative care and common opportunistic illnesses. For example, coun- tries with a concentrated or generalized epidemic could add these med- ications to their list of "essential drugs," which are widely distributed. Governments can also help patients to make informed decisions by ensuring access to reliable information about the efficacy of various treat- ment options, both pharmaceuticals and traditional remedies. The degree of government subsidy for treatment will depend on the country's overall health financing policy. We discuss this issue later in the chapter. Antiretroviral Therapy Is Expensive, Uncertain The treatments discussed above ease suffering and prolong life but ultimately fail to save the patient's life because none attacks the underly- ing cause of illness-the continued spread of HIV within the body and the consequent decline of the immune system's ability to recognize and repel biological threats. A few drugs have reduced the levels of HIV in the patient's blood below the ability of laboratory tests to detect it. Unfortunately, these drugs are expensive and complex to administer, their long-term benefits are uncertain, and their efficacy varies greatly from one individual to another. The first drug that showed evidence of inhibiting the spread of the virus in an infected patient was Zidovidine (AZT, or ZDV). When AZT was introduced in the late 1980s, the cost of a year's dosage was about $10,000 in industrial countries. By 1997 the cost of a year's dose had fallen to about $2,738 in industrial countries, while Thailand and a few 178 COPING WITH THE IMPACT OF AIDS other developing countries had negotiated bulk purchases for as little as $657 per patient per year. However, except for prevention of mother-to- child transmission, AZT rarely provided dramatic benefits, adding per- haps six months of healthy life for the average patient (Prescott 1997; Perriens and others 1997) A more effective therapy involving the use of three antiretrovirals was announced in June 1996. A year later, the U.S. government issued draft guidelines recommending early, aggressive treatment of HIV-infected individuals with triple-drug therapy (Brown 1997). However, it was clear that more time would be needed before the new therapies could be fully assessed. Some individuals taking the medicines in clinical trials have dra- matically improved their health and no longer have detectable levels of viral RNA activity. Yet even among these patients the virus may only be hiding and could re-emerge. And other patients show little or no reduction in viral levels, while still others cannot tolerate the drugs. As of mid-1997 no studies had yet been completed estimating the average percentage of patients who could benefit from triple-drug therapy or the characteristics of patients most likely to respond favorably-or to relapse. Does triple-drug therapy offer reasonable hope for treating the disease in developing countries? Even if the therapy is shown to be generally effective, three substantial problems will remain: the cost of the drugs themselves, the costs and difficulty of the monitoring needed for the therapy to be effective, and problems with patient compliance. Although all of these problems also exist in industrial countries, they are likely to be especially severe in developing country medical settings. Table 4.2 shows the cost of the drugs and necessary monitoring in Thailand, one of the few developing countries where the therapy is avail- able, and the United Kingdom or United States, and hints as well at the great complexity of regimen. Because most of the drug costs and all of the monitoring costs are lower in Thailand than in the two industrial countries, overall costs are a minimum of about $8,000 per year in Thai- land, compared with a minimum of about $12,000 per year in the United Kingdom and United States. These costs are likely to decline over time, perhaps substantially. But even if costs fell to one-hundredth of current costs, or about $80 dollars per person per year, they would still be several times the total annual per capita expenditure on health in many low-income countries. Moreover, effective antiretroviral ther- apy requires a highly trained, specialized physician working in a well- equipped clinic with experience performing a wide range of sophisti- 179 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Table 4.2 Annual Cost of Antiretroviral Therapy, Thailand, and the United Kingdom or the United States (dollars) Dailv .r unit cogt .4Ainiial coSt Inputs Dailv dose U rX or U K or Dnvgf (mg) ThlaiLad U '5. rhaiLnd U S. A'za-/-Z, u;i, REr ,,,/,:,'zg's Zido%udine (AZTi i O O 1.8'! -.;1 1 *!o - 1 .-. 8 Didanosine Iddli -ioO 5.8"1 5 2 ; .1CIq Zaklicabine 'ddC '.'i 5.-i' 6.81 1.9-1 2.-86 Sca%udinc id-JTi .80 - - 2.AQ0 Lanmivudine 1 i'TC 00 - - 2.690 por,;r.i.,iIl;r, S.quina%ir ISQV\ 1 .900 U O I.08 L.b O Riconaj ir IRTV, I 20') 2 .'1i 8.010 Indmna%ir (IDV\ 2.-iOO I 1.8-4 4320 Timre, pe r .4onioring pear Blood coun i 2.'110 2.0'; 21 '1J 2-' 252 Blood heml\rnr% 4.00 1 2.('.i 35 (10 Ii 14i0 CD-i C-Lnr q.00 30 01 I (, .0 I ,, 0 62' RNA vlrai 1usd 3 i() -iI 0'1 I[)i 00 1; 4;l) Additional loUrparieni -.IILS I 2.MJI I I I . W:0.00 11:0 1.2110 AZT. ddl. and I D\.i! I 0,803 AZT. ddl. and RTl' I 3.2 . 21. i)3 - DEir nor s.LLI'leoc noi 3pplc;bl .i Triple -d!. uc her ip. con. , c; n-.- t F r%r grop v.I di..r! pl,li ':.rn o- c, r.nd 4rq.p p!u; mcronir:rinn Drug. tr gsen jail, a l,ch chrr drug. 4hr.uld be cocnbinjd ,. i rn11- I:r, ,1 -cu r r' fc rc ii jirh d pr.;hbabl: b.11: ! pau 1> n R;. "1: " 'i F. errlr- 'i, Prfl-`.r vid orf,uri I 'W-. inJ ',,.i : ind B3irtlkr 100(1,. cated tests and procedures, all of which are in critically short supply in most developing countries. In the event that cost and infrastructure problems could somehow be overcome, patient compliance would continue to pose serious difficul- ties. Patients undertaking triple-drug therapy must swallow up to 20 pills a day according to a complex schedule related to sleep and meal times. Failure to follow the schedule increases the chance that the virus will become resistant or that the patient will be too sickened by the drugs to continue treatment. Even well-educated patients with good clinical support have difficulty adhering to this demanding regimen; moreover, patients in the early stages of HIV infection are sometimes not willing to i8o COPING WITH THE IMPACT OF AIDS take drugs that make them nauseous when they otherwise feel healthy. In clinical trials in industrial countries, for example, as few as 26 percent of patients complied with the instructions (Stewart 1997). Problems with patient compliance are likely to be worse in low-income countries due to lower education levels and the many other problems that poor people in developing countries face. Even with all these difficulties and uncertainties, many patients in developing countries will ask their physicians for triple-drug therapy, just as patients have attempted to obtain AZT. Governments will in turn face pressure to buy these drugs and to subsidize the necessary clinical ser- vices. When very few people have AIDS, total costs will also be small rel- ative to other government expenditures. But as the epidemic progresses, the number of AIDS cases and the cost of the subsidy will escalate rapidly, drawing resources from other pressing social needs. At some point it will become evident that such a subsidy is unaffordable and also unfair to the many people who for a variety of reasons want government help but do not have HIV. Individual Treatment Costs for AIDS Are High, Even in Poor Countries We have seen that medical responses to HIV/AIDS range from a few pennies to thousands of dollars. How much a country actually spends to treat a case of AIDS depends on many factors besides the differing cost of health care inputs. The most important of these is the amount of treatment that the HIV-infected person, his or her family, and any third party payers such as insurance companies or the government are willing and able to buy, and how much the government subsidizes health care and AIDS treatment. Figure 1-8 showed that across countries this amount is strongly correlated with per capita income. An in-depth study of AIDS expenditure in four countries and Sao Paulo State, Brazil, con- firms this general pattern; the average total (public and private) AIDS expenditure varies from 0.6 times per capita GDP in Tanzania to 3.0 times per capita GDP in Sao Paulo; the average is a ratio of about 1.5 (backgroundpaper, Shepard and others 1996). Alternatives to Expensive Inpatient Care Where the AIDS epidemic is severe, health policymakers inside and outside government have sought ways to provide compassionate care at low cost. Three alternatives to expensive inpatient care are outpatient I8i CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC AIDS clinics, hospice care (residential low-technology care for the ter- minally ill), and home-based care. One innovative program to deliver high-quality treatment of symp- toms and opportunistic illnesses without the expense of hospitalization was an outpatient clinic started in 1989 in Sao Paulo, Brazil. Such clin- ics are especially well suited to serve urban HIV-positive and AIDS patients who are able to leave their homes. Later in the course of the disease, when the patient is less mobile, the hospice or nursing home provides a lower-cost substitute for inpatient care in a sophisticated re- ferral hospital. However, since such facilities are rarely available in devel- oping countries, the main alternative to the hospital is care at home. What sort of home-based care is most effective? An analysis of the cost of eight home-based care programs in Zambia found that community- initiated programs were more effective and much less expensive than hospital-initiated programs (Chela and others 1994; Martin, Van Praag, and Msiska 1996). Assuming that the average patient with AIDS would survive six months with either type of care, the benefits of the care must be measured in reductions of hospitalization cost; reduced travel time to the hospital for the patient and the patient's caretakers; increased patient satisfaction and comfort; and ancillary benefits to the community, such as improved understanding of the ways to prevent AIDS and decreased stigma toward HIV-positive people. Since the study found that patients who received home-based care reduced their hospitalization before death by only two days, the expenditure on the hospital-initiated home-based care programs of about $312 (6 months x 2 visits per month x $26 per visit) was much more than the $14.50 saving in hospital charges (2 days x $7.25 per day). On the other hand, the costs for six months of community-initiated home-based care averaged just $26, less than one- tenth the cost of the hospital-initiated program, and could almost be jus- tified on the basis of reduced hospital use alone. The tenfold cost difference between hospital- and community- initiated home care programs was due to the much larger expenditure on transport and staff time for the hospital-based programs. For example, on a typical day a team of trained hospital-based nurses could visit only four to eight patients, about a quarter of whom were away from home when the team arrived. As a result the hospital-initiated teams spent on average about two hours on the road in order to spend only fifteen min- utes with the patient. In contrast, the community-initiated teams walked only a few minutes and spent an average of two hours with the patient. I82 COPING WITH THE IMPACT OF AIDS If the low cost of the community-initiated home-based care program in Zambia can be generalized to other settings, it is possible that such care would be financed by the patients, their families, and their commu- nities. Indeed, the community-initiated Zambian programs function well because of strong volunteer support from the local communities. Since the benefits of the program include the public ones of improved knowledge about HIV prevention and reduced stigma, there may be a government role in financing such programs, at least until their private benefits to patients' farnilies are sufficiently well understood for these families and communities to support such programs on their own. Where policies exist to facilitate access to health care for the poor, they should be extended to include community-based home care programs using the same etigibility critena. Difficult Health Policy Choices in a Severe AIDS Epidemic T HE PREVIOUS SECTION DESCRIBED THE IMPACT OF AIDS ON the individual HIV-infected person and demonstrated that limited treatment of symptoms and opportunistic illnesses, especially when performed partly by community-initiated home care programs, can provide compassionate care at relatively low cost. In this section the need to keep costs low becomes more apparent as we widen the focus from the individual HIV-infected person to the health care needs of all people in a country. To better understand the difficult tradeoffs involved, we first estimate the magnitude of the impact of AIDS on the health sector, and then discuss how government policies can mitigate this impact. How HIV/AIDS Will Affect the Health Setor AIDS will affect the health sector in two ways: by increasing demand and by reducing the supply of a given quality of care at a given price. As a result, some HIV-negative people who would have obtained treatment had there been no epidemic will be unable to do so, and total national expenditure on health care will rise, both in absolute terms and as a pro- portion of national product.2 I83 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Increased demand for care. Most people who develop AIDS are prime-age adults. Without AIDS, this 15-to-50 age group accounts for only 10 to 20 percent of all deaths in a developing country, but these deaths typically generate a disproportionate share of total health care demand (Over, Ellis, Huber, and Solon 1992; Sauerborn, Berman, and Nougtara 1996). Moreover, since several studies suggest that adults with AIDS use more health care prior to death than those who die of other causes, or even of other prolonged illnesses, the percentage increase in the demand for care by adults is likely to exceed the percentage increase in their mortality due to AIDS. As a result of these two factors, in a country where prime-age adults utilized one-quarter of all health care be- fore AIDS, a given percentage increase in their demand for health care will increase totaldemand by at leastone-quarter of that percentage. For example, a 40 percent increase in the mortality rate of prime-age adults Table 4.3 Deaths per Thousand will increase total demand by at least 10 percent, even though total mor- Adults Caused by a Constant Rate tality has increased by only 4 percent to 8 percent.3 If AIDS patients use of HIV Infection expensive antiretroviral therapies, the increase in demand will be much Hf Mleduin greater. preralece tr hme from How much the demand for care increases in the aggregate depends on inection to deathw rate 10 fea to ddy the increase in the prime-age adult death rate, which in turn depends on (percent) 10Years 5 Years the level of HIV prevalence and the median time from infection to death 5) > () I l i) (table 4.3). A stable prevalence rate of 5 percent among prime-age adults 0 103l 22.2 eventually increases their annual mortality by about five deaths per I * 15.8 33.3 1,000 adults if the median time from infection to death is ten years, or 20 21.1 i d.4 by about ten deaths if the median time is only five years.4 A prevalence 50) 53 6 66 rate of 30 percent, such as is observed in Lusaka, Zambia, will increase 100 o lti 3 222.2 the number of deaths per 1,000 adults by 30 to 60, depending on the N,,w. I he d,iah ritre in ikilimni median time to death. In Sub-Saharan Africa, where mortality rates in 2 and tru calculited b% muluph'- this age group were as high as five per 1,000 before the epidemic, even a trij he. pr .-alen rjrc from clIhemnc 5 percent ifecton rate will double or triple the adult death rate. In a I in 1012. io n deah inecinnril d,n Einic trrnm inf-con ri deathL middle-income developing country with adult mortality of one per This formula asiumes a sread% -rlte 1,000, the same endemic level of HIV infection will increase prime-age epidemic irn m %bih incidencu: is c-r.- a sianr 3nd a proportlin ! 2. t adult mortalit five- or tenfold. ihi..e infecred In . givin yvar di In Given these parameters, how much will the epidemic increase the ejch of 2M%uh4equen . In [he demand for care? In a country where adults consume one-quarter of health ilkcnce of HEV. rhe ha.dinr morr.,J- it; rate peC r hcusand .,duics .ge l; careprior to the AIDS epidemic, HlVprevalence is constant at 5percent of t, in range, ticm C.S in indusih.l adults, the median time to death is ten years, and the baseline mortality rate counrrirc rv ai high a s in ,om. piru- of Sub-Sihuran Africa. amongprime-age adults is 5per 1,000, the epidemic will cause a 26percent increase in the demandfor health care at every price.5 If the prevalence rate I84 COPING WITH THE IMPACT OF AIDS is higher, the median time to death shorter, or the baseline adult mortal- ity rate smaller, the percentage increase in demand will be correspond- ingly greater. A final important factor that may increase demand is insurance. This may take the form of private insurance, a government-run insurance pro- gram, or, more typically, health care financed through general taxation. Because a portion of health care costs is often covered by one or more of these types of insurance, the price paid by the patient is usually a fraction of the cost of providing the care. Since insurance enables patients to pur- chase more care than they would otherwise, it increases the demand for care arising from any given level of illness, thus magnifying the price shock of an AIDS epidemic. For example, if the proportion of cost of providing care paid by patients (i.e., the coinsurance rate) is 25 percent, patients will reduce their utilization in response to increased cost by only a quarter as much as they would if they had to pay the full increase. Reduced supply of health care. In addition to increasing the de- mand for care, the AIDS epidemic will reduce the supply available at a given price, in three ways. The magnitude of these effects, discussed below, will generally be larger in the poorest countries with the largest epidemics. The first and largest effect is the increased cost of maintaining a given level of safety for medical procedures. Even without HIV, hospitals and clinics in poor countries may pose a risk to health. Needles and other instruments are not always sterilized, rooms are often overcrowded and poorly ventilated, and care providers may lack rubber gloves and some- times even soap. Without modern blood banks, a transfusion might in- fect the recipient with hepatitis B. In such situations, infections of all types spread rapidly; some, including such common illnesses as pneu- monia, may kill. Before HIV, however, infections picked up in a clinic or hospital were rarely fatal to persons not already in a seriously weak- ened state.6 Because the AIDS epidemic has greatly increased the risk to patients of existing medical procedures, simply maintaining the level of safety that existed before HIV requires additional hygiene and blood screening, both of which increase the cost of care. In middle- to high-income coun- tries, where blood screening and sterilization of injecting equipment are already the norm, the impact of AIDS is confined to the incremental costs of adding an HIV test to existing tests and using rubber gloves and face masks in situations where they were previously not used. In poor 185 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC countries, where blood screening and needle sterilization were lacking before the epidemic, the resources needed to maintain the quality of care in the face of the AIDS epidemic can be substantial. For example, the an- nual recurrent budget of the Ugandan Blood Transfusion Service, which was established in response to the epidemic and meets the demands of the entire Ugandan national health care system for clean blood, is esti- mated to be about $1.2 million, including capital and recurrent costs. This amounts to about 2 percent of national public health expenditures or about 1 percent of total national health expenditures (European Commission 1995a). Despite the potentially high costs of blood screen- ing, HIV has greatly increased the justification for a government role in ensuring a safe blood supply. However, there is no convincing rationale for government to subsidize the entire cost of running such a service indefinitely (see box 4.1). Blood screening and improved collection pro- cedures will protect blood donors and recipients. However, since average donors and recipients do not engage in unprotected sex with a large number of partners, a person infected while giving or receiving blood is not likely to pass the infection to many others. Thus, in developing countries where the cost of establishing a safe blood supply is high, blood screening will not be among the more cost-effective approaches to pre- venting an epidemic based on sexual transmission (see box 4.2). To be sure, blood screening and better hygiene will help to prevent the spread of other infectious diseases besides AIDS. Such measures will also reduce the occupational risk of AIDS and other diseases that health care workers face, and therefore reduce the amount of additional com- pensation needed to offset their occupational risk-an issue we discuss below. A careful accounting of the net cost of protecting patients from HIV by screening blood would need to take into consideration these ad- ditional benefits, for which data are lacking. However, it seems likely that even if these benefits are taken into account, the remaining cost of screening blood and improving hygiene to protect patients from HIV/AIDS would substantially increase the unit cost of medical care. The second factor reducing the suppJy of medical care at a given price is the increased attrition of health care workers who become infected with HIV. Like all adults, health care workers may become infected with HIV as a result of sexual contact or use of unsterile injecting equipment. They also face an additional risk of becoming infected in the course of their work; however, this risk is generally much smaller than the risk from sex- ual contact. Thus whether the AIDS mortality rate among health care I86 COPING WITH THE IMPACT OF AIDS Box 4.1 The Government Role in Ensuring Clean Blood THE Hl\ ,AIDS EPIDEMIC HAS DERANIATICAL[i sidized by the government, then the same argument increased the importance of clean blood. W'here the would apply to clean blood. It: howve%er. one be- most serious ccmmon infection that a transfusion lie%es chat there is no obvious reason to tasor cura- recipient previously had co fear from unscretned tive health care over other necessities, such as blood was hepaLitis B, which is rarel; fatal and com- clothes. hou.ing. and clean "acer. then clean blood municated in only about 2.5 percent ot unscreened should receise as lirtle subsidy from the goternment Eransfusions. recipients in some countries now tace a as other curative health care senrices. one-fourth chance of HI\' infection cEmmaniel. V'et even those xsho believe that most curaci'e WHO, as cited in Fransen. personal communica- care deienes little ubsidy admit chat the treatment tioni. As a resulr of the HI\'AIDS epidemic. the of int'ecious diseases confers positise exrernaliiies transtusion required for a surgical procedure or and thus should be subsidized. This bring; us to an childbirth that mighr hav.e been relatively routine in evaluation of the second consideration. Assumino a developing country ten years ago now% requires the that transfusion recipients recoser from the medical guarantee of clean blood to be equall satc. procedure and then become sexually active. prevent- W-hat should be rhe government's role in thc pro- ting heir infection may- precent the-m froma intecting vision of safe blood? Serting aside povern. shicl is others. How large are rhe.e positice externalities: addressed in the test. live justitications can be identi- For one country'. LUgandi. boX 4.2 sho%si that a lied for the public to subsidize or othenrise play a highly effectise program presented 'i- secondanr role in the provision of blood: HiI to present HRI' infections in 1l9'4 at a cost of $1.68-4 each. \\hile infections in blood recipients: t2) to present infec- this co.c is much less than the liferime trearcnrnt tions in the sexual partners of'blood recipienrt: 13 to costc of an HIV-infected person in an industrial avoid the sudden onset within a communirs of the countn. it is more than any reasonable estimate tf health risk From unscreened blood: i-o to provide the cost of presenting secondarv infections in the economies of scale that apply to a blood bank L.'ganda. Thus, the pievenrion of secondary in'fec- senrice; and (h to avoid the difficulrn that a citizen rions does nor appear to be sufficient to justify gos- w%oald have in udging the qualin ofa blood bank. erniment subsidy of che entire cost ot' the program. WX'hile a high-quality blood bank wvill obviously although it could justify a partial subsidy. be quite effective in preventing the transtusion ot Considerations (3i and 14) appeal to the same infected blood. and thereby in preventing the hospi- economic arguments often used to justif- govern- tal from infecting transfusion recipients. this fact ment infrastructure investments. The sudden does not, by itself. imply chat the governmenct increase in risk from blood transfusion is a shock to should play a role in supplYing the clean blood. the health care sysitem, oo rapid for individuals and Setting aside for the moment considerations 12i prk%are institutions to make newv blood-screening rhrough 15)1 the provision ot'clean blood is compa- arrangements quickly. As the insurer of last resort rable in importance to the provision of clean nee- against catastrophic changes in the environment. the dles. clean bandages, and clean hands of the nurses government has a role in assisting sociery in adjust- who change those bandages. Any arguments for gov- ing to [he new higher cost and complexitn of health ernment financing of decent qualitY of care. includ- care in the presence of AIDS. Furthermore, as ing basic cleanliness in the hospital. also apply to demonstrated by box figure -i., a blood transfusion clean blood. Itone accepts the argument thar hospi- sersice entails substantial economies of scale. Since a tal care is a basic need, vhich should be heavilv sub- ,'Bos coninues on the -follot. : '7gppage.) I87 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 4.1 [continued) single transtusion ser-ice can serve all local needs because of [he economiei of scale. are unliklyc [o Without cxhausting its economies of scale. it would have a choice of blood banks. The g_overnment and be a natural monopoly without the Fear olFcompeti- the public ihould noi assume that any monopol't lion to ensure quality- ise'ice at the best price. It wvhether Lt produces electricity or blood bank ser- .%ould he forced to charge prices above marginal cost ices, and whcther it is "for-profit" or 'nonprofir." in order to co%er it, costs and might charge pric;s w%ill indefinitely perform in the public s best interest. w%ell abotc average costs in order to niaximize its In this situation there is an argument for the esrab- profits. Just as for electric utilities and other natural lihcment of a regulators board to sshom rhe blood monopolies, there is a wcell-e-tablished justification transftusion ser' ice is responsible.' The board should tor government internention to regulate. if not os'n conisist of representati%es ot the miedical establish- and operate. [hem in such circumstances. Hinsjeer. ment. go'ernment. and paitiet' and should produce rhev do not justifY a 100 percent subsidy for blood. an annual report on [he quality of the blood bank Consideration 5i) involves the inability of the serVicc, w%hich should then be ssidlely disseminated public to judge the quality of a blood bank. This in the press. argument Is nor particular to blood trnnshusion ser- In sum, the appropriate role of go'ernmncn in vices. since patients haie an equally difficult imine financing hlood suppliec depends first on one a,ie%s judging the qualiry ot rheir phvsiciins.' Yet patients of the degree of financing the gov-ernment should can choose among many different physiciani but, pro- ide to curatise health services. The argument for curative scr'ices extcnds directly to the provi- sion ot' blood. Thc number of secondary infections Box Figure 4.1 The Cost per Unit of averted through blood screening is unlikely to be a Blood Transfused in Uganda poclcerlul argumentc for gnsetrnment subsidies. Esen so. there is a strong argument for the government to 1994 dollars per unit transfused launch and nurture a blood bank Service as a subsi- sct 1989 dized "infant industry. before subjecting it to the rigors of the financing arrangtments provided [or \ 1990 the rest of [he health circ system. Finally, bccause 6 \ceQonomiec of'scale will tend to make the blood bank 9291 a monopoly in most communities, blood bank ser- vices should he subjec[ to sirict regulatory re-iew. 4 1994 Irlt-rninirion is 1%1nrtltrriealj dit[ribuied bervec,n [he frimn producing .nd n-urn5 bie.'d anid ihe huopakls ph%s,w.ars or 0 _pjarient:s 'he con'iime 'r 0 5.000 10.000 15.000 20.000 25.000 30.000 35.000 - \Vh1k picints should be J3lltd the i[ihi- perx.rageuf t[he rnigisi- coin C-i anin ofbkhtd tha[ tihe, in ch.sr,ed tn, [h,ru Units of blood provided euritI, e citt ' nonmiecri.us d-eas,. ii Jest, nor iiss thal dor or- .hnuld be pAd .-r blt.;.d Th. obscr.ricnr be RchPiiid S .t Ecureran Ii ,mrnr 'il-n r p F-4 N- min] mnourin T.rmril,' i I i-'l rcgirdwn, the hinsnrr; ot recruirtne 5)Llncrirs- ;nl eried to curen! iollars at . 2 dollrs per EC Li and then ro donor, has bKen found Li ippik in many difiereni n iional ] '., dfl!- us.ng ac tei L S cninilncr pr-icer indf. 'ertings. I88 COPING WITH THE IMPACT OF AIDS Box 4.2 Cost of Preventing Secondary HIV Infections through Blood Screening in Uganda Hi is- t Fi 1--FFFEq Tb li BLi 1N D xi HFNv. iX ot thc'e people are quite sii., the e%aluation xtudv- pre%enting secondary HIV mtnlncron^'( One ano.xer csntmatcd that each otrhese adult %%ould ha% e orilk a to thix quetiion can be seen in the resultx of the 'o percenE chance of infecting one other perSon wich LUganda Blood Transfusion Ser'ice ILTBTSi for HRV tEuropean Commiision I')95c. Thus the total 19931. Having established its abilh to supply number of ;econdan' Infections aeried would be Kaampala %%irh clean blood in 1 991. bY 1991 the ICi I If the entire justificarion of the blood supply LUBTS .%as reachinc, out to cm,%er the entire countnr sen icc is prevention of thee secondary intections, Thar vear the ser%ice transfused 20.156 patients the cosr-effectrveness of the service is $9299.00 di- throughout rhc countnr ac an average cost ot .ip- slded bv -41. or $2.2-2O per such infection av-erted. proximately $38 per unit oftblood. and an average If L'ganda had had a sustainable blood supply sys- of 1.2 units per parient. for a total budget otfappro\- term the cost of pre%encing these 415 infections imatel $929.900. Box table -i.2 breaks ouL the HRN' %tould hate been only $319.89-i, or $-1 each. This prevention benefits of the service, sho%t ing that Its much smaller amount is still subsrantial.v larger rhan use averted HIV infection in .an estimated I.863 the cosr of preventing seconlanr infections in other surviving transfusion recipients. "wass (see box 2.6i. But to measure the positive e\iernalities of the program. ajnd thus the rationale t;r gov-ern ment sub- sidies. %%e need to look bevond these primanr infec- rioni to consider secondary infections. Children wv ho are infected by transfusion ire unlikely to Ikie long T'e tthkirs pomcii Ln rh.e o ureling pin. idcd ... birw:,d enough to infect others, but some of the adults niav dn.-r, rv-|it h- aesrind iddir,ond primarx .tit. n Eutopean C..mmss.nr , @i -'. An . ,econdar. ictauon. i,etercd rhrough be sufficiently y.ung and sexually active to engage in ch, r u[t should be idded ,. Lh.- [5 roenomputc tht [,,J pF.L risk-v sexual behavior later in their lit es. otincc man' i.e , tc- rcl!i;.r ..t tin. F-.m Box Table 4.2 Effectiveness of Blood Transfusion at Averting HW Infection, Uganda, 1993 Benotfi Effects of blood nansfiLninsw Chiidirme .Adiultt TrFoal Patients trinsfased i I .i l; 4u i I n. I im Patients expected co die %sithit t,ru nfun i sntun Patients .ho died despite transiuS kS0 i,' I., x.. Number otfdeath preenLted [.'1S 1.2.16 '55 Number of priman HIX infection, presentcd I i.r) .So3 Number of iecundar HI\ infccrcion,pretenred U -i S Sit'. tL Bw.d on the r'euIt. ach,e.-d bt dhi L gcndin [-s-d limr.ttsr- r` Icr. i 4 r±tpu-rrcd n Bna B..nr.n,k. rnd Fja.-n IP-n'n Eure-pemn C.;mm'.s'on i l- i Li and Frin.en j,ur. per- d , tl!ITdflLCItI*flI. I89 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC workers is higher or lower than among the general population depends mostly on the effects of income, education, and social status on sexual be- havior. Two studies of HIV prevalence among health care workers from Africa suggest that doctors and nurses are at least as likely to become in- fected as other people (Mann and others 1986, Buve and others 1994). If this is true elsewhere, a country with stable 5 percent HIV prevalence can expect that each year between .5 and 1 percent of its health care providers will die from AIDS; a country with 30 percent prevalence would lose 3 to 7 percent of its health care workers to the epidemic. This attrition from AIDS deaths may substantially increase the cost of health care. For exam- ple, if labor costs are half of total health care costs, and training or re- cruiting a replacement worker requires a one-time expenditure equal to the worker's annual salary, then a 7 percent increase in attrition will in- crease total costs in the health sector by 3.5 percent. The third way in which AIDS reduces the supply of health care is through the additional risk it imposes on health care workers. Even though most HIV-infected health care workers acquire their infection through sexual contact, in a society with a large proportion of HIV- positive patients, health care work will be more dangerous than if there were no HIV. Some students who would have become doctors and nurses will therefore choose alternative occupations, unless they are compensated with higher pay for the increased risk. A recent survey of medical and nurs- ing students in the United States found that AIDS had indeed reduced the attractiveness of specialties in which contact with HIV-positive patients was more likely (Bernstein, Rabkin, and Wolland 1990; Mazzullo and others 1990).This problem is likely to be most severe in hard-hit develop- ing countries, where HIV prevalence is much higher and rubber gloves and other protective equipment are often in short supply. In Zambia, for example, some nurses have demanded special payments to compensate for increased occupational risk due to HIV (Buve and others 1994). The magnitude of increased costs of medical staff has not been esti- mated. As noted above, improved precautions in hospitals and clinics may reduce these costs. But because people respond to perceived risk rather than actual risk, such improvements may have little impact on the demand for increased compensation. Thus, it seems clear that health care workers' perception of risk will increase the cost of care. The total impact of these three effects-increased cost of preventing infection in medical facilities, attrition of health care workers due to HIV, and additional pay that health care workers demand to compensate them for increased risk-will depend most importantly on HIV preva- I90 COPING WITH THE IMPACT OF AIDS lence and whether modern blood banks and hygiene were already in place. In a country that has 5 percent HIV prevalence among prime-age adults and lacked blood banks and blood screening before the epidemic, a conservative guess is that the cost ofproviding care of a given quantity and quality will rise by about 1 0 percent. Scarce care, higher expenditures. Taken together, increased demand and reduced supply have two related impacts: first, health care becomes scarcer and thus more expensive; second, national health care expendi- ture rises. The size of the increases in health care prices and national health care expenditure depends partly on the price-responsiveness, or "elasticity," of the demand for and supply of care. For most goods, higher prices reduce demand, as consumers switch to substitutes or forgo an intended purchase altogether. This same principle holds true for health care, but the price-responsiveness or elasticity of demand for adult health care is usually small, since there are no close substitutes, and people who are sick and who have the ability to pay will often pay whatever is needed to get well. For the purposes of our simulation, we assume that a price increase of 8 percent would decrease utilization by only about 8 percent, for an elasticity of 0.8.7 Higher prices also generally increase supply. Here, too, however, the nature of the health sector affects the supply response. In the very short run, perhaps a month, the supply of care is unlikely to change much. Over the long run, the supply of physicians and inputs to health care can expand as much as necessary. Over the medium run, five years or so, we would expect the supply of care to respond somewhat to increased de- mand and the resulting higher price. One response observed in Canada, Egypt, India, Indonesia, and the Philippines is that physicians who work in the public sector rearrange their schedules to offer more health care privately, after their obligations to the government have been met. The elasticity of this response has been estimated at about 0.5, meaning that every 10 percent increase in the price of care elicits a 5 percent increase in supply (Chawla 1993, 1997; Bolduc, Fortin, and Fournier 1996). We have argued in the previous two subsections that a constant 5 per- cent seroprevalence rate would eventually increase the demand for care by about one-quarter and the cost of care of a given quality by 10 per- cent. Drawing on the assumptions in this subsection about the elasticities of the demand and supply responses, and assuming that patients pay half the cost of health care, box 4.3 shows that total national health expendi- ture, and also the government's share of expenditure, would both increase by about 43 percent. The increase would be less in a country like India, I9I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 4.3 Estimating the Impact of AIDS on the Health Sector HO¼\X NIL;EH WXILL THE L F-FECTl- E PRICE OF CAPFE sUbsid& on consumersi.) The ligure is constructed so increase as a result of.MUDS' Box figure -d.3 demorn- that the market equilibrium occurs atr a price or 10 srarces h1ow% the appro.imrtc- size of these increases currency% units per unir of health care. at ; hich a can be estima[ed for a hypothetical country- %%ith total of 10 units of care art delivered. Total health elasticities ot demand mnd supply for health care of care c pendirure in this hypothetical country is thus 0.8 and l.i and a government policy tco ubsidle 10 times 10 cr IJ0 currencN units in the abwnce ot half of the cost of care. The rtwo solid lin'c sho% the trhc AIDS epidemic. amoUrnt of health care that is demanded and sup- Nowx- assume thcre is an HIl\IAIDS epidemic plied at each price prior to an HIM cpidemic. iThe that levels off at a constant seropre%alencc of per- demand curse is drawn %%ith an elasticiry of only 0 4 cent of the adult popullation. The argumen[s in the in order to Incorporate [he electi of the government chapter suggest that the amount of health care demanded at escrn- price is likc!l to incrcase by 25 Box Flgure 4.3 The Impact of a 5 Percent percent. wvhile the cort of purchasing any gpiven Infection Rate on the Quantity and Price of aimount of care of a given quaity vhdill increasc b% l'J Heafth Care pcrceint. These cw. Iimpacts of the AIDS epidemic are illuctrated by- a rlghrsvard hilft of the demand Price of health care Demandfor care cUrve bi 2i percent Ito the dashed. do%%nw%ard- ~~~ ~~after AIDS - -> afterAIDS sloping line) and an up% ard shit- of [he supply curve 13.- s > o by I( percent IIt. the dashed. upwvard-sloping inil The impacts on the cquilibriunm price mnd quantity he care A - . can be read from the figure. The price of a unit ot healch care will increase about 31) percent. and the of caie ammount of care provided %%ill Increase about Iii per- 10. h- Suppyaol after AIDS cent. T-otal national experidirure. the price per unit betore AIDS Fot care times the number of units wvill increase -i3 percent to 143 currenn- units isince 1 3s \I I = 14 3i lo 11 Quantlt ot Wec have wecn that a third-parn- paiyment, such as insurance or a gin%ernmenL subsidy for treatrment. Xac The Jlenmnd ind -uppl ciarce. are con,rrucid !o ia.r nmakes peoplc less sensitie to changes in coist ot rhe pril elsr.miire, ai theh porr 0IC) IUl are I.s snd i S. health care By reducing the The mpacr ' -IDS *ilec,irared 1are Sc thelJcag the price elastict of de- cr%e to The nghr b. ot A eDni at er- prie Ill he upf.In mand. such third-parr payments make the demand c -ent LIi hr l IpIcr'-I ir e Ite"<.L]uJn L e (he ,c\ t:t ter a. cLure steeper. both betore and aftier the introduction cmpianatln ..1 thse a;-umpi;n,s1 ot l D)5. where only about one-fifth of the cost of care is paid by the government, and substantially more in countries like those of Latin America and East- ern Europe, where three-quarters or more of the cost are subsidized. Does the available empirical evidence support these conclusions? Although there are significant data problems, the short answer is yes. 192 COPING WITH THE IMPACT OF AIDS Measuring the scarcity of medical care through changes in the price of care of given quality is problematical because of the difficulties in mea- suring quality. This is especially true in developing countries, where a general lack of data is compounded in the health sector by government subsidies and nonprice forms of rationing. In such cases, the effective price of care may rise even though nominal prices remain constant (see box 4.4). Furthermore, because of the lag between infection and death, the time between the attainment of a given HIV prevalence rate and the full impact of that rate on the demand and supply of health care can be ten to 20 years. For these reasons, we cannot accurately assess changes in scarcity of health care in developing countries by observing changes in nominal price. Nonetheless, we can get some sense of the extent to which HIV/AIDS increases the effective price of health care by consid- ering whether the epidemic makes it more difficult to obtain care. Stud- ies of hospital admissions data strongly suggest that this is the case. Table 4.4 shows the percentage of beds occupied by HIV-positive patients in six referral hospitals in developing countries with large epi- demics. The hospitals are the top health care institutions in each coun- try, providing the best care available outside of a few expensive private clinics. Because these hospitals are at the apex of their health care pyra- mids, we would expect that AIDS patients account for a significant pro- portion of their patients. Even so, the percentage of beds occupied by HIV-positive patients is striking, ranging from 39 percent in Nairobi, Kenya, to 70 percent in Bujumbura, Burundi. Box 4.4 The Effective Price of Care SOmE RL.ADtRS MA' OBIEC! Tha i fl-HE Ik(CE Pr-l eY on waiting timt. In others. a patient dissat'sled with patients need not increase in countries %%here che inferior care in a public hfaclin :an pa! for better government guiran[ees free care. However, as we care durine a doctor's pri%arc office hours. In stll have oeen. even with HI\ prc,alence rates of s per- other cases. side pa ymenis [o a nurse or other gate- cent or less. dhe demand for medical care is likelY to ke.per arc necesarn in order to ger access to free' increase faster than the go-ernment s abilirs' to suip- care. The etftective price of health care to rhe con- ply i. W\hen hils happens. nieans ot rationing med- sumer is che %.aue of all of the coinsumer s sacritices. ical care other than price come into plas. People in in time and in mone%. needed to obtain care oft a countries w%here health care is officially 'free" are gien qualirv. The AIDS epidemic increases the familiar with these mechanisms. Some sv,tems rely effectme price. even it care is supposed.l% t'rec. 193 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Table 4.4 Evidence of Possible Crowding Out of HIV-Negative by HIV-Positive Patients, Si Countries, circa 1995 Perrnetage of beds otfY'Epied by City Ho.pital Hf -positile pareiets ChLing hNai Fhailand Pn-.un i J kinh.pa, CuniZo [)R' 1iLna \ eni ilJ KivJ,b. R.sanda (Central (.l Buiumbura. BLIrLrldi Prince Regenr tl N.urobi. Kmn% i Kcn,atta Nau!.nil Hopital i Kaimpala. Uganda Rubaga. Hospiiil it a. Fo,rmctrI Zijri b S,ncu I ..d &nd Gnilk c,ujr,J f.h- i-r- kngrh ..4 rar rE be iznrucal a HIV- po.ito land n*r at,, pcienia. rIc r,Alo 4t HlR *p.'rvt [1 tuIl 3dmri'nun ' , a uicfi.l . irr.ar, :f rh., pr. p.OW-in o.- bce .-Cupwd E,% HV %-F . tric parrtrni Thu rhic enirr o.i uj.ij,d troml ! gurc - _` is ' d ' 1. * i,r . [ I.,,,r h.t prtr i n. P. in m !'k;O rcn. i, H.--pitrl Flod inId Gilk, 1,'Rrm !ubigi Ho4piril Iermn.o mad ochers I'l If the hospitals were operating well below capacity before the epi- demic, they might have accommodated the HIV-positive patients with- out reducing care for HIV-negative clients. Although no data on occu- pancy prior to the epidemic are available for these specific hospitals, bed occupancy rates in such hospitals typically were well above 50 percent even before AIDS.8 The best evidence that AIDS is making it more difficult for people not infected with the virus to get medical treatment comes from an in-depth study of Kenyatta National Hospital (KNH), the premier teaching hos- pital in Nairobi, Kenya. The KNH study compared all patients admitted during a sample 22 days in 1988 and 1989 with all patients admitted dur- ing a sample 15 days in 1992 (Floyd and Gilks 1996). Panel A of figure 4.2 shows that while the average number of patients admitted per day in- creased from 23 to 25, the number of HIV-positive patients more than doubled, while the number of HIV-negative admissions shrank by 18 percent. Since the number of HIV-negative people in the hospital's "catchment area" could not have shrunk by this much, this evidence sug- gests that the AIDS epidemic did in fact result in some HIV-negative pa- tients being dissuaded or barred from admission to the hospital. There are no data on what happened to the HIV-negative patients who were not admitted. But hospital records show that the mortality rates for those who were admitted increased between the two periods, I94 COPING WITH THE IMPACT OF AIDS Figure 4.2 Impact of AIDS on Utilization of and Mortality at Kenyatta National Hospital, Nairobi, 198&'89 and 1992 Average patients admitted per day Percentage of patients who died 30 25 = ~23.0 244 9 - 430 36 35 m988 25 3_ 1___ 20- 10 18.7 ~~~~~~15.3 1 WIV- HIV- 1 5~~~~~~~~~~~~~~~~~~~ 1968/89 1L992 HIV+ HIV- Panel A. Utilization increased for HIV-positive patients but fell Panel B. In-hospital mortality remained constant for HIV- for HIV-negative patients between 1988 and 1992. positive patients, but rose by 66 percent for HIV-negative patients Source: Floyd and Gilks 1996 between 1988 and 1992. from 14 to 23 percent (panel B of figure 4.2). The mortality rate for the Data from Kenyatta National Hospital HIV-positive patients did not increase, and other indicators of the qual- on admissions and mortality suggest ity of care remained constant. Thus, the most likely explanation for the that the increased need to care for people with HW has squeezed out increased mortality rate among the HIV-negative patients is that the some HIV-negative people who would rationing scheme used to allocate increasingly scarce beds had the effect otherwise have received care. of changing the mix of HIV-negative patients toward those with more severe illnesses. 'Whether the rationing was imposed by hospital staff or was a response by prospective patients to their perception of a higher effective price of care (box 4.4), it is likely to have excluded some patients whose lives the hospital could have saved. Since the HIV-infected make up an increasingly large fraction of the sick people in a severely affected country, it is appropriate that they occupy an increasing share of hospital beds and consume an increasing share of health care resources. The pressure of this increased demand for care will naturally be felt by all citizens, whether or not they are HIV- infected. However, the extent of the shift in health care resources away from the HIV-negative can be exaggerated, as indeed it may have been in Kenyatta National Hospital, if the government provides special subsidies for people with HIV9 We discuss this issue, and the broader issue of how the level of government health care subsidies affects the demand for care and health care expenditure, in the next section. 195 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Policies To Mitigate the Impact on the Health Sector Scarcer and more expensive care and increased total health expendi- ture present society with difficult choices. Because a large share of the increased expenditure is typically financed through tax revenues, govern- ments and their constituencies will confront tradeoffs along at least three dimensions: * treating AIDS versus preventing HIV infection * treating AIDS versus treating other illnesses * spending for health versus spending for other objectives. The need to confront these difficult choices can be reduced somewhat if a government is willing and able to increase tax revenues. But few countries will be able to avoid the choices entirely, especially developing countries facing a severe epidemic. Unable to pay for everything, most governments will subsidize some goods and services more than others, thereby disproportionately benefiting certain groups of citizens. As the number of AIDS cases increases, governments are likely to face mounting pressure for two responses that on first consideration seem rational and humane. One is to pay a larger share of health care costs; the other is to provide special subsidies for the treatment of HIV/AIDS. Un- fortunately, these responses can have unintended consequences. For rea- sons discussed below, governments that wish to minimize the impact of HIV on the health sector should try to avoid both courses of action. However, this does not mean that governments should do nothing to help alleviate the suffering caused by HIVIAIDS. The section concludes with a list of compassionate and affordable measures that governments can and should undertake to mitigate the health sector impact of an HIV/AIDS epidemic. No increase in the overall subsidy to health care. One obvious and politically appealing response to the HIVIAIDS epidemic is to increase the government share of health care costs and thus the overall subsidy for health care. Such a course of action may be especially attractive early in the epidemic, when few people are sick with AIDS. There is an argu- ment for it on economic grounds as well: it would fill the gap created by the failure of the private market to offer health care insurance in poor countries. However, increasing the subsidy to curative care increases the demand for a limited supply. As a result, both effective price and total ex- penditure will rise by a greater proportion than the increased subsidy I96 COPING WITH THE IMPACT OF AIDS alone, or the increased demand arising from the epidemic alone, or even the sum of the two, would suggest. As more and more people become sick with AIDS, this effect becomes evident in escalating health care ex- penditures; in a severe epidemic, the burden on the government budget is likely to become unsustainable. To understand how changes in the level of government subsidies affect the impact of the epidemic on the health care sector, we first look at the extent to which governments already subsidize care. Then, taking India as an example, we project the impact of an expanding epidemic at the current subsidy level and an increased subsidy level. As we shall see, increasing the overall subsidy to care can greatly exacerbate the impact of the epidemic on the health sector. Most governments subsidize a large share of health care expenditures. The balance includes payments by private insurers and all "out-of- pocket" payments at private or government-subsidized facilities, whether traditional or modern. The average overall subsidy to health care varies widely but generally rises with GDP. As can be seen in figure 4.3, the poorest countries, with average per capita income of about $600, typi- Figure 4.3 Public Share of Health Expenditure in Selected Countries, by Income Group, Various Years, 199097 Public health expenditure/total 1 0.82T 0.83 0.79 0.76 0.76' 0.75- 0.66 0.65 0.58 0.58 0.5 - 0.46 0.47 0.35 0.25 The public share of health care 0 expenditures tends to be higher in countries with higher incomes. Low Lower-middle Upper-middle High Income Note: The middle line in each box shows the median proportion subsidized; the top and bot- tom of the boxes are the 75th and 25th percentiles, and the "whiskers" give the minimum and maximum subsidy rates observed. Source: World Bank data. 197 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 4.4 Simulated Impact of a Severe AIDS Epidemic on Health Expenditure, India, 1990-2010 Government health expenditure (billions of 1996 dollars) 40 _ _ _ _ _ _ _ _ _ _ _ _ _ -4 Expenditure with AIDS 35 at 50% subsidy * Without AIDS at 30 50% subsidy .- - - - With AIDS at 21% 25 -_ -- 0 subsidy 0 _ Without AIDS at 21% 2: - 00 ~~~~~~~~~~~~~~~~~subsidy 10 5 0 - l l 1990 1995 2000 2005 2010 Year Note: The projections follow box 4.3 in assuming that the elasticity of demand for healthcare is 0.8 and of supply 0.5. If the demand elasticity in India were smaller or the supply elasticity greater than these assumptions, all the expenditure impacts would be correspondingly smaller. Source: Ellis, Alam, and Gupta 1997; authors' calculations. Hf India maintains its current level of cally subsidize less than half of the cost of health care, while upper- health care subsidies, a severe AIDS income countries subsidize about three-quarters of the cost. epidemic would increase government In India in 1990 the government subsidized about 21 percent of total health care expenditure by about $2 billion per year by 2010. If subsidies health care expenditures, a small share even compared with other low- are increased to the 50% level, the income countries. The bottom line of figure 4.4 projects government same size epidemic would increase health expenditure if India were to have no AIDS epidemic and contin- annual government health expenditure by an additional $30 billion. ued to spend 6 percent of a constantly growing GDP on health care, of which the government continued to finance 21 percent. In this baseline scenario, India's government health expenditures grow from $3.2 billion in 1991 to $8 billion in 2010. The second line from the bottom shows the increase in government health spending if India's current steep rise in HIV prevalence continues until 2000, then levels off at a stable 5 per- cent. This is about the growth in prevalence seen in countries such as Zambia and Botswana, where focused prevention was not implemented early in the epidemic. The result in India would be to increase the gov- ernment's expenditure on health in the year 2010 by about one-third, from $8 billion to $10.5 billion. What if India in 1990 had increased health care subsidies to about 50 percent, the level seen in many Latin American countries? The top pair of projections in figure 4.4 shows the impact of the higher subsidy on expenditure. Even without an AIDS epidemic, expenditure more than I98 COPING WITH THE IMPACT OF AIDS triples to $11 billion in 1991 due to the more than doubling of the gov- ernment's share of existing expenditure combined with the demand stimulus caused by the greater subsidy. Subsequent growth of health expenditure proportional to GDP brings health expenditure to $27 bil- lion in 2010 (third line from the bottom). Now again suppose a serious AIDS epidemic that reaches a stable 5 percent HIV prevalence rate in 2000. The fourth line from the bottom of figure 4.4 gives the projected result: health expenditures in 2010 would reach $39 billion. Thus, not only has the increased subsidy tripled health care spending, as might have been expected, but it has also increased the vulnerability of the bud- get to the AIDS epidemic, adding $12 billion (43 percent of $27 billion) rather than just $2.5 billion (31 percent of $8 billion) to government health care expenditures. The large expenditure shocks that will result from the AIDS epidemic will create new pressures on health budgets, especially in countries that enter the AIDS epidemic with higher subsidy rates. For example, al- though Mexico's infection rate was estimated to be only 0.4 percent in 1994 and it subsidized only 49 percent of the cost of AIDS treatment, compared with 76 percent for other sicknesses, AIDS was already con- suming 1.2 percent of its health budget. In contrast, Tanzania has kept the subsidy rate for AIDS treatment down to 28 percent in line with the subsidy it provides to other illness categories. As a result, despite a preva- lence rate of 5 percent, more than ten times higher than Mexico's, the AIDS share of total government health care expenditure is only 3.5 per- cent, just three times larger in Tanzania than in Mexico.10 Although a discussion of the design of health financing systems is beyond the scope of this book, the evidence suggests that countries in the nascent or concentrated stages of the epidemic, like India, should care- fully consider not only the immediate budgetary consequences of any expanded commitment to fund curative care, but also the multiplication of these consequences that would occur if the AIDS epidemic spreads. A prudent course would be to consider any expansion of government- financed health care subsidies or insurance only in conjunction with vig- orous prevention programs that enable people most likely to contract and spread HIV to protect themselves and others. Equal subsidy rates regardless of HIV status. A second common health sector response to the HIV/AIDS epidemic is to offer a different subsidy rate depending on whether or not the person receiving care is infected with HIV. Especially in the countries in the nascent stage of the epidemic, HIV-infected people all too frequently experience discrimi- 199 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC nation, including restricted access to or higher effective prices for health care. As the epidemic advances, however, governments are often pressed to provide special subsidies for the treatment of HIV/AIDS. This sec- tion points out the government's role in limiting discrimination against the HIV-infected in health care settings and then considers the conse- quences of preferential subsidies for HIV treatment. AIDS treatment subsidies vary greatly from country to country. Fig- ure 4.5 presents data on the percentage of AIDS-related and total 1994 health care expenditure funded by the government. In three of the five countries, the subsidy rates for AIDS treatment are significantly different from that for total health care expenditure. For example, although Mex- ico subsidized a generous 49 percent of the cost of AIDS treatment, this was much less than the 76 percent share of total health care expenditure. Brazil and Thailand subsidized AIDS care at a higher rate than all types of care, while Tanzania and C6te d'Ivoire subsidized AIDS treatment and total health care expenditure at roughly the same rate. A bias against those with HIV/AIDS can take many forms, ranging from a singling out of AIDS-specific drug therapies for exclusion from public funding, to outright refusal of service. There are many anecdotes Figure 4.5 Percentage of AIDS-Related and Total Treatment Expenditures Financed by the National Government, Four Selected Countries and Siao Paulo State, Brazil, 1994 Public subsidy as % of total expenditure 80 | I AIDS expenditures 78.1 76 0 All health expenditures | 60 55 48*7 49 so 43 42.0 40 ~~~~~~~~~~~~~36.1 28.5 28 Govemnments often provide differnt 20 20.4 levels of health care subsidies depending on whether or not the patient is infected wifth HIV. 0 Tanzania Cote Thailand Mexico Sio Paulo Average d'fivolre State, Brazil Source: Backgroundpaper, Shepard and others 1996. 200 COPING WITH THE IMPACT OF AIDS about discrimination against the HIV-infected in health care settings. In some hospitals, the HIV-infected were placed in special AIDS wards, which were subsequently shunned by fearful health care workers. In others, the HIV-infected were required to pay extra costs for rubber gloves or a private room. In still other cases, the HIV-infected have been denied treatment for common illnesses, perhaps because doctors and nurses mis- takenly believed that nothing could be done to help a person with HIV/AIDS. Such discrimination is unfair, unprofessional, and unethical. Moreover, it displays ignorance of the many ways, discussed above, in which inexpensive treatments for symptoms and opportunistic illnesses can prolong and improve the lives of people with HIV/AIDS. Govern- ment has an important role to play in training medical personnel in order to eradicate all vestiges of discrimination against HIV-infected patients. Yet it is equally unfair, and also inefficient, for government to subsi- dize a higher proportion of the costs of care for patients with HIV than for other patients. Aside from the issue of poverty, to be addressed in the next section of this chapter, there are three ways to justify government subsidies for curative health care: (1) as an incentive for those with an in- fectious disease to seek a cure and avoid infecting others, (2) as health care insurance with universal coverage and mandatory participation through general taxes, or (3) as government support for a "merit good" or "basic need." No treatment has yet been shown to reduce the infec- tivity of sexual contact with an HIV-infected person (see box 4.5). AZT treatment of HIV-infected pregnant women has been shown to reduce transmission at birth, but is still too costly an approach to preventing secondary infections in the poorest countries (see box 4.6). With the prominent exception of TB, the treatment of which should be subsi- dized in all countries, most opportunistic illnesses that afflict the HIV- infected are infectious only to other equally sick HIV-infected people. Thus the argument for treating the diseases on the grounds that they are infectious is weak. If government subsidy is considered as an insurance payment, efficiency criteria argue for a higher coinsurance (i.e., lower subsidy) rate for any condition in which the patient is likely to be highly price-responsive.1" The first section of this chapter established that the drugs and medical services to treat AIDS can amount to a great deal of money, although some of the most expensive of these treatments pur- chase the patient little additional life span and decrease, rather than im- prove, the quality of life. Thus on efficiency grounds, where the objec- tive is to limit the responsiveness of expenditure to insurance, AIDS 201 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 4.5 Is Antiretroviral Therapy an Effective Way To Prevent Sexual Transmission? L'Ni-lL RF,ENTLY. Ai -iRETR)l)R RkE .\R\, THERAIfl more cases if spent on tbcused prevention in high- could iior be considered aS a possible csis [ prevent risk groups. Furthermore, we saw% in chapter I rhat sewual transmission because the available drugs For esen without dhe expense of antirerromiral rherapy'. treating HIV:AIDS had little impact on infect!virv. current expenditure for treating an AIDS patient The I 9L11 discovery thit protrease inhibitors and wkould bus a sear of primary school for ten stu- triple-drug therapy suppress HIM bcloss the level of dents in most developing countries. In the poorest rhe mostr ensitive blood rescs to derect it has raised countries, che much higher cost of antirerrosiral hopei that these drugs might pre%ent the spread o f therapy would bus a sear ot priman' school for '400 Hi\V. in addition to greatly extending the lite of the students. For this reason. even if che cosc of anti- patient. Ben if this proves true. hossever. polics-- retroviral Lhtrap- is shossn to reduce the infectiviry nakers deciding x'herher to provide public subsidies of sexual contacLs. and even if the cost falls substan- WIll need to consider that the $ 10.000 to 5(0.000 tiall. decisionmakers will still want to consider 'en- cost of treating a 'inglc patient wo.uld prevent mans' careFully btfore initiating such subsidies. patients should face somewhat lower subsidies, not higher ones. The final possibility, that AIDS treatment is a basic need, is difficult to justify in poor countries where the opportunity cost of treating one adult for AIDS may be measles vaccines for 100 to 200 children or, as shown in figure 1.8, ten student-years of primary school. Thus none of these eco- nomic arguments justifies higher subsidy rates for AIDS. What policy recommendations can be drawn from these two obser- vations? The prudent, efficient, and equitable course is to place the financing of health care for HIV/AIDS on the same footing as other dis- eases. The treatment of particularly infectious illnesses striking HIV- infected people, including TB and STDs, should be subsidized relatively generously because of the secondary infections treatment will prevent. Other health care problems of the HIV-infected should be subsidized at the same rate that applies to other adult health problems that are equally infectious. Assuming that Brazil subsidizes about one-third of other health care costs (as figure 4.5 indicates it does in Sao Paulo) and that the infectious proportion of illness episodes is similar among the HIV- infected and uninfected populations, this policy would lead Brazil to re- duce its subsidy to antiretroviral therapy from 100 percent to one-third. Similarly Thailand would reduce its subsidy of antiretroviral therapy from 100 percent to about 20 percent. Mexico, on the other hand, 202 COPING WITH THE IMPACT OF AIDS would increase its subsidy to AIDS patients to approximately the same rate it offers other patients. In mid-1997, none of these countries appeared to be following this recommendation precisely. Brazil and Mexico were continuing their for- mer policies, with a tilt of subsidies toward AIDS treatment in Brazil and away from it in Mexico. Having spent $108 million on antiretroviral medication in 1996, Brazil was projecting an expenditure four times that large for 1997 (Chequar 1997). Thailand had recently embarked on an experiment that held out the possibility of an equal percentage subsidy for the treatment of AIDS and other diseases, on average, if not for the individual patient. In 1996, the Thai Ministry of Public Health found that, given the rising patient load, its policy of a 100 percent subsidy for antiretrovirals and drugs for the opportunistic illnesses would soon con- sume considerably more than the entire budget allocated to the National AIDS Program (Prescott and others 1996). As a result, the government revised its policy to provide free antiretroviral therapy only to HIV- positive pregnant women, where it might prevent mother-to-child trans- mission, and to participants in nationally approved clinical trials, where patients receive the support they need to maximize compliance (Kunan- usont 1997). This policy makes sense for antiretroviral therapy, on the assumption that low levels of compliance outside clinical trials would have little therapeutic effect on patients and might cause negative exter- nalities in the form of drug-resistant strains of HIV Furthermore, par- ticipants in clinical trials produce a positive externality in the form of the knowledge that can be used to benefit many other patients, and therefore should receive a higher subsidy than other patients. Thailand's decision to subsidize AZT for prevention of mother-to-child transmission can be justified as a "merit good," which might be affordable in a middle- income country (see box 4.6). Affordable, humane responses to the epidemic. We have argued that governments should avoid two types of health care responses to the epidemic: increasing the overall subsidy to all types of treatment, and providing disproportionately large subsidies to treatment of HIV/AIDS. There are nonetheless several ways in which governments can intervene to mitigate the health impact of HIV/AIDS on infected individuals and their families and on the overall health sector. Each of these interven- tions is justified on public economic grounds, either because it has large positive externalities, or because it improves the efficiency or the equity of the health care market in other ways. 203 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 4.3 Estimating the Impact of AIDS on the Health Sector HO¼\X NIL;EH WXILL THE L F-FECTl- E PRICE OF CAPFE sUbsid& on consumersi.) The ligure is constructed so increase as a result of.MUDS' Box figure -d.3 demorn- that the market equilibrium occurs atr a price or 10 srarces h1ow% the appro.imrtc- size of these increases currency% units per unir of health care. at ; hich a can be estima[ed for a hypothetical country- %%ith total of 10 units of care art delivered. Total health elasticities ot demand mnd supply for health care of care c pendirure in this hypothetical country is thus 0.8 and l.i and a government policy tco ubsidle 10 times 10 cr IJ0 currencN units in the abwnce ot half of the cost of care. The rtwo solid lin'c sho% the trhc AIDS epidemic. amoUrnt of health care that is demanded and sup- Nowx- assume thcre is an HIl\IAIDS epidemic plied at each price prior to an HIM cpidemic. iThe that levels off at a constant seropre%alencc of per- demand curse is drawn %%ith an elasticiry of only 0 4 cent of the adult popullation. The argumen[s in the in order to Incorporate [he electi of the government chapter suggest that the amount of health care demanded at escrn- price is likc!l to incrcase by 25 Box Flgure 4.3 The Impact of a 5 Percent percent. wvhile the cort of purchasing any gpiven Infection Rate on the Quantity and Price of aimount of care of a given quaity vhdill increasc b% l'J Heafth Care pcrceint. These cw. Iimpacts of the AIDS epidemic are illuctrated by- a rlghrsvard hilft of the demand Price of health care Demandfor care cUrve bi 2i percent Ito the dashed. do%%nw%ard- ~~~ ~~after AIDS - -> afterAIDS sloping line) and an up% ard shit- of [he supply curve 13.- s > o by I( percent IIt. the dashed. upwvard-sloping inil The impacts on the cquilibriunm price mnd quantity he care A - . can be read from the figure. The price of a unit ot healch care will increase about 31) percent. and the of caie ammount of care provided %%ill Increase about Iii per- 10. h- Suppyaol after AIDS cent. T-otal national experidirure. the price per unit betore AIDS Fot care times the number of units wvill increase -i3 percent to 143 currenn- units isince 1 3s \I I = 14 3i lo 11 Quantlt ot Wec have wecn that a third-parn- paiyment, such as insurance or a gin%ernmenL subsidy for treatrment. Xac The Jlenmnd ind -uppl ciarce. are con,rrucid !o ia.r nmakes peoplc less sensitie to changes in coist ot rhe pril elsr.miire, ai theh porr 0IC) IUl are I.s snd i S. health care By reducing the The mpacr ' -IDS *ilec,irared 1are Sc thelJcag the price elastict of de- cr%e to The nghr b. ot A eDni at er- prie Ill he upf.In mand. such third-parr payments make the demand c -ent LIi hr l IpIcr'-I ir e Ite"<.L]uJn L e (he ,c\ t:t ter a. cLure steeper. both betore and aftier the introduction cmpianatln ..1 thse a;-umpi;n,s1 ot l D)5. where only about one-fifth of the cost of care is paid by the government, and substantially more in countries like those of Latin America and East- ern Europe, where three-quarters or more of the cost are subsidized. Does the available empirical evidence support these conclusions? Although there are significant data problems, the short answer is yes. 192 COPING WITH THE IMPACT OF AIDS media channels-for example by issuing press releases and arrang- ing media interviews with credible experts-it can be done quite inexpensively. * Subsidize the treatment of infectious opportunistic illnesses and STDs. Subsidized treatment is especially appropriate for tuberculosis, one of the most common opportunistic illnesses to infect AIDS pa- tients, since curing a single case can avert many secondary infec- tions. Treatment of gonorrhea, syphilis, and the other classic STDs should be subsidized, not only because they are highly contagious, but also because they exacerbate HIV transmission, as discussed in chapter 3. Because few people are susceptible to them, treating toxoplasmosis, cryptococcosis, or one of the other infectious op- portunistic illnesses that develop only in people with severely dis- abled immune systems prevents few secondary cases and thus should be subsidized at a lower rate, closer to the subsidy rate for chronic, noninfectious disease. Whether a subsidy to antiretroviral treatment of HIV itself is justifiable as a way to prevent secondary HIV infections will depend on the efficacy of the treatment and on its cost relative to the cost of other HIV prevention measures. In mid-1997, such treatments were far too expensive and uncertain to warrant subsidies on these grounds (see box 4.5). * Subsidize the start-up costs for blood safety and AIDS care. The AIDS epidemic has increased the willingness of individuals to pay for certain types of services, such as screening of blood for transfusions and care for the terminally ill. Where these services are lacking, government help with the start-up costs is justified, just as governments subsidize other large indivisible investments, such as an electric utility or a water system, so long as the users then pay for the services they receive. Thus, governments in poor countries should establish blood banks but should not indefi- nitely provide free blood. Similarly, government should help es- tablish AIDS treatment facilities, especially community-based home care programs, but should not permanently subsidize the care they provide. * Provide special assistance to the poor. Most countries already make special provision for medical care to the poor. As the AIDS epi- demic increases the demand for care, governments may wish to focus such assistance even more on those who can least afford it. Sliding fee scales and other measures to make care available to the 205 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC poor should apply to people with HIV/AIDS just as they do to people with other illnesses. This principle of providing assistance to those who need it most, regardless of their H1V/AIDS status, is discussed more fully in the next section on ways to mitigate the impact of HIV on poverty. AIDS and Poverty: Who Needs Help? IN ADDITION TO ITS DEVASTATING IMPACT ON INFECTED INDI- viduals, HIV hurts all those who are linked to them by bonds of kinship, economic dependence, or affection. The grief suffered by survivors, and the possible lasting psychological damage, especially to young children who lose a parent, are potentially the most damaging consequences of the epidemic. They are, however, difficult to measure, probably unreachable by public policy, and therefore beyond the scope of this book. In addition, survivors often suffer economically. This harm from a prime-age adult death constitutes the most important economic impact of an HIV/AIDS epidemic and is the topic of the remainder of this chapter. It can be measured by the impact of adult death on such social indicators as orphanhood, child nutrition, schooling, and poverty. By worsening these measures and widening the gap between the poor and others, HIV can exacerbate poverty in poor countries and delay attainment of national economic development goals. We look first at how HIV/AIDS affects poverty, then at the implications of these find- ings for poverty policy in a severe AIDS epidemic. How HIV/AIDS Affects Poverty It is sometimes said that "AIDS is a disease of poverty." In what sense might this be true-or false? First, are the poor more likely to become in- fected with HIV than others? Second, what proportion of people infected with HIV are poor? Answers to these questions are important, for they will influence both the focusing of prevention measures and attempts to mitigate the impact of the AIDS infections that do occur. In considering the impact of AIDS on poverty, we first examine the available evidence to answer these two basic questions; then we ask how the impact of an AIDS death compares with other shocks that households suffer, and how households of different income levels cope. 2o6 COPING WITH THE IMPACT OF AIDS HIV infects the rich and the poor. In developing countries, the rela- tionship between income and HIV infection rates has been best docu- mented in eastern and central Africa.12 Whether the patterns observed in this part of Africa will also emerge elsewhere remains to be seen. Sev- eral factors have exacerbated the epidemic in hard-hit areas of Africa: most people who had HIV at the time of the studies had become in- fected years earlier, when little was known about HIV prevention; more- over, the area is traversed by major transport routes and has suffered from war. Yet each of these factors is also evident in other developing regions to varying degrees: knowledge about HIV prevention is still often scant, and other regions also have major transport routes and wars. Thus, until other data are available, the experience in eastern and central Africa may offer worthwhile clues about how infection rates in other regions are likely to differ across income groups as the epidemic progresses. As we learned in chapter 3, early in the epidemic in Sub-Saharan Africa men and women who travel more, and men who had higher in- comes, were more likely than others to contract the virus. There are rea- sons to believe that this may hold true elsewhere. Studies show that sex is similar to other pleasurable pastimes: the number of partners per year rises with income. Also, a person with a higher income is likely to attract more prospective partners, and will have more money than a person with lower income to compensate sexual partners or to support any offspring. These factors, combined with the fact that HIV, unlike other STDs, can- not be readily cured, has made HIV unique among widely prevalent in- fectious diseases in striking rich people in the same proportion, or larger proportions, than it strikes the poor. That HIV infects the rich as well as the poor is important to keep in mind when considering which house- holds need help the most. Of course, we would expect that more-educated people with higher incomes would be in a better position to learn about the epidemic and alter their behavior to avoid infection. Chapter 3 presents evidence that this is already occurring: in some countries, highly educated people have higher frequencies of condom use than the less well educated. Also, recent studies in developed countries have shown AIDS incidence to be highest among the very poor. If these trends are replicated worldwide, AIDS will become like other infectious diseases, in that the poor will be more likely to become infected than the nonpoor. Ultimately, AIDS may become most prevalent in the poorest urban slums of developing countries. Already, most people with HIV/AIDS are poor. Although lack of data makes it impossible to calculate the precise proportions of poor and non- 207 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC poor who are infected, knowledge of income levels and infection rates across countries suggests that many more poor people are infected than nonpoor people. For example, according to an internationally adjusted standard of absolute poverty, Sub-Saharan Africa has about four times as many poor people as nonpoor people. Thus, even if poor people were infected at just slightly more than one-quarter the rate of the nonpoor, poor people would account for the majority of HIV infections in Africa. Since poor people in many parts of Sub-Saharan Africa have infection rates that are well above one-quarter the rates of the nonpoor, we know that, in Africa, at least, there are many more poor people than rich peo- ple with HIV; Although magnitudes are less striking, the same general principle will tend to apply in other developing regions.13 We have seen that AIDS is already a disease of poverty in the sense that it affects more poor people than nonpoor, and it may eventually become a disease of poverty in the sense of infecting a higher proportion of poor than nonpoor. If we assume that one of government's main responsibili- ties is to make it possible for people to escape from poverty, these findings lead us to new questions. What is the impact on a poor household when the mother, father, or another prime-age adult who is a member of the household dies from AIDS? How do poor households cope with AIDS deaths? We examine these questions in the next two subsections. Box 4.7 describes three sets of characteristics that determine the initial impact of an adult death and how well an afflicted household copes. What is the direct impact of an AIDS death? The death of a prime- age adult is obviously a tragedy for any household. Survivors must con- tend not only with profound emotional loss, but also with medical and funeral expenses, plus the loss of income and services that a prime-age adult typically provides. How serious is the shock of an AIDS death to the economic welfare of the survivors? The direct impact of a death con- sists of the medical costs prior to death and the costs of the funeral. To assess the direct cost of a death from AIDS, we compare the medical and funeral costs of an AIDS death with those of a prime-age adult death from other causes. Finding that the difference is not large, we then con- sider how the death of a prime-age adult, regardless of cause, affects household consumption patterns. Our analysis is based on findings from several household surveys de- scribed in box 4.8. In particular, we rely on the most extensive of these, a study done in Kagera, Tanzania, since detailed data from that study provide a basis for our subsequent analysis of how households cope with zo8 COPING WITH THE IMPACT OF AIDS Box 4.7 Three Factors Determine the Household Impact of a Death ITHE OVERAI ECUNON(MIK iRPA.I F .AN ADULT DFAI H ON THE surv iing household members varies according r0 three sets of characteristics: * Lhoie of the deceased individual. ,uch is age. sex. income. and cause of death * those of the household. such as composition and assets * chose of the communiit,. such as attitudes to%sard helping needy households and the a%aulabilitv ot resources. Thu firs; set of characteristics determines the basic impact of the death on the survi%iny household members: the second and third influence ho%k w%ell the afflicted household copes. AlJhough disentan- gling the three is ven difficult, it is nonetheless impor[ant %%hen [rtemping to assesS che household impact of an adult death to con- sider all three sets of fictors. AIDS deaths. Although the data are very limited, based on the available information, it is reasonable to expect that the impacts and coping re- sponses described in this chapter will prove to be broadly consistent with future findings. In the Kagera study, people diagnosed with AIDS were somewhat more likely to seek medical care than people who died from other causes, and they were more likely to incur out-of-pocket medical expenses.'4 Moreover, household medical expenditures tended to be much higher for AIDS than for other causes of death, as shown in figure 4.6. Strik- ingly, for all groups except men with AIDS, medical expenses were over- shadowed by funeral expenses. On average, households spent nearly 50 percent more on funerals than they did for medical care. Moreover, funeral expenditures for AIDS deaths and non-AIDS deaths differed less than did medical expenditures. Thus, even though a significant propor- tion of funeral costs were covered by gifts from other households (about 45 percent on average), the difference in the household impact of an AIDS death and a non-AIDS death is smaller than the differences in medical costs alone would lead us to expect.'5 209 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 4.8 Studies of the Household Impact of AduH Death from AIDS and Other Causes WITHIN THE PAST FEV' YEARS, FOUR IN-DiEPTH srudv included dearhs from causes other than AIDS -rudies hate examined (he impact of adult dearh as "ell as those from .AIDS. from AIDS on sur'iving household members. Tw%o broad findings emerged from these srudies Compared with ocher studies, these four studies and are discussed in the texr. First, houiseholds use a used more derailed survey in,rruments: applied these variert of informal mechanisms to cope wvith misfor- to larger. more representative samples of households: runes like an adult death in the household. Second. and follovwed the household, over lonpgr periods. although these coping mechanisms cushion the The four studies were carrie.d cuc in thc folloving impact of [he shock-, households are nor entirely suc- locarions tche number of' households surveyed is cessful in protecting their well-being. In general, rhe show-n in parenthestsi: poorer the household. the yreater and more persis- tent the nipact of a prime-age adult death from * Chian,og iai. Thailand 300) AI[)S and similar shocks. * Abidjan. C6ted'Ivoire (107) * Rakai. Uganda II (.&6 7 * Kagera. Tanznmia i -5t)1 1 AILthoLOh thle Ca.rxd I"n!r- rudv did noi include an explici control croup. Bechu Iwqnnn.ewpid pIer 19(6i 1 able ro use dif- IticFrnc ae.s, h0uutholds in the wtventr offthe AIDS cases corn- The studies used many sirilar parameters. Al bined % rh th cquence ol r., ouber.ariont, on eich household io but the Thai study visited the sample households eqtrmac Ehcm irpact cif Fil adutlr dr.L;n oncormmpuon. s r th RAu ttud; u pan .1 siud of rhe cffiect o! rmas STD sev-eral Linie-. All but the Uite d voire srudY Ill- ircairrni o. hen incidence oF.. TIDS The houschold qunuon- cluded households that did not experience an AVIDS nure t£c.a't nn epidemiologtical "ues and onhv asked i ft% sickness as w%ell as tlhoic that did.' Al but the Raklai qu-rion' reIed wit,c1!noinic -1i-bKing srudv, w%ere done ewpresoly to study social and eco- F hi Pa.anon. n nd Jlni3rien , : . . ~~~~~~1 l-11)-) and mlinii enlt'.'7't''>s} 199-51. to -10i<.c '.:l nomic impact and thus had extensive questionnaires 1-hu 0 'ibi. t;r Uganda. Nienon and orher about consumption and other social and economic 19a, an;-i,sr i'96a: and For TanzaiJ, O.ver and nEhers measures of wvell-being.' All but the C6re d'Itoire ifortEC-mVnE In Thailand, where per capita income is 10 times that in Tanzania, households in Chiang Mai province spent more than ten times as much on medical care prior to death as did the Tanzanian households (Pitay- onon, Kongsin, and Janjaroen 1997). The households with an AIDS death spent $973 on average, which in contrast to Tanzania was only about 10 percent more than the $883 spent by the non-AIDS house- holds. But, just as in Tanzania, the households spent much more on fu- nerals than on medical care.16 The relative amounts spent for medical care and funerals will, of course, vary from country to country and even across communities within a district. Nevertheless, two broad observations are likely to apply in most situations: first, medical costs are only a portion of the cost of 2.10 COPING WITH THE IMPACT OF AIDS Figure 4.6 Average Medical and Funeral Expenditures, by Gender and Cause of Death, Kagera, Tanzania, 1991-93 1996 dollars 160 120-so80 25 Medical expenditures were higher for 80 - . people who died of AIDS than people 22 1 38 who died of other causes. But because funeral expenses are large, the overall i difference in expenditure for AIDS 40 . ... 77 ..... 73 ..... deaths and deaths from other causes 56 54 is smaler than the di-ference in medical expenditure alone would O. _ _ _ __ - _ _ _ _ suggest. Males Males Females Females w/o AIDS with AIDS w/o AIDS with AIDS Note: Throughout this report currency amounts have been converted from current Tanzanian shillings to 1996 U.S. dollars. The conversion procedure involves three steps: (1) convert current shillings to 1991 shillings using the project's price deflator; (2) convert 1991 shillings to 1991 dollar amounts at 289 shillings per dollar; and (3) inflate to 1996 dollars by multiplying by 1.15. Sample: deaths of 264 adult household members ages 15 to 50. Source: Over and others, forthcoming. a prime-age adult death; and second, nonmedical costs are likely to be similar, regardless of the cause of death. Where these observations hold true, the direct impact of an AIDS death will not be much different from that of a non-AIDS death, despite higher medical expenditures for AIDS. Thus, the high cost to households from AIDS will usually be due to the large number of deaths caused by the epidemic rather than by the fact that they are caused by AIDS. Given that the impact of a prime-age adult death is likely to be similar, regardless of cause, how does a prime- age adult death affect household consumption? Figure 4.7 shows house- hold consumption during the previous twelve months for two groups of households in the first wave of the Kagera survey: those who had experi- enced a death during this period and those who had not. Households that suffered a death had lower overall expenditures and, as we would expect, devoted a larger share of the expenditure to medical and funeral costs. Also, these households spent one-third less on the "other nonfood" category (i.e., clothing, soap, and batteries). Finally, in households that suffered a death, food produced by the household was a larger share of 211 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 4.7 Consumption in Kagera, Tanzania, Households by Whether the Household Experienced a Death in the Past Year (Results from Wave 1 of Kagera Study) Households not experiencing an adult death In the past year Other nonfood 26% .7 is .Funerals / />\ Housing & L , ' ~~~~~~~utilities ,_ // \ 12U * i Remittances / : i , . _sent Selected nonfood Health care Home- produced food Purchased Education 36% food 13 dcto 21% Household Breakdown of expendi8Ture) selecteddnonfood (392,838 TSh.) expenditure Note: Of the 6,395 TSh. spent on health care in the average sample household, TSh. 748, or 11%, was spent on the care of nonadult household members who had died. Households experiencing an adult death In the past year Other nonfood 16% Housing & / utilities 124, / 12 - U Funerals Selected nonfood Purchased food accounted for a " , , 10% ' Remittances smaller share of total consumption Home- ,e sent in households that suffiered a death produced (bottom pannl) than in households food Health 43% care that did not (top panel). Households --- Purchased - that suffered a death increased their food Educatlon consumption of home-produced food, Household 19n but this only partially offset reduced expenditure Breakdown of consumption of purchased food. (362,826 TSh) expenditure Note: Of the 9,453 TSh. spent on health care in rhe average sample household, TSh. 6,069, or 64%, was spent on the care of household members who had died. Source: Over and others, forthcoming. 2I2 COPING WITH THE IMPACT OF AIDS Figure 4.8 The lime Pattem of the Impact of Adult Illness on Per Capita Household Expenditure, C6te d'lIvoire Study Monthly expenditure per household mernber (1994 CFA) 14,000 14.000 -fr- Total expenditure 12,000 -0--4Basic needs 10,000 -_0-- Other frequent expnses -X- Infrequent 8,000 expenditures 6,000 4,000 2,000 Following an adult death, expendi- tures for the average bereaved house- o | I x t x t hold decline, then partly recover. 0 2 4 6 8 10 Months Source: Backgroundpaper, Bechu 1996. consumption than in households without a death, while purchased food was a smaller share of consumption.17 These differences reflect the fact that the members of households that experienced a death cut back on the number of hours they worked for wages and thus had lower incomes with which to purchase food (Beegle 1996). They were only partially able to replace this lost income with additional production of food at home. In the two studies that followed, detailed household consumption over time, the Kagera study and the one from C6te d'Ivoire, the time pattern of consumption demonstrates the resiliency of the average household to the impact of the death. Figure 4.8, drawn from the C6te d'Ivoire study, shows changes in expenditure per household member for three components of expenditures, and total expenditure during ten months after the households in the survey lost someone to AIDS.18 Two patterns are immediately evident. First, total consumption dips and then partially recovers, still trending upward at the end of the survey. Second, basic needs, which include food, dip less than other categories of expen- diture, then almost fully recover as families reduce other categories of 2I3 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC spending to minimize the impact on necessities (background paper, Bechu 1996). The household surveys from Chiang Mai and Rakai also suggest a partial recovery in per capita consumption but do not have suf- ficient data to confirm this pattern. How households cope with the impact of adult death. The eco- nomic shock of a prime-age adult death, described above, would have been larger and more persistent, except that households use a variety of strategies to cope. Before the AIDS epidemic, prime-age adult deaths were much less common, so these mechanisms were used mostly to cope with other shocks. As a result, early assessments of the household impact of AIDS tended to overlook household coping and assume that AIDS would be catastrophic not only for the infected individual but also for the entire household. Press accounts that presented devastated house- holds as typical contributed to the widespread belief that most AIDS- affected households in developing countries would collapse. To be sure, some households are destroyed by AIDS; this is especially true if both parents become ill or die while their children are still very young. How- ever, such instances may be less typical than is generally assumed because of the typically long lag between HIV infection and death. Moreover, while premature death of a loved one is always tragic, leading to emo- tional pain and sometimes lasting psychological damage in survivors, survey data suggest that when it comes to coping with the economic impact of such a loss, households in general are surprisingly resilient. The degree of household resilience to the economic impact of a prime-age adult death has important implications for society's response to a generalized epidemic. On the one hand, if nearly all AIDS-affected households collapsed, resources for mitigating the household impact of the epidemic would be stretched so thinly that governments and social welfare organizations would be overwhelmed. In such a situation, policymakers might easily conclude that those currently affected were beyond help and that the only reasonable response for the government would be to redouble prevention efforts. On the other hand, if many households were able to cope, governments and NGOs could focus the limited resources available for mitigating the impact of the epidemic on the households that needed help the most. Understanding the variety of household coping mechanisms and how these will affect different groups of households is important. The mix of responses attempted by a specific household in response to a prime-age adult death depends on countless factors, some of which will vary across 214 COPING WITH THE IMPACT OF AIDS countries and communities. Since the available data on household impact comes mostly from Sub-Saharan Africa, the following discussion un- avoidably reflects this bias. Policymakers in all countries faced with the possibility of a generalized epidemic will want to assess the extent to which these responses are evident in their own country. To varying degrees, how- ever, three coping mechanisms observed in Africa-altering household composition, drawing down savings or selling assets, and utilizing assis- tance from other households-are all likely to be attempted whenever households confront the tragedy of a prime-age adult death. This section discusses each of these informal mechanisms in turn and then discusses formal assistance, such as that provided by governments and NGOs. Before this analysis, it would be useful to consider whether income that had been devoted to health care and funeral expenses could be diverted after the death to other expenses. Some household coping cer- tainly involves such responses. However, the potential should not be overestimated, since much of the cost of medical care and an even larger proportion of the cost of funerals is financed by transfers from outside the household. Since these transfers typically cease after the funeral, households must draw on additional coping strategies, described below. Altering household composition. Households everywhere fulfill eco- nomic as well as social functions. In rural areas of developing countries, households are often the main production unit for subsistence farming and, in some instances, for cash crop farming as well. In such a situation, the economic shock of the death of a prime-age adult can be cushioned to some degree by altering household composition. Examples of such changes could include sending one or more dependent children to live with relatives, or inviting an unmarried aunt or uncle to join the house- hold in exchange for assistance with farming and household tasks. Re- sults from three of the four available household surveys-two from Africa and one from Chiang Mai, Thailand-show that the degree to which household composition is used to cushion the shock of the death varies according to the size and flexibility of local household structure. Among the 759 households in the Kagera study interviewed once every six months for two years, 130 household members of all ages died, but roughly nine times as many people left the households alive over the same period and seven times as many joined the households. In addition, about 200 children were born to household members. As a result, the av- erage size of all the households declined only slightly, from about 6.0 to 5.7 members. 2I5 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC During the six months between any two interviews, economically active adults left or joined about one-fifth of the households that did not have an adult death and about 40 percent of the households that did suf- fer a death. Since most households that suffered a death added at least one member, the average size of these households declined by less than one, from 6.4 to 5.7 members-so that the average household size after a death was the same as in households that did not suffer a death. Simi- larly, the dependency ratio rose only slightly in households with an adult death, from 1.2 to 1.4, slightly less than the 1.5 dependency ratio in households without an adult death. A striking fact is that household size and dependency ratios changed very little, even though Kagera has high adult mortality from AIDS. The same phenomenon was observed in the survey in Rakai, Uganda, which, like Kagera, has a severe AIDS epidemic: 15 percent or more of adults in roadside communities are infected with HIV. This suggests that, even in a generalized AIDS epidemic, most African households that suffer an AIDS death will be able to adjust household size and dependency ratios in ways that make them similar to households that did not suffer a death. The Chiang Mai survey reveals that, at least in this area of Asia, households are much smaller and less mutable than in Africa. The 108 households in the sample that had not experienced a death had 432 members, or exactly four per household. In contrast to the no-death households in Kagera, the Chiang Mai households experienced almost no change in membership, receiving among all of them only one new member and losing only 6 members over the reference period. The 216 households that experienced a death had an average of 4.1 members, of which they lost one each because of the death. Unlike the Kagera house- holds, these Chiang Mai households remained a full person smaller (that is, with 3.1 persons per household) at the time of the interview, which was up to two years after the death (backgroundpaper, Janjaroen 1996). There are two points of similarity between the household composi- tion responses to death in the Kagera and Chiang Mai studies. First, the Chiang Mai households with deaths, like their Kagera counterparts, suf- fered an increased dependency ratio due to the deaths. Because of the smaller number of adults in the Thai households, the dependency ratio there almost doubled after the death. Second, in both countries house- holds with a death were twice as likely to experience membership change as the households without deaths. However, the proportion of house- 2i6 COPING WITH THE IMPACT OF AIDS holds that experienced a membership change and the rates of turnover were only about one-quarter as large in Chiang Mai as in Kagera. When we later discuss possible policy responses to adult mortality, the possibility will be raised that households may respond opportunistically by moving people into a household that is benefiting from an assistance program. The evidence here suggests that, even if this turns out to be a problem in Africa, it is much less likely to be an issue in places like Chiang Mai where households are much smaller and apparently less able or willing to adjust their membership in response to outside stimuli. Dissavings and the sale of assets. Drawing down savings and selling assets is an obvious potential mechanism for coping with prime-age adult death. Because assets may have been accumulated as part of a strat- egy to cushion unanticipated shocks, drawing upon them is one of the least painful ways of coping, much less painful than reducing food con- sumption, for example. Evidence from Kagera, Rakai, and Chiang Mai suggests that households do draw down savings or liquidate assets in response to a prime-age adult death. The surveys in Kagera and Rakai both asked respondents about their ownership of three types of durable goods: a car or truck, a bicycle, and a radio. Less than 2 percent of the households owned a car or lorry, and changes in ownership did not show any clear pattern in relation to whether households suffered an adult death. However, ownership of bi- cycles and radios, which is much more widespread, did reveal a pattern. Table 4.5 shows how ownership of these assets changed over the course Table 4.5 Asset Ownership in Households with and without an Adult Death (percentage of total households) Rikiji Diarri,r. L gaudi £aLgerm Rtegioii, Tanzanria HousebroLdi Househ/ok, HovsteIokljS Hoimebaol&d le{/ wirtI u /o uiihih .4L;set addlr Xdt-at adf//t dfeatb aifeltr ebth adult da2tly Bi, ,Je Firqt VmI 4 At) 19 2/) Last %isit 41 i;, 28 Radico FIr;r vi-mic 4ti i Lait 'i,c 3- i6 3 3i 217 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC of the surveys, depending on whether or not the household suffered an adult death. In both surveys, radio ownership increased among house- holds that had no death and decreased among households that had a death.19 A similar pattern can be seen for bicycle ownership in Rakai, although not in Kagera. Thus the evidence from Rakai, partially sup- ported by the Kagera data, suggests that some households that suffer an adult death may be selling durable goods as part of their coping strategy. The alternative is that these goods may have belonged to the deceased and been willed to someone outside the household. Even in this in- stance, however, the loss of the asset may help with coping efforts, if the recipient feels an increased obligation to assist the bereaved. Additional evidence of households drawing down savings to cope with an adult death can be seen in Kagera data on membership in tradi- tional rotating savings and credit associations (ROSCAs). In wave 1 of the survey, 51 percent of the 80 households that would experience an adult death during the eighteen-month survey period were members of a ROSCA; by the end of the survey, participation had dropped to 36 per- cent. Among households that did not experience a death during the sur- vey period, ROSCA participation varied less, from 41 percent in wave 1 to 36 percent in wave 4. Although the Chiang Mai survey does not permit comparisons of fi- nancial variables over time between households that experienced a death and those that did not, fully 41 percent of households with a death re- port having sold land, 57 percent report some dissavings, and 24 percent report borrowing from a cooperative or revolving fund (i.e., a ROSCA) to finance the adjustment to the death. Perhaps it is the greater wealth of the Thai households that permitted them to cushion the shock in these ways, rather than through readjustment of household composition.20 Assistance from other households. For all households confronted with an adult death, help from relatives and neighbors is a potentially impor- tant supplement to the household's own efforts. Policymakers consider- ing how to best use the limited resources for mitigating the household impact will want to avoid displacing such private transfers. To do so, they will need information about assistance from households in the spe- cific communities concerned. The discussion that follows is not in- tended to substitute for this information, but only to suggest the types and possible relative magnitude of such responses. An important feature of the social organization of Kagera households, and indeed of most African communities, is interdependence in time of 2,I8 COPING WITH THE IMPACT OF AIDS need. In Kagera, both bereaved and nonbereaved households were very likely to receive cash or in-kind assistance from other households. (About three-quarters of the nonbereaved households received such assistance, compared with 80 to 90 percent of the households that suffered a death.) But among households that received private transfers after a death, the median amount received during the half-year of the death ($53) was more than twice that received during the year before the death, as well as twice that received by households that did not suffer a death. 21 New organizations established to help cope with the costs of AIDS death may be one explanation for this large difference. Focus group interviews in 20 of the sample villages found that besides traditional sav- ings and mutual assistance associations, such as ROSCAs, residents of many villages had launched associations specifically to help families affected by an AIDS death. Most of these associations were launched and operated by women; many have regular meetings at which members make contributions in cash or in kind (Lwihula 1994). Figure 4.9 shows the amounts of private transfers received by house- holds according to whether or not they suffered an adult death. Note the dramatic response of private transfers between wave 1 and wave 4 for the households that experienced a death in that interval (i.e., during the "panel"). The figure also shows the much smaller amount of program transfers, an issue we turn to below. Figure 4.9 Median Value of Assistance Received among Sample Households ReceMng Transfers, by Source, Wave, and Occunence of Adul Death, 1991-94 1996 dollars 60 53 111W.. ave4 50. ..... 40 . ------ ----------------.. . ....------ Households that suffered a death between the first and fourth "wave" of 30 . ., ---------........ . ........... data colection received larger private 21 20 and program transers than house- 20,.. ........ 13 2 ...holds that did not suffer a death dur- 10 . . . . .. .. . ......... .. V7 --.-] - ing this period. For these bereaved lo ------ -- ..........|......... d.. families, private transfers were much 0 _. ____I larger than program transfes No Panel No Panel death death death death Private transfers Program transfers Source: Over and others, forthcoming. 219 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Assistancefrom government and NGOs. Regarding assistance from gov- ernment and NGOs, two types of questions must be asked. First, which families receive such assistance and how much do they receive? Second, how much does it cost to provide this assistance? Cost considerations did not figure in our discussion of private assistance, since only private re- sources are involved. But governments and NGOs use public resources, whether generated by taxes or by voluntary contributions. Accordingly, we should ask whether these funds are used in the most effective way possible. In this discussion, assistance from government and NGOs is referred to as program transfers or formal assistance, to distinguish it from the private and informal assistance provided by households and village associations. According to the Kagera survey, program transfers reached fewer households and provided a smaller amount of assistance than private transfers. In the last wave of the survey, one-fifth of the households that had not had an adult death in the past eighteen months had received assistance from an organization in the past six months; almost two-fifths of the households that had experienced a death received such assistance. The median value of program assistance received was small relative to total household expenditure, and to the amounts of private assistance received. But while, on average, households that had experienced a death re- ceived larger amounts of assistance, as shown in figure 4.9, this was not always the case. Nor were program transfers always small relative to annual income. In one village, 50 percent of the households, including some that had never reported a death, received more than $110 in pro- gram transfers during the six months before wave 4. To analyze the relative costs of various types of programs for assisting households affected by AIDS and other causes of adult death, the survey collected data from one governmental and eleven nongovernmental organizations operating in the Kagera Region. Figure 4.10 presents the average cost per year of operating each of the programs where cost data were available from at least two agencies. In considering the implied cost comparisons, it is important to remember that the services provided may be very different; for example, home care focuses on the sick household member, while educational support helps dependent children, who may not be sick, to attend school. Furthermore, even the programs averaged in a single category often contain disparate program elements and vary in quality. Despite these caveats, the figure reveals that there can be very large differences in the cost per beneficiary of different types of programs. A 220 COPING WITH THE IMPACT OF AIDS Figure 4.10 Average Cost per Year of Survivor Assistance in 1992 by Government and Nongovernmental Organizations, Kagera, Tanzania Average cost per unit of output (1996 dollars) 250 | I1 Fixed cost Variable cost] 200 ........................................................ 91 ± 0 0 -- -- - - - - - - - - - - - - - - - - - - - - 126 126 50 .-........... - - - . - .- 7. Some types of assistance for house- holds affected by AIDS consume much 0 _ _ 11 14 / more msources than others. Home Foster Feeding "Basic Educational care care post needs" support for PWAs to HHs Note: PWA = person with AIDS, HH = household. Other costs are per child per year. Source: Over and Koda, forthcoming. particularly telling comparison is between the cost of supporting a child in a foster home, estimated at $107 per year, and the cost of supporting a child in an orphanage, which averages $1,063, or ten times larger (not shown on the figure). For children who cannot be placed in a foster home, it may be necessary to consider the alternative of an orphanage. However, policymakers and NGO providers should keep in mind that every child sent to an orphanage will consume the resources that could have been used to support ten children in foster homes. The economic impact of AIDS is larger in poor households. We have seen that households that experience an adult death draw on their assets to cushion the shock of this catastrophe. It follows that households with lower levels of assets can be expected to have more difficulty coping with the death than households with more assets. In this section we examine the effect of a household's initial assets on its ability to cope with adult death. First we show how a household's assets affect the short- term impact of an adult death on per capita food consumption; then we consider the long-term harm to children, through worsening malnutri- tion and reduced school enrollments. 221 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC In examining this evidence, it is useful to keep in mind the key question that policymakers are likely to face in deciding how society in general and governments in particular, can mitigate the impact of a generalized epidemic: who needs help? The impact on food consumption. The greater impact of a prime-age adult death on poorer households appears most starkly in changes in food expenditure and food consumption. Figure 4.11 shows the changes in per capita food expenditure and consumption (which includes both purchased food and home-produced food) for the poorer half of the Kagera households and the less-poor half of the Kagera households dur- ing the six months when the death occurred. For the better-off house- holds, both measures of food intake increased. The picture is quite dif- ferent for the poorest 50 percent of the households: food expenditure, which was already lower in these households than in the others, dropped by nearly a third. The resulting drop in per capita food consumption was cushioned by an increase in the consumption of home-produced food (not shown). Even so, per capita consumption in the poorer households Figure 4.11 Short-Tenn Impact of the Death of an Adult Household Member on Food Expenditure and Consumption per Adult Equivalent Member, Kagera, Tanzania, 1991-93 Percentage change after a death 20 - 12| Nonpoof 10- 0 -10 -- Among households in Kagera, -15 Tanzania, that suffered a prinme-age -20 - adult death, food expenditure and consumption declined for the poorer -30 - half of households but increased for -32 the others. 40__- Food Food expenditure consumption Note: The poorest 50 percent of households are those with less than the median value of assets per member in wave 1 of the survey, which was about $415 per adult equivalent member in 1996 dollars. The sample is 64 households that experienced an adult death between the first and last waves of the Kagera survey. Source: Over and others, forthcoming. 222 COPING WITH THE IMPACT OF AIDS fell by 15 percent. Even if these households eventually return almost to the predeath level of per capita food consumption, as did the Ivoirian households, lack of adequate nutrition for a year or more can have a pro- found effect on the development of children. We turn to this topic. The impact on child nutrition. Childhood malnutrition is potentially one of the most severe and lasting consequences of a prime-age adult death. The death of a parent or other adult may lower the nutritional sta- tus of surviving children by reducing household income and food ex- penditure, and by reducing adult attention to childrearing. Because childhood malnutrition can impede intellectual development and thus reduce a person's long-run productivity, improving childhood nutrition has long been an important development goal. Policymakers seeking to mitigate the impact of the AIDS epidemic will therefore be particularly concerned about minimizing the impact of the increasing number of prime-age adult deaths on childhood nutrition. The impact of adult death on childhood nutrition is likely to vary according to many factors, not least of which is the nutritional status of children in the overall population. Little information is available on how adult death affects child nutrition. Moreover, the impact is likely to dif- fer across countries and communities. The following discussion of the findings in Kagera illustrates some of the issues that policymakers will want to consider in attempting to mitigate the impact of the epidemic. In this discussion, the term "orphan" is used to indicate a child who has lost one or both parents. We would expect that the drop in food consumption among the poorer bereaved families described above would result in an increase in malnutrition among children in these households, since these children are likely to be malnourished or at risk of malnutrition before the adult death. As figure 4.12 shows, among the poorer households in Kagera, stunting (very low height for age) among children under 5 is indeed sub- stantially higher for orphans (51 percent) than for children whose par- ents are both alive (39 percent). What is surprising, however, is that the difference between orphans and nonorphans in the better-off house- holds is even larger; indeed, orphans in the better-off households are stunted at almost the same rate as orphans in the poorer households. This unexpected result raises difficult operational issues. If orphans in the poorer households were much more likely to be stunted than or- phans in the less-poor households, as we might have expected, the policy prescription would be straightforward: to minimize childhood malnutri- 223 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 4.12 Stunting among Orphaned and Nonorphaned Children under 5, by Household Assets, Kagera, Tanzania Percentage of children stunted 60 - 50 51 _ 39 _|3~~3 40 4| Nonorphans 30 - 29 | i lil1_bl 1* Orphans 20 In Kagera, Tanzania, half the children who had lost one or both parents were 1 stunted, regardless of the level of household assets. 0 Households with Households with more assets fewer assets Source: Kagera data, authors' calculations. tion, focus nutritional assistance on the poor households that suffer a prime-age adult death. Instead we find that, at least in Kagera, half of the children who have lost one or both parents are stunted, regardless of whether they live in a poorer household or a less-poor household. There are several possible explanations for this surprising observation. One is that stunting among both groups of orphans is due in part to pediatric AIDS and to other illnesses that a child may contract from an HIV-infected adult, such as tuberculosis, which would not necessarily be closely linked to household asset levels. Another possible explanation is that some stunted orphans in the households with more assets originally resided in poorer households, and their stunting is a legacy of that earlier poverty. Finally, the fact that stunting is about equal among both groups of orphans suggests that, for this population, stunting of 50 percent may be approaching an upward limit beyond which any additional deteriora- tion in childhood nutrition results in increased child mortality rather then increased stunting. In all three cases, child nutrition could indeed be worse in the poorer bereaved families than in the less-poor bereaved families. It may also be the case, however, that childhood nutrition does deteri- orate sharply after a prime-age adult death, even in households with com- paratively high levels of assets. This could happen, for example, if grief and psychological depression in the surviving parent interferes with child- rearing, including obtaining food and providing meals. If this is true, 2Z4 COPING WITH THE IMPACT OF AIDS young orphans stand a high probability of being malnourished, irre- spective of the economic status of the household in which they are living. One policy approach that could be appropriate in either case would be to focus nutritional assistance on young children who show evidence of being malnourished or who are likely to be at risk of becoming mal- nourished (by virtue of losing one or more parents). There are several advantages to such an approach. First, because the percentage of children under 5 who have lost a parent will be small, even in a generalized epidemic, such a response is likely to be much less costly, and therefore more feasible, than the alternative of providing assistance to all households that suffer a prime-age adult death from AIDS. Furthermore, because of the long illness that often precedes death from AIDS, it will often be possible to identify young children who will soon be orphans before the mother or father dies and to enroll them in programs to minimize the nutritional impact. In cases where the mother is HIV-positive, this supplemental feeding could perhaps simultaneously reduce the risk of mother-to-child transmission through breast milk. In addition, programs that provide food directly to malnourished children and to orphans, rather than to households that include orphans, may avoid the problem of households fostering children primarily to obtain benefits intended for the orphans. While creating an incentive for households to foster children may be desirable in a severe AIDS epi- demic, too large an incentive can increase the number of children shifted between households, to the detriment of their welfare. A better approach may be to include children in households where a death is anticipated in community-based nutrition monitoring and feeding programs on the model of the UNICEF-sponsored "village feeding posts." Finally, including orphans in a program designed to address malnu- trition more broadly is more equitable than focusing assistance on AIDS orphans alone. This is particularly true in very poor countries, where an alarmingly high proportion of all children are malnourished. In Kagera, for example, even in households that are less poor and have both parents alive, nearly one-third of children are stunted. In such a situation, pro- viding assistance only to AIDS orphans would neglect a large number of children who are also very needy. The impact on child schooling. Besides increasing childhood malnutri- tion, a prime-age adult death in a household is likely to reduce school enrollment. This lack of schooling, perhaps exacerbated by inadequate nutrition, will make it particularly difficult for child survivors of a prime- 225 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC age adult death to escape poverty. The effects of a prime-age death that could decrease enrollment among children in the household indude: * reducing the ability of families to pay for schooling * raising the demand for children's labor * reducing the expected return to adults of investments in children's schooling. We have already seen how changes in income and expenditure that occur before and after a death would tend to reduce the ability of families to pay school fees and other education costs. Children may also be with- drawn from school to work outside the home, help with chores and farming, or care for an ailing family member. In addition, where prime- age adult mortality is high, parents may be less willing to invest in their children's schooling, either because they fear that the children will not live long enough to realize the higher earnings schooling promises, or be- cause the parents themselves do not expect to live long enough to bene- fit from their children's future earnings. Similarly, relatives who take in an orphan may be less willing than the parents would have been to invest in the child's schooling. For all these reasons, children who have lost one or both parents are likely to have lower enrollment rates than those whose parents are alive. Data from the Demographic and Health Surveys on enrollment and orphanhood in nine countries generally support this premise. Figure 4.13 shows the predicted enrollment rates for children by orphanhood status from a regression that holds constant within each country the children's age, gender, urban residence, and quality of housing-a crude proxy for wealth. All of these areas are characterized by low incomes and all except northeast Brazil are in the midst of a generalized HIV/AIDS epidemic. Orphans have significantly lower enrollment rates in every area except Uganda and Zimbabwe; differences in enrollment rates of orphans and nonorphans are greatest in the five sets of bars on the left. But while the data support the view that orphans are less likely to attend school than other children, they also clearly demonstrate that in most of these countries a very large proportion of children who are not orphans are also not attending school. This indicates that at least in these low-income areas orphanhood is not the major reason for nonenroll- ment of children; other demand- or supply-side problems in the educa- tion sector or the labor market are leading to low enrollments among 226 COPING WITH THE IMPACT OF AIDS Figure 4.13 Enrollment Rate for Children Ages 7 to 14, by Orphan Staus, Nine Countries Percentage enrolled 100 _ Both alive * Father dead 80 -_ 11 so * L E * * Mother dead * Both dead 60 40 -- - 20 - 0 N.E. Burkina Central Cote Haiti Kenya Tanzania Uganda Zlimbabwe Brazil Faso African d'ivoire 1994-95 1993 1994 1995 1994 1991 1993 Republic 1994 1994-95 Source: DHS data, authors' calculations. children irrespective of orphan status. As with nutrition, orphans appear Orphans often have lower school to be specially disadvantaged in terms of education, but because enroll- enrollment rates than nonorphans; but orphans are not the only children ment levels are generally low in these low-income countries, special mea- who do not enrol. sures to boost enrollment among orphans would neglect the needs of the many nonenrolled children who are not orphans. Significant improve- ments in enrollments are therefore likely to require a systemic approach; this is beyond the scope of this book but is already among the educa- tional goals in these countries. Using the Kagera data, it is possible to consider how other factors may influence the relative enrollment levels of orphans and nonorphans. Using the same distinction between households according to assets, fig- ure 4.14 shows that children in the poorer households are less likely to be enrolled than children in the less-poor households, regardless of orphanhood. Also, the difference in orphan and nonorphan enrollment rates is significant only among the poorer households. Most striking, however, are the low enrollment rates of children ages 7 to 10 regardless of the level of household assets. Thus, although orphans in the poorer households have the lowest enrollment rates, enrollment rates among all young children in Kagera are disturbingly low. Higher enrollments in the 1 1-to-14-age group are due mostly to the widespread practice of 227 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 4.14 Enrollment Rates by Age, Orphanhood, and Household Assets, Kagera, Tanzania, 1991-93 Enrollment rate last year (percent) 100- 87 80- 76 =66 0 More assets, nonorphan 60- * D 56 II More assets, orphan D5* i~49 1 Fewer assets, nonorphan 42 42 Fewer assets, orphan In Kagera, Tanzania, orphans in 40 - households with fewer assets were 28 least likely to be enrolled in school; but even nonorphans in households 20 with more assets and both parents alive had low enrollment rates. 0 7-10 11-14 15S19 Age of child Source: Ainsworth and Koda 1993 and authors' calculations. delaying enrollment, which results in many older children attending primary school. These observations, combined with the fact that school fees are lower in Tanzania than in neighboring Kenya and Uganda (both of which have higher enrollment rates), suggest that for some reason households in Kagera are choosing not to enroll their young children in primary school. Whatever the reasons for nonenrollment, they are prob- ably not solely financial, even among the poor. While household asset levels had only a relatively minor impact on whether or not orphans were enrolled, the study also found significant differences according to whether the deceased household member was female or male. In households where a prime-age female had recently died, children had lower enrollment rates and were more likely to engage in activities in which women typically specialize-cooking, shopping, laundry, cleaning, and collecting water and firewood. For younger chil- dren, enrollment was often merely delayed, while older children were likely to drop out and not return to school. Even those children in households with a female death who did not drop out nonetheless spent fewer hours in school than children in other households. Because these effects were not observed in households where a prime-age male died, it appears that children are dropping out of school to perform tasks that had been done by the woman before her death. 228 COPING WITH THE IMPACT OF AIDS Are these findings from Kagera likely to be true elsewhere? Baseline levels of enrollment vary enormously within and across regions, accord- ing to the financial costs of schooling, levels of household income, the opportunity costs of children's time, and the economic benefits of schooling. For low-income countries such as those shown in figure 4.13 it is reasonable to expect that the enrollment rates for orphans and nonorphans in the poorer households will be lower than for children in less-poor households, and that the differential enrollment rates between orphans and nonorphans in poor households are greater. Similarly, we would expect the death of a prime-age female to have a larger impact on enrollments than the death of a prime-age male in any community where women provide households with crucial services that, in the event of a woman's death, could be provided by children. In middle-income countries, where enrollment rates among all children are much higher, orphans in low-income households may account for a larger share of nonenrollments than in poorer countries. In Thailand, for example, as of 1992, 93 percent of primary-school-age children were en- rolled and the enrollment rate for secondary schooling was growing rapidly (Shaeffer 1995; Brown and Sittitrai 1995). Although we lack country-wide data on the extent to which orphans are underenrolled in Thailand, one small study found that 13 percent of the school-age children in families where someone was ill and dying of AIDS were withdrawn from school to help support the family (Pitayanon, Kongsin, and Janjaroen 1997). Two broad policy conclusions can be drawn from this evidence. First, in areas where enrollments are very low, a systemic effort to improve overall enrollment will be more fair, and likely to yield larger benefits, than special programs focused only on orphans. Second, as enrollment rates improve, it becomes increasingly likely that poor orphans will have much lower enrollment rates than other children. Yet even in these situ- ations, special programs for orphans may not be the fairest or most effective answer. So long as a significant proportion of poor children are not enrolled in school, interventions that aim to raise enrollments among the poor will address the schooling of the neediest children, including the neediest orphans. Poverty Policy in a Severe AIDS Epidemic The preceding analysis has highlighted several key factors of the im- pact of AIDS on poverty. First, HIV infects both the rich and the poor. 229 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Although it already infects more poor people than rich people and will probably eventually infect a higher proportion of the poor, many HIV infections continue to occur among the nonpoor. Second, the short- term impact of a prime-age adult death is higher for AIDS than for other causes of death, primarily because of the long illness that often precedes an AIDS death. However, because other costs are incurred in all cases of adult death, the overall difference in the short-term impact of AIDS deaths and other deaths is not large. Third, rather than being destroyed by AIDS, households use a variety of mechanisms to cushion the short- term impact of an AIDS or other prime-age adult death. Fourth, we have seen that these coping mechanisms are much less effective in poor house- holds, where more of the coping is at the expense of children's school enrollment and nutritional status. These latter effects are a permanent legacy of the AIDS epidemic that will harnper national efforts to achieve development goals for years to come. These findings are based on a combination of theoretical analysis and empirical observation. Because the problems are relatively new and there are very few comparable data across countries, or indeed even across regions within a country, we have relied heavily on a single survey in Kagera, Tanzania, with additional observations from three other recent surveys in other countries. As experience with the AIDS epidemic in- creases and additional data become available, our understanding will cer- tainly improve and some of the observations here will be challenged. Despite these caveats, the broad findings described point out a key question that policymakers must consider in deciding how to mitigate the impact of AIDS on poverty: which households need help? The short answer, of course, is that the poorest households are most in need of assistance, and these are not necessarily the households hit by AIDS. In poor developing regions, many households that have not been affected by AIDS are likely to be very poor. Among the Kagera households that had not experienced an adult death, one-third of the children under 5 were stunted. Similarly, even in households where both parents were alive, 50 percent of the children under 11 were not enrolled in school. The lack of a clear correlation between poverty and AIDS is strikingly evident in figure 4.15. The figure presents estimates of the percentage of the population in each district of Kagera living below the absolute poverty line of $124 dollars per person per year in 1991.22 In parenthe- ses below the name of each district is an indicator of the severity of the AIDS epidemic in that district during that year, the mortality rate of 230 COPING WITH THE IMPACT OF AIDS Figure 4.15 Poverty in Kagera Region, by Distnct and Adult Mortali Rate, 1991 Percentage of population below absolute poverty line 60 57 40 30 AIDS is not necessarily more wide- 20 _ spread in poorer districts. Among the districts comprising Tanzania's 10 11 _ Kagera region, some have severe epi- 7 demics but low rates of poverty; the 0 n district with the most poverty (Ngaral 0 ______ has a smaller epidemic. Bukoba Bukoba Muleba Biharamulo Karagwe Ngara Urban Rural (13.7) (5.1) (5.0) (5.0) (15.2) (14.7) District Note: Adult mortality rate is per 1,000 for 1988. Source: Gupta, Mujinja, and Over, forthcoming. adults ages 15 to 50, as calculated from the 1988 census data. The fact that AIDS is not the most important cause of poverty is obvious from the fact that two of the districts where the AIDS epidemic had been ex- tremely severe for a decade, Bukoba Urban and Muleba, had the least problem with poverty, while one of the districts with almost no AIDS, Ngara, had the most poverty.23 While the households affected by AIDS are not necessarily poor, the poor households that are affected are much less able to cope than the nonpoor households. Less-poor households in Kagera actually experi- enced an increase in consumption per capita after the death, while the poorest households experienced a sharp reduction in consumption, espe- cially food consumption, as a result of the death. While the increased consumption in the nonpoor households will not necessarily be ob- served in other locales, it is reasonable to assume that nonpoor house- holds faced with an adult death will be much better placed to smooth consumption than poor households. 231 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Finally, we have seen that even households that lack access to formal credit and insurance markets nonetheless use a variety of measures to cope with AIDS deaths and other misfortunes. Thus, even if the shock from an AIDS death is larger than that from other misfortunes, and poor households are more vulnerable to the shock than nonpoor households, governments seeking to mitigate the impact of AIDS on poverty must ask themselves whether it is possible to design and implement, either di- rectly or through NGOs, assistance programs that are more efficient and equitable than the informal coping strategies already in place. These broad findings can be further distilled into three general rec- ommendations for policymakers: * Not all households experiencing an AIDS death need assistance. * If survivor assistance is to be offered, it should be targeted to all very poor households that suffer a prime-age adult death, regard- less of whether the death was due to AIDS. * Assistance will do the most good immediately before and after the adult death, during the period when per capita food consumption has fallen but not yet recovered. It need not be permanent. In addition to these three points, the findings also suggest that there is potentially important synergy between AIDS mitigation and anti- poverty programs. For example, the finding that poor households are more vulnerable to the impact of an AIDS death implies that general antipoverty policies can also be AIDS mitigation policies. If general anti- poverty policies are effective in reducing the number of poor households, then AIDS deaths will occur in stronger households that can cope at smaller cost to the survivors. Similarly, the finding that an adult death depresses per capita food consumption in the poorest households by 15 percent implies that AIDS deaths that occur in poor households exacerbate poverty. Thus, when AIDS mitigation policies are targeted to households that were poor be- fore the AIDS death, they are likely to prevent the affected household from slipping further into misery as a result of the death. In this case, AIDS mitigation policies could be effective at limiting the depth, if not the extent, of poverty.24 In sum, the results of these studies suggest that antipoverty programs and mitigation programs be integrated. When an antipoverty program is designed for a community with low living standards, consideration 232 COPING WITH THE IMPACT OF AIDS should be given to including components that specifically address the needs of the poorest households hit by AIDS deaths. For example, sup- pose the antipoverty program in a specific AIDS-affected community consists of a labor-intensive public works program. Components that would generate such synergy might include: * home-based or hostel-based care for the terminally ill to enable healthy adults who would otherwise have provided this care to take advantage of the jobs * day-care centers or "feeding posts" to enable single parents to take the jobs. Examples of targeted antipoverty programs in developing countries that could be modified to use adult death in the household as an addi- tional targeting criterion are presented in box 4.9. Conversely, when a mitigation program is established, locating it close to and combining it with a conventional antipoverty program will improve its effectiveness. In an area severely affected by an AIDS epidemic, implementing either program in isolation from the other sacrifices an opportunity for effec- tive development policy. How Governments Can Cope with the Impact of HIV/AIDS on Health Care and Poverty A LTHOUGH THE SPECIFICS OF THE IMPACTS ON HEALTH and poverty differ, the analysis leads to broadly similar con- clusions in both areas. Special government assistance for people infected with HIV/AIDS and their survivors must be weighed carefully against the many other pressing needs that governments face. Well-intentioned government efforts to assist individuals with HIV/AIDS and their families may divert resources from other families that have not been afflicted by HIV/AIDS but are nonetheless suffering from illness, poverty, or both. In particular, HIV-infected patients should be responsible for the same portion of the cost of their care as other patients with similar income and likelihood of infecting others. Because of HIV's long incubation period, governments may initially underestimate the cost of programs to provide special assistance to those 233 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 4.9 Using Adult Death as a Targeting Criterion for Antipoverty Programs THE CENERALLN ACtCEPTED STRXTE(CY FO:IR REDU-- Using prinme-age adult death ai a targeting crite- lionot long-term povery comprises three componenLs: non is lLkeir to hase several advantages. Compared pro-grovw-th macroeconomic policies. human capital %%ith providing help [o families w%-ith a deach from development, and social satety net programs. \V'hile HI\'iAIDS. i is fairer. since it w-ill include families most countries depend primarily- on [he frt rt-o. some with prime-age adult deaths from other causes. Com- countries, including some %ets poor countries, also bining this criterion %,ith others that identit' the have substantial safrv net programs char arrempt to household as poor may help identit ihle neediest lam- provide assistance directly to the poorest households. ilies. Since the death of a prime-age adult is usually An important question ftor all ,uch programs is how% to w%ell known to everyone in the community, using rhis identify the hou.eholds thar need help most. Esen a as a targeting criterion may help program administra- brief summary' ot'he c-y-rensive literarure on targeting is tors identi dtitute railies that might otherwise be beyond thc scope of this volume. Howevcr, it is impor- misscd. For th ,ame reason. such a criterion may be tan[ to note that a number of countries at different in- effecti%e in minimizing opportunistic responses: feign- come le%els and at different stages of the epidemic ing a household death to ob[ain the benefits of the already have in place targeted safetn net programs that targeted program tbir survniing household members could help poor households thar suffer a prime-age would be very difficult Finally, including prime-age adult death. In some oLf these. including prime-age adult death as a targeting criterion mas help to in- adult death as a specific targeting criterion along with crease the political acceptabiliry of saters ner prcgrams other crirenra maY help to Identify the neediest families, among those wvho do not benefit. since many people Box table -j.9 describes five such programs. will readils' understand that poor households suffering Prior to the AIDS epidemic. prime-age adult death such a death-and especially the children in such was rare. perhap- too rare to warrant including it as a households-arc lik-ls to face sev-ere hardship. targeting criterion. Sadly, it is now. common enough that countrics ssirh targeted poverny reduction pro- gram, should con%ider whether and how to include it as a targeting crirerion. Because this is a new% area. mi- pact evaluation of programs that attempt to do so Sira Besles and karnbur j'i3. subbarac. ind others 1946; would generate important new- knowledge. -.an '\j all mand Nedt l1It04 Box Table 4.9 Social Safety Net Programs in Which Prime-Age Adult Death Could Be Used as an Additional Targeting Criterion, Five Countries Counitry and sntn of ie epidemnic Program and ex isting targeting criteria Zimbabwe (generalized) Feeding pro.,izrin. Targeted to children in drought-prone areas using nutritional sur- Veillance data. U'es locally grown Food and include, nuTnition education. India icon;cInrared I hood guai d:imrinbuirn Through publicl! operated ration" hops. the states have dis- tributed grair to an%ont requesting it. but under a ne's gotetnmcnt program they are required to limit dt Lribution to those beloas the po;errt line. Honduras iconceritratedi FR3wei cr-nip.; Distribured rhrough healrh centers to lo, -income childten under 5 3nd pregnant and lacrating mothers, and rhrough schools to poor murhert and their chil- dren grades 1-i. Bangladesh inascenrtil lievucreet preigr.e Targ ctcd r-. hou;eholds cemning les thin It)i acre c. land; group lending and peer monitoring serve the poor %kithout collateral and cnsure repaymntnt. Ch&l nascenti Ca.?,h nu.4tr. Targeted to rural and urban poor based on their ans-wers to a computer- scored quetr'onnaire. 234 COPING WITH THE IMPACT OF AIDS affected by HIV/AIDS. As the number of people who get sick and die from the disease increases, these programs will absorb a growing share of resources that could have been used to address other problems. Because AIDS may divert resources from other pressing problems and commit governments to expenditures from which it will later be politically diffi- cult to withdraw, policymakers in developing countries should be wary of programs that provide special assistance to people with HIV and their families solely on the basis of an HIV diagnosis. At a minimum, they should consider the long-term cost of such programs based on a range of likely assumptions about the course of the epidemic. Notes 1. Impacts on other sectors maybe substantial in some 7. Gertler and van der Gaag (1990) show that the countries. See Ainsworth and Over (1994b). poor are more responsive to price (measured as travel time) than the less poor. Lavy and Quigley (1993) and 2. The AIDS epidemic will increase costs, and thereby Mwabu, Ainsworth, and Nyamete (1993) provide recent reduce supply in all sectors of the economy, and will re- evidence on the elasticity of demand with respect to qual- duce the domestic demand for nontradable goods. Since ity. See Carrin, Perrot, and Sergent (1994) and Gertler the value of health care output increases while that of and Hammer (1997) for reviews of the literature. other sectors declines, AIDS will increase the share of health care in national expenditure and product. 8. Barnum and Kutzin (1993, tables 3.3, 3.4) give oc- cupancy rates for the developing world that range from 31 3. The increased number of deaths among prime-age percent in Belize and 46 percent in Fiji to 116 and 129 adults is eventually offset by decreased numbers of deaths percent in Malawi and Lesotho, respectively. But the pat- in older age groups. Since terminally ill elderly patients tern observed in Kenya (Collins and others 1996) of a pos- use little care in poor countries, this offset can be ignored itive correlation between the occupancy rate in a public fa- in the developing countries that are the subject of this cility and its degree of medical sophistication is common report. within many public health systems, especially if the user 4. A population that goes from 0 to 5 percent HIV in- charge is equally low at all levels of care. fection in one year will not experience any increase in mortality the first year. Assuming the incubation period 9. As part of a national health care reform process, the has a median of ten years, mortality from HIV would Kenyan government decreed a fee increase throughout the begin to rise in the second year, reach 2.5 per thousand in country in December 1989, which is after the 1988/89 the tenth year and 5 per thousand in the twentieth year. data on Kenyatta Hospital patients were collected in the Floyd and Gilks study. Since AIDS patients were ex- 5. The calculation is: 100 x 0.25 x (5.3 /5. 0). empted from fees imposed during this period, care at the hospital simultaneously became more expensive for the 6. For example, one source assumes that only about HlV-negative and less expensive for the HIV-positive. 2.5 percent of those African transfusion recipients who re- This change in relative prices between the two groups ceive blood infected with hepatitis B would contract the probably accounts for part of the change in admission mix disease, which would then require 20 years to kill them seen in figure 4.2. Collins and others 1996 describe health (Beal, Bontinck, and Fransen 1992, p. 116). financing reform in Kenya. 235 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 10. Since Tanzania's national prevalence rate only women, the percentage incurring out-of-pocket medical recently attained 5 percent, the epidemic's full impact on expenses was higher for those who died of AIDS (70%) mortality and health care expenditures are still in the future. than for those who died of other causes (59%). Expenditure data on Mexico and Tanzania are from figure 4.6 and Shepard and others (backgroundpaper, 1996). 15. Since everyone must eventually be buried, that portion of the funeral costs that would have occurred any- way, discounted to the present, should not be attributed 11. Insurance policies suffer to varying degrees from to the prime-age death. However, when the death occurs "moral hazard," when the amount of the loss incurred de- to yearimege it Howeve whenwise den octed pends upon whether or not the individual is insured. For many years before it would have otherwise been expected, pend upo whtheror nt te inividal s inuredFor as is the case with most AIDS deaths, the discounted value example, insured houses are somewhat more likely to burn of the fuue fuea osts IS qithe dsm compario than uninsured ones. The result of the problem is that wt the actual cost of th funeral. insurance against a specific risk becomes more expensive per dollar of risk coverage and, in the extreme, may not be 16. The authors of the Thai survey analysis did not available at all (Arrow 1963). The problem arises regard- analyze direct costs by the gender of the deceased. less of whether the insurance is private or public and is particularly severe with health insurance, where it is con- 17. Analysis of a single wave of data such as this does trolled in practice by coinsurance provisions. These provi- not show the direction of causality: were the expenditure sions typically specify higher coinsurance rates on highly patterns the result of the death or were households with price-elastic services like outpatient visits or psychiatric certain expenditure patterns more likely to suffer a death? care than on less-elastic ones like inpatient services. Figure 4.11 demonstrates from an analysis of changes in consumption over time that the differences between the 12. Areas of Africa studied include areas surrounding two pie charts are largely due to the impact of the death. Lake Victoria-Rakai (Serwadda and others 1992; back- groundpaper, Menon and others 1996b); the Masaka dis- 18. In most of the 29 households, the person with tricts of Uganda; Kagera (Killewo and others 1990) and AIDS died; in a few cases the AIDS patient moved to Mwanza (Barongo and others 1992; Grosskurth and oth- another household ers 1995a,b) regions of Tanzania; and Kigali, the capital of 19. The difference is statisticallysignificant at the 0.01 Rwanda (Allen and others 1991). level on the Rakai sample, but not significant in the Kagera sample, perhaps because it is a smaller sample. 13. One study converts consumption in developing countries to parity with the U.S. dollar (using the purchas- 20. The initial differences between Kagera and Rakai ing power parity indices). It estimates that two of every households that later suffered an adult death and those three persons in the developing world, and, in Africa, four that did not are intriguing. The former began the survey out of five, spend fewer than two 1985 U.S. dollars a day period with somewhat lower ratios of dependents to (Chen, Datt, and Ravallion 1994), a level of consumption adults, slightly more household members and assets, and that suggests substantial poverty by almost any standard. greater participation in ROSCAs. All of these drop after the death, causing the households suffering a death to 14. Surveys of survivors in the Kagera study found resemble the other households more after the death than that, of the 264 household members age 15 to 50 who before. There are two possible explanations for this. On died during or in the year before the survey, 82 percent the one hand, it is possible that households that anticipate sought treatment, while 15 percent sought no medical a death prepare by accumulating assets, recruiting addi- care at all (survivors were unsure about the other 3 per- tional household members, joining a ROSCA, and so on. cent). Among men, about 90 percent of those who died of Such adaptive coping in a risky environment is undoubt- AIDS were reported to have sought medical care, com- edly part of the explanation. On the other hand, there is pared with only 66 percent of those who died of other also evidence that AIDS-affected households in the sam- causes. (About 85 percent of women sought care, regard- ples were on average somewhat less poor than their neigh- less of the cause of their death.) For both men and bors. The initial differences cited above between the 236 COPING WITH THE IMPACT OF AIDS households that later suffered a death and those that did and Goodhart (1995) and World Bank (1996b) for a dis- not could simply be indicative of the greater affluence of cussion of poverty in Tanzania and details on the deriva- the average AIDS-affected households. tion of this and other poverty lines for Tanzania. 21. Because averages are pulled upward by extreme values and some households received as much as $5,000 in 23. The recent iflux of refugees to Ngara DistrIct private transfers, the mean amount of assistance received from Rwanda and Burundi may be exacerbating the AIDS in wave 4 by households that had experienced a death was epidemic t S and ODA (1994). $192, much larger than the median. 24. That is, AIDS mitigation policies targeted to the 22. The figure of $124 is the 1996 dollar equivalent poor may reduce the poverty gap if not the poverty head of 31,000 Tanzanian shillings at 1991 prices. See Ferreira count. 237 CHAPTER 5 Working Together To Confront HIV/AIDS I F DEVELOPING COUNTRY GOVERNMENTS, DONORS, AND multilateral organizations were already following the policies advocated in previous chapters, HIV would be a serious but manageable health challenge, perhaps not unlike cancer or other difficult-to-treat diseases, rather than a global epidemic. Unfortunately, national and international efforts to fight AIDS are far from optimal. Fifteen years into the epidemic, many developing country governments still lack adequate surveillance systems and have yet to enable a large enough share of those who are most likely to con- tract and spread HIV to protect themselves and others. Moreover, many countries also lack the society-wide policies to support such prevention interventions. Despite the willingness of nonprofit private groups to contribute to the fight against AIDS, some governments have difficulty providing the right mix of support and supervision. Donor governments and multilateral organizations, which provide much of the financing for national AIDS programs, have not always encouraged recipient govern- ments to set and address appropriate priorities and they have invested too litde in international public goods: knowledge and technology for fighting the epidemic in developing countries. Why has the national and international policy response to the AIDS epidemic not been more satisfactory? Much of the explanation involves an understandable lack of knowledge. AIDS is a relatively recent chal- lenge. Faced with a potential emergency, governments, donors, and mul- tilateral organizations responded as best they knew how using the infor- mation then available. As knowledge about the epidemic and ways to combat it increases, public responses are likely to improve. Like all public policies, however, AIDS policy is not made in a vac- uum. Indeed, because the spread of HIV involves private behavior that 239 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC many people deplore-frequent changes in sexual partners and the in- jection of addictive drugs-governments that adopt programs to reduce the riskiness of these activities may be accused by their constituents of facilitating socially deviant or immoral behavior. Thus HIV/AIDS pol- icy may be subject to strong political pressures, some of which work against the policies most likely to contain the epidemic. In considering these issues, this chapter looks beyond the national policies discussed in previous chapters to consider how the main actors in the AIDS policy arena can work together to more effectively confront the epidemic. First we examine the evolving roles of national govern- ments, donors, and the multitude of other nonprofit and for-profit organizations which we refer to collectively as NGOs. We conclude that many low-income countries should confront the epidemic more force- fully, both directly and in collaboration with NGOs. Turning to a de- tailed examination of donor funding and policies, we argue that bilateral donors and multilateral organizations, despite their substantial contribu- tions, have focused too little on fostering new knowledge and technol- ogy, such as information about costs and effectiveness of alternative pre- vention strategies and research on an HIV vaccine. Finally, the chapter discusses how public opinion and politics shape AIDS policy and how developing country governments can work with a variety of partners to overcome the obstacles to sound policies for fighting AIDS. Government, Donors, and NGOs N ATIONAL GOVERNMENTS BEAR THE RESPONSIBILITY FOR protecting their citizens from the spread of the HIV epidemic and of mitigating its worst effects once it has spread. But they are not alone in the effort. Bilateral and multilateral donors have pro- vided both leadership and major funding for national AIDS prevention programs, especially in the poorer developing countries. And both local and international nongovernmental organizations have stepped forward to help against the epidemic, sometimes prodding reluctant govern- ments into action. The challenge for national governments is to define their role in the struggle against the epidemic, not in isolation from or in passive response to the other actors, but in active collaboration with them. Only the gov- 240 WORKING TOGETHER TO CONFRONT HIV/AIDS ernment can claim to represent and act on behalf of the national popula- tion. Among the three types of actors, it has the unique ability to autho- rize implementation of an intervention by a donor or NGO. However, a donor cannot be commanded to finance or implement a program in which it has little interest. NGOs, too, have preferences and technical strengths or weaknesses. Thus the government cannot simply assign tasks to itself and the other actors. Instead it must learn the preferences and judge the comparative advantages of donors and NGOs. If there are im- portant tasks that public economics considerations assign to the public sector, but that donors and NGOs either cannot or will not perform, then the government must undertake them directly or subcontract them. What roles have the three sets of actors played in the struggle against the AIDS epidemic? How can cooperation be improved to make the most of the strengths of each? To answer these questions, this section first describes the important role that donors have played in funding AIDS interventions in most developing countries. Although the data on fi- nancing are incomplete and imprecise, they present a coherent picture of the relative roles of national governments and donors: donors have assumed the major financing burden in the poorest developing coun- tries, and bilateral donors show a preference for countries suffering from generalized epidemics. Since the available cross-country data speak only to financing, the analysis of the NGOs' role in implementation is based on examples. Although no generalization regarding roles will apply to every country, the analysis suggests that many national governments and NGOs should assume a somewhat larger share of the funding of preven- tion activities, leaving donors to focus on the international public goods discussed in the next section. Furthermore, anecdotal evidence supports the claim that the donors frequently work at cross-purposes at the coun- try level. Efforts would be more effective if donors would improve their coordination with one another and with national authorities without slowing the speed with which they deliver assistance. Most of the National Response Is Funded by Donors The total amount of donor funding for AIDS was estimated at approximately $300 million in 1996. The largest contributor of new funds in that year was the United States ($117 million); the European Union ($55 million) and Japan ($40 million) provided the next largest amounts of grant funding, and the World Bank provided approximately 241 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC $45 million in new loan commitments that year, most of which was at concessional rates. This seemingly large amount of money is, however, only about 6 percent of total donor health assistance to developing coun- tries.' Nevertheless, since AIDS expenditures represent a substantial fraction of total public spending on health in some developing countries, observers have asked whether too large a proportion of health resources is devoted to AIDS in these countries relative to other health problems. The WHO Global Programme on AIDS, predecessor of UNAIDS, collected data on donor, national, and NGO funding of AIDS programs in participating countries for the period 1991-93. While this database is imperfect and underrepresents funding from national and NGO sources, it provides the only detailed view of AIDS funding for a signifi- cant number of developing countries. By matching it with data on total national health expenditures from the World Bank (1993c) and data on domestic AIDS spending collected by Mann and Tarantola for AIDS in the Workl H1(1996), it is possible to measure the relationship between expenditures from each source and total national health spending in a country, and to compare the importance of national government and donor funding across countries. Average annual 1991-93 spending on AIDS by donors recorded by the GPA funding database exceeded 10 percent of 1990 public health spending in only seven countries: Uganda (59), Tanzania (36) Zambia (27), Malawi (16), Central African Republic (13), Guinea (11), and Rwanda (11). In these seven countries, all of which are in Africa and have some of the most serious AIDS epidemics in the world, interna- tional AIDS spending is large enough to overshadow all other prevention programs operated by the ministries of health. International spending was greater than 1 percent of the public health budget in 32 additional countries, including the non-African countries of Haiti (7 percent), Vietnam (3 percent), Thailand (3 percent), Lao PDR (3 percent), Bo- livia (3 percent), Bangladesh (2 percent), Sri Lanka (2 percent), Pakistan (1.4 percent), Honduras (1.1 percent), and Chile (1.01 percent). However, table 5.1 reveals that countries with large ratios of donor- funded AIDS spending to total national health spending are the excep- tion rather than the rule. The average country in fact received a little less than 2 percent of its 1990 health budget for AIDS. Even in low- income countries with generalized epidemics, the average percentage of the national health budget received for AIDS was only 8.5 percent. Looking across the table's three country income groups (see "Average" 242 WORKING TOGETHER TO CONFRONT HIV/AIDS Table 5.1 Average 1993 Donor-Funded HIV/AIDS Expenditures, by Stage of the Epidemic and Income Level (percentage of 1990 national health expenditure) hircome level (GDP per capita) Stage of Low Louer-middle Upper-middle the epidemic (< $725) ($726-S2,985) ($2.986-$8,955) A4rerage Nascent 07. 0.2 0.01 0(. Concentrated o. 0.4 0.1 1.2 Generalized 8.5 n.a. 0.1 -. Unknown 0.4 0.1 ° A (0 2 Average 3.2 0.2 0.1I 1.8 n.a. Not applicable. AViote See cnd ol chapter 2 for definitions of -nascenr. -oncencrawed. and ,er.eral- tzed.- Donor assistance data art exuacted by Pvne i bactvwind,r'ape, [II -I rn.r, ihc GPA hfnding database deve]opcd br rhe predecessor prcgram ro UNAIDS. Ehs WHO Global Prograrnmc on .UDS. Nauonal health e:.pendirure data tor 1990 are ftrc.m World Bank (1993c;. row), we see that the average percentage of the health budget received from donors for AIDS declines from 3.2 percent among the low- income countries to one-tenth of a percent among the upper-middle- income countries. This is due to higher total public health spending in the higher-income countries, as well as to lower donor allocations to these countries. Looking across stages of the epidemic (see "Average" column), we see that donor funding as an average percentage of na- tional health spending rose consistently from one-third of a percent in the nascent countries to nearly 8 percent in the countries with general- ized epidemics. However, even the higher figure does not threaten to overwhelm ministries or to overshadow other public health programs in the average recipient country. If donor-funded AIDS spending is occasionally large compared with the national health budget, national AIDS program spending never exceeds 10 percent of total national health spending and only rarely exceeds 1 percent, as judged from the survey of national program spend- ing conducted for AIDS in the Work II(Mann and Tarantola 1996). Fig- ure 5.1 shows that only three developing countries reported a figure for 1993 national AIDS program spending that was above 1 percent of their 1990 public health spending: Thailand (5 percent), Mali (2 percent), and Malaysia (2 percent).2 Twenty countries reported spending nothing 243 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 5.1 Comparison of Average Annual AIDS Spending by Donors with That of National AIDS Programs, 1991-93 (percentages of 1990 government health spending) Average donor AIDS spending, 1991-93 100 Tanzania Uganda . 'Zambl Guinea . Rwanda 10 Madagascar d'ne *wad Congo CrCte dilvoirel Senegal D.R' Mali Thahthnd NplCameroon. Chad * *Tojo Nepal* * * Benin Chile India 1 Brazil Burkina Faso Marocc 0.1 ~A ll T E Jordan G ea ga Azerbaijan * E Tunisia. GuOatemala cuador .Uri iguay 0.1 * El Salvad rF Intenational and national AIDS ex- Peru * Me(CoHungary *Colom la e Malaysia penditures are not correlated across Bulgarla Venezuela countries and are usually snnall in LUthuanla Argentina reltion to total public healt Czechoslovakiab spentding; 0.001 - I 0.001 0.01 0.1 1.0 10.0 Spending by national AIDS programs a. Formerly Zaire. b. Now Czech Republic and Slovak Republic. Note: Since both axes are scaled in logarithms, 16 countries with zero national AIDS program spending are omitted from the scatter plot. Donor funding for AIDS is the average of 1991-93 funding from the GPA funding database. National AIDS program expenditures are from the sur- vey conducted by Mann and Tarantola (1996) and are typically for various years between 1990 and 1993. The denominator for the ratios on both axes is 1990 public health spending as esti- mated in World Development Report 1993 (World Bank 1993c). See Pyne (background paper, 1997) for further discussion of the data. at all of their own funds through their national AIDS program in the rel- evant fiscal year, although seven of these received donor contributions for AIDS in excess of 1 percent of their national public health budgets. Since international and national AIDS spending are expressed as a percentage of the same denominator in figure 5. 1, the scatter plot would reveal any tendency for spending from these two sources to be correlated. However, the distribution of points is almost spherical: there is no rela- tionship, either positive or negative, between donor funding and na- tional AIDS program spending. This and further evidence presented below suggest that in the average country, the national program budget was not primarily determined by donor spending decisions. The 45-degree diagonal line in figure 5.1 represents equal allocations of donor funding and domestic funding in response to HIV/AIDS. The thirteen countries below the diagonal received less in donor funds to 244 WORKING TOGETHER TO CONFRONT HIV/AIDS combat AIDS than they spent of their own resources. The 26 countries above the diagonal, plus another 16 that reported spending none of their own resources for AIDS and are omitted from the figure, received more from donors than they spent through their national AIDS programs. Thus in roughly three-quarters of developing countries, donor spending on AIDS exceeded domestic allocations over this time period. This analysis suggests that, although donor allocations for AIDS are not large enough to overwhelm the domestic health care system in most developing countries, these allocations are remarkably large relative to national spending on the same problem and probably in comparison with current international spending on any other disease. Perhaps only the international campaign to eradicate smallpox in the 1 970s benefited from such a large preponderance of donor funds. But the amount of both types of funding varies a great deal from one country to the next. The next section seeks to explain this variation. Donor Favor Lower-income Countries That Have Larger Epidemics As discussed in chapter 3, the severity of the epidemic and the avail- ability of resources should be the two primary determinants of the extent of HIV/AIDS interventions in a developing country. Furthermore, the absolute size of a country's population will affect the scale of activities and therefore of expenditure. How does spending vary across countries by infection rate, GDP per capita, and population size? To avoid double-counting the HIV- infected, the analysis below breaks the population into two components, the number infected and the number not infected. It then examines their separate influence, and that of GDP per capita, on aggregate levels of na- tional and international AIDS expenditures in a country. More than 60 percent of the cross-country pattern of domestically financed AIDS ex- penditures can be explained by these three variables. The analysis focuses on the number of people infected and on GDP per capita to see how these two variables affect national and international allocations for con- fronting AIDS. It is not surprising that national and international decisionmakers respond to the severity of the AIDS epidemic. Figure 5.2 presents the relationship between the number of HIV infected in the country and the amount of national and international AIDS expenditures in a country, after controlling for the number of people not infected and for GDP per 245 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Figure 5.2 ReWlionship between the Number of HNWinfected People in a Country (in Millions) and the Amount of National and Inlernational AIDS Expenditures National AIDS expenditure Intemational AIDS expenditure (millions of 1992 dollars) (millions of 1992 dollars) 100- 100- 10 ± 0 1.0 * * 1.0 0.1 *0.1 0.01 * 0.01 . The expenditure of intemational donors is more responsive than that of 0.001 0.001 national governments to the number I af HIV-infected people in a country. 0.0001 0.001 0.01 0.1 1.0 0.0001 0.001 0.01 0.1 1.0 Number of HIV-Infected people Number of HlWnfected people (millions) (millions) Note: The country data plotted in the graphs have been adjusted for the influences of the country's GDP per capita and the number of uninfected people in its population. See note 9 in chapter 1. Sources: Expenditure data: see note on figure 5.1. HIV infection data see Pyne (background paper, 1996). capita. The relationships are positive (and statistically significant) in both cases, but the international donors were responding much more to the number of HIV-infected people than were the national govern- ments. Every 10 percent increase in the number of HIV-infected people (after controlling for the other factors) is associated with a 6 percent increase in international spending in the country, while national spend- ing rises by only 2 percent. Although it is reasonable for governments to respond to evidence of HIV infection with increased funding for both prevention and curative programs, national governments that view the epidemic with urgency might be expected to respond more strongly to HIV infections than would international donors, not less so. One possible explanation for the weakness of the national govern- ments' spending response to the epidemic might be the availability of international donor funding. If this were true, one would expect that some of the variation in national expenditure, after controlling for HIV 246 WORKING TOGETHER TO CONFRONT HIV/AIDS infections and the other variables, could be explained by the receipt of donor funds. However, as we have seen in figure 5.1, there is no statisti- cally significant correlation between national and donor funding in a country. Moreover, this is true even if we correct for the influence on national expenditure of GDP per capita, the number infected with HIM and the number not infected.3 Although some individual governments undoubtedly made their decisions on national funding levels based on what they were receiving from outside, this evidence suggests that this is not the case for the average country. Turning to GDP per capita (figure 5.3), we see in the left-hand panel that national spending is extremely responsive to national income: of two countries of the same size and the same number of HIV-infected people, the one that is 10 percent poorer spends about 12 percent less on managing its AIDS epidemic. And the fit to that relationship is quite good. This striking sensitivity to income level could be explained as the rational response by decisionmakers who have full information about the danger of AIDS and the role of the public sector in confronting it but are not convinced that government intervention can slow the Figure 5.3 Relationship between GDP Per Capita and National and International AIDS Expenditure National AIDS expenditure International AIDS expenditure (millions of 1992 dollars) (millions of 1992 dollars) 100- 100- 10- 10- 1.0- -- @.**. 1.0 - * -. .V, ~ ~ . 0.1 * 0.1 *, . * -. 0.01- * 01 With an AIDS epidemic of a given size, countries with more natonal 0 .001 * 0.001 resources spend more on AIDS, while 0.001 - 0.001 m-~~~~~~~~~~~~~rceivng less from donors. 100 316 1,000 3,162 10,000 100 316 1,000 3,162 10,000 GDP per capita (1992 dollars) GDP per capita (1992 dollars) Note: The country data plotted in the graphs have been adjusted for the influences of the country's number of HIV infections and the number of uninfected people in its population. Source.: See note on figure 5.1. 247 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC epidemic and are acutely aware of the many other demands on extremely scarce public resources. Under this interpretation, national decisionmak- ers view AIDS expenditures as a luxury, affordable only at higher income levels. Alternatively, it could be that decisionmakers in lower-income countries have less complete information about AIDS than those in other countries and perhaps are handicapped to a greater degree by con- servative constituencies. Either of these interpretations suggests that donor assistance is acutely needed in the lowest-income countries to en- able significant national activity against AIDS. But the latter interpreta- tion, which is supported by the discussion of the political economy of AIDS in the last section of this chapter, further argues that low-income countries should strive to increase their national efbort againstAIDS in order to ensure that people most likely to contract and spread HIVare able to pro- tect themselves and others. The right-hand panel shows that the level of donor spending is also re- lated to the recipient country's income, but in the opposite direction. This bias of the donors in favor of the poorer countries compensates to some degree for the much smaller national expenditures there; poorer countries receive somewhat more in donor funding than less poor ones, after correcting for population size and epidemic severity. However, donors do not fully compensate for the reduction in national spending: the country that is 10 percent poorer receives only 3 percent more donor resources. Furthermore, although the relationship is statistically signifi- cant, the fit is not very good. Thus many factors besides population size, epidemic severity, and GNP per capita influence international assistance to a country's AIDS program.4 While donors must and should take many other considerations into account, this evidence suggests that donors should give somewhatgreater consideration to per capita income than they did from 1991 to 1993 when determining how to allocate resources to confront AIDS across countries, so that low-income countries with severe epidemics would be sure to receive the resources neededfor the essential core functions of an AIDS program. Box 5.1 gives a detailed breakdown of AIDS funding by source for four countries and the Brazilian state of Sao Paulo. These detailed data from in-depth background studies performed for this report confirm the patterns discussed above (backgroundpaper, Shepard and others 1996). First, donor and national spending on AIDS both vary a great deal, even within this small sample of five countries: national government AIDS spending ranges from only 5 percent of AIDS spending in Tanzania to 248 WORKING TOGETHER TO CONFRONT HIV/AIDS Box 5.1 Government, Private, and Donor Expenditures on AIDS in Frve Countries SHEPA.RD AND OTHERS I 1q%ni EX;AMINED THE LEVEL \Variations in incidence explain why Tanzania, and source ot'expenditures on HIV/AIDS in Tan- wvith the highest AIDS incidence 114.3 per 100,0001 zania. C6te dIvoire. Thailand. Mexico, and Sao has moderately high expendirures per capita despite Paulo State. Brazil.' Box table 5.1 retlects [he inter- the lowest per capita GNP. while Mexico, with the narional dollar amount and percentagze of funding lowvest incidence. also has the low%est expenditures by source for each of the countries. despite the second highest GNP. With rhe exception of Nle-ico. public funding per Finally, political factors within the country and capita rises steadily %'ith per capita GNP. Donor the donor communiry are also important.Tanzania'k funding is by Far the largest share iS) percenti ot egalitarian ideals and relatively honest administration resources in Tanzania: in other countries. it is no hav-e long earned respect from the international more than 12 percent of total AIDS expenditures. donor communitY. and helped the country gain The importance of donor Funding outstrips its nione- international support For its ctff,rcs to contcrol AIDS. tanr %alue. First, it is insulated from domes[ic political Thailand's openness in addressing AuDS through its pressures tfrom patients and health care providers National AIDS Task Force. chaired by the prime to,,ard curative care. pocenriallv at the expense of pre- minister, has also brought support for that country s vention. Second. it may plav a catalytic role. shoaxing progaram. the efleti%encas of prc%entike expendirures and spark- ing contributions from other source. of funding. The shares of public expenditures devoted to AIDS differ from total expenditurc for AIDS by morc than I percentage point onls in Tanzania and 'in the urme ajadAbi for rho, 'rud,. 1rairlri, disa on AIDS Thaland.I In F1nz'nia. the overall share is m h * I epcndirurs ,oWld bc obraind onh, (or the 4iare ot Sao Pauto. Thailan3 . In 'anzan2am,, (he o--erall shnare is nmuch \X i'h i I9-) I populairen :A1 33 riliJonihn ti. 'te i larger thin higher due to substantial donor funding. In Thai- t%,- ,F rhe Fire iQdlic,t. in the ,rud% iTananroa ani Cote land. the overall share is smaller because Thailand's d'CItir` AS it contains ;-s iol Bruzil' r-portid AIDS eaSes. Ehc p-sia s AIDS siruairorn i belie ,d ic CJjPELLI the eralhn' oF AIDS extensire pre%encion program Is prcdominandl) pub- e,sp:-nduure, n Brjdr Economic and cxpendirur, dani sere liclk funded. inferrcd from naii'nal 'iJtJiLu . Box Table 5.1 Per Capita AIDS Expenditures Broken Down by Source of Financing in Four Countries and Siao Paulo State, BraSil t1993-95 international dollars) Souirce Tanzania Cole d 7Ioire TIaiLaznd Jwexico 5o Pauilo. Brazil .4Aerage Public 0.2u 1.34 3.-iS 07. ;i.-8 2. 31 iPercenti Si I i i2 1U I2, i6C'-, lPs, Private 0.39 1.69 0.76 1-1-0 2.65 1.2 3 Perceii.i ,1rl i3i I 1 s-C' i 31, '31 I Donor 3.12 0.16 0.56 0.02 0.26 0. 82 IPerentr 18M1I iSi 112 III 131 2 Toral 3.cS 3.18 s4 L- i. Si o.(') 4.16 AVr, - C lIrImni arr in£ed leh io r ight fro m lo%tc-r to highe'r C] DP ein Bro .Ih.'IIi paper S;hepard and thari 1''96 249 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 72 percent in Thailand. Second, donors clearly favor the lower-income countries over the higher-income ones, while national government spending on AIDS is positively correlated with income. And last, the low spending on AIDS in Mexico reflects the trend, discussed above, for the level of spending to be associated with the number of HIV-infected. Bilateral and Multilateral Funding and the Stage of the Epidemic H AVE BILATERAL DONORS AND MULTILATERAL ORGANIZA- tions responded differently to the epidemic? Figure 5.4 shows the allocation of approximately $1.2 billion in donor funding recorded in the GPA funding database over the period 1991-93 according to the type of donor and the stage of the epidemic in the recipient country.5 While bilaterals allocated the lion's share of their AIDS assistance ($316 million, or 63 percent) to countries in the gener- alized stage of the epidemic, multilaterals allocated the bulk of their as- sistance ($379 million, or 62 percent) to countries in the concentrated stage of the epidemic. International NGOs (not shown) accounted for only a small portion of total funding reported in the database; the 16.4 million they provided was about equally split between countries with concentrated epidemics and countries with generalized epidemics. This discrepancy between the funding patterns of bilaterals and multilaterals may have been only temporary and was perhaps partly due to two large World Bank loans made to India and Brazil during this period, both of which have concentrated epidemics. However, the pat- tern casts doubt on the frequent assertion that countries with advanced epidemics will encounter donor "fatigue" from bilateral donors and be forced to turn to multilaterals as the funders of last resort. The observed pattern suggests instead that the bilaterals are particu- larly concerned about countries where the caseload is highest. Such behavior is consistent with two views of the motives of bilaterals. Perhaps they are altruistically responding to the suffering of countries with gen- eralized epidemics. Or perhaps they view their self-interest as jeopardized most acutely by countries where there are large numbers of infected peo- ple. Whatever their motive, the bilateral focus on countries with gener- alized epidemics has left multilaterals to fund countries at the nascent 250 WORKING TOGETHER TO CONFRONT HIV/AIDS Figure 5.4 Donor Funding for HIV/AIDS Interventions in Developing Counties in 1993 by Type of Donor and Stage of the Epidemic Expenditures (millions of dollars) 400- . _ * Nascent 3505- Concentrated 315.7 * Generalized 300- * _ * Unknown 250 - k Bilateral donors devoted the largest 200 - _ share of their AIDS funding to coun- tries with generalized epidemics; 150 - 1706 U 128.5 7multilateal institutons focused their 100 - _ funding on countries where the epi- .s m 59.6 -38.6 demic was still in the concenbtated 5032 3 _ _ * ... :stage. Bilateral Multilateral Source.' GPA funding database as analyzed in Pyne (backgroundpaper, 1977, table 8). and concentrated stages. The outcome of this division of responsibility is that countries where the epidemic is nascent or concentrated pay higher costs for their external AIDS funding than those where it is generalized, but it does ensure that funds are available for all countries. It also permits donor governments to focus their resources on hard-hit countries where their constituencies are likely to most support spending. NGOs Extend the Reach of Government and Donor Programs Achieving the most cost-effective response to the HIV/AIDS epi- demic requires cooperation between governments and NGOs, both non- profit and for-profit. But working with NGOs can be costly for govern- ments. Governments need to develop and apply guidelines and procedures to assure that the collaborative relationship operates with minimal friction and maximum effect. Many of the highest-priority interventions require delivering highly differentiated services to small distinctive groups of clients, such as sex workers or poor AIDS-affected households. Effective service delivery requires the ability to learn from, and respond quickly to, the changing needs of a specific subpopulation. Because the needs of one client group are different from those of the next, unit costs are likely to rise rather than fall when the same organization attempts to deliver to multiple groups. In this situation, service delivery costs less if undertaken by 251 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC many small entities rather than by one large one, such as a government agency. When a highly differentiated service has the attributes of a public good (as chapter 1 argues is the case for many AIDS-related prevention and mitigation services), local communities often spontaneously create a nonprofit, grass roots NGO to deliver them, endowing it with finances and volunteer labor (Weisbrod 1977, James 1982). However, in devel- oping countries many local communities lack the internal organization or resources to create their own NGOs, and few have sufficient incen- tives or resources to subsidize services whose benefits extend beyond their boundaries. Thus, governments cannot expect spontaneously cre- ated NGOs to tackle the epidemic alone. NGOs need the public man- date, technical information, financing, and sector coordination that gov- ernments can provide, while governments need NGOs for their diversity, flexibility, potential cost-effectiveness, and credibility with marginalized people. By working together, NGOs and governments can be a formidable force in the struggle against HIV/AIDS. How should governments select an NGO partner to deliver an AIDS- related service? Characteristics of the service to be delivered can often indicate the type of NGO that will be most appropriate, but ultimately governments will have to judge the qualifications of competing NGO candidates for the specific service delivery contract in question. Once government has identified an AIDS-related service that is un- dersupplied by the private market, it must ask the question whether it is possible to (1) precisely specijy in a written contract the quantity and qual- ity of the service to be provided, and (2) monitor the contract for compli- ance. It might be difficult to specify a complete contract either because the quality of the service depends on subjective aspects of its delivery (for example, the kindness and solicitude of an individual delivering home- based care), or because even the physical aspects of service delivery are difficult to measure (for example, whether the condoms reportedly sold to sex workers really went to them or were instead sold to pharmacies catering to the middle class). Monitoring might be difficult for technical reasons (for example, the presence of a government representative in the room to watch an NGO member educate prisoners or sex workers might destroy the rapport between educator and clients), or a government might simply not have sufficient trained and motivated personnel to monitor a large number of NGO contracts. When the contract can be specified accurately and monitored for compliance, the government can choose among all the available NGO WORKING TOGETHER TO CONFRONT HIV/AIDS candidates, induding both for-profit and nonprofit firms, depending only on their technical qualifications for the task at hand. In such cir- cumstances the winning candidate will often be a for-profit firm with no direct ties to any client constituency, because they typically have the capability to mobilize the best expertise in the country, to produce out- puts to international standards of excellence, to follow government or donor guidelines regarding records and accounts, to minimize costs, and to raise capital for expansion as needed to fulfill the contract.6 However, since it typically has no other constituency to satisfy, a for-profit firm will reduce its costs, not only by eliminating waste, but also by reducing the quantity or quality of any unmonitored dimension of service. When the contract for a highly differentiated service is difficult to specify or monitor, there is a strong argument for the government to favor an NGO that has its own constituency with a stake in the quality of the service. In contrast to the for-profit firm, which might divert resources in order to maximize profits, the nonprofit NGO has the incentive to divert resources toward the other services it provides or toward advocacy. Thus, governments considering delegation to an NGO as a solution to their in- In selecting an NGO partner, a ability to completely monitor performance must consider the conformity government may face tradeoffs of the NGOs' overall objectives with the public interest. between credibility with clients, on Different types of nonprofit NGOs have different overall objectives, the one hand, and accountabiliy Different types of nonprofit NGOs to majority preference and its own Broad-based public charities have large constituencies drawn from the objectives and procedures, on general public and are therefore likely to have objectives in broad con- the other. formity with the general public interest. However, such broad-based public charities are likely to be less credible with the dient groups than an organization composed of members of that group. Therefore, in se- lecting an NGO for a specific contract, governments may face a tradeoff Fgure 5.5 Differng Sbteghs of between the degree to which the organization's objectives conform to Four Types of Nonprofit NGOs those of the general public and the organization's effectiveness in work- Credibility wlth clients ing with the specific client group. Figure 5.5 shows the differing strengths of four stylized types of nonprofit NGOs-dient affinity groups, social service clubs, nonprofit firms, and broad-based private * Client afflnty groups charities-along these two dimensions. NGOs of all four types may be indigenous or may be the local affiliate of an international organization. *Social service clubs These tradeoffs can be best understood by considering the two types v Nonprofit firms of organization at either end of the spectrum. Because the management charities and staff of a client affinity group is typically drawn from and selected by Conformity of objectives to majorit the dient group it is serving, this type of NGO will have the most cred- preferences ibility with its dients. Furthermore, to the extent that the services for Source: Authors' construction. which the government pays the organization are perceived by its mem- 253 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC bers as in their interest, the clients will themselves monitor its perfor- mance, greatly reducing government monitoring costs. For these rea- sons, client affinity groups can be very cost-effective in delivering highly differentiated services, such as peer counseling of sex workers. However, the government's ability to delegate to a client affinity group is limited by the fact that the group's interests and objectives will sometimes diverge from those of the general public.7 For example, it might not be a good idea to subcontract to a client affinity group of sex workers the collection of data on the proportion of its members who are HIV-positive, since the group might perceive that publication of such a number would be against its best interest. Moreover, some of the mem- bers of the affinity group may have socially unacceptable objectives that could be cross-subsidized from government resources. Thus, the diver- gence of objectives between the client affinity group and the government implies that a service contract with this type of organization will entail a risk of resource diversion toward the group's own objectives. Addressing this problem will increase monitoring costs. Examples of client affinity groups that attain international recogni- tion and receive international funding are multiplying. Perhaps the two earliest and best known such groups are WAMATA of Tanzania and TASO of Uganda. Founded by female relatives of people with AIDS, these organizations began as grass roots self-help groups providing basic home care services to home-bound or bedridden AIDS patients. Later, with outside support, they began to offer counseling for other HIV- infected individuals, as well as other services. At the other end of the spectrum are the broad-based private charities. These organizations may be religious or secular but typically represent a large body of dues-paying members. Therefore, their constituency mir- rors an important portion of the entire public, those who are willing to contribute regularly for charitable causes. Except for divergences due to religious belief, the interests of this mainstream constituency are likely to conform quite closely to those of the general public. However, private charities may not have credibility with all of the subpopulations that the government wishes to reach with its message and thus might be less ef- fective delivering services to them. An example of a broad-based charity is the Thai Red Cross Society, which organized the first HIV/AIDS sup- port group for affected individuals and their families in 1991 and only later developed the expertise to reach out to sex workers. That first sup- 254 WORKING TOGETHER TO CONFRONT HIV/AIDS port group was a powerful example that led the Red Cross and many different types of NGOs to create 80 such groups by mid-1996 (Phoolcharoen and Phongphit 1996). Two other types of nonprofit NGOs fall between the affinity group and the large private charity. Social service clubs are local charitable organizations typically composed of middle class and elite community members who volunteer their time in order to improve their community. Their direct constituency, to whom they are primarily accountable, is their peer group within their own community. Social service clubs may be able to establish trust in client groups because service providers live in the same community with the clients and are volunteering their time. The members of such social service clubs have skills and education that can enhance the NGO's utility as a delivery organization. Although the inter- ests of the typical social service club will conform to those of the local elite society, they may not exactly match those of the government or general public. For example, a social service club NGO providing AIDS infor- mation to truck drivers in Lahore, Pakistan, mentioned only one possible source of infection: blood transfusion ("Signs of Change. . ." 1996). Nonprofit firms constitute the majority of NGOs in most countries. The distinction between a nonprofit and a for-profit firm varies from country to country and depends both on the tax laws of the country and on the vigor with which they are enforced. The most sophisticated non- profit firms are like for-profits in that they can draw on the best national expertise and be held accountable to international standards. But non- profit firms can, with greater ease and legitimacy than for-profit firms, develop a constituency, independent funding sources, and their own agenda of objectives. However, the nonprofit firm typically is con- strained by a much smaller constituency, perhaps consisting only of the members of its board of directors and their immediate acquaintances. The rapid proliferation of nonprofit firms observed in some countries in response to the availability of service contracts suggests at least some degree of profit motive. For example, in the four years after the Brazilian government initiated a grant program for AIDS-related services, the number of NGOs registered with the Ministry of Health jumped from 120 to 480. A 1996 evaluation report that compared the earlier with the later group of NGOs found a change toward a more consolidated and formal organization structure, greater dependence on government fund- ing, and a tighter focus on service delivery at the expense of public ad- 255 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC vocacy. This change suggests that the profile of the average Brazilian NGO working against AIDS is now closer to that of a nonprofit firm than to either a client affinity group (which would typically be less for- mally organized) or a broad-based charity (which would be less depen- dent on government funding). Of course an NGO can embody the characteristics of more than one stylized type and some have objectives that conform closely to the pub- lic interest while also having high credibility with its clients. Box 5.2 de- scribes one such program in Sonagachi, one of the largest red-light dis- tricts in Calcutta, India. The program combines the characteristics of the nonprofit firm and the social service club. How well have governments done in delegating preventive or mitigat- ing service delivery to NGOs? A program in Burkina Faso, one of the four West African nations with a generalized epidemic, offers an example of how government and NGOs acting in concert can extend the reach of their AIDS prevention and mitigation efforts, achieving better quality and access than if either had acted alone (Van der Gaag 1995). The proj- ect, which is supported by the World Bank, seeks to increase the use of condoms and other contraceptives and change behaviors that facilitate the spread of STDs. The government and NGOs share responsibility and Box 5.2 Helping Calcutta Sex Workers Avoid AIDS iN 1''102 i HE INDIAN COV ERNNENT. INTERNA- anmong sex workers in Sonagachi have declined sig- tional donor,, three local NGCOs. and sex w%orker§ in nificanlk. And srriliingly. HIV prevalence among Sonagachi. onc of che largest red-light district, in the iex %%orkers has remained at less than I.S percent. Calcutta. joined to,eiher to launch a rvmarkabl-, Nlu ch of the program, success is credired ro 1uccessfLI1 STDIHIV Interxention Program. The the iex wvork-ers w% ho have become peer educators, program. known as SHIP. has trained sex %torkers as since other sex w,orkers regard them as trustvorthy peer educators, pro%icding them vsith knowledge advocates or behav-ior change. Moreover, their em- about S FD,. the use of condomi. and negotiation plo-mcnt in the programn his brought communinr rec- slkill. which are essential if sex u%orkers are to ccjn- ognition. self-respect. and dignit. w-hich have encour- vince their clients to LEu condoms without the sup- aged other sex workers to become peer educators. thu-s port ot pimrp and brothel ovwners. helping to ensure that the program will contnue. Ilhe success ot [his approach can he wcen in sct- SHIP h3s been expanded into four other red- era] indicators. The nurnber of condoms distributed light districts n Calcutta: by 199- it was reporied to through the program per month rose from I .S:Ou at cmoer are.as that include more than 80) percent of the the starr of the programn to 65.000 -it the end of se% orkiers in rhe city. 1 ). The numrrber of abortions and [he STD rate 5,,,rcc Singh V'0i 256 WORKING TOGETHER TO CONFRONT HIVIAIDS costs. Government roles include providing supplies at a subsidized rate; launching a national media campaign to promote the purchase of con- doms; and teaching traditional healers to fill prescriptions, diagnose STDs, and refer cases to health dinics. Treatment of STDs will be han- dled primarily by NGOs, which are both nonprofit and for-profit firms; NGOs will also provide training to traditional healers. Both NGOs and public dinics will provide free condoms to people with high-risk behav- ior. The government has also provided encouragement and access to grant money for NGOs to enable them to offer additional services. This type of collaboration lays the foundation for increased coordination in the future between the two actors and fosters an environment of trust. The largest and most elaborate effort to subcontract AIDS services to NGOs is probably the annual competition for service grants in Brazil. Supported by a World Bank loan, the program has funded all four types of nonprofit NGOs, induding client affinity groups such as an associa- tion of transvestites in Rio de Janeiro, and nonprofit firms such as a university-affiliated research center in Sao Paulo. Clients have included children, hemophiliacs, pregnant women, feminists, transvestites, pros- titutes, drug users, prisoners, truck drivers, and men who have sex with men. While grant competitions are managed centrally by an NGO liai- son office attached to the Ministry of Health in Brasilia, state and mu- nicipal as well as federal government agencies provide complementary funding to, and collaborate actively in the execution of, funded pro- grams. In the recent evaluation of this program, only 7 percent of the 111 current grantee NGOs were deemed to be falling short of their project objectives, and only 2 percent were having serious difficulties reaching their target populations. The financial control mechanisms em- ployed by the liaison office, which include an annual visit to each grantee and audit of its accounts, has identified serious mismanagement in con- nection with less than 1 percent of projects. While the NGO liaison of- fice constitutes a substantial new and expensive function for the Min- istry of Health, in the four years of its existence it has facilitated the funding of 308 projects and disbursed a total of $14 million. Although the total impact of this activity on HIV infection rates in Brazil has not been assessed, it is clear that no government agency could have carried out directly so many diverse and precisely focused activities with these resources. Unfortunately, to our knowledge there is no systematic study that compares the merits of alternative government procedures for evaluating 257 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC NGO proposals for an AIDS-related service delivery contract; nor are we aware of any study that compares ways for governments to monitor NGO performance under such a contract (National Research Council 1996, appendix to chapter 6). A starting place for such a study would be a comparison of the lessons learned in the recent service contracting experiences of Brazil, Burkina Faso, and Thailand. The availability of a set of standard, transparent, internationally recognized procedures for governments to follow in delegating service provision to NGOs could greatly facilitate government-NGO cooperation and minimize the disappointment of all parties concerned. AIDS donors, NGOs, govern- ment policymakers, and indeed the entire international health commu- nity would benefit from studies of the costs and effectiveness of alter- native procedures for identifying effective NGOs to be service providers and for monitoring their performance. Such studies are but one example of an urgently needed international public good, the topic of our next section. Who Will Invest in New Knowledge and Technology? D ONOR SUPPORT FOR NATIONAL AIDS PROGRAMS IS IMPOR- tant and, in a nascent epidemic, often critical; yet there are other crucial activities in which donors have a greater compara- tive advantage and a clearer public economics mandate. Because the benefits of prevention programs accrue primarily to a country's own population, all but the poorest national governments can and should finance a significant share of these costs. In contrast, donors are in a unique position to mobilize international support for the creation and dissemination of knowledge and technology that is transferable across countries. This section first discusses the organizational response and financial contributions of bilateral donors and multilateral organizations since the start of the epidemic. It then explains why knowledge and technology should be regarded as international public goods that the donor community alone is likely to provide. Finally, it discusses the need for specific types of knowledge and technology, induding a vaccine, and organizational innovations for tapping the creative energy and resources of private firms. 258 WORKING TOGETHER TO CONFRONT HIV/AIDS The Evolution of Donor Policy Although AIDS was first diagnosed in 1981, a systematic international and national response to the epidemic was not evident until the late 1980s. In many parts of the world, NGOs led the way in providing care and prevention services for individuals and communities affected by the epidemic (Mann and Tarantola 1996; background paper, Pyne 1997; Sittitrai 1994). The incremental and relatively limited response of WHO in the early years has been attributed to resistance by many mem- ber states to addressing the problem of HIV/AIDS (Panos Institute 1989). The establishment of the WHO Global Programme on AIDS (GPA) in 1987 helped to generate momentum for global prevention and mitigation efforts; that same year the U.N. General Assembly adopted a resolution encouraging U.N. agencies and other members of the U.N. family to initiate their own HIV/AIDS activities (Mann and Tarantola 1996). During its early years GPA focused on helping national governments develop strategies to curb the spread of the epidemic. The year GPA was established, 170 countries requested assistance; by 1989 GPA had helped 151 countries to establish national AIDS programs, 102 coun- tries to develop short term (6 to 12 month) plans, and 30 countries to develop medium-term (3 to 5 year) plans (Panos Institute, 1989). Largely as the result of the prodigious efforts of GPA, almost all coun- tries today have national AIDS programs; most of them were formed be- tween 1985 and 1990. Meanwhile, in response to the U.N. General Assembly Resolution, UNDP, UNICEF, UNFPA, and UNESCO developed a joint HIV/ AIDS strategy document that specified the resources and staffing that each would allocate to combating the epidemic. UNDP played the most prominent role, devoting 2.1 percent of overall agency resources and 0.43 percent of overall agency staff (Garbus 1996, as cited in background paper, Pyne 1997). Other multilaterals also initiated AIDS programs. In 1987 the European Union established the AIDS Task Force in order to fund AIDs-related programs in developing countries. The World Bank, which made its first loan exclusively to combat AIDS in 1986, had financed 61 projects in 41 countries, for a total commitment of $632 million by the end of 1996 making it the largest source of funds for confronting HIV/AIDS (background paper, Dayton 1996; World Bank 1996a). 259 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC In the late 1980s, the wealthier donor countries, in addition to mak- ing contributions to the GPA and providing support through the other multilaterals, also launched their own bilateral HIV/AIDS programs. By 1993 the largest of these was the U.S. program; launched in 1988, it includes the centrally funded AIDS Control and Prevention Project (AIDSCAP) as well as other activities initiated and funded by country USAID missions.8 Other countries with large bilateral AIDS programs include Canada and Norway (launched in 1987); Denmark, Germany, the Netherlands, Sweden, and the United Kingdom (1988); Japan (1989); Belgium and France (1990); Australia (1991); and Switzerland (1993). Table 5.2 gives the total amount spent in 1993 by twelve major donor countries. Under the leadership of the GPA, many national plans were written, many AIDS interventions were launched, and many national leaders became aware of the severity of the AIDS epidemic. For the first time, senior policymakers discussed high-risk sexual behavior and how gov- Table 5.2 International AIDS Expenditures through Bilateral and Multilateral Channels, by Major Donor Countries in 1993 and Net Immigration in 1992 (millions of dollars except as indicated) Aet m lulti - inigravtion Coetn' Bilkueral Literal Botlh Total (ihousanhI.l Linired Scares 8.(l 3-4.0 1.1) H l-O -c)4 France I 8.5 1. 0.! 20.0 86 LUniced Kingdom .8 .S.q4 n.a. 1'.2 14- Germany .8 )9 4.1 12.8 S Canada S.' 3.1 0i.3 11 .6 I .Seden 3. I. 1.0 9). 2e Norv,dy 4.6 2.5 2.3 '1.4 WO Denmark 2.1 2. 4.1 X.9 ' 2 Au,crafia 4. t.' t. 1 S Netherlands 2.- 2.4 0 t..I -d3 jap3n 1.0 -.i n a. Lu.Xen1burg [0 0.3 n a. 1.2 6 Toral Ior I2 dccnors 146.4 65.9 14.1 216.3 216 n.a. Nor ipplicable. V'r: Funding IoraIE cIude Lhe AIDS shafe o1 nflaonlil nrr1buuu,n% ru the rnuldrlir- eTal lending igencIes. r,.r La's LP e'0. cabl 3s I . ind ECD C i" rabkl 1.1. p. -'4 26o WORKING TOGETHER TO CONFRONT HIV/AIDS ernments should respond. However, the epidemic continued to spread. In the early 1990s a group of member states, especially the donor gov- ernments then funding the GPA, became concerned that, as a part of WHO, it had insufficient mandate to coordinate the expanding efforts against the epidemic across the U.N. system. The donor community perceived that the GPA was unable to restrain donors from competing vigorously with one another instead of cooperating around a mutually agreed plan of action and came to believe it necessary to create a special- ized international institution with an explicit mandate to coordinate the work of the other U.N. agencies at the country level. As a result they worked with UNDP, the World Bank, and other multilaterals to create a new special-purpose U.N. program dedicated uniquely to combat- ing AIDS. The Joint U.N. Programme on AIDS, widely known as UNAIDS, officially began operations on January 1, 1996. It is based in Geneva and works most closely with its six cosponsoring agencies: WHO, UNDP, UNICEF, UNFPA, UNESCO, and the World Bank. It is governed by a Programme Coordinating Board (PCB) of 22 mem- ber states and 6 cosponsors, plus, for the first time in the U.N. system, 5 rotating nonvoting representatives of NGOs The PCB has assigned UNAIDS four roles: first, policy development and research, which are to account for a larger share of UNAIDS' activi- ties than it did of the GPAs; second, like the GPA before it, UNAIDS is to take the lead among U.N. agencies in providing technical support to national AIDS programs around the world; third, the program is com- mitted more formally to advocacy on behalf of HIV/AIDS prevention and mitigation than was the GPA; and finally, UNAIDS is charged with the difficult task of coordination of its cosponsors and of other U.N. agencies. In this last role, it can potentially address the needs described in the next section by serving as a forum within which multilateral and bilateral donors can agree to donate more to AIDS research, prevention, and control than they otherwise would. Since cooperation with other donors at the country level entails substantially increased costs to each donor and deprives each of being able to claim sole credit for supporting the government on a specific activity, the incentives for such cooperation are weak. Since UNAIDS lacks the power to enforce cooperation from its co-sponsoring multilaterals, much less from the bilaterals, the hope for this form of donor cooperation lies in the good will of the staff of the various donors working at the country level-perhaps reinforced by the insistence of the national government.9 26I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Donors Should Focus More on International Public Goods One explanation for international assistance to help developing coun- tries combat their AIDS epidemic is altruism. Just as famine and flood overseas can elicit an outpouring of generous assistance from more favored countries, the disease problems of low-income countries have often been the cause of generous government and private contributions. However, in the case of an infectious disease that even the most sophisticated medical technology cannot always cure, like drug-resistant tuberculosis, the Ebola virus, or HIV, it is also in the self-interest of higher-income countries to help poorer ones combat the disease. Chap- ter 1 argues that there is a compelling role for government in the pre- vention and control of infectious disease. Figure 5.6 illustrates that in in- death from infectious disease in dustrial countries HIV is estimated to have caused 65 percent of adult industrial countries, HIV could deaths from infectious disease in 1990 and, unless new antiretroviral account for twice as many adult therapies are effective and become widely available and affordable, is deaths by 2020, unless new treat- projected to account for over 96 percent of such deaths in 2020.10 This ments prove to be effective and widely affordable. is much higher than the HIV share of deaths from infectious disease in developing countries (see chapter 1). The current and future magnitude of HIV's contribution to the in- Figure 5.6 Deathseof Adults irom fectious disease burden within industrial countries' borders provides in the Established Mafket them with two reasons to spend money on HIV control in the low- Economies, 1990 and Projected income countries. First, any lessons learned about how to slow the to 2020 spread of the epidemic, whether through behavioral modification or Annual deaths technological advances, are potentially applicable at home. Second, be- (thousands) cause HIV is infectious and the higher-income countries exchange thousands of tourists and attract thousands of legal and illegal immi- 100- * HIV grants to their shores every year, a reduction in HIV prevalence in 80 Other 84 low-income countries has a secondary effect of protecting the citizens of higher-income countries. Evidence suggests that countries are 60- _ -'-6 already aware of these arguments: the five countries that provide the ii i l | 3 most support to the global effort against AIDS also receive the most 40 ---I i . --1131 immigrants. Assuming that self-interest is at least part of the explanation for the 20 - high-income countries' contributions to AIDS prevention in the devel- 21 l _ oping world, will this be sufficient to generate the globally optimal 0 - 990 2020 expenditure on AIDS control in developing countries? Recall the discus- sion in chapter 1 of the difficulty in coordinating the contributions to Source: Murray and Lopez 1996. mosquito control of all the individuals who inhabit mosquito-infested 26Z WORKING TOGETHER TO CONFRONT HIV/AIDS land. Once the mosquitoes are gone, even people who have contributed nothing to the effort will benefit. Since each individual can hope to "free-ride" on the others' contributions, each holds back from giving as much as he would be willing to pay to end the mosquito infestation. A similar free-rider problem threatens to prevent the donor countries from voluntarily donating as much to the AIDS effort in developing countries as the abolition of the epidemic would in fact be worth to them. Because it suffers from this international free-rider problem, the effort to combat AIDS can be viewed as an international public good. Another good on which it is easy to free-ride is new technical infor- mation, such as that generated by frontier medical research on treatment of AIDS and opportunistic illnesses, AIDS vaccines, or, to the degree that the results are transferable across countries, by operations research on the best way to market condoms to the people most likely to contract and spread HIV The solution to problems of local or national public goods is typically government intervention. At the local level it is in the interests of all the individuals concerned to support a government that taxes them all and uses the taxes to control mosquitoes and fight other infectious diseases. A similar argument can be made for an international government with the power to tax countries and spend the proceeds on international pub- lic goods such as the control of HIV/AIDS. However, since countries are unlikely to surrender their sovereignty to a supranational body for this or any other reason, another solution to the international free-rider prob- lem must be found. As an alternative to government, the individuals who live on the mosquito-infested land could negotiate with and persuade one another ("I agree to give more if you will") until sufficient money was raised among them all to solve their joint problem. While requiring more time and effort from individuals than the simple solution of a tax, the negoti- ated solution is potentially workable. At the international level, the United Nations is a forum for such negotiation and persuasion. Through it, countries can potentially be persuaded to donate their "fair share" to international public goods, such as AIDS control. Thus from the public economics point of view, it is not surprising that donor countries have been willing to donate to AIDS control and to research on AIDS. However, given the free-rider problem, it is unlikely that the donor countries have committed as much as it would be in their joint best interest to provide. 263 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Investments in International Public Goods Information that can be generalized beyond the country in which it is produced can originate in either the social or the physical sciences. This section discusses both types of knowledge and a third type of interna- tional public good: international institutions. The medical and social sciences of epidemiology, sociology, econom- ics, and operations research are necessary to track the epidemic and to learn what sort of interventions prevent the most secondary cases of HIV infection per government dollar spent. Applied social science research offers the greatest hope for immediately slowing the spread of AIDS and of improving the well-being of the hardest-hit survivors. The biological sciences, including microbiology, immunology, and virology, are making slow progress toward a vaccine and a cure. However, market imperfections mean that only a small share of biomedical re- search is designed to produce products or knowledge that will benefit low-income countries. WHO's Ad Hoc Committee on Health Research estimates that 95 percent of spending on health research and develop- ment is directed toward solving health problems that mostly affect the richest 10 percent of the world's population; only 5 percent of such spending is directed toward the diseases that account for most of the dis- ease burden of the remaining 90 percent of the world's population (Ad Hoc Committee 1996, p. 102). An important role of governments, especially of donors, is to tilt incentives for medical research somewhat more in favor of the low-income countries. A third important type of public good is the international institution that enables a group of countries to coordinate their efforts in their mutual best interests. Two types of international institutions are relevant to the AIDS epidemic: those among low-income countries in a region, and those that bring poor and high-income countries together in a com- mon struggle against HIV/AIDS. Information from the social sciences on behavioral interventions. Any successful preventive intervention among individuals who are very likely to spread the virus will produce positive spillover effects for the host country, in the form of reduced secondary transmission, which to some degree will also benefit other countries. But the most valuable out- put of such an intervention for the outside world is knowledge that can be applied in other countries. Donors who fund behavioral interventions 264 WORKING TOGETHER TO CONFRONT HIV/AIDS have a responsibility to ensure that the opportunities for the generation of new knowledge that arise from such programs are not wasted. Although the imperative to learn from interventions seems self- evident, surprisingly little is being done in this regard. Recent literature reviews found that publicly available written evaluations exist for only about 10 percent of donor-funded interventions. Worse, of the few hun- dred published studies, very few were conducted with sufficient thor- oughness to determine whether or not the intervention actually changed the risk behavior or HIV incidence (Choi and Coates 1994; Oakley, Fullerton, and Holland 1995; National Research Council 1991).11 The reviewers noted many deficiencies in the available studies. In some cases the lack of baseline data made it impossible to know whether a measured difference between a control and experimental group was due to differences in the two groups present before the intervention. In others, baseline data were collected but there was no control group against which the intervention group could be compared. Few studies attempted to determine whether changes in behavior were due to the intervention or to a placebo effect arising from the existence of the study. To be sure, ethical considerations and the complexities of research with human subjects often make it impossible to use a true experimental approach. An alternative is to have copious baseline data and implement quasi-experimental research designs (Moffitt, 1991). However, very few studies attempted such an approach. Differences between the standards of knowledge for pharmaceutical products and those for behavioral interventions against HIV are striking. Since pharmaceutical products can be patented, private firms have a strong incentive to win the race to the market with a new drug. Govern- ments have responded by requiring that companies prove the safety and the efficacy of new drugs, typically at a cost of millions of dollars. These sums are spent even on such relatively minor drugs as a new headache pill in order to ensure very high standards. The government does not hesitate to require such expenditures, knowing that firms will spend this money on any drug they think will pass the market test. In contrast, preventive interventions that have the potential of pro- ducing far more public benefits, in the form of secondary HIV infections averted, are held to much weaker standards. Since these kinds of inter- ventions can not be patented and they produce positive externalities, the public sector typically must finance them. If governments held them- 265 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC selves to standards as rigorous as those they set for pharmaceutical man- ufacturers, HIV prevention interventions would be forced to meet stan- dards of rigorous design and data collection methods that would enable the public to learn whether an intervention will be safe and efficacious in subsequent application. Although it might seem that safety would not be an issue, the exam- ples of needle exchange programs and counseling and testing for HIV infection suggest the contrary. It is precisely the fear that the provision of clean needles might encourage injecting drug behavior and that an HIV testing program, even with the accompanying counseling, might reduce the propensity to practice safe sex among those who are told they are positive that often undermines public support for these programs. The public has an interest, and indeed a right, to know the magnitude of any such "side effects," as well as the efficacy of the intervention, before it finances its continuation or expansion. Information from the biological sciences on medical interven- tions. With potential profits protected by the patent system and a large potential market in industrial countries for an AIDS cure, research by both private firms and nonprofit institutes has been intense in the in- dustrial countries. The most recent product of that research is the triple- drug therapy discussed in chapter 4. As shown, the high costs of provid- ing this therapy mean that it will not be of immediate benefit to the 90 percent of HIV-infected people who live in low-income countries. Some observers, aware of these prohibitive costs and pessimistic about the prospects for successful behavioral interventions, believe that the only hope for reducing the impact of HIV on low-income countries is a vaccine. But vaccine research of all types faces serious impediments.12 These include the increasing complexity and expense of vaccine research, the need to sell perhaps 40 million doses before production processes attain economies of scale; the inability of people in developing countries to afford expensive vaccines; and, perhaps most serious, companies' vul- nerability to damage claims in the millions of dollars, if even one dose of a vaccine causes the disease it was designed to prevent (Ad Hoc Com- mittee on Health Research 1996, Robbins and Freeman 1988). Partly as a result of these impediments, worldwide public and private sector in- vestment in vaccine development totaled a mere $160 million in 1993, compared with an estimated $1.3 billion spent on other approaches to prevent HIV infection and about $5 billion spent on HIV-related health care (FitzSimmons 1996). 266 WORKING TOGETHER TO CONFRONT HIV/AIDS In order to achieve the substantial international public benefits of vac- cines for the diseases of the developing world, governments must play a role. The May 1997 announcement of a U.S. goal to produce an effec- tive AIDS vaccine within ten years as a U.S. national goal is welcome news not only for people in the U.S. but for people everywhere, includ- ing developing countries. His choice of a ten-year target date, which some experts believe to be too optimistic, is a sobering reminder that no vaccine will solve the AIDS problem in the developing world in the near future (see box 5.3.) The need for government involvement is apparent not only for an AIDS vaccine, but also for other medical advances, which would sub- stantially benefit people in the developing world who lack the purchas- ing power to motivate the pharmaceutical companies of the industrial countries. Examples include vaginal microbicides and simple inexpen- sive diagnostic kits for classic STDs such as chlamydia and chancroid that are currently difficult and expensive to properly diagnose (Ad Hoc Committee on Health Research 1996; Elias and Heise, 1994). As the example of the hepatitis B vaccine discussed in box 5.4 makes dear, once a vaccine or other drug has been invented, tested, and pro- duced on a large scale, its price is likely to fall to the point at which com- mercial firms can profitably manufacture and distribute it in large quan- tities at prices that are affordable in developing countries. Thus, the need for government involvement is likely to be temporary, but critical. International institutions can produce international public goods. We noted above that the United Nations and other multilateral organizations can provide forums in which countries can persuade one another to contribute more than they otherwise would to the production of an international public good. Two additional types of international institutions that could solve specific kinds of international free-rider problems are private-public alliances for health research and regional cooperation bodies. Public-private alliance for health research. WHO's Ad Hoc Committee on Health Research has recently proposed a "Health Product Develop- ment Alliance" between the public and private sectors whose mandate would be tightly focused on the development of a limited number of products for major causes of disease burden that are currently neglected by existing efforts (1996, p. 101). Such an alliance would use a variety of approaches to improve the incentives for private firms to develop phar- maceutical and other health care products urgently needed in developing 267 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 5.3 Challenges To Be Overcome in Developing an HIV Vaccine THIS BOOKARGUES TEAT DONOR COUNTR]ES AND lntriguinglv. some indiv iduals do seem to have multilateral institucions have a comparative adv an- protective responses that enable them to ward off tage in creating incentives for HIX' vaccine research either HNI' infection or the effecrs of rhe virus. and that doing so would be in their own self- Examples include the apparent absence of HIV in interest, as well as the interest of developing coun- half to three-quarters of babies born to HIV-infected tries. Policymakers asked to provide such support. morhers and apparent resistance to HR' infection in either directly or by supporting mechanisms to gen- a few individuals who remain uninfected, despite erate appropriate incentives, have a right to ask: Is repeated exposure to the virus. Similarly. a few indi- an HIV/AIDS vaccine really possible? What chal- viduals. called long-term nonprogressors. have car- lenges must be overcome? ried the virus for ten or more years but have not The short answer is that many scientists believe that become sick with AIDS. In addirion. trial HIX' vac- avacine is indeed possible. but that the challenges are cines appear to have been cffective in protecting very substantial. The most basic challenge involves the chimpanzees from HiV. while other vaccines appear question of whether human immune responses can to protect monkeys from rhe simian immunodefi- prevent HIV infection or prevent illnesses in a person ciencv virus or SI'. All of these responses could be infected with H-1' after vaccination. Although most due at least in part o a strong immune response. people infected with HIV develop a broad range of A second set of challenges involves the high anti-HR' immune responses (antibodies are one degree of genetic variability in HIV: there is no example), these responses are generally not capable of guarantee that a vaccine developed to protect against eliminaring the infection or preventing progression to one strain would necessarily protect against the oth- disease. Nobody knows s'hether these same immune ers. HR' strains from different parts of the world responses would be more effective if the) were induced have been grouped into ten generic subtypes: A. B, by a vaccine, before exposure to Hn'. C, D. E. F. G. H. I and 0. hlost of these subtnpes countries. These mechanisms, some of which require changes in the tax codes and legislation of participating countries, include: * direct support for the costs of the early stages of product develop- ment * analysis of the potential market for a specific new product that would primarily benefit people in the lowest-income countries13 * tax relief and or streamlined regulatory controls for the develop- ment of products for low-income countries * worldwide tax breaks for pharmaceutical companies and extended periods during which they have the exclusive right to sell the drug (provisions similar to those of the U.S. Orphan Drug Act of 1983) 268 WORKING TOGETHER TO CONFRONT HIV/AIDS are present in Africa, subn-pe B is most common in is unthinklable. information on pro[ecti-e elticacy can developed countfies. Encouragingly. recent research onl be ob[ained from large-scalc Phase Ill field trials. indicares chat the genetic differences among sub- Multiple Phase Ill trials w%ill he necesianr to e-al- types may not necessarily affect the was- that they uare the protective efricacv of different vaccine con- respond to a vaccine. Nonetheless. Lhe issue remains ccpti. against different HPV subntpes, against differ- high on the vaccirLe research agenda and is of parnic- ent routes ot transmission, and under diffcrc-nr ular importance for those developing countries health, nutritional. and/or genetic condirions which where several sub- pes are present. may be present in differcnt countries where the vac- The rhird set of challenges involves the need for cine is Lo be used. In order to zather the necessan- human trials; and a related need ro ensure that these information, these trials must be conducted in in- are conducted according to accepted standard, of duscrial and developing countries. The United States medical ethics. Despite the progress in testing sac- has announced it. intention to proceed w-ith a U.S. cines on chimpanzees and monkeys, human trials are Phhaie Ill trial w%ithin the nex;t n'o sears. and discus- essential to determine the safer- and effecriveness of sions are undersvav to conduct Phase Ill rrials in an HIV vaccine. Mtore than 20 candidate HI[' vac- selected developing countries. The results of these cines have been resred in Phase I and Phase 11 trials trial may become available earl. next centur-. Of with more than 2.000 HIV-negarive volunreers. course, there ik no guarantee thar these trials 'sill mostly in the United Srates. These trials have indi- lead to an etfectist %accine. Ho%%e%er. w%ithout pro- cated that candidate vaccines are safe (Phase 11 and ceeding to Phase Ill trials. an HIV viccine will ne%er that at least some of them induce HIV-specific imi- be as atlable. mune responses (Phase 11 ) which could confer pro- F.pa... Hoy; ud ard Osnano% i Cl. FirThimrron tection against HIX infection or diseasc. Hosvever. c- Gold !`0tirbn[cn[I'.n}iAiW \atne Inamnisni6' because deliberarely exposing trial volunteers to HEX' jAhmmn ')ti,. and tl;mniro.o ' * advance guarantee of a market for a health product that meets cer- tain objectively verifiable criteria. The last idea is a particularly innovative approach to solving the incentive problem. One way to implement the idea would be for one or more low-income countries with a specific disease problem not being adequately tackled by biomedical research to offer to buy a large quan- tity of the first drug or other medical product that meets precise specifi- cations, regardless of the identity of the developer. To be credible, this offer would be guaranteed by a consortium of international donors and lenders. The total financial package could include, for example, grants from bilateral donors and a mixture of soft or market-rate loans from multilateral institutions or even commercial lenders. In the purest form 269 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 5.4 Can Companies Make a Reasonable Return from the AIDS Vaccine? -1 CAN ONLY TELL YOIU ABOUT THE EXPERIENCE Companies must realize thar the porenrial mar- with hepatitis B vaccine. which Aas developed 20 ket for an HIV' vaccine in the developing world is years ago. In the first couple of years the price of the tremendous, bur it can only be captured by using vaccine was as high as $25 to S40 a dose, with three nto or three price tiers. The high price would be for doses needed [US$50 to $80 at 199- prices]. So industriali7ed countries, while developing councries companies aimed for the upper end of the market would have another price. Companies must profit and rhe market was stalled. The price could not go from their investment. And developing countries too high despice the global need for the vaccine. In must be able to afford the vaccine. Figuring out how, china alone. xsith 1.2 billion people. the hepatitis to work this out is a ver' important challenge Lrr carrier rate is 10 percent. Y'et many countries were government. business, sciencists and international practically out of the purchasing market. organizations.' 'When the recombinant heparitis B v-accine %as developed, the price decreased a little. Now the price has come down to probably vUS1.00 per dose. [less than rwo percent of the initial price]. And four vears Dr. Nath Eharnprim. chairman. Subcnmn e on H1' ago, Thailand puc the hepatitis B vaccine on the \AJccinc Tris deNeloper o' a .acrine agunt deneue hemorrhagic general program of immunization. So all bibies in fever. and former pre;i&dnr of iahidril Llni%rsity. Bangko.L, our country nowx receive the vaccine. Thali3nd i-ronm inttr%rw. pubiThed in L\% R i l091 i. of the guarantee, none of this financing would be released until the de- sired product was approved by independent testing laboratories. Only then, as the culmination of a period of research and development that might last as long as five or ten years, would the financial instruments be executed, the donor contributions delivered, the international loans dis- bursed, and the delivery and distribution of the product initiated. The most urgently needed anti-AIDS technology-a vaccine to pro- tect against HIV infection-is already the subject of a public-private partnership. Established in 1996, the International AIDS Vaccine Initia- tive (IAVI) is the first attempt to organize a health product development alliance along the lines recommended by the Ad Hoc Committee on Health Research. First proposed by the Rockefeller Foundation, the LAVI has attracted support from the Merieux Foundation, UNAIDS, and the World Bank, and Until There's A Cure, an NGO with links to the AIDS community. Its mandate is to accelerate the development of HIV vaccines appropriate for worldwide use by reducing obstacles to vaccine development and filling gaps in the current effort. In 1997, IAVI's first full year of operation, participants expected to devote a total of $2 million to $4 million in direct support of research on an AIDS vac- 270 WORKING TOGETHER TO CONFRONT HIV/AIDS cine (IAVR 1997a). In light of the U.S. initiative to produce an AIDS vaccine, the challenge to IAVI will be to assure that vaccine development programs do not neglect the needs of low-income countries, where 90 percent of HIV infections occur. Other important potential applications for a public-private alliance in the AIDS arena lie in the development of vaginal microbicides and viru- cides, which would allow a woman to protect herself from HIV infection without asking her partner to wear a condom. Poor women in develop- ing countries are often at a particular disadvantage in negotiating con- dom use with their partners; yet these same women are also most likely to lack the purchasing power to buy a vaginal virucide. By guaranteeing the market, a public-private alliance would provide an incentive for pharmaceutical companies to develop such products. Public-private partnerships could also offer incentives for development of reliable, in- expensive diagnostic tests for STDs such as chlamydia, which are preva- lent in developing countries and accelerate the spread of HIV, and to stimulate research that would lower the cost of antiretroviral therapies. Regional cooperation. When an AIDS epidemic first comes to public attention, many people and some governments blame neighboring countries or "foreigners" generally for the introduction of the virus. But every infectious event, whether through sex, or needle sharing, or trans- fusion, involves two people. One of them must be a national resident if the epidemic is to enter the national population. For the epidemic to then spread within the country, there must be additional infectious events involving residents. Thus, in any country with a serious AIDS problem, the national population must have taken an active role in its spread. Blaming foreigners for the spread of the disease within the national population is not only illogical, it also undermines efforts to confront the epidemic. First, blaming foreigners gives people who are not in direct contact with foreigners a false sense of security, thus hindering efforts to encourage safer behavior. Second, blaming foreigners may worsen rela- tions with neighboring countries, making it more difficult for govern- ments to coordinate prevention and mitigation efforts. Soured relations may also threaten other benefits of economic cooperation, such as those from migrant workers or from commerce. Rather than casting blame, a more productive approach would be for governments in neighboring countries to discuss ways in which they can cooperate to overcome the sharedproblem of AIDS. For example, neigh- boring countries can agree not to attempt to screen out HIV-positive 271 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC migrants; to share information about prevention measures and the course of the epidemic; to coordinate policies on AIDS-related social issues such as prostitution and drug addiction; and to offer similar levels of subsidies for AIDS treatment and assistance to affected households, to avoid creating incentives for HIV-positive people to migrate in search of higher subsidies. Donor governments and multilateral institutions can play a useful role in supporting such regional dialogues. Overcoming Political Impediments to Effective AIDS Policy T HE POLICY MESSAGES OF THIS REPORT ARE NOT STARTLING new findings. The call for preventing infections among people most likely to contract and spread the virus is a reiteration of arguments for the control of the sexually transmitted diseases that were already recognized 20 years ago (Brandt 1987). The warning that subsidizing AIDS treatment more generously than, say, cancer treat- ment endangers the quality and accessibility of health care for everyone is familiar from discussions of health sector reform (World Bank 1993c). The finding that the poorest households are most vulnerable to the shock of an AIDS death is consistent with previous work demonstrating that poor households have difficulty weathering other kinds of shocks. The condusion that "survivor assistance" provided by the government or NGOs should be targeted to the poorest AIDS- affected households follows logically. The advantages of decentralizing and privatizing government service programs are well established. At the level of international public goods, the need for better knowledge and technology for developing countries has been glaringly apparent for years. If these messages are familiar, why are they not being followed in countries around the world? The answers clearly lie outside the technical discussions that have occupied this book so far and fall instead into the domain of political science, a less-developed discipline than either epi- demiology or economics, with fewer guiding principles. However, the examples of countries that have achieved modest success in confronting AIDS suggest some lessons. 272 WORKING TOGETHER TO CONFRONT HIV/AIDS Interest Groups and AIDS Policy Many groups with divergent interests affect the design and imple- mentation of HIV/AIDS policy, and the mix of groups and their relative strength changes over the course of the epidemic. At the outset, few groups are concerned. However, as the epidemic progresses, the number of interest groups increases and the politics of AIDS becomes ever more complex. Early in the epidemic, physicians and medical suppliers have an in- terest in learning how to treat AIDS and how to protect the safety of health care workers from needle sticks and other accidental infection on the job. A group that emerges at about the same time is people infected with HIV. Although the number of these individuals is initially small, and they may at the outset lack political influence, they are often highly motivated to lobby government, since their very lives may depend on persuading the government to subsidize AIDS treatment and care. As the epidemic spreads, the size of this group and its potential to influence government policy increase. Often overlapping with this group are indi- viduals who practice high-risk behavior but who are not infected-or hope they are not infected. Although these people have a strong interest in government-subsidized prevention for themselves, in the early stages of the epidemic they are rarely well organized enough to lobby on their own behalf Yet increasingly the AIDS epidemic has induced people with the highest risk behavior to organize in order to promote their interests. Furthermore, NGOs working on HIV prevention and AIDS care be- come advocates for the populations they serve. Finally, as the number of AIDS cases increases, insurance providers and employers will become concerned about rising health care costs and increased sickness and death among employees. At all stages of the epidemic, the largest interest group is the one least motivated to learn about the issues or lobby on its own behalf: the gen- eral public of HIV-negative individuals who rarely practice risky behav- ior. Like most of those with riskier behavior, these people at low risk have an interest in marriage, in conceiving and raising healthy children, and in seeing them married, all without the risk of HIV infection. Although not themselves suffering from AIDS, in a generalized epidemic these people find that the price of health care has greatly increased because of higher demand and increased costs. Some of these people are the poor who have never suffered an HIV infection or death but nevertheless need 273 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC help in order to escape poverty. Some suffer dangerous chronic diseases other than HIV, such as cancer, kidney disease or diabetes, and cannot afford the treatment to keep themselves alive. To be truly democratic, a society must find ways-for example, opin- ion polls or elections-for the many with a small interest in an issue to express their views inexpensively and influence the course of events. Politicians facing a ballot box have an incentive to seek the opinions of ordinary people and consider these together with the views of smaller, more vocal interest groups. A government that is responsive to the na- tion's political leadership will follow suit. However, in the case of HIV/AIDS, the policies that will best protect the average citizen are not necessarily popular. Politicians and govern- ment officials, who may themselves be unsure of the best policies for confronting the epidemic, have the difficult task of explaining to the public why taxes should be spent subsidizing condoms and STD treat- ment for prostitutes and clean needles for injecting drug users. Conserv- ative social and religious groups, perhaps not fully appreciating the great harm that can arise from failing to prevent the spread of HIV, may op- pose efforts to reduce the risks involved in commercial sex or injecting drug use, or to encourage condom use generally, out of concern that these efforts will encourage behavior they regard as immoral. Business interests, having immediate profits in mind, may apply the kind of pres- sure to government that was dramatized in Henrik Ibsen's 1883 play An Enemy of the People: a physician who discovers that his Norwegian town's polluted public baths are a threat to tourists' health is pressured to keep silent by the democratically elected mayor and his supporters, and ulti- mately declared to be an "enemy of the people" himself Mexico and Thailand offer two dramatic examples of AIDS policy- making in the midst of all these conflicting pressures. The former coor- dinator of Mexico's National Committee for the Prevention and Control of AIDS (CONASIDA), Dr. Jaime Sepulveda, has summarized the re- sponses of government, NGOs and mass media during three periods from 1985 to 1992 (Sepulveda 1992). As shown in table 5.3, the gov- ernment response evolved from "erratic and medicalized" in 1985-86 to "reactive and participatory" in 1989-92. Strikingly, organizations of homosexual and bisexual men and liberal NGOs were initially silent and then actively opposed to the AIDS control program. Through continued efforts to engage these interest groups, government policymakers even- tually won them over; by the third period they were active participants 274 WORKING TOGETHER TO CONFRONT HIV/AIDS Table 5.3 Responses to the AIDS Epidemic in Mexico: Government, NGOs, and Mass Media AVongov'rmwenti' respoi&.e Gay and Pro- lida and Ahmss liberal otber right- media Goremment responice AXGOs wiing gronips response Erratic, medicalizcd, Slight 98S-S6, Silencc oppo)itiun Alarmiit Reacrime onlh Planned technocratic. Anger. Stiong to sensa- 10)8--88 prorei Oppo.:ri.on Lional neAs Reacl, . parucipatorV. Protcet Ljv. ,uit. participation marches F atlge in carrying out prevention programs. Meanwhile, Pro-Vida, a conserva- tive religious group, and other right-wing organizations became increas- ingly outspoken, if ultimately ineffective, in their opposition. Sepulveda includes the mass media among the actors in the shaping of Mexican AIDS policy, but he describes their role as only occasionally helpful. As late as 1992 he characterizes media coverage as continuing to focus on the number of AIDS cases, while neglecting other crucial in- formation about the disease: "In spite of the constant presence of infor- mation about AIDS in the mass media, specific aspects of the disease are not addressed so that collective accurate knowledge about AIDS is not generated nor is participatory discussion promoted." He points out that television and radio do a somewhat better job than print media, some- times using live programs with interviews, phone-in questions, and au- dience participation to generate discussion (Sepulveda 1992, p. 143). However, he condudes that, by the third period covered in the table, the media have passed from "alarm" to "fatigue" without ever providing the information that the public needs to understand the epidemic. An authoritative case study of Thailand highlights other political problems that can arise in designing and implementing an effective response. In the second half of the 1980s, as evidence that HIV was spreading rapidly among Thai sex workers and injecting drug users accumulated, a government official insisted that the situation was under control: "The general public need not be alarmed. Thai-to-Thai trans- 275 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC mission is not in evidence." In keeping with this sanguine view, the gov- ernment spent only $180,000 on HIV prevention in 1988 (the GPA committed $500,000 to Thailand that same year). The study suggests that during this period of democratic rule, in a pattern reminiscent of that described by Ibsen in Norway 100 years before, "high-level cabinet pressure was brought to bear on the ministry of public health not to pub- licize the emergence of increasing HIV in the population" (Pora- pakkham and others 1996, p. 8). Although Thai national funding increased to $2.6 million by 1990 (and donor funding reached $3.4 million), the government did not ini- tiate a high-profile, aggressive campaign to control HIV until 1991-92, when the country was led by Premier Anand Panyarachun, who had been appointed by the leaders of a military coup. The new prime minis- ter took several important steps that have since been credited with help- ing to slow and perhaps reverse the epidemic in Thailand. First, he shifted control of the AIDS control program out of the Ministry of Pub- lic Health to the Office of the Prime Minister, giving it added political clout. Second, he increased the budget almost 20-fold, to $44 million in 1993. Perhaps most important, he initiated the "100 percent condom program" focused on brothels, as described in chapter 3. Since then Thai funding to AIDS control has continued to increase, reaching more than $80 million in 1996, a sum equivalent to more than one-quarter of the entire international donor commitment to AIDS control in developing countries that year. The high-profile campaign was initially unpopular with the influen- tial tourism industry, and tourism indeed temporarily declined. How- ever, once AIDS had a prominent place on the national agenda, opposi- tion to the measures gradually faded-and support increased. "There were too many vested interests in maintaining the high status of the national AIDS program to make a policy reversal," the case study noted. "In particular, the enormous budget allocated to the HIV/AIDS preven- tion and control campaign was vigorously coveted by a wide-range of participants" (Porapakkham and others 1996, p. 17). Thus, the policy situation in Thailand had come full circle, from one in which special interest groups used their influence to oppose a vigorous prevention pol- icy, to one in which the participants in the prevention program assumed the role of vested interests in sustaining it. Since all programs that involve significant public expenditure develop their own constituencies, policymakers must be careful at the outset to initiate programs that are 276 WORKING TOGETHER TO CONFRONT HIV/AIDS in the interest of the general public, as appears to have been the case in Mexico and Thailand. Donor Assistance and Public Consensus Although the politics of AIDS will differ greatly across countries, bilateral donors and multilateral organizations can help to encourage public consensus on effective, low-cost responses to HIV through direct funding and through a judicious use of encouragement and conditional- ity. For countries that are still in the nascent stage, where citizens are not sufficiently aware of the epidemic to support funding activities from public revenue, donor funding can be critical in gathering surveillance data or establishing a demonstration project. Sometimes donors can require certain actions as a condition of the receipt of an aid package. However, the leverage afforded by conditionality is often quite limited and may depend on all donors agreeing to the desirability of a given con- dition. Conditionality is more likely to work if the government (or im- portant elements of it) intends to carry out the action in any case but has not yet made it high enough priority to get it accomplished. One example of the effective application of conditionality occurred during negotiations of the $70 million World Bank loan to India. In 1991 the government's initial posture was that there was no need for spe- cific interventions with sex workers and their clients in Indian cities. One influential government figure asserted that "in India AIDS is not sexually transmitted." As a result of a position taken jointly by GPA and the World Bank, the government of India agreed to double the size of its proposed AIDS program to include interventions with those most likely to contract and spread HIV, to be implemented by NGOs. Since then the extent of the sexually transmitted AIDS epidemic in India has be- come obvious to the highest levels of government, as evidenced in a 1997 speech by Prime Minister Deva Gowda. Attention has turned from whether interventions with those who practice the riskiest behavior are necessary to how best to implement them. These instances suggest that donors can significantly improve the timing and quality of country-level responses to HIV/AIDS. However, the evidence cited in chapter 3 and earlier in this chapter suggests that donors have often waited until AIDS has moved beyond the nascent stage before providing support. Although the data suggest that multilat- eral institutions are more likely than bilateral donors to direct resources 277 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC to countries at the concentrated stage of the epidemic, neither supports countries sufficiently at the nascent stage, when the largest benefits can be achieved with the smallest expenditure. We return to this issue in the policy recommendations in chapter 6. Individuals Who Make a Difference Although this chapter, and indeed most of the book, has focused al- most exclusively on national governments, donors, or groups, sometimes a courageous individual changes the way an entire nation or society thinks about HIV/AIDS, opening the way for a more effective and com- passionate response. These individuals may be national political leaders or other well-known figures, such as athletes or movie stars, who are not themselves infected. Or they may be individuals, famous or not, who are infected with HIV and summon the strength and courage to serve as ad- vocates for a sound national response. Examples of such individuals in the industrial countries are known worldwide. Actress Elizabeth Taylor has made fundraising for AIDS a nearly full-time occupation. Others, such as the late Princess Diana of Britain, have reduced prejudice and fear simply by being photographed embracing a child with AIDS. Among U.S. athletes, diver Greg Louga- nis, the late tennis star Arthur Ashe, and basketball's Magic Johnson have each helped to raise awareness of the disease by coming forward with the news of their infection. But while these figures are widely known and often admired around the world, the fact that they are from industrial countries means that their high-profile activities have only a limited ability to overcome denial in developing countries. People in a poor country learning that a movie star or athlete in a rich country has become infected may continue to think, "It can't happen here"-even though 90 percent of HIV infec- tions occur in developing countries. Because of this, every country and all societies need local individuals with the courage to advocate an effec- tive response to HIV/AIDS. Where such individuals have stepped for- ward, their efforts have often had a significant positive impact on public awareness and attitudes. Fortunately, as the understanding of the epidemic increases, a grow- ing number of individuals in developing countries are demonstrating such leadership. To mention just three examples: the speech by Indian Prime Minister Deve Gowda naming HIV/AIDS as a national health problem helped to overcome the idea that India was somehow not 278 WORKING TOGETHER TO CONFRONT HIV/AIDS threatened by the virus. Zambian President Kenneth Kaunda, in ac- knowledging publicly that his son had died of AIDS, helped to energize his country's response to the ravages of a widespread epidemic. Finally, Marina Mahathir, daughter of Malaysian Prime Minister Mahathir Mohamad and the president of the Malaysian AIDS Council, a non- governmental organization, has spoken out in her own country and internationally for greater political commitment to mobilizing the re- sources necessary for effective prevention. Some of the most compelling advocates of an effective response to the epidemic are people who are themselves infected with HIV Phily Lutaaya, an enormously popular Ugandan singer and songwriter, be- came the first prominent African to acknowledge that he was infected with HIV. He spent his remaining healthy time writing songs about his battle with AIDS and touring churches and schools throughout Uganda to spread a message of prevention and hope. After Lutaaya's death at age 38, the Philly Lutaaya Initiative continued his work. With assistance from UNICEF, the Initiative sponsors lectures in schools and communi- ties across Uganda highlighting personal testimonials of hundreds of people infected with HIV A 90-minute television documentary on Lu- taaya's struggle with AIDS released in 1990 reached millions of television viewers around the world (Graham 1990, Kogan 1990, McBrier 1995). But a person need not be a celebrity prior to infection for personal tes- timony to have a powerful impact. Perhaps the most courageous individ- uals are otherwise ordinary people who, after becoming infected, step forward to acknowledge their disease and, in the face of discrimination and persecution and with very limited personal financial resources to draw upon, speak out for a more effective public response. All these indi- viduals serve as a powerful example to those who meet them, a few be- come nationally known. Box 5.5 describes how one such individual, a factory watchman, raised awareness about HIV/AIDS in Thailand. a * U This chapter has analyzed the roles of governments, donors, and NGOs in financing and implementing effective policy responses to HIV. It has argued that each of these types of organizations has particular strengths and that for an effective global response to HIV/AIDS, all of these groups, plus countless exemplary individuals, must work toward a com- mon goal of overcoming the epidemic. As the chapter relates, much has already been done; yet the analysis also identified some key shortcomings. Governments have the unique 279 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 5.5 Someone with AIDS Who Made a Difference WHEN CHA-ON SLrLLI.NI C ONTRACTED HI% FROM A turning point for sociern as a n%hole in dealing with blood transfusion he was fired from his job as a fac- che epidemic. \X'hile Cha-on w.a, still well, his activ- torn watchman; his wife, Aho %%orked in the same ities focused attention on discrimination against facoro,. was alsc fired. In 198-. Cha-on decided to those infected with HlI\'. Howe%er. as the entire make his case public and accepted a job as an AIDS narion witnessed his rapid progression to clinical educator with the Population and Community AIDS and finally his death. another critical issue Development Association of Thailand, a non- Was brought to the fore. Thais began to understand gov-ernmental organizacion. that AIDS was real and chat the; themselves could Cha-on soon appeared on national television becomne infected and die. Cha-on's lasting legacy talk shows and on the front pages of Thailand's wvas stronger support for and receptivity to HIV biggest newspapers. The result was an outpouring prevention efforts throughour Thailand IPorapak- of public sYmparhY for his own hardship, and a kham and others 19961. responsibility for coordinating their country's overall response to the epi- demic. As part of that responsibility, many governments, especially in developing countries, should take on greater responsibility for basic epi- demiological surveillance and prevention activities. NGOs have often played an important role in prodding governments into action; govern- ments that select appropriate NGO partners can often greatly increase their reach, especially in working with marginalized groups to help peo- ple who practice the riskiest behavior to protect themselves and others. Donors and the multilateral institutions they support have provided significant financing and other assistance for all of these efforts. But donors need to do a better job of focusing attention and resources on countries where the epidemic has yet to attract policymakers' attention, especially countries with nascent epidemics, where prevention is most cost-effective. Moreover, international donors have the unique ability to mobilize financing and other support for international public goods, such as evaluation of alternative approaches to preventing HIV and mit- igating the impact of AIDS, as well as research on a vaccine that would work in developing countries. Such efforts are in the donors' own best interest, as well as the interest of developing countries, and deserve much greater attention and support. Finally, donors have the responsibility to coordinate their activities at the country level, both among themselves and with the national government. 280 WORKING TOGETHER TO CONFRONT HIV/AIDS Although there are no easy solutions to the technical and political problems posed by the HIV/AIDS epidemic, examples from countries around the world offer hope that people of good will, working together, can overcome this global epidemic. The next and final chapter in the book summarizes its main policy recommendations and looks toward the future. Notes 1. This calculation uses the estimate of $4.8 billion less-poor countries, so that the equivalent grant would be for total health assistance in 1990 (World Bank 1993c, significantly smaller. For the purposes of this analysis, the p. 166). difference between loans and grants has not been taken into account. 2. Although national AIDS program spending under- represents total national spending on the AIDS program, 6. Tax law typically forbids nonprofit firms from sell- it probably captured most of it in 1993. ing shares of the firm to raise capital, as for-profit firms are free to do. 3. The effect of donor spending on national spending was estimated under the maintained hypothesis that 7. See, for example, "NGOs Flout AIDS Control Pol- national funding does not affect donor funding by includ- icy" (1994). We set aside the fact that the government ing donor funding as a fourth regressor in the equation may imperfectly represent the interests of the public. to predict national funding. After controlling for (the logarithms of) GDP per capita, population and the num- 8. In late 1997, USAID was preparing programs to ber of HIV-infected people, the coefficient of (the loga- succeed AIDSCAP. rithm of) donor spending is 0.01 with a t-statistic of 0.08. The instruments necessary to identify a model of simulta- 9. Given a constant total donor AIDS budget in a neous causation between national and international recipient country, the country would benefit if its AIDS funding are not available; therefore, such a pattern cannot programs were developed as a coherent whole and all be ruled out. donors agreed to pay a share of the total. However, expe- rience shows that the amount of any given donor's expen- 4.l Thes twational ionshipsdo to be the re- diture in a country is not usually fixed. In those cases splt of a national dersion to reduce AIDS funding insr where the donor's budget for the country is fixed in the sponse to the perception that donors are providing such short run, it is still likely to be fungible across sectors. Thus, the amount of AIDS financing from a given donor 5. According to the available data, multilateral fund- depends upon how much its representatives want to fund ing totaled $605.7 million during this period, 22 percent the AIDS projects that the government allows it to fund. more than bilateral funding. The World Bank made two It is often alleged that donors prefer to 'put their flag" on large loans, one to India for $70 million and one to Brazil a project, so that they can claim credit for it in the inter- (the total size of the Brazil project was $250 million, of national community and to their domestic constituency. which $160 million was borrowed from the World Bank). These incentives lead to a situation in which no donor The loan to India was at a concessional interest rate ac- wants to fund the overhead costs of an AIDS programs or corded to the lowest-income countries and thus approxi- a portion of any part of the program. Any attempt to mately equivalent to a $50 million grant (Arias and Serven coordinate donors, whether bilateral or multilateral, must 1997). The loan to Brazil was at higher rates accorded to struggle against these perverse incentives. 28I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC 10. This projection is based on the assumption that 12. A "decrease in the willingness of pharmaceutical the number of incident cases will stabilize in every region companies to become involved in vaccine research, devel- of the world once incidence falls to half of its peak value. opment and manufacturing" was observed as early as "The choice of an equilibrium value for incidence that is 1985 (Institute of Medicine 1985, p. viii). 50 percent of peak incidence is entirely arbitrary and does not take into account advances that may be achieved in 13. The behavioral sciences can also contribute infor- behavior modification or technological breakthroughs mation that will increase the profitability of a medical such as a vaccine or more effective chemotherapy. Conse- intervention. For example, the European Commission is quently considerable caution is required in interpreting sponsoring a "market perspectives study" on vaginal these HIV projections, particularly for the years beyond microbicides in Brazil, C6te d'Ivoire, Egypt, India, 2005" (Murray and Lopez 1996, p. 347). Kenya, the Philippines, Poland, and South Africa. A finding that women are willing to pay for this drug would 11. See the examples in chapter 3 and the summary of improve the incentives for private pharmaceutical firms rigorous evaluations of preventive interventions in devel- to invest in their development (AIDS Analysis Africa oping countries in appendix A of this report. 1996). 282 CHAPTER 6 Lessons from the Past, Opportunities for the Future VER THE PAST TWO DECADES THE HUMAN immunodeficiency virus has spread silently throughout the world, profoundly affecting the lives of men and women, their families, and soci- eties. It has not respected international boundaries ---~ =- or spared the elite. By the time that researchers un- derstood how HIV spreads, how it can be prevented, and the behaviors that put people at risk, HIV had already infected millions of adults in the industrial and developing world. In the hardest-hit countries in Sub- Saharan Africa, poverty, illiteracy, poor health, low status of women, and political instability fueled its spread. By the time East African health au- thorities identified the mysterious "slim" disease as AIDS in the early 1980s, HIV had already widely infected those with the riskiest behavior and had a firm foothold in the general population. On the medical frontier there have been many advances, but there is still no vaccine for HIV and no cure for AIDS. Medical researchers have succeeded in substantially prolonging the lives of some people living with HIV and AIDS in industrial countries. However, these treatments are still very expensive, they are not always successful, and no one knows for how long they prolong life. The costs of these new therapies are so high and the requirements for their implementation are so demanding that they are simply not feasible in low-income countries and would bankrupt the health systems of middle-income countries. 283 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Lessons from Two Decades of Experience STILL, WE HAVE LEARNED MUCH IN THE PAST TWO DECADES that is cause for optimism as we confront the epidemic. We now know that HIV is not spread easily and that it can be prevented through behavioral change. Other STDs signal risky behavior, and pre- venting or treating these STDs can slow HIV transmission rates. Low- cost, cost-effective interventions to prevent HIV/AIDS in poor coun- tries are now known to exist. Behavior change has reduced incidence of HIV among specific groups in countries as diverse as Australia, Thailand, and Uganda. And there are many opportunities to alleviate suffering and prolong the lives of HIV-infected people in developing countries, for example, through low-cost treatments of common oppor- tunistic infections, particularly tuberculosis. We can also learn from the policy mistakes of the past. No country, rich or poor, is insulated from the risk of HIV. Governments should intervene as soon as possible; if policymakers wait until AIDS is killing many peo- ple, HIV already will have spread widely, interventions will be less cost- effective, reducing infection will become more difficult and, absent a cure, the epidemic and its terrible impact are likely to persist for decades. Be- havioral change must focus first on people with high-risk behavior who are most likely to become infected and unknowingly infect others. But discrimination against such individuals makes behavioral change more difficult and inhibits efforts to cope with the impact of AIDS. The Role of Government E XPERIENCE HAS ALSO SHOWN THAT ACTIVE GOVERNMENT involvement is crucial if AIDS is to be overcome. Only govern- ments have the means and mandate to finance the public goods necessary for the monitoring and control of the disease-epidemiologi- cal surveillance, basic research on sexual behavior, information collec- tion for identif,ving high-risk groups, and evaluation of the costs and effectiveness of interventions. Private individuals left to their own devices would not invest adequately in these activities. Governments also have a unique responsibility to intervene to reduce the negative externalities of high-risk behavior, while preventing discrimination that 284 LESSONS FROM THE PAST, OPPORTUNITIES FOR THE FUTURE would inhibit behavioral change. Without these government efforts, those at high risk of contracting and spreading HIV are unlikely to reduce risky behaviors enough from the perspective of the rest of soci- ety. The government role extends to ensuring equity in access to HIV prevention and treatment for the most destitute. Other key functions that most governments are already attempting to perform can also make an important contribution to slowing the spread of HIV: promoting labor-intensive economic growth to reduce poverty; assuring basic social services, law and order, human and property rights; and protecting the poor. Investing in female schooling and ensuring equal rights for women in employment, inheritance, divorce, and child custody proceedings are part of this broader mandate. These policies yield large development benefits in their own right but are also important for pre- venting an HIV epidemic and coping with its impact. Reform of health systems, as outlined in the World Development Report 1993 (World Bank 1993c), will improve the efficiency of health care delivery, including HIV and STD prevention, and will reduce the impact of AIDS on the health system. In areas where there is a severe epidemic and targeted poverty re- duction programs already exist, these can often be combined with efforts to ease the impact on surviving household members, especially children, in the most destitute families that suffer a prime-age adult death. We know that certain policies can work, yet developing countries face many financial, political, and managerial obstacles to implementing them. Financial resources are scarce. In low-income countries, annual health spending from public and private sources averages only $16 per person.1 This is one-tenth the resources available in middle-income countries and only 0.7 percent of the $2,300 per capita annual health spending in high-income countries. Many developing country govern- ments also have limited capability to implement complex or multi- faceted programs. The World Development Report 1997 (World Bank 1997a) makes a compelling case that the government role must be matched with its capability. In fighting the spread of HIV and mitigat- ing the impact of AIDS, developing country governments will be most effective if they focus their financial and other resources on a limited set of feasible activities that have the potential to be highly cost-effective. Pressures from the public and from international donors can lead gov- ernments to try to do too much with too few resources, reducing the effectiveness of programs. Governments can sometimes expand their effectiveness by involving the private sector, reputable NGOs, those 285 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC most severely affected, and decentralized community organizations in the design and implementation of high-priority HIV/AIDS prevention activities. However, coordination and management of these activities can also stretch the capabilities of government. Opportunities To Change the Course of the Epidemic O NCE LAUNCHED, AN HIV/AIDS EPIDEMIC CAN TAKE decades to unfold. Epidemiological models predict that between 1996 and 2001, between 10 and 30 million new infections will occur in developing countries. But the future of the epi- demic is not carved in stone. One reason that such projections are very uncertain is that nobody knows the extent to which individuals, espe- cially those most likely to contract and spread HIV, will change their behavior in response to the virus. Concerted, focused action in devel- oping countries, where more than 90 percent of HIV infections occur, can save millions of lives (box 6.1). Preventing the Expansion of Nascent Epidemics Public action can make the greatest difference for the 2.4 billion peo- ple who live in areas where the epidemic is nascent. Developing areas with nascent epidemics include half of the world population, two-thirds of the population of developing countries, and nearly 40 percent of the population of low-income countries (table 6.1). Half of India, all of China except Yunnan Province, Indonesia, the Philippines, most of Eastern Europe and the former Soviet Union, North Africa, and a third of the countries in Latin America and the Caribbean are at this stage. In these areas, HIV has not yet spread widely even among those whose be- havior puts them at risk. But countries with nascent epidemics cannot assume that they will never be affected; every country that now has a generalized AIDS epidemic went through a phase of denial that gave the virus time to gain a foothold. These nascent-stage areas present an enormous opportunity for gov- ernments and donors to prevent an HIV epidemic by intervening ac- tively and early. Epidemiological surveillance of those who practice the 286 LESSONS FROM THE PAST, OPPORTUNITIES FOR THE FUTURE Box 6.1 Estimating the Power of Prevention in Three Countries THE PREVIOUSI CH kP ilERi HA F HIGHL.1IG,HTED the population). The simulation, iho%% the impaci ot simulations of the epidemic in h)pothetical popula- raisin, condom use in rheic uxto groups suth [he ions. \What might happen in a real country'? most partners from 2u tr) SO percent and tfom t o Mliodeling the pocential bentfirs ot' incer'ennions 15 percent. respectix-el. bernsten 19'- and 2000. in for a specific countr requires detailed information Brazil. %here needle sharing ha' played an important about behitioral and biological chiraclerisrics or rhe role in spreading HNB'. the share of inlectioris with population-the types and distribution of risk clIen necdles amonp inicLring drug users is assumed behavior. the number of' peoplc in%olved, sexual to rise From 210 to 80 pcrcent. Finall). t;r comparc- mixing patterns. and [he pre%alence ot' other STDs 'on. w%e show thc effect of raising condom use among in specitic population groups. SuCh intormation is vomen in stable relatic.nships. fronm I to 3 percent in rarely av ailable ind is urgently needed. Efforts are C6tr d'loire and from > to 10 percent in Indonesia underway to calibrate rhe STDSI NI model tor bermeen 1()5-o and 21(100. The simulations showv the Nairobi. Kenva. for example. and the isgAIDS and results of these intcrxentiors through 20 10O.2 SimuL_UDS models hate been calibrared to predict the impact of interventions in Kampala. Liganda U In Indonesia. HIV previlence i- still 'er i Bernstein and others 1 99-T. low-less [han 10.0I percent of the population Nevertheless. w%e can get some senme of the likely is infected. Among se\ %%orkTers. homoseLuak. impact of interventions that change high-risk beha%- and transtestites. homcxer. HN\ prevalence is ior bi applying lmired country-specific parameters as high as 3 percent. A rapid expannsion of to existing models. Simulation results t'or three condom use among the rmxo groups vsith the countrics at differenc stages of the epidemic- highest rates of partrcr change can preenr Indonesia inascenti. Brazil lconcentrated;. and Ctec the level of HNI incction in the Fenerd pop- d1voire (generalizedl-have been deriBed from a ularion from reaching abo%e 0.2 percent. model de%eloped at thc World Bank.' Like the Increacd condoLm use amrong w-omen in sta- STDSINI model used earlier in this repor[. the ble relarionships has, er lirtie impact. W\'orld Bank model simulates the spread of HI\ U In Brazil. 'sith a concentrated epidemic. an through heterosexual contact-i and from mothers to increase in condom usc amongz the i%s their children. and it takes into account tactors. like groups s ith thc highest ratcs of partner the presence of ST[)s 3nd condom use. thar affect changc i sufficient to bring HlI' preialence the probabilir- of HI\ transmission. In addition, it dovn to about 2 percent in 2011) Arn increase models transmission through blood transfusion. in the use of clean ncedlcs accelerates this needle sharing, and homoexual corntat. trend, bur by itself i4 insutficicnt to reduce Con nrrv-specifie pari merer tfor these simulationi prevalernec ushianriallk. "Uver based on information from sur%evs and otiher U In C'ore d Isoire. whcre HNIX presalcnce in the 'udic, in each counmrr. as v,tell as on informed esri- general population has already reachcd 13 per- mares. \W'e sho%% belo%s the simulated impact of rais- cent. the epidemic %tould, in rhe absence of ing condom use among two segments of the popula- any beha. ior change. continue to increase. tion w% ith high rates of partner change: w%omen who reaching lot percent of the population by- have 500 new% partners per year I percent or less orit' 2010. Interxentions to dramatically increase all womeni, and women tand men, in countries condom use among those %tith the highest where homosexual transmission is modeledi %ho rates ot' partner change mould lower preva- hate one nev partner per month i" to Ii) pcrcent of ,B' -:m "1, * r' . t. L: 1, Iprg .....'.".. . g 287 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Box 6.1 (continued) lence to 9 percent bh 2010. In contrast, in- or as the result of intenrentions. More derailed infor- creased condom use among, s,omen in stable mation on se\ual beha-.ior in these countries is nec- relationships would hase a negligible impact. essan- to generate more accurate models. These results arc suggcstive of the impact that - . P~~~~~~~~~~~~~1 1I-hc mT.-.dcl cdld PFli)iLmng, AIC)' or PRAY. ! dr;.,nbed programs can hase if they succeed in changing clic alitaoI NI 9I I behavior of the population with thc highest ratcs of '. Condom u, C.L.,iounuc4 Er clrmb ar the same race unril 2020 partner change. How\ever. thexL undcrstace the impact Lhc end or che iinr-ljrion ptrrod It reiche, %8 pcrcen[ and t) to) dhc extenti that ocher segments of the population -re up h rhc eh.t and nexr hxgheo ries AO r pariner change XVrnong c&h rtr,mcn w-th the fewsril pinru- ; t may alo change. their behavior. ctiher spontaneously reachc' onl Iri-24 pcrc-cn Box Frivre 6.1 Projected Impact of Behavioral Interventions in Three Counties Adult HIV prevalence (percent) Indonesia 1.0 Baseline 0.8 I _ _ _ _ _ _ _ - Increased condon use among women In 0.6 - -__ ~ - - - - -stable relationships to 10% by 2000 0.4 _ _ _ _ _ _ - _ Increased condom use among sex workers to 0.2 -_ _ _ _ _ _ _ _ 80% and others with man) partner to 20% 0.0 by 2000 1985 1990 1995 2000 2005 2010 Adult HIV prevalence (percent) Brazil Baseline 5 _ _ - Increased ure of sterIlIzed Injecting 4 _ _ _ _ _ _ _ equipment to 80% by 2000 Increased condom use 2 _ _ _ _ _ _ _ _ _ _ _ * ' " ,. among sex workers to 80% and others with 1 __ _ _ _ _ _ _ _many partners to 20%. by 2000 1980 1985 1990 1995 2000 2005 2010 .... Increased condom use and sterIlized Adult HIV prevalence (percent) Cote d'lvolre Injetng equipment 16 14 __ _ - Baseline 12 ._ Increased condom use 10 _ _ _ __ - _ _ _ _ _ _ among women In stable relatlonships to 3% by 2000 Increased condom use 4 _ among sex workers to 2 * - ~ ~-~ - - -80% and others with manvy partners to 20% 0 by 2000 1975 1980 1985 1990 1995 2000 2005 2010 288 LESSONS FROM THE PAST, OPPORTUNITIES FOR THE FUTURE Table 6.1 Distfibution of Developing Counbty Population by Stage of the Epidemic and Income Louer-iniddke C (per-mniddil lou' income' itnrome incrome Total Stage of tle Popuilarion Population Pop.Ld4tiotl Population epidernic 0iiillions) (0 {millions) °o Irmillioi..) ' oo (millionss) %0 Nascnt3 I 739 3 ;(3 11 21 2.265 i9 Concentrated I .,Is 2 2 320 3 311 - 1.640 3 5 Generalized IX1I - 3 O 42 1 226 5 Unknown IiI 3 3.. -12 I 500 11 Tlrdtl popu!auonb 3.3)S i 6 1.133 2 _422c 9 -4,63( 100 Nwumkber o countric; (.) 46 1 V 123 i Tht popularion otf Ch;ni ind India, borh louw incom ccounurir. hv'c bccn dj,urJtu1cd I:rwceftn nantra and concenraraod srsii ot rh, .pi&dmi, ba:.e. rhc .i h e in spsc!fi prc rnc a and cl 1aw.. re'p;cw.el\ b Anx dis repci, .-n rioiai .ah.a r f ro,undinr nurnberc ,,zo.e.e Inwine rnolip. i ifrom rhe tt ;rld/ De.A.r,p. *. Rpv.r / 99 IW.rld B3nk 1 );'i. ,rag ot The epidemic and l10i populadctn ir.! trom rable [ .:.f th nitri.:alappendi.. [,-. thie cpor riskiest behavior and studies of risky behaviors in the general population and specific subgroups have high payoffs at this stage. An HIV/AIDS epidemic can be pre-empted, for little cost, by promotion of safe inject- ing behavior among injecting drug users and of safe sex and STD pre- vention through condom use among those with high levels of sexual ac- tivity. We know that this can be done. In chapter 3 we highlighted the example of five cities in which early intervention kept infection levels among injecting drug users below 5 percent, even as HIV prevalence soared among injecting drug users in nearby cities. Experience has shown that early interventions focused on groups at high risk of sexual transmission can be equally effective. Prevention efforts focused on those who practice the riskiest behavior may be politically controversial, especially if such efforts are perceived by some constituencies as facilitating antisocial or immoral behavior. Policymakers who encounter such opposition have an obligation to make clear that preventing infections among those with risky behavior is the best way to protect everyone. Containing Concentrated Epidemics Developing countries with concentrated epidemics-where HIV prevalence exceeds 5 percent in one or more groups with high-risk be- 289 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC havior but not in the general population-are a diverse group of low- and middle-income countries, with a variety of risk factors. In Latin America, Ukraine, Yunnan province of China, much of Indochina, and the northeast of India, the epidemic has reached concentrated levels among injecting drug users; in many countries in Latin America, the epidemic has also reached concentrated levels among homosexual and bisexual men. In addition, HIV has infected more than 5 percent of high-risk heterosexuals, among them sex workers, in southern India, Indochina, and much of Africa. Once HIV has reached high levels among those who are most likely to contract and spread the virus, containing the epidemic is difficult and requires drastic action-but is nonetheless possible. Thailand undertook such a massive effort when injecting drug users and prostitutes were dis- covered to have high infection rates. A policy of heavily subsidized con- dom promotion and STD treatment programs for prostitutes and others with high-risk behavior, supplemented by widespread dissemination of information to the general population, brought down the prevalence of HIV among military conscripts within a few years. Not all countries have the same institutional setup or implementing capacity as Thailand. Each country will have to find its own way. But whatever tactics are adopted, the underlying strategy of massive interventions to change the behavior of those most likely to contract and spread HIV is crucial. Adapting this strategy successfully will require better information about the cost-effectiveness of alternative interventions to prevent the spread of HIV. Research documenting the effectiveness of such inter- ventions in preventing secondary infections can be very valuable in gen- erating and sustaining support for these measures. Governments also have a role in ensuring that basic information about HIV is presented to the general public in ways that will minimize irrational fear and persecu- tion of individuals who are infected with HIV or thought to engage in high-risk behavior, since such responses make it harder to reach those with risky behavior and encourage safer behavior. As people infected early in the epidemic begin to get sick and die from AIDS, governments will face growing pressure to spend public resources on care and treatment. Responding to these needs compassionately, while keeping them in perspective with the many other pressing human needs and demands upon public resources, is one of the most difficult chal- lenges posed by the epidemic. Pressure for spending for AIDS care and treatment will be stronger in a generalized epidemic, when the disease has spread into the general population and people infected with HIV are a 290 LESSONS FROM THE PAST, OPPORTUNITIES FOR THE FUTURE large and highly motivated constituency. By then, subsidies begun during the concentrated stage may be unsustainable and yet very difficult politi- cally to withdraw. The concentrated stage of the epidemic is therefore the time when policymakers and their constituents need to consider how government can best respond to the medical needs of people with HIV. The fair response in terms of health care, advocated in chapter 4, is to offer the same level of subsidy for the care and treatment of people with AIDS as for the care and treatment of people with other diseases that are expensive and difficult to treat. Denying care to individuals simply be- cause they have HIV/AIDS is unjust to those who are infected and to their families. By the same token, providing a higher level of subsidy for AIDS care than for other illnesses is also unfair to the majority of people who are not infected with HIV. Choices about the appropriate overall level of public subsidies for health care will vary across societies. Govern- ments and their constituents should be aware, however, that high sub- sidy levels will be extremely difficult to sustain in the face of a large epidemic. Since it is unfair and impractical to deny care and treatment subsidies to people with HIV while providing them to people with other illnesses, any changes in subsidy levels should apply equally to the HIV infected and the uninfected. Policymakers need also be aware that the care and treatment of AIDS, in sharp contrast to preventive interventions focused on those most likely to spread HIV, is primarily a private rather than a public good: most of the benefits of the care and treatment of AIDS accrue to the per- son who receives the care. There are important exceptions to this general rule. Treating tuberculosis, STDs, and other infectious diseases in people with HIV can prevent these infections from spreading to others, indud- ing people who are HIV-negative; these "externalities" are a sound ratio- nale for public funding of such treatments, regardless of whether the recipient of the treatment has HIV. Similarly, outreach programs that include care for those infected with HIV who practice high-risk behav- ior may be a justifiable use of public funds if the program results in behavior changes that reduce the spread of the virus. Often, however, de- mands for publicly funded care and treatment threaten to drain scarce resources that could have been used for preventing new infections. Maintaining Focused Prevention in Generalized Epidemics Countries with a generalized epidemic will face two related sets of challenges: establishing or maintaining prevention programs focused on 291 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC those most likely to contract and spread HIV, while expanding preven- tion efforts to those with somewhat lower risk of transmitting the virus; and mitigating the impact of AIDS sickness and death, especially among the poor. Except for Botswana and South Africa, all countries that currently have generalized epidemics are low-income, with 1995 per capita in- come of $765 or less. Scarce financial and managerial resources mean that these governments must be especially vigilant in implementing the most cost-effective prevention programs. Although prevention measures for the general population become increasingly cost-effective as preva- lence rises, interventions for those practicing the riskiest behavior con- tinue to have the greatest impact on incidence per dollar spent and must be maintained even as prevention programs are expanded to others. Condom social marketing programs and other forms of prevention sub- sidies aimed at poor people who would otherwise be unable to afford to protect themselves are an appropriate government response at this stage, where resources are available. But these programs are no substitute for reaching the highest-risk groups. Indeed, one of the greatest threats to effective prevention in generalized epidemics is pressure to divert re- sources from highly targeted cost-effective interventions to politically popular interventions with lower cost-effectiveness. Even where prevention measures are very effective, declines in preva- lence will occur only gradually, as people already infected die and are suc- ceeded by younger cohorts. But declines in incidence-the number of new infections-can be achieved relatively quickly, even in the face of a generalized epidemic. Recent declines in HIV incidence among young people in Uganda are an encouraging sign that even the worst-hit coun- tries can make progress against the epidemic. The second challenge to governments in a generalized epidemic is mitigating its impact, especially on the poor. A widespread epidemic will greatly increase the number of households that suffer a prime-age adult death. In poor households, such deaths can have a severe and lasting impact on surviving children, who may suffer further declines in already inadequate nutrition and schooling. But not all households that suffer a prime-age adult death are poor. Indeed, in many of the countries hardest hit by AIDS, while most of those infected may be poor, it is still the case that nonpoor people are more likely to be infected than the poor. Confronted with demands to finance programs to help households affected by AIDS, policymakers need to balance the needs of poor 292 LESSONS FROM THE PAST, OPPORTUNITIES FOR THE FUTURE households hit by AIDS with the needs of other poor households that are more numerous and often poorer. In approaching this task, they should ask two questions: Which households need help most? How best can they be helped? If many households are very poor and children are malnourished and not in school, government's priorities must include such basic development policies as fostering labor-intensive economic growth, improving nutrition levels, and increasing school enrollments, especially of girls. Where targeted poverty reduction programs are al- ready in place, modifying these programs to make assistance available to very poor families that suffer a prime-age adult death can help to im- prove the targeting of assistance to the households that need help most. Challenges for the International Community I NTERNATIONAL DONORS HAVE BEEN GENEROUS IN THEIR support for AIDS prevention in developing countries, but their support has not always gone to those interventions that are most cost-effective from the perspective of government. To have the largest impact now on the pandemic, donors need to consider two main strategies. First, in terms of bilateral and multilateral assistance, donors should support major interventions in countries at the nascent stage of the epi- demic, including epidemiological surveillance, surveys of risky behavior, and programs to change behavior among those who practice the riskiest behavior. Among countries at the concentrated and generalized stages, particularly the low-income countries, ensuring prevention of infection among those who practice the riskiest behavior would be the most cost- effective strategy. Moreover, donor funds could help promote such pro- grams when they might be politically unworkable if openly sponsored by government. With respect to mitigating the tragic impact of AIDS on society, donors must not lose sight of the myriad development problems faced by the low-income countries with generalized epidemics. The AIDS epidemic will increase poverty and will undermine household in- vestments in human capital. Countries with generalized epidemics are therefore likely to need renewed support for core public programs to raise levels of human capital and reduce poverty. In addition, there may be some scope in specific hard-hit areas for assistance in integrating tar- 293 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC geted poverty reduction efforts and AIDS mitigation. However, govern- ments and donors need to be careful that such assistance does not dis- place household and community efforts to cope or, worse, drain time, energy and money from prevention measures focused on people who practice high-risk behavior. The second important strategy for international donors is to finance key international public goods that poor countries cannot afford to sup- port collectively. Two important public goods stand out: knowledge about the costs and impact of interventions on the incidence of HIV in differing environments; and development of vaccines and low-cost pre- ventive medical technologies that will be effective under conditions pre- vailing in developing countries. * * E The poet and philosopher George Santayana said, "Those who can- not remember the past are condemned to repeat it." This maxim is nowhere truer than with the AIDS epidemic. Country after country responded to evidence of the first infections by saying "We are different. AIDS cannot strike us." Each has been proven wrong. When countries discovered that they indeed did have a fatal, sexually transmitted disease spreading rapidly in their midst, one after another responded by clean- ing up the blood supply or conducting general awareness campaigns, while avoiding or devoting insufficient resources to efforts to encourage safer behavior among people most likely to contract and spread the virus. But recent history also offers valuable examples of success. Experience demonstrates that enabling people who practice the riskiest behavior to protect themselves and others can be extraordinarily effective. National policymakers now face the challenge of applying this strategy in the cul- tural and political context of their own countries. Note 1. When China and India are excluded, average health spending in low- income countries is even lower-$11 per person per year (1994 World Bank data). 294 Appendix A Selected Evaluations of Interventions To Prevent Transmission of HIV In Developing Countries 295 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Appendix A. Selected Evaluations of Interventions To Prevent Transmission of NW in Developing Countries Aw4o s u l oSdu d (learl (%nnrrsn Irtrerzention design' Sanspi size Inrertencion Condom promnorions'ateuSe Bha%e and oiher' India Condom- distributioin and HR resr- QE 9-41 I%ex vorker, and 3- madams i 1995 i ing and counselIing Fotrd anid tither, Indones-ia HI\ education, pirr.p training, and QE 31A) ,c-x %sorkers and 300 clientsi ii I 199)C61 condomr 'ales arnd disrnbution ~ircs: 2 case. I controlh Fo\ and niihers, Ho)ndtura C ondojm distribucion and HIV PC 134i female sex ssorkers i19931 i ucarco'n Nueui and others Ken%. a Tjrcerecd condomn promonrio QE 366 Female sex "torkers P.aUn and otlici Nicair-cui .Iturc I AID'S edu-11i- QE Rxcsidecnr I,es;c I 2-f I 6I2I66 a I199I6, baseline and 2.2'l an tollon%-up Inrertrnnion ~T[D zrcarrren. rrcstmcn[ onla Coheni anid onhc r NI alav. i Antibiotric irtatnicni.tIcr me-n to ich E I3' HRN *posirl'c. mtn. 86 w%ith I 1Ci9- i urethrinis ureibrirk and -s9' contro)ls x% ithour urrd-hritLs SJrc.,5ku1rih and Tanzartia cT[D irtarmeni I 3100i( adult~ ini each o-f rx%o ran- otfieri I IA5ai~1 doml% selected communiries \\awer and crhcr' L'ganda Sl, TUH treatmentr E 0%er -0 TO adult:s ages I5-99 per Ii Qn16b I drm: Ni illagc~ in I I) clusier~ ran- do,mis ao~igned to intervention or :ontrcl arnm I nrerenriuon Coni:,bincd S) ID ire aiment and condom prLnMOnconl Eiriene-Traur,: and Coic Tv.o 'TD crrtarvner metrhod,. stEan- E (irnup I: 2I sex workers: GrouIp 2:. other, I ~Io d ut dadad4rni e l rci"dhar 23 ; x %%orkers iselected from a CS educirion Free condorm,, anid ftre ol exctrkers aind randumized no. nreatmntcn. groupi J4Ckson and others Kcn.c, sTED tcrecnint, 4nd trearnmenr and PC 56 HR[ -negative miale emploYees i I 9I condomir promotion :1 truckinig conmpans~ Laga and orhers Congo, DR STD 'crrening and rrcarmCnr; con- PIC 43I niriaIlk' HIV-neganire t'emale I I "~~~ii i ~i t Ir n erl Jinii prvmn,ii*nse okr Zacrc- I Le'. ine and orbere Boli% ia STD rre3crnenr mnd condom PC I5SO iniria.1l% HIV-negarie female 'I 11N0 i) promnOtion sex cc orker, 296 APPENDIX A L.ngthl of [.Ofcl71;o obiemw,,lOi Resui"t/ LUrbri 2n mLonths Amc.ng rhe inren rn[ion fgroup. an increase in akavs- u;ing condom; Fronn 3 cii 28'.. lij...OO. Jt!Ii 1; c,mpared w-rh nochinge 'n thecon[rol group: inerrcas in iomeime' using condoms from 31 to Ni'¾, Iu.0I i and from 36 to 5° it'- iJ) (.-l . rtspec.ivcek in mhc rv grol Pb. Lirban 6 month; Condom usc %% dh chent' increased from 1. to [i";r, and 29 to ;)62"i., i.p(1.0) I i the no vintr- 'enrion sires and tronmi - [o 6(1l9.. ip/.C05 I in dhe contrrol irc Lirban 6 months Increast in mran condom nl ul frm ( tn [C' -r,in ip i).5 u i: Condom use reporrcd In diaric, dLIr- ing p[n tr im a e%en higher 0t1911, I. I.rban I 2 moith, Tho, % ho receied enel idual and group counseling gr.Lup Ii increased occasional cond..rr uLc tronm 1tli to oi",. rhose v.ih (roup counreline group 2i increased from '1 Lo -0'.. and eontrol group g-roup .I from in to : niean condom uLe vsa. 3g",, 3,"'u. and '-.-.. rspee- tO Il I I. 2. p -t i.i102. 2 N 3. pti 1. xi ondom u,I resuled In ihree -foid reducri o t riske ')R =- 14 3 N. pl1)0I Urban 12 nionih, Condoim ui ;ncreiscd trom L o lo'., .p = 0 tl003) among _inLerention "iomen. hut onlv from ci to I P'-. tp =. Si In contL1 lnomcri: condom uwe amung men tinreacad tromi lO tc 1 ' 1'. / ii 0111 a and From 31i1 to 3i-1-,= Illimi iimonf intercnmion and control groups. respeciv-d'. Urban 2 V-I.cLk Men % ith HIV and ureihriti avc HIV\- I RNA o.incetr.ttiins m seminal plasnla ,c ht-f..'ld higher than mein %% thour urcrhrsti; tand chc sam CLE)t-i T-ell counti Too "eelk atrer antbiotic trtamcnt t;or urcthiiuis. the le'cl oF RNA in .emrn deerevsed signifle inrls. (Zonorrh.ea ca-use-d the greitest increas in . iral shedding Rural 2- mo nth' Hl\ 'eroconcisioni rarct %%ere 1.2' . in mienrcntion communitisi and I1.9'. In control com- mrunitie: r!sk dtiio tor .crocon%ers!on %a.s, Oi-'.. l (1. I' ro) O 's . p = 0.000Th Rural U-K rriOnths ii . OL m.nrh' after the nmass treatment there "a% a sa|tariicallv sipmficant decline in STI longo-ingi ,ssmptovm a_nd pre%incc in the- intentinon ari and not in control arm.. Ongoing research' U rbIn o mr.!'rthS Ind!' id ual] i "ere randomized to itcher rrLiditional STD rreatr in t nI tecatd o-l!i hen s nimpto- longo'iLng matic) or irten e ne treamnenL -,a minrd cerv mLonth and tri ted .LIc-rding tO an IiLenci'.e rlherapsI To date, no- Ig nificant ditterenc' I mean number I'F vIsIts or In sTEl pr'. alence haic barn detecred Urban 12 m.n[h' [D)ecline in c'rramarj[al 'c;x from P1 to 56".- ip.:Ii.Ou I declne in se:. '.ih sex S %orl.ers frion 12 i-. 6''.. i/p = fl.tll i. significant drtrcars in incidence o gornorrho.a. nongonwiococal urethri- liS and enital ullcr diwaee. bur no cihange in reporred condom use. Urban 3i1 mcnths Dccline n HIV ericon' ersion (torn 1 I. per t1:.0 %%:inian-sear, to -4 -i p r I-II.) cconian-% cars (p = 10 rIi3i. incrc asN in regular condom u,r %t[h client. from II) to 68...) U'rban -d' monthi Self-reported condonm use inc rea,ed Front 36 to 4I -t-.l.tIl H; SITE) prevalence decreased: gonorrih-a (rum 21 to IIJ',:.- 1p0 Il. Ii syphils trom 19 Lo T-ip = 1.03i3: geni-1d ulcer d!s- cav -i to 2'. p, =s ..i)iS ' r.I . i'd. '' ,,,i. ,, ." , .ig I 297 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Appendix A. (continued) Auith/or study, (Year) Couitry ldenn etion &cig,a 5ample size Inter'enaion: \'oluntarn counneliing and testing Alien and orhers R manda HR counseling for discordant PC '; cohabirating di;cordant couples I' b' couples Allcin and othcrs R.anda AIDs dlucation. HINI testing PC 460 HRV-positive and 998 HIVW- I 1'N2'ar and counseling, free condoms and negative women spermicides AlIen and oTher, RPtanda Hl\',ounselinpg nd testing PC I.58 childbearing women Deschimpes and HNii tCounseling. tesriLig. and frcc PC -i discordant heterocexual coupleo other' 14A I ondomrs Kam.cna and orher, Congo DR HI counseling for discordanr PC 14[1 discordanr married couples I IL00 II It;.rmcrIv Zairel coupls Mloort and others iL'ganda HI\ c:ounmdrn; and cotnmg PC 3.011) clhents at the .IDS I1.'11 Informticon C entre in Kampala Pickenn, and others The Gambia HNI' ouneling and tes[ing PC 31 (C5SW 1 I N HWV-positive and I ICs''1 19 HKR'-negative) Rsder arnd oher' Ccngo) DR HIR counscling and resting Pc 238 HR'-posiri%e and 315 NHT'- I1''1 %l Inegart'c women Tcmtnetman arid Kenya HRV conweling ar,d rttiing P(C 2i HRV-positive and 33 HIV' others 1il4i negatie childbearing \omen Ineerctraion Harm reduc:ion Peak and orher' Nepal Needle and ,\ ringe exchinge cS/PC 42-4 inijecing drug users CI Coafdun, riraenv- O R (tadd, r trio Xi., Sr d.- .-.re 'ic:lu.dd ,i dhen hdLl ar icnr a -monrh ftll.oiw-up and if thn- reporn-d the s;arianacal ligtiicance of the resulti. ;rirh the e;cpn. .in ,:.t the l-,cii,i - ' n.pienr bh Cohen .id nt[ho- I *,i - E=e%.pcr,mL,ra-.. QF =qua! spcran1cntal. PC=prospecra'e cohort. RC=rc[ro,spec;tie cohort. CC=casc-conirol. CS=Crois-section; L: St±ti't,l !ig.fi-ihncei; ;nd;catcd h,re Lhe au[hor, ha.e reported it i.. d& Li..rnp!ILd ha lulla. Di a .r! anhd hacl Nlcr,n. )al iL nUmcrsir Nchc.l.i 4 NlediCLnt. Department ' Ep,dcrnmologs and Publc H, ilt 298 APPENDIX A Len, e/h o Loottion obsenation Rent/n/ Urban 2o mronths Condom u,e incrcascd from t to s oin- alier t %c. \cr; condomr. u,e le;, common anmong sero- conmerred i v)l)'i, s. 5;ij, j = tU.01 in men: (C' . v , 2;', ,p = 0.1 i in %omrnenl . LUrban 24 month, Repori of c%er usingcondom increased from -to 22"':' after I yeAr: HI\-posici wromen *%cre more likely To use condoms r n 6;: vs 16'...: ti,I) l)i) HI V scioconversion rates decreased '13 To p5''.. p.: 41 In somcn v.hchse partners "terc resred ard counseled. L'rban -i months Tu%o-vear mcidcnrce f prcEgnancY wa, -fl(.u in HI\-p-esit, c %onmen and ;9"i. in Hl\-oneatis o omeni ipr.i.Oii . H IV-p(sitirL- oen ir oith te%ter than -i children mcrc nwre like1v to become pregnan[ [han [hose wi[h -4 or more children. Urban 6i months Safe sex practices or abstinence %%ere ftllo%%ed b o 4 f Of coupks: HIV incidenrc %\a 1.1' per Ii1IJ person-,cars 195" ! (CI: 0.80 to I. I1) among those practicing safe sex: .of couples cn[mnled si hase unpro[ected set. s-orh HIV incidence clI p. per iii) person-sears IC 1, .. C, CI o.X S3 r, X-Ii Lrban I rmonihs Betlre norificailon ot'fserchtut. lec" thin ot couple' reported uing condoms: renemonth after notilicarion. - .-:. ofcoupie,~ repenrted condonm use in all sesual intercourse: 1.R monthq atier notilic unLion. condom use rse- v. .-. Iriteonsi%e counseling tollov ing ni,Lticati,,n of HV I sratus led to lom erocon. ersion in pariner ( S. I !' during I 11(l peron- scars cibser%ed. Urbar. 6 nonrths Significant increase bietween ba,eline and - and 6-month t.lIAw-Up in risk reductuon strate- gies: e.g.. refraining frrn sexual ativirx faithfuliri,c. and condom u'c. Urban 2-; muncth, 0xerall pererntagr inc:itca in 5ondom uc in tirsi mrnonh " a, I L i, , I -2.8 to (.ut and between I and 2-s montis fC11 by 6 -i'" l)i. (i -I 1A ro I 2I U-irban i, monchs Condom use rarci aftcr 3t) monh,i dilikrcd i-ording to ,ero,ratu. I t. for HI -po.[itie and 3'.. for HIV- ncatL.1' 1 .. I) 111, . Adiustcd fcrrihrl rate! alo differdd 2-i i lI. birLh, per 1.000 o.omen for HI\-polU[se and i31 lI li o fi-r I iIVnegatie ip( .(r in ULrbarbn 12 m.nth, Lise of condcms vwas inf lrequent i1s- ;.Ir HiV-poitiiie and 641,. tr HI\ -nearisEe i and not sig- niFicantlv diftercnt ac;.;rding To sercsrNatia: prcenacnc rdtsr verc I ti'i. and I". rC,pcCti%Ck. a. no sratistccal diHfercnce bs ,enrc u,. Urban i sears HIl\' sropresalcnct remained loi- I u o1.. i L') I and i1in I m unsafe injection reduced: no change in un,afe e. 299 Appendix B Selected Studies of the Cost-Effectiveness of Preventive Interventions in Developing Countries HE TABLE BELOW SUMMARIZES STUDIES OF THE cost-effectiveness of interventions to prevent HIV in developing countries and several other studies that have measured only the costs of program outputs, without measuring their effects.1 Readers are advised to use this table cautiously. The cost-effectiveness of a particular intervention is not a constant; the costs, effects, and ranking of different interventions are very likely to differ across countries because of the degree to which the intervention is targeted to those with high rates of partner change, the prevalence of HIV in high- and low-risk groups, the length of time that an interven- tion has been in the field, the labor-intensity of the intervention, and the local cost of labor and other inputs (backgroud paper, Mills and Watts 1996). Thus, the results of the different studies in the table are not directly comparable with each other. Ideally, we would like to have measures of cost-effectiveness across multiple interventions for a single country (see box 3.9 of the text). The cost per HIV infection averted is available for only four interven- tions-one targeted to people with very high rates of partner change (sex workers) in Nairobi and three others addressed to those with lower rates of partner change. As discussed in chapter 3, government has a strong interest in supporting interventions that prevent the most secondary 301 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC infections per dollar spent. However, except for one study, secondary infections were not included in the benefits. * The annual operating costs of the Nairobi, Kenya, sex worker pro- gram came to roughly $70,000, or $140 per sex worker per year (Moses and others 1991). At the beginning of the program, 80 per- cent of the participants were infected with HIV and they had a mean of four clients per day. The annual cost per case of HIV averted came to $8, under the assumption of 80 percent condom use, or $12, under the assumption of 50 percent condom use. The number of cases of HIV prevented among the clients of sex work- ers and among sex workers themselves were included in the calcu- lations, but infections prevented among the partners of clients were not included. Had they been, the intervention might have been even more cost-effective. Reportedly, the largest share of the program's costs was for STD treatment, although most of the ben- efits arose because of increases in condom use (Mills and others 1993). However, the availability of STD treatment may have been a major factor in obtaining the cooperation of participants. * The Mwanza, Tanzania, STD intervention is the only one in the table for which the cost per DALY saved has been calculated- $10-11 (Richard Hayes, personal communication). The cost- effectiveness of this intervention is understated because the authors did not include prevention of any secondary infections in their analysis. The intervention might also have been more cost-effective had it been implemented in an urban area, where the number of secondary infections prevented might have been greater for each primary case. Of the total cost of $10.08 per treated case of STD, $2.11 was for drugs (Richard Hayes, personal communication). The incremental annual cost of this intervention, which served a catchment population of about 150,000 people, was $59,000, or $0.39 per capita. By comparison, the recurrent health budget of Tanzania in 1993 amounted to $2.27 per capita. * The cost-effectiveness of safe blood programs is strongly dependent on the level of HIV prevalence in the population and on the extent of risky behavior among transfusion recipients. The Ugandan study included only averted primary infections, that is, infections due directly to transfusions (European Commission 1995a,b). It assumed a prevalence rate of 16 percent among blood donors and of 40 percent and 9 percent, respectively, among adult and child 302 APPENDIX B Table B.1 Annual Costs per Infection Averted, per Condom, and per Contact for Interventions To Prevent HIV location, imiplementipg C(os per H(ll Coit per condom Cost per Intervenwtioni agenw' and year oeftkernci infection weired dfiscribmted contact Intenrt nijon S ru rgeted ti peciple i.iif/r ii.gli-r.k tk i/fl e r Informarion. condoims. Nairobi Kenva. rscearch project I 9s5 -9 I $I li_s 12 STD treatment for female se,- sor kers Peer educ3tion and Prostitute peer educ3tion proiect. 'i'aounde. i condoms. e'emale sex k%orkers Cameroon: Nlinistrn of He3lth I 19_9I Education and condoms. Pegap,o programme. Rio de Janeiro BrazLI. iJl 013 -3 male ser. workers aged I I -2 3 Social Health Guid;nce LUnit iNOSS.V8 Peer education and condoms. Bulanasc peer education project. Buda%%asc. $u. li 51)- female iex %%orkers and Zimbabw%e: Bulaasoc CCin Health clients, others4 Department. LUni'esiri of Zirbabsi. AIDSTECH I 1989i Needle exchange, bleach, K.arhmandu. Nepal. LifceQavng and 3.321 education. condoms. Lifegi%ing Sociec% i"K i2i health care to IDLJC Treatment of'sympomratic Research project. six rJral communlniei ot 52S .' I o.0, STD5 lMwanza Region. Tanzania. earlk 1ot1), Miaputui cirx and pronince. N%lorambiqc S'. 4O Johannesbur.g. South Africa 10. 16 Condom social marketing Ten programs IBolivia. Congo DR.' 50.02-Si). 3 Cote d'Itoire. Dominican Republic, Ecuador. GIhana. Indonrsia. Nlc-xiac. Morocco. Zimrbabwe Safe blood supplY liganda SIt-2 W31 ( per unit ol blood,I Shore-course ALT iherapy Hpothetical inter%enion In Sub-Siah ir in S to pre%ent mother-ro-ch,ld Africa assuming ptrinaral tran,nisicrio I transmission reduced from f'rom 2''3 ... to I '6-. a L.v.cr tiguret as;ime, S!"l-. condom u. highet figurc a,ut,1 m- .. ,nd.rri ;c I'. ircnludc he . due ;. dinarid cond,-m. c . e excluding tcchmical a,sistmncc from AIDSTECH is SO.Y)- pet tindorn distributed. d AI.o incudcd pc,ipL r, i br .torkcr 51TD peiu.nt, e Formerl Zaire r P20 Fer init collected S33- 5i per unit wLed in I )i. :*MiIl¶ mand oihcrt I 0,li) frnd a co-i of `il per un,t .- duced It-r the s..rne progrim but po'ibls lor an arlier yeaf NnatLsti Nairobi srud- MNoue, and othe'i 1Ni l Camirroon. Bractl. Zimbab,:c ,ux vrket pcer coun.cimin and condJor progr3m, Zre- dom .ocial mtrkeung programs. and S-TD treatmeor in NIc.-nmbiquc arid South Mirica. NIilU 3nd :eher'I 1I 9I SI. sr2sa §[E cre3rrment. Richard HaYe'. pcr'onal communeinri and Gilson and ciher, I 1 i%i6! LgIrndin af'e blood Eurp.ean tComrcrs;:.n I lCCi% AZT ther- apy in Sub-Saharan Attica Nlamnergh and others i 303 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC transfusion recipients. The calculations concerning the number of primary infections averted are in box 4.2 of this report. The cost per HIV infection averted was obtained by dividing the total addi- tional costs of HIV screening in 1993 ($319,894) by the total number of infections averted (1,863). * The effectiveness of short-course zidovudine (AZT) therapy to prevent mother-to-child transmission is not known as of this writ- ing; clinical efficacy trials are under way in a number of countries. The cost-effectiveness numbers in the table are, therefore, hypo- thetical. The calculations assume that the therapy would reduce transmission from 25 percent to 16.5 percent, or half the effect of longer-course therapy. Program costs were estimated from the lit- erature and are based on those in Sub-Saharan Africa, where most mother-to-child transmission occurs (Marsergh and others 1996). The authors calculated that a national program in a country with a 12.5 percent HIV seroprevalence rate would lower incidence of HIV by 12 percent. Since infants and young children are very un- likely to transmit HIV to others, there are virtually no secondary cases generated by this intervention. Preventing infection of chil- dren is one of the important external benefits of preventing infec- tion in their mothers (see box 4.6 of the report). The cost per case of HIV averted or per DALY has not been calcu- lated for the other studies in the table; only the costs are available. A nee- dle exchange and bleach program serving injecting drug users in Kat- mandu, Nepal, cost $3.21 per contact after only one year of observation and was organized with community-based outreach. A second program in Lubljana, Slovenia (not shown), was based in a fixed facility and had been operating only 5 months when costed at $12.59 per contact (Mills and others 1993). The cost per condom distributed varied from $0.10 to $0.70 for three highly targeted programs that had peer education and condoms for sex workers. Costs were much lower for ten condom social marketing programs-from $0.02 to $0.30 per condom distributed, in- cluding the value of donated condoms. Note 1. For examples of studies of cost-effectiveness and cost-benefit analysis of HIV/AIDS interventions in industrial countries, see Holtgrave, Qualls, and Graham (1996) and National Research Council (1991). Key principles of the economic analysis of health projects are reviewed in Hammer (1997). 304 Statistical Appendix T t 8 his statistical appendix assembles information about the levels and determinants of the HIV/AIDS epidemic and selected policy variables for low- and middle-income countries as classified in the 1997 World Development Report (World Bank 1997a). Table 1: HIV Infection Rates by Subpopulation Table 1 assembles the most recent information on HIV prevalence by subpopulation for all low- and middle-income countries as defined in the 1997 World Development Indicators (World Bank 1 997b). Country- level adult HIV-1 seroprevalence estimates are for adults age 15-49, as estimated by the World Health Organization for December 1994. The remaining data in table 1 on HIV seroprevalence by subpopulation, for all regions except Eastern Europe and Central Asia, are from the HIV/AIDS Surveillance Database of the U.S. Bureau of the Census (U.S. Bureau of the Census [database], 1997).1 Except where otherwise noted, data for Eastern Europe and Central Asia are from the WHO/EC Collaborating Centre on AIDS (1996). 1 The complete HIV/AIDS Surveillance Database can be obtained from: Interna- tional Programs Center, Population Division, U.S. Bureau of the Census, Washington, D.C. 20233-8860 USA. The e-mail address is ipc-hiv@census.gov. 305 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC The second through fourth columns of table 1 show the U.S. Bureau of the Census's "best estimates" of HIV-1 prevalence in "high-risk" and "low-risk" subpopulations in the capital or major cities and outside major cities. Their summary estimates are based on the detailed data they have compiled from published and unpublished sources in developing coun- tries in the January 1997 version of their HIV/AIDS Surveillance Data- base, and usually represent the results of a specific study that is judged to be of good quality. As a rule, "high risk" is defined as sex workers and clients, STD patients, or other people with known risk factors. "Low risk" is defined as pregnant women, blood donors, or other people with no known risk factors. The figures for urban high- and low-risk groups in the capital/major city were used in the cross-country analysis of the soci- etal determinants of HIV infection in chapter 1 of this report. The data in columns 6 through 17 on HIV prevalence among specific high- and low-risk groups of men and women are derived from analysis of the approximately 24,000 individual data records, drawn from 3,100 publications and presentations, contained within the Census Bureau's HIV/AIDS Surveillance Database. The figures in these columns repre- sent either the result of a specific study, if there was only one suitable study for a given year, or an average of prevalence rates from several suit- able sources or sites within the same country for the most recent year available. This table uses only seroprevalence surveys measuring HIV-1, except where otherwise noted as including HIV-1 and/or HIV-2. The data in these columns were used to define the stage of the epidemic for chapter 2 of this report, as summarized in table 2 of the statistical ap- pendix, below. Whenever possible, the entries for columns 6 to 17 represent the results of one or more studies with a sample size of at least 100 people. If more than one study satisfied these criteria for the same year and covered comparable geographic areas, an unweighted average was taken. Surveys with exceptionally small sample sizes (<100) were not taken into ac- count, except as a last resort where they were the only estimate available. Estimates based on small samples have been noted and should be used with caution because of their potential unreliability. Because of the difficulty of establishing samples of individuals with certain characteristics, studies based on samples of "high-risk" subpopu- lations may be unrepresentative. The same caveat applies to samples of pregnant women, when such data are collected from a non-random sample of clinics. Further, self-selection of some of the individuals in 306 STATISTICAL APPENDIX these groups-such as women attending antenatal dinics and STD pa- tients-may be serious and the HTV prevalence rates should not be taken to be representative of individuals who do not use these services. Similarly, while military populations are characterized by tighter-than- usual health surveillance in general, these populations are selected by age and other characteristics. For these reasons, results from the various high- and low-risk groups should not be considered to be indicative of prevalence in the general, low-risk population. Table 2: Indicators of Socioeconomic Development and Government AIDS Policies by Stage of the Epidemic Table 2 presents indicators of socioeconomic development and gov- ernment AIDS policies for 123 low- and middle-income countries with at least 1 million population. The countries in this table have been ordered alphabetically according to the "stage" of the HIV/AIDS epidemic-the extent to which it has spread among those practicing high-risk behavior and outward to low-risk populations-as used in chapter 2. In countries with a nascent epidemic, HIV prevalence in high-risk subpopulations is less than 5 percent. A concentrated epidemic is defined as one in which HIV prevalence in high-risk subpopulations is 5 percent or higher, but is still less than 5 percent among women attending antenatal dinics. A gen- eralizedepidemic is one in which HIV prevalence among women attend- ing antenatal dinics is 5 percent or higher. Classification of individual countries is based on the data in statistical appendix table 1, columns 6-17, generally using data from 1990 onward. If the only available data on high-risk subpopulations were from before 1990 and indicated low prevalence, then the country was classified as "unknown." Countries were also classified as "unknown" if there were no data on high-risk subpopu- lations and data on antenatal women were either missing or below 5 per- cent. Note that there are many countries with missing data, particularly on high-risk subpopulations, and many with outdated data. Therefore, these are conservative estimates of the spread of HIV; many countries may in fact be further along than indicated by this typology. China and India have been classified as having concentrated epidemics, based on a con- centrated epidemic in at least one province or state, respectively.2 307 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC The first ten columns of table 2 are socioeconomic factors likely to affect the spread of HIV3 The mid-1995 population, the 1995 gross na- tional product (GNP) per capita, and the growth rate in GNP per capita are from the World Development Report 1997 (World Bank 1997a). The Gini index is a measure of income inequality; an index of zero indicates perfect equality, while an index of 100 indicates perfect inequality. Data are for various years in the 1980s, from Deininger and Squire (1996), and are based on either income or consumption data from household surveys. The urban population as a percentage of the total and the growth rate in urban population are from the WorldDevelopmentIndica- tors 1997(World Bank 1997b), table 3.6. The 1990 urban male/female ratio is the ratio of adult men ages 20 to 39 to women of the same age in urban areas, calculated from United Nations (1993). Foreign-born as a percentage of the total population in 1990 is from United Nations (1995). The 1995 adult illiteracy rates are for adults ages 15 and older and come from the World Development Indicators 1997 (World Bank 1997b), table 1.1. The last six columns are indicators of HIV/AIDS policies on infor- mation and prevention. Year of the first reported AIDS case is as of De- cember 31, 1995, from WIHO/GPA data (UNAIDS/Country Support 1996). Note that in virtually all countries AIDS cases occurred before they were first reported and that the first case of HIV would have oc- curred years before the first AIDS case. The four categories of sentinel surveillance implementation as of 1995 are: planned sites, limited sites, many sites, and extensive sites. This information comes from Sato (1996), in Mann and Tarantola (1996); unfortunately, the definition of these categories was not provided in the original source. Government condom distribution in 1992 is from the AIDS in the World II survey, appendix table D-7.3, column PN5, of Mann and Tarantola (1996). The presence of a condom social marketing program in 1996 is from 2 Countries with population of less that 1 million that do not appear in the table but could be classified by stage of the epidemic include: Bahrain, Cape Verde, and St. Lucia (nascent stage); and Djibouti and Swaziland (concentrated stage). 3 Many of these variables were used directly or in some transformation in the national-level regressions of urban HIV infection in chapter 1. Note, however, that for the regressions in chapter 1, data for the same variables for earlier years were often used. Other variables used in those regressions are not reported in this table, such as the purchasing-power-parity-adjusted GNP per capita. The complete data set used for the chapter 1 regressions is available on request from the authors. 308 STATISTICAL APPENDIX personal communications with Philip Harvey (DKT) and Guy Stallwor- thy (PSI). The number of socially marketed condoms sold per adult in 1995 is derived from the data on total condom sales for 1995 from table 3 of the statistical appendix (see below) and the number of adults 15 to 49 from the World Development Indicators 1997 (World Bank 1997b) on compact disk. Table 3: Socially Marketed Condom Sales in Developing Countries, 1991-96 The 1991-95 data are from DKT International (1992-96) and 1996 data are from personal communications with Philip Harvey (DKT) and Guy Stallworthy (PSI). Several countries have now or had in the past more than one social marketing program: * India I is the government program offering Nirodh brand con- doms; India II is implemented by PSI/India (Masti and Pearl brands); India III is operated by Parivar Seva Sanstha (Sawan, Bliss, and Ecroz brands); India IV is implemented by DKT (Zaroor and Choice). The total for India I includes Nirodh condoms sold by PSI/India. * Indonesia I is the project that sells Blue Circle and Gold Circle con- doms, and is implemented by the National Family Planning Coor- dinating Board (BKKBN); it has devolved to the private sector; In- donesia II is implemented by SOMARC. * Nigeria I is currently supervised by PSI; Nigeria II was started by Family Planning International Assistance and Sterling Products. They were consolidated into a single program in 1993. * Philippines I is operated by DKT and Philippines II by SOMARC. 309 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 1 HIV Infection Rates by Subpopulation Sumnlry.: Hfll :I Preanilwice Women Adult H ' Cpifdl/mljOrthy Outsde maiore 'yAn4n,eutal prevalence, Loliw High Lou' High clinic, St Country 12/94 ri.k risk risk risk urban enar workers iear AMngola 1.(1 I .1: 11.5 L.0i S)5 Benin I.2 1.s 5S ,c,Lb 4L (O A 93 S3 i 93.-9 Bociwana 18.0 32.4 i.) 1 1)0 34.2': )s Burkina Fano 6.- 1 2.0 60.-i j 2.tl 95 60.-i' 4 Burundi 2- 20.0 i.S i .2 ')3 Cameroon 3.0 ;.' -; 3 2.0 tfl 1.; 21.2' Cape \'erde 2 01A S- Cenrral African RepLiblic i_s I 31.1L .; 10 9 1 .SI 0 " ) C hbd 2. d. I -t. 'C2 Comoros t!. Congo, DR 3.7 s 3'0. Z2 2 '.' '4 -i.6 03 30 3 0i Congo, Rep. -.2 I I!- 6: 2.6 o.1 '14 A).2' 1' J s- Cored'l1'oire 6.8 123'> II .ji ')S_')1 (',6J 94-9'; Dliburl ;.iJ 41i. '0 26.9- 91 Equarorial Guinea 1.1 Er,trta 3.2 1.6 83 Erhiopia -6 i 2_ 8. (- * * ( 9I Gabon I 3 1 L 9-4 Ganibia. Th 2.1 1.6 1t.6' I I.- 3s 34.- 93 Gh,na 2.3 3.2t i.2t l-u: 2.'' 15 0.8- S-b8 G uinea (1.o iJ,-t 36 I {. 3 0)-' "09 1 36.6 9-t Guinea-Bksalu 53.1 2.6k 0.9h 6.' 9 367. -' 8- Ken-3 8.3 18.1 i;S 101.3 I 3; S .; 92 LesoLho 31 6..i 11.1 t2 21.m 6.1 ;i Libcria 1.3 i3i Nladava;car ii.! 0.1 . i 1.17' 0.4 t!.1 I -IS 0.2 nS Niajaw I" Li 3 2.8 tJ.4- 11.8 32.8 9o T8.0- 9 - Mlali 1.3 1.4 -S 4.1' 52.8 S: 94 S;.j 95 Nlauriiania 17.- 0) 5. 0.0 0.5' B.13-c-s NlaurltIu. 0 I O.S 0.0 86 NInwzamb,que i 27- - 1' I~ r5 0s*-: Nanibia 6. i 1- 6 10 . 1- '-h. Niger I.(I 131 12.t L 1. t ')3 1S.-i ) Nigeria 2.2 o 20.1 -.S 93-9-s 223 93-0-s Rwanda -.2 2i.- 3.2 2.3' 'i SL.% .'5 Sao Tom& & Principe Senegal l.4 -F Ii) I' l.t . 1'- l 22.1' ' S vc hellec Sierra I tone 'ii 'o - ni ' iU 26. 95 Somdlia u.S 3 (I-% 2.4 90 310 STATISTICAL APPENDIX Men Both Homo- Injecting seua7ia drug STD ,1filitary Yenr bisexual 1ear u.ser. Year patients )ear 0.83 9)> 2-t.7 88 3.' 93 t 2.&; 95 41.8 9' IB.5' 8 6.2" 93 S.i' 9-t 15 O' 88 34.0 9.f-9; 0.0 88 0.1 8--88 9. I 86 I 2.0 92 20.1 92 I o.uj 9lo 104 91 2.6 9'1 '3.6 90 i 2a 68 6.8' 91 5.6d 9-a 0:1.0i 86 4.5 9 5 I1.8 86-8 6.0:3 wR 16.0 9.T 15.2c 93 0.OJd 8- 0.0 89 0.0" 85-86 0.3 95 4W.8c 95 0.9 9'-94 0.0 88-91 0.8 88-91 3.- 9(i 24.0' 94 .2; 92n 8.2 93-9'i 61.1 88-91 0.81 9 1.8 89-111 O.4 9 1 3.3 92 0O 90 0.0 90 Table ,rinnlEei on i/.6e-oul'n, 'z.g page.) 311 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 1 (continued) Sumnitiry: Hil :l Preralemce IWto,nenr Adult HNl' CGpi:lalln:jo? gifv O,nmide major fi .mn,n,a:ll pint a/chnow. Lou' Higb, Lrw Hfijli clinic, Sex Coinnj 121/94 ri-k ri k r.fik risk mrban Kw workers )iwr SOuthAfrica 3.2 1 S.' 2(1.1: t'i i0.q' yS 2O Sudan 1.0 3.0' 3* U'~ COS - 6 Sv aziland ..8 ' I I 5 ) Tan:anma 6.4 l - 49 j j U3 .3i I l tlS% .jti s 93 lo-go 8; 31 3.0! .3' Uganda -i. 1 82.5 d i o 1! 21 5 ipJIS 86.u 87 Zambia I'. I i. s S 1.1 12 - . ' t Zimbab%ie -2.o s lf -,.'' S*b 9514l I.' (..95 1..,.'. ,:a.4 O L (jU.bbd.ii.L' Argentina '5i .S '.b 2At d ' Barbados 2.8 I I- 0 ii" , Belizc 2.o Bvli l-a t 816% |za . Brazil I 2. 133 I 3. i. 3 1.' c72-91 Chl c 1. .1 1 11.1 4 Co1,*mbi 0.2 1.1 3i ''; S i IL j j t Cosra Rica IJ.i j- i 1.6 1.0 C2 ( Cuba 11.1)2 11.11 I) 1 0J I.' kIt) Domrinc.mn Republic I 0 2.0 6 2' 2.5'- 'iS - i EcuadUr u.. 3 . i6 9.3 9 1 i*i .- El all, ador t). i) 6) 0oh 94-95 2.0 95 Grenada ii 1 2 -o,h 91 GuLarermnala 0f. 0 0 S 0.0 .IIIl! 0.2c 89 LuI% ana I. ' 2i 16 ') ' 25.0 03 Haj 4.4 i.- -I..j i) S b3 I -j 1. 8') Hondura, lW i I' 2i it 9O 'tl i larnmijca Ot U. 2- 6 1. U(, 24 6 '6qOS NIeicc li6 'Ij I-' (Jl % C .I L',i( Nicaragu., t) I .lt .6' I .o LII LI I Panama %)6.1 1)0 3. jh 04 I*.I 6 P'aragUia% 0.I 11.11 . I) '2 (I 1 ' -". Pleru 8 3iI Sr. Kirrb & Nc'is s Sc I uca - I.' 2t1. SV. \,ncencr & clheCrenadines 0. I - iJ.2 'rc surinamc I .2 1118h 'I ?.1 *2 Triniclad x I ob.J 0L i - 4)3' I0; I ;.il SS lUrugua 1).- I I I I 1 3 \nrlczuda t). ^ 111 6.1 d 1. 312 STATISTICAL APPENDIX fmn Botlh Homo- In jecting iexuall 4d7tg S TD Aliditan' rea.7 bisexmwd }ar ser; )ear patideint ov 'go w.6 S(b J1.1 4 GAS 89 _j 89) IC1 IC cS' * 8() tl.(} .h(! 211. 1 ;14 1.9 S9-J rind M,) 2. Cl .6 3.1 9 3 -.z3 941 6E!1; 'Si 0.6 92 I | 4it 72 41 -i 96 ').le 1; 0,.0) 8& S., l d b 0.11'- 88 0.0 88 9.1- '~~~~~~ ~S 0. Id d 8 9 96 -i li - '4.9-5)i t 61 %18 0 9-s 26.2z 94 . bi-q- d.9 94i ().d COfl l-1 2 0. ) 9-_ _ iOd OY 4--85 6. 4; 0.0 88 28.8 d 88 3.6 "3 .4I 88-91 6.0 ');-'( 1.2d 9 1 0.- Q-7 5.3 '10-93 76.i- ' '! . 2 1.6 932 30.0' k' II.2 ,1 * 8 0 8- 6 2*3.1q I). 1 86 .72.- 1d 1.3 7 1 0.0 9(J 3.1 8-i-S 8.8 8--90 0.S 80L9(0 , 41.0) S'1)0) 28. I d 5ltC IS Si c1- A2 12 -4f.0 83-8-4 4.- 88-8i9 12.0. '1' 77 m. Ii 1 30 Cj(. u.w tAt 3;.0e 3 '- 3I3 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 1 (continued) Sumnmnr: Hll'-l Prevalmee ncee .Adult HI, Capiwalswio fiittr Outside major rz4' 1ntetl prevahlence Lowi Highs Lott, Higl e linic, Ser Counn-y 1219q risk risk rrsk risk urban Year workers Year Banggladeh 0.03 0.tl I 2"i 0x) o l3huta1 o3.01 (.I (-I Cambodia 1.) 3.2 -1.6 .0 '6 - 436-U' i6 Chin.a 0.0 0.0 73.2" ' f 0l31 03 Fil, 0.01 t:. 0.1 Iridia 9.-. .5d S6.3 t) .j lndoneSet t a (!. j.nl s(I- tl e Kribatri 1. I K(--rca. Den' Rep u.1 f0 Korea. Rep. U.1I1 0.8 ).0 10.1 qS Lao PDR (I. 03 0.8 1.2 1 Nlaldi,ci (1. 10 26 1 I NI a1idaxe 0!.6 Jq klcingulia 0 0)1 0it 11.8 l -3 ''' -93 NhIanmar !.S Ci I, 11: 0 I 3 1);2 NepIl w.1tS ).0" I Id 0 0: ' tl Pakiran. 0.06 11.6 3 ''(1 '; 5 Papua N1 w(Guinca 11.2 00I I.) 00 ''W I,... C Philippines 0 I.0 (.tv &olmon Isntad s 0.0 Sri Lanka O ..uu u.i 03 Th.i.and 2.1 2.8 21.6 2 -i ' 9C I 9.S \iernan- 0. (1.2 43.6 (I iFI C' .K.'14l, E.,,r Vo-n,r, . .,!, Afghanisran 0. 001I .lgeria .,, 'iI Bahrain 0j 2 Egvpr. Arat. Rep. (I.03 0.. .n oil (1.0 ' I '51-1 Irar. klamik Rep ()J 003 I raq 0 t13 lordan Lebinon I ibis tl 1-16 Miorocco 1l tl4- tl.2 I.4 0.2 ); - IA t- (-)man 0.1 saudi Arabia 0.01 cSyrian Arab Rep. 0.0)I Tunisia 0).0- 0. ( it ii.iti emcn. Rcp. 0.01 314 STATISTICAL APPENDIX Alen Both Homo- Injecting ;exuayli diitg ST1D ,liilitsirv Yeat bisexual Year wer; Year patients }etr 6. 9 90 12.6 q 854;' t1.4 tI.oj 00 66.5' 9 -i 01.0'J 13 1) 'c S6-91 1 9-9 61 92 'b.3' ')i 0.3 1!2-93 ()o 88-80 ii 0 (1-9 . I'.'' ')4 2S 92 19 n4 Q.~~ L'~j (*I*I*I SS...Iqi S 'S '*1, 'Thi .0~~~~~ 1 ')2 11.5 ½: ~~~~~~~IS1 f0.0 Ss-8S9 0 o ') "5' "2 g 02 0.0 'i - ;. , I'j. 0).0i b8-8'i 2.91 9( JII) "° 0I I S It 2-S 2.9 90 ~~0.111 8S-8 -.6~~~~~~,.{1' . i Tank 1' 1i!?ltjfC " II; tic fOllOJ' *,Y pl3I 3I5 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 1 (continued) Summary: HIT/ - Prevalence Women jdult Hj-n Capital/major rcir Outside major rit Antenatal prevalence. Lou l Highl Lou, Highl clinic. Sex Counltr 12(94 risk risk risk risk urban Year workmers ear E.ofrcn. Flttv.;r Lt ( r zi.k: Albani (1.01 Armenia 0i Azerbaijan 0.01 Bclarus 0.10 Bulgaria 0.01 0.0 93 Croatia 0.01 Czech Republ 0.i.)i 1 0 93 Esmronia o.in I Georgia 0.02 Hunganr 0.06 kaiakhsian n.u K-KTg,- Repuhbli.c 0.0 Larvi3 0.01 Lithuania 0 01 0.0 02'_ Nlaced,mna. FYR 0.Oi Nioldo' a 0.0) O3 Poland 0.05 Romania' 0.1) Russian Federation 0.10 0.0 95 5lo%Aa Republic 0.01 0.0 92 Slov-enia (0.0 Taiiki;an o.0 Turkey 0.0 0I 0 1.0I 8 --S8' 0.0 I Turkmenistan U.u Ulkraine 0.01 ('.0- 03 Llzbck-sran 0.0 Yugosla,ii. FR iSerb.rNlonit. 0.09 Bl.ni-11,l = Data n-. a%dlaijtle HR - I indu or HVP-2. b Rare repre;enrs I nfrcion wich HI\ -1 nnl and duil infcctuon H R'I and H!V 2i c Dlac- a. trap.d d ita are heir .valabk but axe nrt rece;y arilv rrhlble jlut tc mll.d Lmple ize c Sample Sic Linkn.,wn. t Natonal dca e P'.tc, h Nor specificila% u;ban LUrbin m.al pFr. imej in Si. Paul:. had a -iS.Lv', pr,%aicree rate in 1,10; i. For i unnar Pr.'ancc. K. Tamil Nidu Starc. 3I6 STATISTICAL APPENDIX Men Bfth Htnamo- injecting sextall tines STD Mlilizary ear bijexual Year ussers Jar patients Year 0.0 C.) 3 0.0 93 C'.') 93 ri. 9'3 0.1' C91 0).r0 ,i5 C.1)1 C) 0.0 0 0 'i 0 92 ~~U.0 ~ o.) 0 93 (1.9 85-90 130 6 ' 0 0.0 93 i lor Ni.snpwr Saic. m F ii the c' *.t I l dr r incIudes .11rr;- 'hr',rhel-hi.ed . ndireci irn-bri,rhel.basca. and milc pr.iii.arw tX. L'a...n qI Epliderri,ks. ai'8ubI H.lk ThaiLand. and Arm'. inrimmuic of Pathrhlog Ro; .l[Ial Arm;. . p \X1- lO FL( C:oIlab'.rarrnzCentre ar UIDt I9oi., UMAID'Ž' datd r In Ronmarua. twn P'9i rudic LI irphaned children ftorud 1-11 prc"dJence rare .' - S ind 21 tie I l Hri h and orherr 1901, ['-,kruv:}' arid ofher: 1I9 I r Ifl, AlT)S Al f lt3 CONFRONTING AIDS: PUBLIC PRIORITIES .i- A GLOBAL EPIDEMIC Statistical Appendix Table 2 Classification of Countries by Stage of the Epidemic, with Selected Economic and Policy Variables Affecting the Spread of HIV Fen,,omui, g7 o r li AIy. I (rbanizarion ,initel Foreigni- GA'Pper .Ag. aniual L'rbln born. G(;VP per fapitd U ribanl nIrban, malel 1990 Popuaioni. nZpira. rowt/i Gi, pop. gronwth rate fiemale r of' mid- 1)'95 1995 1%"), i,tdek 199j OtnJ, ratio, totzl Counnmi (lmillio,nJ ido/tn:rs 1985-95 +1980s.' r,;I 1990-95 1990 pop.) Nasceni Epidemic AJCt'rid - ';.13 1.0.11)l -2.4 t 11.392 S1 -tLU 1.03'4 1.5 \z:crbaiian-l -; ISIl 61;,' I 16. (-) Bang1Ld- h S. .' 2 il ' i 1 i; !b 0 t 5 0.7 BuLgai ria 89 -' 1.33' -2.o ''.23 I -0 I 1.1)0 o.2 Chilc I4 2 t.10 I u.' So I 'I '.'I(I 0.8 Co,ra Ric a 4 2.11 ri 2 8 i S - 1.91 S ( Cubd j j. ; - - o i.; 0.9U 0. C.:cch R,puhIb IhI- ,.. 1. -I - ('.2 A - ! 1l - E,uajdor 111; II S 1).-i S 55s i.6 09U t:t.8 EIrod.lo 1. e.S 23t -1. - - 4 "I 1i*l - Lio I'DR t- Sfl 27 I' i0 22 u' Q6'l6 t L[lhu.anua . I 'lii -1I> - 2' 1.0 I.96 - \I.idagbcir 1li.> 'il -'.2 -1 ti S .~ 0I'? 1!.3 N\3laur,rad 2 41111 ,,.; UI.-, 5-4 ;,i l.3it 3.3 Nl.atririu1.1 3 >1' l 4 ' -1 I, l 4! 1.3 I 1. lungIia 2 i 310 -i S - (-tl - 3.3 MI.:.roLLL. 2(1 I.1 .0 U.%r II.0 3'; -i. 1.1!*6 1. 2 Ncpal 'I ' 24 1)1j t - !. .111 ' .1 NiCMaEU4 d .5) -; Jt 0I30 2 4. 1 i] - I 2.1 Pipua NL - (JLIr.caI J 2 I Itli ' - il i - 0 Phdirpn,-; u-S (i) A 1; -I 4 nOS t1.1 Po[nd 2 iS.I _ I 2 11 ' 6i 1.0 0 I6 3 6 Ru-ian Fedrar.min I -t.s. 2.2' l - b - - 2 - - SIu.akd Ri-public ;.ek 2.05 -2 S - ;' 1.1 _ _ SIc..cni' 2.0 S.2iii - _ I U - - sonli_ia 's t- - - - - - sri Li-L.a Is'.' -hli4 2.1' ('2 22 2.1 1.1 CJ.I Nurinnmc U1 8.SI - - - _ - 2 \enin I ; ('11 - - 4 9. I 129 0.6 Concencrated Fpidemic Anguila I' -11' -4'.I - 41. - Ar2cfnina i. - S.tl1 I S - S. I - 0.9) Bra zil ;'I ' 2.0-41-1 -1-1S - S 2 J.0fF 07 3I8 STATISTICAL APPENDIX /'rev'ntioDe polin' W[omen'ns ,IatWcs CS'll altfae Fermle fear of Sn,:. Go' t fir SMr adrlt adult 1st Sur'. condone# Hl condoms illil., illit, reported Implern. distri- preven- sod pi-r 1995 1995 AIDS statews, budion, tion. adult, 1%°) r0 od caste 1995' 1992 1996 1995 26 S I S6 I Y - - 4 2 N il -4 'l90 I - Y ' - - 8- I - i i Si 3 V . 5 i 83 Y Y 3.28 -t " 86 2 - - - 24 ' N 8 !2 86 2 N V '.12 - - 92 2 - _ _.. I ) 22 8- 2 - 3 1 -( sI - - 8X8 2 iO 1 88 2 N 1 3 21 8 _ V _ _ _ I - 86 2 0S( 2.IS 8 6 ss 2 V V- 0.8) 3 5 33 8 2 - 1 9 3- S- i 6 846 4 - - - 86 2 - .. - - 86 2 - V . - - Mf 2 N - - S~ - _, _ 1 if 2- 1 i*8iJ ; ') .7; - - _ ~~~~~) - - S;S I V' -1 4 82 - r_ V_ 80j - V ; .'1.2 I Ld'/c$ ,c'tc'nI:,s'. 'il :S'r ,b!;'1{' '~ m-r.z ' 3I9 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 2 (continued) Economic grouwth A y. Urbanization annual Foreign- GVP per Avg. tnnuael ('rban born, GA'Pper capita Urban urban malel 1990 Poplaikiioni, capita. growth Gini pop. growh rae fi7nale (On of mid- 1995 1995 (Oo) index 1995 0j, rario. total C0fnuy (millions) (dollars) 1985-95 (1980s) (0o) 1990-95 1990 pop.) Cambodia 10 0 F(9 - - 21 0. i .95 0.3 C arner-.n 13.3 65)1 -6 (6 O.-.9 d iS5.3 1.1)5 '.4 Chad 6.0 180 0.6 - 21 3.5 1.21 0.3 China 1200.2 1a2(U 8.3 0.33 30 3.8 1.11 0.0 C.olombia 36.8 1.910 2.6 0.;' 3 2.6 0.91 0.3 Congo. DR -43-. 120 - - 2') -0 1.05 2.8 Domninican Repuhlic -.S 1 .40tI 2.1 0.4- C 6S 3.4 O.2)- 5 Eg-pr S-.8 -9H 1.1 0.38 45 2.5 l.WI- 0.3 El Saltador 5.6. 1.610 2.8 0.-8d 45 27 0.81 1.0 ErirtEa 3.6 I - - - - - - Fthiopia ;6 e 100 -0j 3 - 13 3.'I 0.92 1.6 t labon 1.1 3..')(0 -8.2 11.61 51) .5 1.33 8.9 -ambia 1.1 320 - - 26 6.8 1.13 11.2 Ghana 1-.1 3'%0 I.4 0.35 36 -i.3 1.01 0.9 Guaremala 11.0 1.3-jO 0.3 0.56 -.2 -..0 0.9-i4 0 G uinea - 0 i5o 1.d 0.4- 30 5.8 1.25 I. Honduras >.9 600 0.1 0.5.- 48 -jA. O.QO 0.- India 9ll).4 3N4O 3.2 0.33 ' 2.9 I).2- 1. 0l Jamaica 2.5 1.510 3.6 0.i3 55 2.1 (.89 0.8 NIala\sia 20.1 3.89 0 i.- 0.i5 ¾d t.u 1.01 t.2 Mli .8 2) 050 i!.8 - F 57 1.10 1.2 Nklexic'.) 91.8 3.3 20 0.1 0.5-i i 27 0.') d 'lvanniir 45.1 - - F 3.3 0.99 0.2 Niger ).C 220t - 0.36 23 69 - I 1 Nigeria 111.3 26t0 1.2 0.39 39 5.3 1.1' 0t3 Paki,rin -1tot) 60 1.2 0.32 35 C 1.23 6.1 r'Jraizu1% -4.8 1.690 1.2 - S-i -d 4 0 06 4 3 Pcrio 23.8 2.310 -I (6 0.48 2 2. 1.02 t 3 Sencgal 8.- 600 - 0.5-4 -2 4.0 1.US 2S Sicrra Leonc -.. ISO -3.6 0.61 39 -..9 - 5.0 Sudan 26.- - U1.39 26 -i.6 1.35 3.3 Thailand 58.2 27.-0i 8.4 t045 36 2.3 0I.92 0.6 Tooo -.1 310 -2. 0. ;-I 3 -i.8 1.23 4.1 Trinidad & [obago 1.3 3 0 -1 - 0.-46 68 1.5 0.9S 5.0 LUkrjine SI 6 !.o.t30 -'j. _0 1.0 -- 320 ]~~~oa T -a'i . j. sC tjs's --_Ji I z C. '£ n |1 ;a oC x O oS2 >* Z;~ |~~04 *.J .na s!o1$1r-. _ ,J -Cr.Js ; 5. |2r 0 t. _ hxi oA - >> £00 £000 DC X£ £ £0 acc X £ £0 X 0x X X o rc X or£ x x cr. Cf. "I Q ':r ,) .Z ML 1t CO I |. t r z I I I I | -fS v Z Z | / ~ Z | ~ < I - | _ -< < Z -: qCI .C Or t-I~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~0 ,,,> r . a = -2 0 ~ - 3 = _ C _:a-c a : -< - - - '.0 o ~~ '¢ * * *z ia I.j.']'I^ -_5b". . :r=j .-- t._- .. . vos k!!;;S; 0 * C .. . n .v 5aa. - - ~j £0. cr. IC_' . C X ICU -0 l}> W ><~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~, CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 2 (continued) Etonoajic grout/ 4 ,g t~~~~~~~ 'rblsiani.aion anna.11fi/1 Foreign- Gc Aper Avg. 4 rn0nua.1l (rtbau born, GAPpei capItr.a Urb/n: unhait ma/el 1990 Population. capita. gaout/t GitCi pop -rowuth rate filkle f -lo of jatid- 1995 1995 4.) indexe / 1995 '-). rat7o, totai Coua,rri .'id/iain' (do/law, 1985-95 '1980si I' 19909-95 19.9 pop.) LUriguŽ' 2 ' ii S i j.3'' 'I 0" 31.30 3 0 \enat,utls 'I ntlll _1 3ua . a5 2." 1.01 S \Vcnhrn '.s 241. _ 131(i 1 3 I 1 9')J .O. Generalized Epidemic Bajnir 5.i .i lS - 2 d 1.0') I Au B a. m.,% ina I I i.S4. 31 X- I . 2-i I .8 Burkina Fasca Iai 23u - ' - I6 I -i-i BUrundi 6i lotu -j.3S - i S .5cl 6.1 t-xnrral African kr i 3ia - 2 4 1') 2 2 _Y;3 2.a) aoringaa. FRP. 2R 1 - ' - s 1 .1l C i' cl r,, I j_jIO 4a d . A') .a 1.3 2_ ?.3 (,uInca-Bl.siil I!. 250: 2 i ('Snl 22 -t.2 - 1 -S Guam.il II S 'h'l - Ii if - - I l.'l (a. Hal U 2i'' -;. - . 0 ' 0.3 Kenw i . 2s ' I I.-i 2 '( I l il Lesotho 2.'' 1.2 ').6 23 6 2 .I 1.L 1''' l 'I b I , - _,, ; 1 gI.l a 3 5, 'a .d ic 12.1 NtI'!inih!q[XA~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~3 3i 1 i 4 |. i3 11. | N1zarnib-qu 16. ii ;( S1.613-1 Nalmk'a I c, 2'I"t ' K) _ 3 6.2n I .6i '- 6 R; and J i 1.I i Ji I . 16 .i) '9!Ljrh -Vr.-a 1. i o': I I i.(5i .1IlS 3 I TrSouc Atv _ 4t 1 In') 1 I: ''.62. _l d> 1.;1)8 _ 31 Tan,-jan ia 2'. ' I 0.-il)' 2-4 Ia S 3.3 Uganda I.2 2-si ' ) - '' V 1 _i( .1 I 9) 7ambia i It I -1 Al .'. 1. 37 di .1.I ZrnmbaEv -I I I - S 8.0 Unknoimn Stage of Epidemic AIgh.niiran 'A _ _ _ _ 1.12 ' ).2 Xlbmana 3.3 (j1-- 3- 31 ; 3 .1 AXrcn!a 1 3 S. - I - A 21 _ ; BcIarus'. 2; 5 ' 1l C rB - a r I a 4. 3 01 K' .i2 ib t 2 'IA-i IQ rcar.ia d.b 4-_I - - I-d 1.i - - H u ncarn al -vtdd -1 PI' - 55 6-3 Hunc:arv I ' 2 -.1";' -] t - '' l(, 090t) 0.3 Iran 1-di I4 - d.-l) I (1 6.2 322 STATISTICAL APPENDIX Prevenrionpolri't Mrio,nim satm. wS Mlakc Fem/ale Yea of Sn,:r. Grz,e finp SMI adult ddul lIs Sun'. condom H11' con1doms i/lir.. illit., repo rted Implemi diarr,- pmernc- '0old r 1995 1995 .AUDS Stains, bhntion. tiol. .dnlh, (.oo (0b) rete 1995' 1992 1996 j 95 2 3 2 ' .S II) 83 2 V V t,,, -I (I iI 2 .8-*' 9! 4 _V N 3.1. SI - U, x - V V.4-I 32 -*8 8 - I~~~~~~~~~~~~ - V 33 86~~~~"I I I x s 2 - 5 4 -b - V I I d 3') Ms - _ 1c-l Xt,8 b X 'Y 'S 55v. 4 - j I t 2 _(% - N 1.> I - 1 -t 1's 4S b2 X d N \ t, 'N 2 :t _ ] .i i -. \ E t . d N 2f1 43' .- 4 I N' 11 !1I SX, 5') 8 NN V 532 SS1 - - -- '14 -- - 3 - _ "I 2 N 1d S _i I N _ '3-5 --_ _ - - - ~ S'. _- N 323 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 2: (continued) Econromic growth A t. Urbanization annual Foreign- G.Pper -lr. annual Urban born, GAP per capita I (rban urban nialel 1990 Populatioa i. capita, groa-ath Gini pop. growth rate female Vio of muid-1995 1995 (Io), index 1995 %06), ratio, total Counry (rlmillionis) (dollars) 1985-95 (1980s) (so.l 1990-95 1.990 pop.) Iraq 20.1 - -8 x .5 1.11 2.8 Jordan 4.2 -f 3.39 ''2 .9 1.11 26.-i KAzakchscan 16.6 1.330 -8 - 60 1 2 - Korea. Dem. Rep. 23.9 d ' - 61 2.-i - 0.2 Klirgyz Republic 4-5 s0) -.t - 39 1 1 - Larvia 2 -6.65 - -3 -0.6 0.96 Lebmnon -t 266f0 - - 8- 8 0.88 12.2 Libena 2. - - - - 1.31 S.0 Lib%a 5.-i - - 86 dt 1.12 12 .3 Mlacedonia. FYR 2.1 86i - - 60 1.6 - - NloIdo.a ±43 920 - - _2 1.5 - Oman 2.2 4.820 0.3 - 13 S.6 - 33.6 Panama 2.6 27,50 -n0.4 0.52 56 27 0.9i 2.6 Romania 227. 1.4801 -3.8 0.26 55 0.0 1.03 0.6 Saudi Arabia 1 ____ .uiO _ -1.9 - _9 4.0 - 25.8 Sy rian Arab Rep. I q.1 112I 01.9 - 53 -t3 1.10 6.6 Tajiikisan 5.8 340 - - 32 2.1 - - Tunisia 9.0 1.X20 1.9 0.-i3 2. 8 1.Oo 05i Turkev 61.! 2.-hu 2.2 u-iD -(1 47 1.1-4 0 Turkmenistan 4.i 920 - - - LTzbekisLan 21.8 9Th _3.-)h - -0 2.6 - - YugolJavia. FR lSerh.!o1onr.i 1 0 d 0. 33 j 1.5 0. 98 1_ -Nor a%aulable. Nor applic,blk iNI so|li markered V = Xes. N = N., a. I Planned. 2 Lirmrcd. i Nlany,rres. qi ELacnmie. b. E'rniarc; kr conumis ot rhe torTTier Sovier L[ni-n are preliMi.nasn. c. Eonmawd to bc Iu % income iSni5 or lewt. d. Fstimared r be louer-middle income (5-66S ro 53.3SX i e Eritmared to be upper-middle income 53.036 1t) 324 STATISTICAL APPENDIX Pretvention polico Wometnt' status CSA Male Female Year of Sen. Go' t fAr SMl adult adult Ist Sun. condom HITl' cottdond il/it. illit. reported Inplem. distri- preven- sold per 1995 1995 AIDS status, bution, tion, adult, (96) (qo) case /995i 1992 1996 1995 29) ts 9 1 1 2 1 86 2 N 3 . - - - 2 _ - - 90 2 5 10 8( 2 - 46 -s 8O - 12 8 9 - - I 3- SC2) - - 86 2 N 10 86 2 Y* - - 85 3 - 29 SO 86 _ 1 -54 s I - 2I S 86 1 Y 8 2S 8S l - Y 0.30 _ _ 1 _-, - - ~~~ ~~'12 2 V _ - 8S 2 N 325 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Statistical Appendix Table 3 Socially Marketed Condom Sales in Developing Countries, 1991-96 iearprograw Social/v markeed conidom sales igoausands) Con,,rn t'launcled 1991 1992 1993 1994 1995 1996 .AJbsnis 1'3')% . ...... i' BanglideTh H1' 5; 11 .36u_1 I.'52-i8 I;O 05'9-i 1 615i3b8 1l5.999 Bcnin 1(184 6S sSI I 34-i 1L5 2.66(3 2 506 Bo0 1:88 b 3.; 5-B 6-Si (632 1 338 2,53-4 Bof'r anra I)3 .. . . 1,0o 1'283 2.233 1.625 BrazO 119I -'Jlo 3.08-t 6 -i I I.;(b IS.2._2 2 6. ( 88 Burkin.s Fad 11 -i 2)I 2 252 1 ;6 ( ;Si3 - 5-i) Burundi 1q933 lo .I 1.12 5iS 1.27 -S L itobodi3 1 .)C . . 7'N 'i 032 9 16 (arnr:.on 5'38;C3,19-i ".lli 5.i .20i -.563 9.254 CentrraI .AiriLan Republic 1Cie, I II) 7 1.381 I .1 2.32 2 000J C had 1'969 . . . . . Chind 1'N,$ .. .. II.--b (C'nlonbi,, 1'i i )i j ,0i 6._22' 4. 10 6,.391) Congio. D)R IshS. 1.81") .1I(fl 2.3Si 3.190 82 17,-6 ConFco Rep ot' . .. . . . 5 Loqa RR.z I 2,33 Y x .s'kiS i 2') '.9 S.bIS - c.rc d I ro,re 'NI) 1 S28 o,3s; i.8j b.! 9 I0.hO0 12.3I1 Dorninic.in RcpUblic I .i) 1S{s' !.i. 1,81'l 1.2-0 8i 1.I-0 Ecuador I )b6 21-i 8 .S 2-i-i - 69-i I.-19-i E, pi I 1')-' 1-,t668 121.3%l' .L92 I -j9 13$ 3 - El >l%ador 10-6 1 >o9 2.241 '.1 1.512 1.51, 1.5Xi Eihipll Iti j -9 2 _3 6 11 I-Ss - 2ci;3 I ).N 20.n62 Glhn.d I3 )S! 4.S 3t3 -i)s') - 3s6 3.411 4,66S (Lua[rnial3 1 CIS.; i 61 2 2!0 ' .46-3 '.0'2 2,90 1,993 , i, l nc 3 I i2 2 l ] . 'I' i 2 SI 1.988 I 398 27 3,1'9 Guine_-BisiLl I J! . . . 54'; Haimi 19(1 . . I I %3.; 'I 3.2 i. -iJ 4 2o Honduras 19')i1 A.1- 92)1 ,-i 834)' 6'b8 IndilA i' I 8 2 4 1.2-i _'3) 43- 2s - d . I I I .50s I s i -5-i 19"6.-i 10 Indid 1I l9bb -S.-i-i 2i.96 -, 824 21.1 9) 3-i-,44 41.380 Ilndia III I% i.91 8 7.68 8.8h, o S 1 ;.' 3.3621 I i,821 indja IN I'N) - 4 2 61 .ISI) ,!'98i Ind.-,n,'i 1 ait. darcc >.2 28 2 ') ob Sqqo)(t 3.Mb Ind.linc ia il 1'I"6 . . . , . . 1.4-.3 lanius 3 19-4 19S I ,1(- s.(I' IS.'Sl . 1.325 Kens a I 'b -i4) Sib I .6 -9 6.009 4762 Leyn>cho 199( 9C . .I. 82 21'] 2S Nladasvar I )Ku ;.. . . . ..? 4.6 3 i. 1 NIlaIaE^,a 10'11l . I .258 3.',(-l 5,053 .152 8.i83 Malil; I'"'2 .) . S9 .87 3.051 3,053 .NIarz oeI .8'.;' . I ' _10 2 .4i 2.326 2.ir 1,958 3z6 STATISTICAL APPENDIX Statistical Appendix Table 3 (continued) leTr progEZ#ra //. l .7i ,iSociall/ ,',a wd w saie, I,/'sadip Co(unn Llhum-bed /991 1992 199S 199,t 1995 1996 Niozambique 1')'1 . . . . 2.13 4.i)S6 Nivanmar I 99i .3. . . . . 368 Nepal i-4 4T) A.i-6 5.uS. - 203 -S.46l 6t - I-i Niger I99 . . . - 1,120 ' 216 Ni,eria P I,Ss.1 23614 t S.243 SS6i 4 . I .1S Nigeria 11 mid-S-i _ 31!12 Pakisran I'j6 3.3b; 3t- 11-i"' 3'1 (;23 -48 340 i7S,-i l SS-3 Peru 16 - 13i1s A.2 - i,33o Philippines 1P I "tI, , I.2(o 2'.c1,2 S `3 136 'IJII6 I I 832 Philippine 11 I"q2 . _ - liii i 1. - Russian Federarion I '.t:. . . .. .. 2.212 R%%anda 003 . . . I 'Ib3 S(,1 iW; 2.s i Senegal 1. .r. . 1311 SouCh Africa 192 . Ili4 s:O 1.3 I.;I I (11 Sri Lanka j4 .. t n; - 3 - .t.t - ..S2 Tanzania I '" . . - i -i II TOrJO 1,-)c12 _ _ 2.22 .i.-4i 1W') TLurkey' Lt' j ?32S i 6 5'6! .64 '261.q4 j iJ!i)Ii Ligindi .'1I - ;1.1S 12 I I2S''Si '1951 LUzbekj,i'n I'64{ . . . . Venezuela Ilo2 - - 2' A2 -0 Vietnam S1'.1 3'K-44 Zambia I'4N2 . . st3 -0 ThYi u.6'I ,.2~ - iii) Zimb3b%ke ''". .l 2s, I!.'2 .0t'2 S; (:JI -NI -Nmr it alIabk * . Nor ippiicablc progr*.rn nor in ..in raron. i tourni has or hN. ni.re rhnz .hr, .wc..n, .ci il rnri,.erirc pilgidri b Program deCId r iht pr,. I t. rc rtcr.. Nivr! (I .;nd 11 ; -.e .i.J.,rcd f nj,n .1 pnglc-Irdo ,.s s '. ±nd ir.. ..d d u.i..r N Igo, 327 Selected Bibliography Background Papers Jones, Christine, and Allechi M'bet. 1996. "Does Struc- tural Adjustment Cause AIDS: One More Look at the Ahlburg, Dennis, and Eric Jensen. 1996. "Economics Link between Adjustment, Growth, and Poverty." of the Commercial Sex Industry in Developing Countries." Kremer, Michael. 1996a. "AIDS: The Economic Ra- tionale for Public Intervention." Ainsworth, Martha, and Innocent Semali. 1997. "Socio- economic Correlates of Adult Deaths in Tanzania." . 1996b. "Optimal Subsidies for AIDS Prevention." Bechu, Natalie. 1996. "Les Effets du Sida sur L'Eco- nomie Familiale en C6te d'Ivoire : Analyse empirique Menon, Rekha, Maria J. Wawer, Joseph K. Konde-Lule, de l'eolution de la consommation des menages tou- Nelson K. Sewankambo, and Chuanjun Li. 1996a. ches en C6te d'Ivoire." "The Impact of Adult Mortality on Households in Rakai District, Uganda." Biggs, Tyler, and Manju Kedia Shah. 1996. "The Impact of the AIDS Epidemic on African Firms." . 1996b. "The Socioeconomic Correlates of HIV Infection among Household Heads in Rakai District, Dayton, Julia. 1996. "World Bank AIDS Interventions: Uganda." Ex-ante and Ex-post Evaluation." Mills, Anne, and Charlotte Watts. 1996. "Cost- Deheneffe, Jean-Claude, Michel Carael, and Amadou Effectiveness of HIV Prevention and the Role of Noumbissi. 1996. "Socioeconomic Determinants of Government." Sexual Behaviour and Condom Use: Analysis of WHO/GPA Surveys." Morris, Martina. 1996. "Sexual Networks: What Is Their Relevance to Modeling the Spread of HIV?" Filmer, Deon. 1997. "Socioeconomic Correlates of Risky Behavior: Results from the Demographic and Health Over, Mead. 1997. "Societal Determinants of Urban Surveys." HIV Infection: An Exploratory Cross-country Regres- sion Analysis." Janjaroen, Wattana. 1996. "The Economic Impact of AIDS on Households in Thailand." Perriens, Joseph H. 1996. "Treatment of HIV Infection." 329 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Pyne, Hnin H. 1997. "National and International Paper 6. Population Services International, Responses to the HIV/AIDS Epidemic in Developing Washington, D.C. Countries." Ahituv, Avner, V. Joseph Hotz, and Thomas Philipson. Riehman, Kara. 1996. "Drug Use and AIDS 1995. "The Responsiveness of the Demand for in Developing Countries: Issues for Policy Condoms to the Local Prevalence of AIDS." Journal Consideration." ofHuman Resources 31(4): 869-97. Saba, J.' and J. Perriens. 1996. "Mother-to-Child AIDS Analysis Africa. 1996. "Vaginal Microbicides: Transmission of HIV." How Do You Kick-Start a Global Market?" Vol. 6 Shepard, Donald S., Justine Agness-Soumahoro, Richard (5): 1. N. Bail, Charles S. M. Cameron, Antonio, C. C. Campino, Roberto F. lunes, Jose Antonio Izazola, AIDSCAP/Family Health International, Harvard School Tiecoura Kone, Sukhontha Kongsin, Phare Mujinja, of Public Health, and UNAIDS. 1996. "The Status Jeffrey Prottas, Jorge Saavedra, Adele Silue, Laksami and Trends of the Global HIV/AIDS Pandemic." Suebsang, Paula Tibandebage, and Samuel Wangwe. Final Report of a Satellite Symposium of the Eleventh 1996. "Expenditures on HIV/AIDS: Levels and International Conference on AIDS, Vancouver, B.C., Determinants, Lessons from Five Countries." Canada. July 7-12. Family Health International. Stover, John. 1997. "The Future Demographic Impact Ainsworth, Martha, and Godlike Koda. 1993. "The of AIDS: What Do We Know?" Impact of Adult Deaths from AIDS and Other Causes on School Enrollment in Tanzania." Paper Van Vliet, Carina, King Holmes, Burton Singer, and presented at the annual meetings of the Population Dik Habbema. 1997. "Effectiveness of HIV Preven- Association of America. Cincinnati, Ohio. April 1-3. tion Strategies under Alternative Epidemiological Scenarios: Evaluation with the STDSIM Model." Ainsworth, Martha, and Mead Over. 1994a. "AIDS and African Development." World Bank Research Observer 9 (2): 203-40. Bibliography . 1994b. "The Economic Impact of AIDS on Africa." In Max Essex, Souleymane Mboup, Phillis J. Adler, Michael, Susan Foster, John Richens, and Hazel Kanki, and Mbowa R. Kalengayi, eds., AIDS in Slavin. 1996. "Sexually Transmitted Infections: Afiica. New York: Raven Press. Guidelines for Prevention and Treatment." Overseas Allard, R. 1989. "Beliefs about AIDS as Determinants of Development Administration Health and Population Preventive Practices and of Support for Coercive Occasional Paper. Overseas Development Adminis- Measures." American Journal of Public Health 79 (4): tration, London. 448-52. Ad Hoc Committee on Health Research Relating to Future Intervention Options. 1996. Investing in Allen, Susan, A. Serufilira, J. Bongaarts, P. Van de Perre, Health Research and Development. Geneva: World F. Nsengumuremyi, C. Lindan, M. Carael, W. Wolf, Health Organization. T. Coates, and S. Hulley. 1992a. "Confidential HIV Testing and Condom Promotion in Africa." Journal Agha, Sohail. 1997. "Sexual Activity and Condom Use of the American Medical Association 268 (23): in Lusaka, Zambia." PSI Research Division Working 3338-44. 330 SELECTED BIBLIOGRAPHY Allen, Susan, J. Tice, P. Van de Perre, and others. Andriamahenina, Ramamonjisoa. 1995. "Enquete 1992b. "Effects of Serotesting with Counseling on Epidemiologique sur l'Infection a VIH et la Syphilis Condom Use and Seroconversion among HIV Dis- dans les Villes d'Antananarivo, de Toliary et de Toa- cordant Couples in Africa. British Medical Journal masina (Mai-A6ut 1995)." Laboratoire National de 304: 1605-09. Reference MST/SIDS, Madagascar/Futures Group International, Madagascar. Allen, Susan, C. Lindan, A. Serufilira, P. Van de Perre, A. C. Rundle, F. Nsengumuremyi, M. Carael, J. Aral, S. O., and T. A. Peterman. 1996. "Measuring Schwalbe, and S. Hulley. 1991. "Human Immuno- Outcomes of Behavioral Interventions for STD/HIV deficiency Virus Infection in Urban Rwanda: Demo- Prevention." International Journal of STD and AIDS graphic and Behavioral Correlates in a Representative 7 (suppl. 2): 30-38. Sample of Childbearing Women." Journal of the American MedicalAssociation 266 (12): 1657-63. Archavanitkul, Kritaya, and P. Guest. 1994. "Migration and the Commercial Sex Sector in Thailand." Health Allen, Susan, A. Serufilira, V. Gruber, and others. 1993. Transition Review, 4 (suppl.): 273-95. "Pregnancy and Contraception Use among Urban Rwandan Women after HIV Testing and Coun- Arias, R., and Luis Serven. 1997. "Measuring Aid Flows: seling." American Journal ofPublic Health 83: 705-10. A New Approach." Working paper. World Bank, Policy Research Department, Washington, D.C. Anderson, Roy M. 1996. "The Spread of HIV and Sexual Mixing Patterns." In Jonathan Mann and Arrow, K. J. 1963. "Uncertainty and the Welfare Eco- Daniel Tarantola, eds., AIDS in the World II. Global nomics of Medical Care." American Economic Review Dimensions, Social Roots, and Responses. The Global 53 (5): 941-73. AIDS Policy Coalition. New York: Oxford University Press. Asiimwe-Okiror, Godwil, Alex A. Opio, Joshua Mu- singuzi, Elizabeth Madraa, George Tembo, and Anderson, Roy M., and R M. May. 1988. "Epidemio- Michel Caradl. 1997. "Change in Sexual Behaviour logical Parameters of HIV Transmission." Nature 333 and Decline in HIV Infection among Young Preg- (6173): 514-19. nant Women in Urban Uganda." Forthcoming in AIDS. Anderson, Roy M., S. Gupta, and W. Ng. 1990. "The Significance of Sexual Partner Contact Networks for Asiimwe-Okiror, Godwil, J. Musinguzi, George Tembo, the Transmission Dynamics of HIV." Journal of Alex Opio, Elizabeth Madraa, B. Biryahwaho, S. Acquired Immune Deficiency Syndromes 3 (4): 417-29. Okware, C. Byabamazima, and P. Turyguma. 1995. "Declining Trends in HIV Infection in Rural Areas in Anderson, Roy M., B. Schwartlander, F. McCutchan, Uganda." Abstract WeC206. Ninth International and D. Hu. 1996. "Implications of Genetic Vari- Conference on AIDS and STDs in Africa, Kampala, ability in HIV for Epidemiology and Public Health." Uganda. December 10-14. Lancet 347 (9018): 1778. Bagenda, D., F. Mmiro., F. Mirembe., C. Nakabito., D. Anderson, Roy M., R. M. May, M. C. Boily, G. P. Mugenyi, and L. Kukasa. 1995. "HIV-1 Prevalence Garnett, J. T. Rowley. 1991. "The Spread of HIV-1 Rates in Women Attending Antenatal Clinics in in Africa: Sexual Contact Patterns and the Predicted Kampala, Uganda." Abstract MoCO16. Ninth Demographic Impact of AIDS." Nature 352 (6336): International Conference on AIDS and STDs in 581-89. Africa, Kampala, Uganda. December 10-14. 33I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Baggaley, R., P. Godfrey-Faussett, R. Msiska, D. Rasamindrakotroka. 1996. "High Syphilis and Low Chilangwa, E. Chitu, J. Porter, and M. Kelly. 1994. but Rising HIV Seroprevalence Rates in Madagascar." "Impact of HIV Infection on Zambian Businesses." Lancet347 (9004): 831. British Medical Journal309 (6968): 1549-50. Bernstein, Robert S., David C. Sokal, Steven T. Seitz, Bankole, Akinrinola, and Charles F. Westoff. 1995. Bertran Auvert, John Stover, and Warren Naamara. "Childbearing Attitudes and Intentions." DHS Com- 1997. "Simulating the Control of a Heterosexual parative Studies 17. Calverton, Md.: Macro Interna- HIV Epidemic in a Severely Affected East African tional. City." Forthcoming in Interfaces. Barnett, Tony, and Piers Blaikie. 1992. AIDS in Africa: Berstein, C. A., J. G. Rabkin, and H. Wolland. 1990. Barnett and Piers Bmaie. 192 AIDS Gilforia "Medical and Dental Students' Attitudes about the Press. AIDS Epidemic." Academic Medicine 65 (7): 458-60. Bertozzi, Stefano. 1995. "Global Challenge of AIDS: Barnum, Howard, and Joseph Kutzin. 1993. "Hospital Ten Years of HIV/AIDS Research." In Y. Shiokawa Costs and Efficiency." In Public Hospitals in Devel- and T. Kitamura, eds., Proceedings of the Tenth Inter- oping Countries: Resource Use, Cost, Financing. Bal- national Conference on AIDS/International Conference timore, Md.: Johns Hopkins University Press. on STD. Yokohama. August 7-12, 1994. Tokyo: Kodansha. Barongo, L. R., M. W. Borgdorff, F. -F. Mosha, A. Nicoll, H. Grosskurth, K. P. Senkoro, J. N. Newell, J. Besley, Timothy, and Ravi Kanbur. 1988. "The Princi- Changalucha, A. H. Klokke, J. Z. Killewo, J. P. ples of Targeting." University of Warwick, Develop- Velema, R. J. Hayes, D. T. Dunn, L. A. S. Muller, ment Economic Research Centre, Department of Eco- and J. B. Rugemalila. 1992. "The Epidemiology of nomics, Coventry, U.K. HIV-1 Infection in Urban Areas, Roadside Settle- ments: Villages in Mwanza Region, Tanzania." AIDS Bhave G., C. P. Lindan, E. S. Hudes, and others. 1995. 6 (12): 1521-28. "Impact of an Intervention on HIV, Sexually Trans- mitted Diseases and Condom Use among Sex Work- Bastos, F. I. 1995. Ruina e reconstrucfdo: AIDS e Drogas ers in Bombay, India." AIDS 9 (suppl. 1): S21-30. Injetaveis na Cena Contemporanea. Brazil: Relume Biggar, R. J., P. G. Miotti, T. E. Taha, L. Mtimavalye, Dumara. R Broadhead, A. Justesen, F. Yellin, G. Liomba, W. Beal, R. W., M. Bontinck, and L. Fransen, eds. 1992. Miley, D. Waters, J. D. Chiphangwi, and J. Goedert. "Safe Blood in Developing Countries: A Report of the 1996. "Perinatal Intervention Trial in Africa: Effect of EEC's Expert Meeting." EEC AIDS Task Force, a Birth Canal Cleansing Intervention To Prevent Brussels. HIV Transmission." Lancet 347 (9016):1647-50. Birdsall, Nancy. 1988. "Economic Approaches to Pop- Becker, Gary S. 1981. A Treatise on the Family. Cam- ulation Growth." In Hollis Chenery and T. N. Srini- bridge, Mass.: Harvard University Press. vasan, eds., Handbook of Development Economics. Vol. 1. New York: North-Holland. Beegle, Kathleen. 1996. "The Impact of Prime-Age Adult Mortality on Labor Supply." Michigan State Birungi, Harriet, and Susan Reynolds Whyte. 1993. "In- University, East Lansing. jections, Control and Community Needs in Ugandan Health Care." In M. Bloem and I. Wolffers, eds., The Behets, Frieda, Ramamonjisoa Andriamahenina, Joce- Impact of Injections on Daily Medical Practice. Amster- lyne Andriamiadana, John May, and Andry dam: Free University Press. 332 SELECTED BIBLIOGRAPHY Birungi, Harriet, Delius Asiimwe, and Susan Reynolds Bourdeaux, Richard. 1996. "HIV Ravages Drug-Torn Whyte. 1994. "Injection Use and Practices in Town in Belarus." LosAngeles Times, Sept. 29, 1996, Uganda." WHO/DAP/94.18. World Health Orga- p. Al. nization, Action Programme on Essential Drugs, Geneva. Brandt, Allan M. 1987. No Magic Bullet: A Social History of Venereal Disease in the United States since 1880. Blix, 0., and L. Gronbladh, 1988. "AIDS and IV NeYrOfor Unves .pres Heroin Addicts: The Preventive Effect of Methadone New York: Oxford Unversity ress. Maintenance in Sweden." Paper presented at the Brown, David. 1997. "U.S. Advocates Triple Therapy Fourth International Conference on AIDS, Stock- To Fight AIDS." Washington Post, June 19, p. Al. holm, Sweden. June 12-16. Bloom, David E., and Peter Godwin, eds. 1997. The Brown, Tim, and Werasit Sittitrai. 1995. The Impact of Economics of HIV and AIDS: The Case of South and HIV on Children in Thailand. Research Report 16, South EastAsia. Delhi: Oxford University Press. Program on AIDS. Bangkok: Red Cross Society. Bloom, David E., and Ajay S. Mahal. 1997. "The AIDS Brown, Tim, and Peter Xenos. 1994. AIDS in Asia: The Epidemic and Economic Policy Analysis." In David Gathening Storm. Asia Pacific Issues 16, Program on E. Bloom and Peter Godwin, eds., The Economics of Population. Honolulu: East-West Center. HIV and AIDS: The Case of South and South East Asia. Delhi: Oxford University Press. Brown, Tim, Werasit Sittitrai, Suphak Vanichseni, and Usa Thisgaborn. 1994. "The Recent Epidemiology of Blom Dai E. Ti'rw,Pu erlr n eai HIV and AIDS in Thailand." AIDS 8 (suppl. 2): Sittitrai. Forthcoming. "Employment and Earnings of H131a41s Commercial Sex Workers." Bobadilla, Jose Luis, Julio Frenk, Rafael Lozano, Tomas Bulatao, Rodolfo A. 1991. "The Bulatao Approach: Frejka, and Claudio Stern. 1993. "The Epidemiologic Projecting the Demographic Impact of the HIV Transition and Health Priorities." In Dean T. Jami- Epidemic Using Standard Parameters." In United son, W. Henry Mosley, Anthony R. Measham, and Nations and World Health Organization, The AIDS Jose Luis Bobadilla, eds., Disease Control Priorities in Epidemic and Its Demographic Consequences. New Developing Countries. New York: Oxford University York: United Nations. Press. Bulterys, M., A. Chao, P. Habimana, A. Dushimimana, Bolduc, Denis, Bernard Fortin, and Marc-Andre Four- P. Nawrocki, and A. Saah. 1994. "Incident HIV-1 nier. 1996. "The Impact of Incentive Policies on the Infection in a Cohort of Young Women in Butare, Practice Location of Doctors: A Multinomial Profit Rwanda." AIDS 8 (11): 1585-91. Analysis." Journal of Labor Economics 14 (October): 703-32. Burr, Chandler. 1997. "The AIDS Exception: Privacy vs. Bongaarts, John. 1996. "Global Trends in AIDS Mortal- Public Health-The Case for Reinstating Traditional ity." Population and Development Review 22 (1): Rules for Fighting Epidemics." Atlantic Monthly 21-45. June). Bongaarts, John, P. Reining, Peter Way, and Francis Buve, S. D. Foster, C. Mbwili, E. Mungo, N. Tollenare, Conant. 1989. "The Relationship between Male Cir- and M. Zeko. 1994. "Mortality among Female cumcision and HIV Infection in African Popula- Nurses in the Face of the AIDS Epidemic: A Pilot tions." AIDS 3 (6):373-77. Study in Zambia." AIDS 8 (3): 396. 333 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Buwembo, Joachim. 1995. "Uganda Government Media Chakraborty, A. K., S. Jana, A. Das, L. Khodakevich, M. Move To Lift Ban on Condom Ads." The East Afri- S. Chakraborty, N. K. Pal. 1994. "Community-based can, March 6-12. Survey of STD/HIV Infection among Commercial Sex Workers in Calcutta (India). Part I. Some Social Caldwell, John C., and Pat Caldwell. 1993. "The Nature Features of Commercial Sex Workers." Journal of and Limits of the Sub-Saharan African AIDS Epi- Communicable Diseases 26 (3): 161-67. demic: Evidence from Geographic and Other Pat- terns." Population and Development Review 19 (4): Chawla, Mukesh. 1993. "Physician Moonlighting in 817-48. Egypt." Ph.D. diss. Boston University, Department of Economics, Boston, Mass. 1996. "The African AIDS Epidemic." Scientific American 274 (3): 62-68. . 1997. "Dual-Job Holdings by Public Sector Physicians in India." Working paper. Data for Caldwell, John C., Pat Caldwell, and I. 0. Orubuloye. Decision Making. Harvard School of Public Health, 1989. "The Social Context of AIDS in Sub-Saharan Department of Population and International Health, Africa." Population and Development Review (2): Boston, Mass. 185-234. Chela, C. M., Roland Msiska, Anna Martin, M. Carael, Michel. 1995. "Sexual Behaviour." Chapter 4 of Sichone, B. Mwiinga, C. B. Yamba, S. Anderson, E. John Cleland and Benoit Ferry, eds., Sexual Behaviour Van Praag. 1994. "Cost and Impact of Home-based andAIDS in the Developing World London: Taylor & Care for People Living with HIV/AIDS in Zambia." Francis. Unpublished report. World Health Organization, Global Programme on AIDS, Geneva. CARE (Cooperative for American Relief to Everywhere) Chen, Shaohua, Gaurav Dart, and Martin Ravallion. and ODA (Overseas Development Administration). 1994. "Is Poverty Increasing in the Developing 1994. "Refugee Inflow into Ngara and Karagwe World?" Review of Income and Wealth series 40 (4, Districts, Kagera Region, Tanzania: Environmental Worl> 359W764 Impact Assessment." Environmental Resources Man- December): 359-76. agement Report. ODA, London. Chequar, Pedro. 1997. "Implications of Anti-retroviral Carr, Jean K., Narongrid Sirisopana, Kalyanee Torugsa, Treatments-Brazil." Paper presented at the Informal Achara Jugsudee, Thippawan Supapongse, Cheodchai Consultation on the Implications of Antiretroviral Chuenchitra, Sorachai Nitayaphan, Pricha Singharaj, Treatments." World Health Organization, Geneva. and John G. McNeil. 1994. "Incidence of HIV-1 P Infection among Young Men in Thailand." Journal of Ch'ien, J. 1994. "Street Surveys on Drug Users' Be- Acquired Immune Deficiency Syndromes 7: 1270-75. haviour and Peer Counselling for Preventive Edu- cation and Crisis Intervention." Abstract PC0473. Carrin, G., J. Perrot, and F. Sergent. 1994. "Influence of Paper presented at the Tenth International Con- Financial Participation by the Population on the De- ference on AIDS, Yokohama. August 14-17. mand for Health: An Analytical Tool for Countries in Greatest Need." WHO, Macroeconomics, Health Chin, James. 1990. "Public Health Surveillance of AIDS and Development series. Geneva. and HIV Infections." Bulletin of the World Health Organiz;ation 68 (5): 529-36. Carswell, J. W., and Lloyd G. Howells. 1989. "Prev- alence of HIV-1 in East African Lorry Drivers." AIDS Choi, Kyung-Hee, and Thomas J. Coates. 1994. "Pre- 3 (11): 759-61. vention of HIV Infection." AIDS 8 (10): 1371-89. 334 SELECTED BIBLIOGRAPHY Choopanya, K., A. Vanichseni, D. C. DesJarlais, K. Connor, E. M., R. S. Sperling, R. Gelker, P. Kiselev, G. Plangsringarm, W. Sonchai, M. Carballo, P. Scott, M. J. O'Sullivan, R. Van Dyke, M. Bey, W. Friedmann, and S. R. Friedman. 1991. "Risk Factors Shearer, and R. L. Jacobson. 1994. "Reduction of and HIV Seropositivity among Injecting Drug Users Maternal-Infant Transmission of Human Immuno- in Bangkok." AIDS5 (12): 1509-13. deficiency Virus Type 1 with Zidovudine Treat- ment." New England Journal of Medicine 331 (18): Cleland, John, and Benoit Ferry, eds. 1995. Sexual Be- 1173-80. haviour and AIDS in the Developing World. London: Taylor & Francis. Cuddington, John T. 1993. "Modeling the Macro- economic Effects of AIDS with an Application to Clemente, M., C. Ferreros, and M. E. L. Fernandes. Tanzania." World Bank Economic Review 7 (2): 1996. "The Brazilian Condom Market: Positive 173-89. Effects of Social Marketing." Abstract No. Tu.D.2887. Paper read at the Eleventh International Conference Curtis, S. L., and Katherine Neitzel. 1996. "Contracep- on AIDS, Vancouver, B.C, Canada. July 7-12. tive Knowledge, Use and Sources." DHS Comparative Studies 19. Calverton, Md.: Macro International. Cohen, Myron S., Gail E. Henderson, Pat Aiello, and Heyi Zheng. 1996. "Successful Eradication of Sex- Dabis, F., P. Msellati, P. Lepage, M. L. Newell, C. ually Transmitted Diseases in the People's Republic Peckham, and P. van de Perre. 1993. "Estimating the of China: Implications for the 21st Century." Journal Rate of Mother-to-Child Transmission of HIV." A of Infectious Diseases 174 (suppl. 2): S223-29. report of the Workshop on Methodological Issues, Ghent, Belgium. February 17-20. AIDS 7(8): Cohen, Myron S., Irving F. Hoffman, Rachel A. Royce, 1139-48. Peter Kazembe, John R. Dyer, Celine Costello Daly, Dick Zimba, Pietro L. Vernazza, Martin Maida, Dallabetta, Gina, Marie Laga, and Peter Lamptey, eds. Susan A. Fiscus, Joseph J. Eron, Jr., and the 1996. Control of Sexually Transmitted Diseases: A AIDSCAP Malawi Research Group. 1997. "Reduc- Handbookfor the Design and Management of Programs. tion of Concentration of HIV-1 in Semen after Arlington, Va.: AIDSCAP/FHI. Treatment of Urethritis: Implications for Prevention of Sexual Transmission of HIV-1." Lancet 349 Dallabetta, G. A., P. G. Miotti, J. D. Chiphangwi, A. J. (9069): 1868-73. Saah, G. Liomba, N. Odaka, F. Sungani, and D. R. Hoover. 1993. "High Socioeconomic Status Is a Risk Coleman, Susan, Mary Lyn Gaffield, Frederic Robin, Factor for Human Immunodeficiency Virus Type 1 Goussou Koudou lazare, Rebecca Dillingham, and (HIV-1) Infection but Not for Sexually Transmitted Carol Squire-Diomande. 1996. "PSI Research Report Diseases in Women in Malawi: Implications for Summary: C6te d'Ivoire Condom KAP Survey, June. HIV-1 Control." Journal of Infectious Diseases 167 (1): 1995." Population Services International, Wash- 36-42. ington, D.C. Dames and Moore. 1996. "Environmental Impact Collins, D., J. Quick, S. Musau, and D. Kraushaar. Assessment for Chad." Report. Washington, D.C. 1996. "Health Financing Reform in Kenya: The Fall and Rise of Cost Sharing, 1989-94." Stubbs Mono- De Cock, Kevin. 1993. "TB and HIV: An Overview." graph Series 1. Boston: Management Sciences for Excerpts from speech at the 1993 International AIDS Health. Conference in Berlin, reproduced in "TB and HIV." SIDALERTE 27, October (suppl.). Conant, Francis. 1995. "Regional HIV Prevalence and Ritual Circumcision in Africa." Health Transition De Cock, Kevin M., and Fran,oise Brun-Vezinet. 1996. Review 5 (1): 108-12. "HIV-2 Infection: Current Knowledge and Uncer- 335 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC tainties." Chapter 12 of Mann and Tarantola, eds., . 1993. "1992 Contraceptive Social Marketing AIDS in the World IT Global Dimensions, Social Roots, Statistics." Washington, D.C., August. Unpublished and Responses. The Global AIDS Policy Coalition. data. New York: Oxford University Press. . 1994. "1993 Contraceptive Social Marketing De Gruttola, Victor, George R. Seage III, Kenneth H. Statistics." Washington, D.C., July. Unpublished Mayer, and C. Robert Horsburgh, Jr. 1989. "Infec- data. tiousness of HIV between Male Homosexual Part- ners." Journal of Clinical Epidemiology 42 (9): 849-56. . 1995. "1994 Contraceptive Social Marketing Statistics." Washington, D.C., July. Unpublished Deininger, Klaus, and Lyn Squire. 1996. "A New Data data. Set Measuring Income Inequality." World Bank Eco- nomic Review 10 (3): 565-9 1. . 1996. "1995 Contraceptive Social Marketing Statistics." Washington, D.C., August. Unpublished Demographic and Health Surveys, final reports, various data. countries and years. ______. 1997. New Breakthroughs in Social Marketing: De Muylder, X., and J. J. Amy. 1993. "Caesarean Sec- DKT International 1996-97 Progress Report. tion Rates in an African Country." Paediatric Peri- Washington, D.C. natal Epidemiology 7 (3): 234-44. ______. n.d. How Social Marketing Changes Lives: The Des Jarlais, D. C., S. Friedman, K. Choopanya, S. Stories Behind the Statistics. Washington, D.C. Vanichseni, and T. Ward. 1992. "International Epi- Dolin, P. J., M. C. Raviglione, and A. Kochi. 1993. demiology of HIV and AIDS among Injecting Drug "Estimates of Future Global Morbidity and Mor- Users." AIDS 6 (10): 1053-68. tality." Morbidity and Mortality Weekly Report 42 (49). Summarized in HIV and TB 2, January- Des Jarlais, D. C., H. Hagan, S. R. Friedmann, P. March 1994. Friedmann, D. Goldberg, M. Frischer, S. Green, K. Tunving, B. Ljungberg, A. Wodak, M. Ross, D. Dunn, D. T., M. L. Newell, A. E. Ades, and C. S. Purchase, M. Millson, and T. Meyers. 1995. Peckham 1992. "Risk of Human Immunodeficiency "Maintaining Low HIV Scroprevalence in Virus Type 1 Transmission through Breastfeeding." Populations of Injecting Drug Users." Journal of the Lancet-340 (8819): 585-88. American Medical Association 274 (15): 1226-3 1. Elias, C., and L. Heise. 1994. "Challenges for the Devel- Deschampes, M., J. W. Pape, A. Hafner, and others. opment of the Female-Controlled Vaginal Microbi- 1996. "Heterosexual Transmission of HIV in Haiti. cides." AIDS 8: 1-9. Annals ofInternalMedicine 125: 324-30. Ellis, Randall, Moneer Alam, and Indrani Gupta. 1997. de Vincenzi, Isabelle, and T. Mertens. 1994. "Male "Health Insurance in India: Prognosis and Prospec- Circumcision: A Role in HIV prevention?" AIDS 8 tus." Boston University, Boston, Mass. (2): 153-60. Esparza, Jose, William L. Heyward, and Saladin DKT International. 1992. "1991 Contraceptive Social Osmanov. 1996. "HIV Vaccine Development: From Marketing Statistics." Washington, D.C., June. Basic Research to Human Trials." AIDS 10 (suppl. Unpublished data. A): S 123-32. 336 SELECTED BIBLIOGRAPHY Ettiegne-Traore, V., P. D. Ghys, M. 0. Diallo, and Ferreira, Luisa, and L. Goodhart. 1995. "Incomes, In- others. 1996. "HIV Seroincidence and STD Prev- equality and Poverty Alleviation in Tanzania, 1993." alence during an Intervention Study among Female World Bank, Africa Technical Family, Human Devel- Sex Workers in Abidjan, C6te d'Ivoire: Preliminary opment Division, Washington, D.C. Findings." Abstract Mo.C. 442. Eleventh Inter- national Conference on AIDS, Vancouver, B.C., Ferry, Benoit. 1995. "Risk Factors Related to HIV Canada. July 7-12. Transmission: Sexually Transmitted Diseases, Alcohol Consumption and Medically-Related Injections." In European Commission. 1995a. Safe Blood in Developing John Cleland and Benoit Ferry, eds., Sexual Behaviour Countries: The Lessons from Uganda. Edited by Rex and AIDS in the Developing World. London: Taylor & Winsbury. Development Studies and Research. ancis. Luxembourg: Office for Official Publications of the Francis. European Communities. Fitzsimmons, David. 1996. "International AIDS Vaccine _. 1995b."Safe Blood in Developing Countries: Initiative Launched. " International AIDS Vaccine Principles and Organisation." Development Studies Initiative Report 1 (summer): 7. and Research. Written and edited by C. Gerard, D aondReseag-Thu,. Written Watson-Wedil. and D. Floyd, Katherine, and Charles F. Gilks. 1996. "Impact Sondag-Thull, E. J. Watson-Wlloiams, PubLi of, and Response ro, the HIV Epidemic at Kenyara tions ofLthemEuropean Communities. National Hospital, Nairobi." Report 1 (April). Liverpool School of Tropical Medicine. . 1997. Considering HIV/AIDS in Development Assistance: A Toolkit. Produced for the EC under con- . 1997. "Cost and Financing Aspects of Providing tract B7.5046/94/06: Study of the Impact on the Anti-retroviral Therapy: A Background Paper." Paper Social and Economic Development of Developing presented at the Informal Consultation on the Impli- Countries. Brussels. cations of Antiretroviral Treatments, World Health Organization, Geneva. April 29-30. European Study Group on Heterosexual Transmission of HIV. 1992. "Comparison of Female to Male and Ford, K., D. N. Wirawan, P. Fajans, P. Meliawan, K. Male to Female Transmission of HIV in 563 Stable MacDonald, and L. Thorpe. 1996. "Behavioral Couples." British MedicalJournal304: 809-13. Interventions for Reduction of Sexually Transmitted Disease/HIV Transmission among Female Com- Expert Group of the Joint United Nations Programme mercial Sex Workers and Clients in Bali, Indonesia." on HIV/AIDS. 1997. "Implications of HIV Vari- AIDS 10: 213-22. ability for Transmission: Scientific and Policy Issues." AIDS 11 (4):1-15. Foster, Susan, Peter Godrey-Faussett, and John Porter. Farwell, Byron. 1989. Armies of the R.aj From Mutiny to 1997. "Modelling the Economic Benefits of Tuber- Independence, 1858-1947. New York: Norton. culosis Preventive Therapy for People with HIV: The Example of Zambia." AIDS 11: 919-25. Feachem, Richard. 1995. Valuing the Past, Investing in the Future. Evaluation of the National HIV/AIDS Fox L. J., P. E. Bailey, K. L. Clarke-Martinez, and oth- Strategy, 1993-94 to 1995-96 AIDS/Communicable ers. 1988. "Prevention of Transmission of HIV in Diseases Branch, Commonwealth Department of Africa: Effectiveness of Condom Promotion and Human Services and Health, Australia. Canberra: Health Education among Prostitutes." Lancet 15: Australian Government Publishing Service. 887-90. 337 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Futures Group International. 1995a. "SOMARC Peru's Services as a Preventive Intervention against HIV in PIEL Condoms Soar to the Top of the Market." Mwanza Region, Tanzania." Abstract Mo.C. 444. SOMARC [Social Marketing for Change] Highlights 3 Paper read at the Eleventh International Conference July): 1-2. on AIDS, Vancouver, B.C., Canada. July 7-12. Johns Hopkins School of Public Health. 1995b. "Working with Religious Opposition in Africa." SOMARC [Social Marketing for Change] Giraud, Patrick. 1992. "Economic Impact of HIV/AIDS Highlights 3 Uuly): 1-2. on the Transport Sector: Development of an Assess- ment Methodology." Consultation on Economic Garbus, L. 1996. "The UN Response." In Jonathan Implications of HIV/AIDS, U.N. Development Pro- Mann and Daniel Tarantola, eds., AIDS in the World gramme, Bangalore, India. May 25-28. II: Global Dimensions, Social Roots, and Responses. The Global AIDS Policy Coalition. New York: Oxford Gluck, Michael, and Eric Rosenthal. 1995. The Effective- University Press. ness of AIDS Prevention Efforts. Washington, D.C.: Office of Technology Assessment. Garnett, Geoff P., and Roy M. Anderson. 1995. "Strate- gies for Limiting the Spread of HIV in Developing Gold, David. 1996. "Progresss in AIDS Vaccine Devel- Countries: Conclusions Based on Studies of the Trans- opment." International AIDS Vaccine Initiative Report mission Dynamics of the Virus." Journal of Acquired 1 (summer): 4-5. Immune Deficiency Syndromes and Human Retrovirol- ogy 9 (5): 500-13. Golz, J. 1993. "Current Experiences with Methadone in the Treatment of Injection Drug Users with HIV In- Garrett, Laurie. 1994. The Coming Plague: Newly Emerg- fection in Germany." Abstract WS-D12-6. Paper pre- ing Diseases in a World Out of Balance. New York: sented at the Ninth International Conference on Penguin. AIDS, Berlin. June 6-11. Gerard, Christine, Danide Sondag-Thull, Edward-John Graham, Ron. 1990. "One African's Tragedy Focuses Watson-Williams, and Lieve Fransen. 1995. Safe Attention on AIDS." New York Times, April 1, sec- Blood in Developing Countries: Principles and Organi- tion 2, p. 35. sation. Brussels: Office for Official Publications of the European Communities. Greenberg, Jerome. 1972. "Venereal Disease in the Armed Forces. Medzical Clinics of North America 56 Gertler, Paul, and Jacques van der Gaag. 1990. The (5): 1087-1100. Willingness To Pay for Medical Care: Evidence from Two Developing Countries. Baltimore: Johns Hopkins Grosskurth, H., F. Mosha., J. Todd., E. Mwijarubi, A. University Press. Klokke, K. Senkoro, P. Mayaud, J. Changalucha, A. Nicoll, G. Ka-Gina, J. Newell, K. Mugeye, D. Gertler, Paul, and Jeffrey Hammer. 1997. "Strategies for Mabey, and R. Hayes. 1995a. 'Impact of Improved Pricing Publicly Provided Health Care Services." Treatment of Sexually Transmitted Diseases on HIV World Bank, Policy Research Department, Wash- Infection in Rural Tanzania: Randomised Controlled ington, D.C. Trial." Lancet 346 (8974): 530-36. Gilson, L., R. Mkanje, H. Grosskurth, J. Picard, P. Grosskurth, H., F. Mosha, J. Todd, K. Senkoro, J. Mayaud, J. Todd, D. Mabey, and R. Hayes. 1996. Newell, A. Klokke, J. Changalucha, B. West, P. "Cost-Effectiveness of Improved STD Treatment Mayaud, A. Gavyole, R. Gabone, D. Mabey, and R 338 SELECTED BIBLIOGRAPHY Hayes. 1995b. "A Community Trial of the Impact of Notes in Biomathematics 56. New York: Springer- Improved Sexually Transmitted Disease: The HIV Verlag. Epidemic in Rural Tanzania-Baseline Survey Re- sults." AIDS 9 (8): 927-34. Hien, N. T. 1995. "Drug Use and HIV Infection in Viet Nam." Report of the WHO Drug Injecting Project Gupta, Indrani, Phane Mujinja, and Mead Over. Planning Meeting, Phase II. Bangkok, Thailand. Forthcoming. "Distribution of Poverty and Mortality in the Presence of AIDS Epidemic in Kagera, Tan- "Ho Chi Minh City Worries about Its Shooting Booths." zania." Working paper. World Bank, Policy Research 1996. AIDSAnalysis Asia 2 (5): 1. Department, Washington, D.C. Hammer, Jeffrey. 1997. "Economic Analysis for Health Holtgrave, David R., Noreen L. Qualls, James W. Cur- Projects." World Bank Research Observer 12 (1): ran, Ronald 0. Valdiserri, Maty E. Guinan, and Wil- Projects." Wol akRsac Osre 2() liam C. Parra. 1995. "An Overview of the Effec- 47-7 1. tiveness and Efficiency of HIV Prevention Programs." Hampton, Janie. 1991. Meeting AIDS with Compassion: Public Health Reports 110 (2): 134-45. AIDS Care and Prevention in Agomanya, Ghana. Strat- egies for Hope 4. London: ActionAid, in assoc. with Holtgrave, David R., Noreen L. Qualls, and John D. AMREF and World in Need. Graham. 1996. "Economic Evaluation of HIV Pre- vention Programs." Annual Review of Public Health Hassig, Susan E., Joseph Perriens, Ekungola Baende, 17: 467-88. Mbindule Kahotwa, K. Bishagara, and N. Kinkela. 1990. "An Analysis of the Economic Impact of HIV Hook, E. W. III, R 0. Cannon, A. J. Nahmias, F. F. Infection among Patients at Mama Yemo Hospital, Lee, C. H. Campbell, Jr., D. Glasser, and T. C. Kinshasa, Zaire." AIDS 4: 883-87. Quinn. 1992. "Herpes Simplex Virus Infection as a Risk Factor for Human Immunodeficiency Virus In- Hatty, Suzanne E. 1993. "Australia." In Nanette J. fection in Heterosexuals."Journal of Infectious Diseases Davis, ed., Prostitution: An International Handbook 165 (2): 251-55. on Trends, Problems, and Policies. Westport, Conn.: Greenwood Press. Human Rights Watch/Asia. 1995. Rape for Profit: Traf- ficking of Nepali Girls and Women to India ' Brothels. Haverkos, Harry W., and Robert J. Battjes. 1992. New York: Human Rights Watch. "Female-to-Male Transmission of HIV." Journal of the American MedicalAssociation 268 (14): 1855-56. Hunter, Susan, and John Williamson. Forthcoming. ershB.S,F.opovici. C. pereiL. Zootusca,N. "Developing Strategies and Policies for USAID Inter- Hersh, B. S., F. Popovici, R. C. Apetrei, L. Zolotusca, N. vnin o upr f I/ISIfce n Beldescu, A. Calomfirescu, Z. Jezek, M. J. Oxtoby, A. ventions for Support of HIV/AIDS Infected and Gromyko, and D. L. Heymann. 1991. "Acquired Im- Affected Children." U.S. Agency for International munodeficiency Syndrome in Romania." Lancet 338 D W (8768): 645-49. Hurley, Susan F., Damien J. Jolley, and John M. Kaldor. Hethcote, Herbert W. 1976. "Qualitative Analyses of 1997. "Effectiveness of Needle-Exchange Programmes Communicable Disease Models." Mathematical Bio- for Prevention of HIV Infection." Lancet 349 (9068): sciences 28: 335-36. 1797-1800. Hethcote, Herbert W., and James A. Yorke. 1984. Ingham, Roger. 1995. "AIDS: Knowledge, Awareness Gonorrhea Transmission Dynamics and ControL Lecture and Attitudes." In John Cleland and Benoit Ferry, 339 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC eds., Sexual Behaviour and AIDS in the Developing Jalal, Fasli, Hadi M. Abednego, Tonny Sadjimin, and World. London: Taylor & Francis. Michael J. Linnan. 1994. "HIV and AIDS in Indo- nesia." AIDS 8 (suppl. 2): S91-94. Institute of Medicine. 1985. New Vaccine Development: Establishing Priorities. Vol. 1. Washington, D.C.: James, Estelle. 1982. "The Nonprofit Sector in Inter- National Press Academy. national Perspective: The Case of Sri Lanka." Journal of Economic LiteratureO053 (123): 99-129. IAVI (International AIDS Vaccine Initiative). 1996. "A View from Uganda: An Interview with Dr. Edward Janz, Nancy K., Marc A. Zimmerman, Patricia A. Wren, Mbidde." InternationalAIDS Vaccine Initiative Report Barbara A. Israel, Nicholas Freudenberg, and Rosa- I (summer): 3. lind J. Carter. 1996. "Evaluation of 37 AIDS Preven- tion Projects: Successful Approaches and Barriers to IAVR (International AIDS Vaccine Research). 1997a. Program Effectiveness." Health Education Quarterly "IAVI Launches Scientific Program." International 23 (1): 80-97. AIDS Vaccine Research Report 2 (1): 1-3. 1997b. "Studying AIDS Vaccines in Thailand: Johnston, Margaret. 1996. "Why an HIV Vaccine is An Interview with Natth Bhamarapravati." Interna- Scientifically Possible." International AIDS Vaccine tionalAIDS Vaccine Research Report 2 (1): 4-5. Initiative Report (summer): 1. Ismail, Rokiah. 1996. "Changing Pattern of HIV Infec- Jones, Clair. 1997. "Current Costs of HIV/AIDS- tion as Seen at the University of Malaya Medical Malawi: A Case Study." Business Exchange on AIDS Center." Eleventh International Conference on AIDS. and Development [BEAD], meeting notes. London, Abstract Mo.C. 1502. Vancouver, B.C., Canada. July March 14. 7-12. University of Malaysia, Department of Medi- cine, Kuala Lumpur Jones, Gavin W., Endang Sulistyaningsih, a-nd Terence H. Hull. 1994. "Prostitution in Indonesia." ANU Izazola-Licea, Jose A. 1996. "AIDS: The State of the Art: Working Paper 52. Department of Demog- A Review Based on the Eleventh International Confer- raphy, Research School of Social Sciences, Canberra, ence on AIDS, Vancouver, B.C., Canada. July 7-12. Australia. Fundacion Mexicana Para La Salud, A.C., Mexico. Kahn, James G. 1996. "The Cost-Effectiveness of HIV Jackson, D., J. Rakwar, B. Richardson, and others. 1997. Prevention Targeting: How Much More Bang for the "Decreased Incidence of Sexually Transmitted Dis- Buck?" American Journal of Public Health 86 (12): eases among Trucking Company Workers in Kenya: 1709-12. Results of a Behaviour Risk-Reduction Programme." AIDS 11: 903-909. Kamali, A., J. F. Kengeya-Kayondo, S. S. Malamba, A. J. Nunn, J. A. Seeley, H. U. Wagner, and D. W. Mul- Jacquez, J. A., J. S. Koopman, Carl P. Simon, and Ira M. der. 1992. "Orphans and HIV-1 Infection in a Rural Longini, Jr. 1994. "Role of the Primary Infection in Population Cohort in S.W. Uganda." Poster presented Epidemics of HIV Infection in Gay Cohorts." AIDS at the Eighth International Conference on AIDS, Am- 7 (11): 1169-84. sterdam, the Netherlands. July 19-24. PosterAbstracts, vol. 2, PoD5159. Jain, Manoj K., T. Jacob John, and Gerald T. Keusch. 1994. "Epidemiology of HIV and AIDS in India." Kambou, G., S. Devarajan, and M. Over. 1992. "The AIDS 8 (suppl. 2): S61-75. Economic Impact of AIDS in an African Country: 340 SELECTED BIBLIOGRAPHY Simulations with a General Equilibrium Model of "Physicians' Experience with the Acquired Immuno- Cameroon." Journal of African Economies 1 (1): deficiency Syndrome as a Factor in Patients' Survi- 109-30. val." New England Journal of Medicine 334 (11): 701-706. Kamenga, M., R W. Ryder, M. Jingu, and others. 1991. "Evidence of Marked Sexual Behavior Change Asso- Kogan, Rick. 1990. "Final Crusade." Chicago Tribune, ciated with Low HIV-1 Seroconversion in 149 Mar- April 3, section 5, p. 7. ried Couples with Discordant HIV Serostatus: Experi- ence at an HIV Counseling Center." AIDS 5: 61-67. Konde-Lule, Joseph K., Maria J. Wawer, F. Nalugoda, Ronald H. Gray, and Rekha Menon. 1997. "HIV Kang, Lai-Yi. 1995. "Economic Growth and Urbanisa- Infection in Rural Households, Rakai District, tion in China Leads to an Increase in STDs." AIDS Uganda." Preliminary draft presented at the IUSSP Analysis Asia 1 (5): 14-15. [International Union for the Scientific Study of Pop- ulation] Conference on the Socio-Demographic Kaplan, J. E., K. K. Holmes, H. W. Jaffe, H. Masur, and Impact of AIDS in Africa, Durban, South Africa. K M. De Cock. 1996. "Preventing Opportunistic In- Ferary 3-6. fections in Human Immunodeficiency Virus-infected February 3-6. Persons: Implications for the Developing World." Krueger, L. E., R. W. Wood, P. H. Diehr, and C. L. American Journal of Tropical Medicine and Hygiene Maxwell. 1990. "Poverty and HIV: Seropositivity: 55: 1-11. The Poor Are More Likely To Be Infected." AIDS 4 Killewo, J., L. Dahlgren, and A. Sandstrom. 1994. (8): 811-14. "Socio-geographical Patterns of HIV-1 Transmission Kunanusont, Chaiyos. 1997. "National Anti-retroviral in Kagera Region, Tanzania." Social Science and Medi- Program: Thailand Experience." Paper presented at cine38 (1): 129-34.Prga:TalnExeine"aprrsntdt the Informal Consultation on Implications of Anti- Killewo, J., K. Nyamuryekunge, A. Sandstron, U. retroviral Treatments, World Health Organization, Bredberg-Raden, S. Wall, F. Mhalu, and G. Biber- Geneva. April 29-30. feld. 1990. "Prevalence of HIV Infection in Kagera Region of Tanzania: Population-based Study." AIDS Laga, M., M. Alary, N. Nzila, A. T. Manoka, M. Tuliza, 4 (November): 1081-85 F. Behets, J. Goeman, M. St. Louis, and P. Piot. 1994. "Condom Promotion, Sexually Transmitted Diseases Kin, F. 1995. "Injecting Drug Use among Heroin Users Treatment and Declining Incidence of HIV-1 in Malaysia: Summary of Research Findings." Report Infection in Female Zairian Sex Workers." Lancet 344 of the WHO Drug Injecting Project Planning Meet- (8917): 246-48. ing, Phase II. Bangkok, Thailand. Laga, M., A. Manoka, M. Kivuvu, B. Malele, M. Tuliza, Kirby, Douglas, Lynn Short, Janet Collins, Deborah N. Nzila, J. Groeman, F. Behets, V. Batter, and M. Rugg, Lloyd Kolbe, Marion Howard, Brent Miller, Alary. 1993. "Non-ulcerative Sexually Transmitted Freya Sonenstein, and Laurie S. Zabin. 1994. "School- Diseases as Risk Factors for HIV-1 Transmission in Based Programs To Reduce Sexual Risk Behaviors: A Women: Results from a Cohort Study." AIDS 7 (1): Review of Effectiveness." Public Health Reports 109 95-102. (3): 339. Lamptey, Peter, and Peter Piot, eds. 1990. The Hand- Kitahara, M. M., T. D. Koepsell, R. A. Deyo, C. L. book for AIDS Prevention in Africa. Durham, N.C.: Maxwell, W. T. Doge, and E. H. Wagner. 1996. Family Health International. 34I CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Laumann, Edward O., Christopher M. Masi, and Ezra and Brazil." InternationalJournal of STD and AIDS 4 W. Zuckerman. 1997. "Circumcision in the United (3): 135-41. States: Prevalence, Prophylactic Effects, and Sexual Practice." Journal of the American Medical Association Linden, Eugene. 1996. "The Exploding Cities of the 277 (13): 1052-57. Developing World." Foreign Affairs 75 (1): 52-65. Lavy, Victor, and John M. Quigley. 1993. "Willingness Lockheed, Marlaine E., Adriaan M. Verspoor, Deborah To Pay for the Quality and Intensity of Medical Care: Bloch, Pierre Englebert, Bruce Fuller, Elizabeth King, Low-Income Households in Ghana." LSMS 94. John Middleton, Vicente Paqueo, Alastair Rodd, World Bank, Policy Research Department, Poverty Ralph Romain, and Michel Welmond. 1991. Improv- and Human Resources Division, Washington, D.C. ing Primary Education in Developing Countries. New York: Oxford University Press. Laws, Margaret. 1996. "International Funding of Global AIDS Strategy: Official Development Assistance." In Low-Beer, Daniel, and Seth Berkley. 1996. "HIV." In Jonathan Mann and Daniel Tarantola, eds., AIDS in Christopher J. L. Murray and Allan D. Lopez, eds., the World II: Global Dimensions, Social Roots, and Health Dimensions of Sex and Re > Heath Dmenszns o Sex nd Rproduction: The Responses. The Global Policy Coalition. New York: Global Burden of Sexually Transmitted Diseases, HIV, Oxford University Press. Maternal Conditions, Prenatal Disorders and Congeni- LziAANicolosi. talAnomalies. Cambridge, Mass.: Harvard University Lazzarin, A., A. Saracco, M. Musicco, and A. Press.l. 1991. "Man-to-Woman Sexual Transmission of the ress. Human Immunodeficiency Virus." Archives of Inter- Lowenthal, Nancy, Angela Chimwaza, Saifuddin Ahmed, nal Medicine 151 (12): 241 1-16. and Judith Timyan. 1995. "Malawi Contraceptive Lee, S. S, W. W. Lim, and S. H. Lee. 1993. "Epidemiol- Consumer Knowledge, Attitudes, and Practices Sur- ogy of HIV Infection in Hong Kong: Analysis of the vey, June-July, 1995." Population Services Inter- First 300 Cases." Abstract PO-C08-2780. Paper pre- national, Washington, D.C. sented at the Ninth International Conference on AIDS, Berlin. June 6-11. Lurie, Peter, and Ernest Drucker. 1997. "An Oppor- tunity Lost: HIV Infections Associated with Lack of a Leiner, Marvin. 1994. "AIDS: Cuba's Effort to Con- National Needle-Exchange Program in the U.S.A." tain." In Sexual Politics in Cuba: Machismo, Homo- Lancet 349 (9052): 604-608. sexuality and AIDS. Boulder, Colo.: Westview Press. Lurie, Peter, M. Fernandes, V. Hughes, E. Arevalo, E. Levine, W. C., G. Higueras, R. Revollo, and others. Hudes, A. Reingold, N. Hearst, and the Instituto 1996. "Rapid Decline in Sexually Transmitted Dis- Adolfo Lutz Study Group. 1995. "Socioeconomic ease Prevalence among Brothel-based Sex Workers in Status and Risk of HIV-1, Syphilis and Hepatitis B La Paz, Bolivia: The Experience of Proyecto Contra Infection among Sex Workers in Sao Paulo State, SIDA, 1992-1995." Abstract Mo.C. 441. Eleventh Brazil." AIDS9 (suppl. 1): S31-37. International Conference on AIDS, Vancouver, B.C., Canada. July 7-12. Lurie, Peter, A. L. Reingold, B. Bowser, D. Chen, J. Foley, J. Guydish, J. G. Kahn, S. Lane, and J. Libonatti, O., E. Lima, A. Peruga, R. Gonzalez, F. Sorensen. 1993. The Public Health Impact of Needle Zacarias, and M. Weissenbacher. 1993. "Role of Exchange Programs in the United States and Abroad. Drug Injection in the Spread of HIV in Argentina Vol. 1. San Francisco, Calif.: University of California. 342 SELECTED BIBLIOGRAPHY Lwihula, George. 1994. "Variability in Saving and Mastro, Timothy D., and Isabelle de Vincenzi. 1996. Assistance Behaviour in Households across Ethnic "Probabilities of Sexual HIV-1 Transmission." AIDS Groups in Kagera Region, Tanzania: A Focus Group 10 (suppl. A): S72-82. Study." University of Dar es Salaam, Muhimbili University College of Health Sciences, Institute of Mastro, Timothy D., Glen A. Satten, Taweesak Nop- Public Health, Department of Behavioural Sciences. kesorn, Suebpong Sangkharomya, and Ira M. Lon- gini, Jr. 1994. "Probability of Female-to-Male Trans- Maharjan, S., J. A. Peak, S. Rana, and N. Crofts. 1994. mission of HIV-1 in Thailand" Lancet 343 (8891): "Declining Risk for HIV among IDUs in Kath- 204-07. mandu: Impact of a Harm Reduction Program." Ab- stract 561C. Paper presented at the Tenth Interna- Mauskopf, J. A., J. E. Paul, D. S. Wichman, A. D. tional Conference on AIDS, Yokohama, Japan. White, and H. H. Tilson. 1996. "Economic Impact August 14-17. Lifesaving and Lifegiving Society, of Treatment of HIV-Positive Pregnant Women and Kathmandu, Nepal. Their Newborns with Zidovudine: Implications for HIV Screening." Journal of the American MedicalAsso- Mann, Jonathan, and Daniel Tarantola, eds. 1996. AIDS ciation 276 (2): 132. in the World II Global Dimensions, Social Roots, and Responses. Global AIDS Policy Coalition. New York: May, M. M., and R. M. Anderson. 1987. "Transmission Oxford University Press. Dynamics of HIV Infection." Nature 326: 137-42. Mann, Jonathan, D. Tarantola, and T. W. Netter. 1992. Mazzullo, J., S. Wroblewski, R. Rudd, and P. Fairchild. AIDS in the World. The Global AIDS Policy Coali- 1990. "Influencing Choices of Medical Students to tion. Cambridge, Mass.: Harvard University Press. HIV-related careers." Abstract no. S.D.910. Sixth International AIDS Conference, San Francisco. June Mann, Jonathan, H. Francis, T. C. Quinn, K. Bila, P. K. 20-24. Asila, N. Bosenge, N. Nzilambi, L. Jansegers, P. Piot, K. Ruti, and J. W. Curran. 1986. "HIV Seropreva- McBrier, Page. 1995. "Children's-Health: Giving AIDS lence among Hospital Workers in Kinshasa, Zaire: a Human Face." Inter Press Service, Unicef Feature Lack of Association with Occupational Exposure." 00153.UGA, September. Journal of the American Medical Association 256 (22): 3099-102. McCoy, Clyde B., Shenghan Lai, Lisa R. Metsch, Xue- ren Wang, Cong Li, Ming Yang, and Li Yulong. Mansergh, G., A. Haddix, R. Steketee, P. Nieburg, D. 1997. "No Pain No Gain: Establishing the Kunming, Hu, R. J. Simonds, and M. Rogers. 1996. "Cost- China, Drug Rehabilitation Center." Journal of Drug Effectiveness of Short-Course Zidovudine To Prevent Issues 27 (1): 73-8 5. Perinatal HIV Type 1 Infection in Sub-Saharan Afri- can Developing Country Setting." Journal of the Metzger, David. 1997. "Drug Abuse Treatment as AIDS American MedicalAssociation 276 (2): 139-45. Prevention." In NIH Consensus Development Con- ference, Interventions to Prevent HIVRisk Behaviors: Martin, Anna L., Eric Van Praag, and Roland Msiska. Program and Abstracts. Bethesda, Md.: National Insti- 1996. "An African Model of Home-based Care: Zam- tutes of Health. bia." In Jonathan Mann and Daniel Tarantola, eds. AIDS in the World II Global Dimensions, Social Roots, Miller, Norman, and Rodger Yeager. 1995. "By Virtue and Responses. The Global AIDS Policy Coalition. of Their Occupation, Soldiers and Sailors Are at New York. Oxford University Press. Greater Risk." AIDSAnalysis Africa 5 (6): 8-9. 343 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Mills, Anne, Jonathan Broomberg, John Lavis, and Neil Epidemic in Africa. Health Transition Review 5 (1): Soderlund. 1993. "The Costs of HIV/AIDS Preven- 100-103. tion Strategies in Developing Countries." World Health Organization, Global Programme on AIDS, Moses, Stephen, F. A. Plummer, E. N. Ngugi, N. J. GPAIDIR/93.2, Geneva. Nagelkerke, Ao Anzala, and Jo Ndinya-Achola. 1991. "Controlling HIV in Africa: Effectiveness and Cost of Minon, Jeffrey A., and Jeffrey Zwiebel. 1995. "The Eco- an Intervention in a High Frequency STD Trans- nomic Case against Drug Prohibition." Journal ofEco- mitter Core Group." AIDS 5 (4): 407-11. nomic Perspectives 9 (2): 175-92. Msamanga, G., E. Urassa, D. Spiegelman, E. Hertzmark, Moffitt, Robert. 1991. "The Use of Selection Modeling S. Kapiga, D. J. Hunter, and W. W. Fawzi. 1996. To Evaluate AIDS Interventions with Observational "Socioeconomic Status and Prevalence of HIV Infec- Data." In National Research Council, Evaluating HIV tion among Pregnant Women in Dar Es Salaam, AIDS Prevention Programs. Washington, D.C.: Na- Tanzania." Abstract TuC2464. Eleventh Interna- tional Academy Press. tional Conference on AIDS, Vancouver, B.C., Can- ada. July 7-12. Muhimbili University, College of Moore, M., E. Tukwasiibwe, E. Marum, and others. Health Sciences, Dar-es-Salaam, Tanzania. 1993. "Impact of HIV Counseling and Testing in Uganda." Eleventh International Conference on Mulder, Daan. 1996. "Disease Progression and Mortality AIDS. Abstract WS-C16-4. Vancouver, B.C., Can- Following HIV-1 Infection." In Jonathan Mann and ada. July 7-12. Daniel Tarantola, eds., AIDS in the World II Global Dimensions, Social Roots, and Responses. The Global Moore, Richard D., and John G. Bartlett. 1996. AIDS Policy Coalition. NewYork: Oxford University "Combination Antiretroviral Therapy in HIV Infec- Press. tion: An Economic Perspective." PharmacoEconomics 10 (2): 109-13. Mulder, Daan, Andrew Nunn, Anatoli Kamali, and Jane Kengeya-Kayondo. 1995. "Decreasing HIV-1 Sero- Morris, Martina, Chai Podhisita, Maria J. Wawer, and prevalence in Young Adults in a Rural Uganda Co- Mark S. Handcock. 1996. "Bridge Populations in the hort." British Medicaljournal 311 (30): 833-36. Spread of HIV/AIDS." AIDS 10: 1265-71. Murray, Christopher J. L., and Allan D. Lopez. 1996. Morrow, Richard, Robert Colebunders, and James Chin. "The Global Burden of Disease." Global Burden of 1989. "Interactions of HIV infection with Endemic Disease and Injury Series, vol. 1. WHO, Harvard Tropical Diseases." AIDS 3 (suppl.): S79-87. School of Public Health, World Bank. Cambridge, Mass.: Harvard University Press. Moses, Stephen, Janet Bradley, Nico Nagelkerke, Allan Ronald, J. 0. Ndinya-Achola, and Francis A. Plum- Mwabu, Germano M., Martha Ainsworth, and Andrew mer. 1990. "Geographical Patterns of Male Circum- Nyamete. 1993. "Quality of Medical Care and cision Practices in Africa: Association with HIV Sero- Choice of Medical Treatment in Kenya: An Empirical prevalence." International Journal of Epidemiology 19 Analysis." Journal of Human Resources 28 (4): 838-62. (3): 693-97. Mwizarubi, B. K, C. L. Mwaijonga, and 0. Laukamm- Moses, Stephen, F. A. Plummer., J. E. Bradley., J. 0. Josten. 1992. "HIV/AIDS Education and Condom Ndinya-Achola., N. J. D. Nagelkerke, and A. R. Promotion for Truck Drivers, Their Assistants, and Ronald. 1995. "Male Circumcision and the AIDS Sex Partners in Tanzania." Paper presented at the 344 SELECTED BIBLIOGRAPHY Global Programme on AIDS, Effective Approaches to Nold, A. 1978. "The Infectee Number at Equilibrium AIDS Prevention, Geneva. May 26-29. African for a Communicable Disease." Mathematical Bio- Medical and Research Foundation, Dar-es-Salaam, sciences46: 131-38. Tanzania. Normand, J., D. Vlahov, and L. E. Moses, eds. 1995. National Research Council. 1989. AIDS: Sexual Behavior Preventing HIV Transmission: The Role of Sterile and Intravenous Drug Use. Washington, D.C.: Na- Needles and Bleach. A Report of the Panel on Needle tional Academy Press. Exchange and Bleach Distribution Programs, Com- mission on Behavioral and Social Sciences and Edu- 1991. Evaluating AIDS Prevention Programs. cation. National Research Council and Institute of Edited by Susan L. Coyle, Robert F. Boruch, and Medicine. Washington D.C.: National Academy Charles F. Turner. Washington, D.C.: National Press. Academy Press. Oakley, A., D. Fullerton, and J. Holland. 1995. "Be- 1996. Preventing and Mitigating AIDS in Sub- havioural Interventions for HIV/AIDS Prevention." Saharan Africa. Washington, D.C.: National Aca- AIDS 9 (5): 479-86. demy Press. OECD (Organization for Economic Co-operation and Nelson, Kenrad, David Celentano, Sakol Eiumtrakol, D. Development). 1995. "Trends in International Mi- R. Hoover, C. Beyrer, S. Suprasert, S. Kuntolbutra, gration: Continuous Reporting System on Migra- and C. Khamboonruang. 1996. "Changes in Sexual tion." Annual Report 1994. Paris. Behavior and Decline in HIV Infection among Young Men in Thailand." New EnglandJournal of Medicine Ong, Jin Hui. 1993. "Singapore." In Nanette E. Davis, 335 (5): 297-303. ed., Prostitution: An International Handbook on Trends, Problems and Policies. Westport, Conn.: Greenwood "NGOs Flout AIDS Control Policy." 1994. The Tele- Press. graph (India), Metropolitan section, September 8. ONUSIDA (UNAIDS). 1997. Diagnostico Epidemia Ngugi, E. N., F. A. Plummer, J. N. Simonson, D. W. VIHISIDA Internacionaly Nacional. Santo Domingo, Cameron, M. Bosire, P. Waiyaki, A. R. Ronald, J. 0. Dominican Republic. Ndinya-Achola. 1988. "Prevention of Transmission of Human Immunodeficiency Virus in Africa: Effec- Oppenheimer, Edna. 1995."Drugs and HIV/AIDS in tiveness of Condom Promotion and Health Educa- Myanmar: Land of the 'First Inhabitants of the tion among Prostitutes." Lancet2 (8616): 887-90. World."'AIDS Analysis Asia 1 (4): 5-6. Nicoll, Angus, D. Bennett, M. Catchpole, B. Evans, Osmanov, Saladin. 1996. "Implications of the Human O.N. Gill, J. Mortimer, P. Mortimer, and K. Paine. Immunodeficiency Virus Variability for Transmis- 1996. "HIV, AIDS, and Sexually Transmitted Infec- sion: Scientific and Policy Issues." Expert Group of tions: Global Epidemiology, Impact, and Preven- the Joint United Nations Programme on HIV/AIDS. tion." Health and Population Occasional Paper. UNAIDS, Geneva. Overseas Development Administration, London. Ou, C. Y., Y. Takebe, B. G. Weniger, C. C. Luo, M. L. Nolan, Peter. 1993. "Economic Reform, Poverty and Kalish, W. Auwanit, S. Yamazaki, H. D. Gayle, N. L. Migration in China." Economic and Political Weekly Young, and G. Schochetman. 1993. "Independent 28 (26): 1369-77. Introduction of Two Major HIV-1 Genotypes into 345 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Distinct High-Risk Populations in Thailand. "Lancet Padian, N. S., S. C. Shiboski, and N. P. Jewell. 1991. 341 (8854):1171-74; errata corrected in 342 (8865): "Female-to-Male Transmission of Human Immuno- 250. deficiency Virus." Journal of the American Medical Association 266 (12):1664-67. Over, Mead. 1992. "The Macroeconomic Impact of AIDS in Sub-Saharan Africa." AFTPN Technical Padian, N. S., T. R. O'Brien, Y. Chang, S. Glass, and D. Working Paper 3. World Bank, Africa Technical P. Francis. 1993. "Prevention of Heterosexual Trans- Department, Population, Health, and Nutrition mission of Human Immunodeficiency Virus through Division, Washington, D.C. Couple Counseling." AIDS 6 (9): 1043-48. . 1997. "The Public Interest in a Private Disease: Pal, S. C., S. Sarkar, T. N. B. Naik, P. K. Singh, S. I. Why Should the Government Play a Role in STD Tuchi, A. S. Lal, and S. P. Tripathy. 1990. "Explosive Control?" In K. K. Holmes, P. F. Sparling, P-A Epidemic of HIV Infection in North Eastern Stares of Mardh, S. M. Lemon, W. E. Stamm, and J. W. India, Manipur and Nagaland." CARC 3: 2-6. Wasserheit, eds., Sexually Transmitted Diseases. 3d ed. New York: McGraw-Hill. Panos Institute. 1989. AIDS and the Third World. Phila- delphia: New Society Publishers. Over, Mead, and Godlike Koda. Forthcoming. "Average Cost of Survivor Assistance Programs in Kagera, Tan- Parker, Richard G. 1996. "Historic Overview of Brazil's zania." Working paper. World Bank, Policy Research AIDS Programs and Review of the World Bank AIDS Department, Washington, D.C. Project." Family Health International/AIDSCAP, Arlington, Va. Over, Mead, and Peter Piot. 1993. "HIV Infection and Sexually Transmitted Diseases." In Dean T. Jamison, Pauw, J., J. Ferrie, and R. R. Villegas. 1996. "Controlled W. Henry Mosley, Anthony R. Measham, and Jose HIV/AIDS-related Health Education Programme in Luis Bobadilla, eds., Disease Control Priorities in De- Managua, Nicaragua." AIDS 10: 537-544. veloping Countries. New York Oxford University Press. Peak, A. S., Rana, S. H. Maharjan, and N. Crofts. 1994. 1996. "Human Immunodeficiency Virus Infec- "An Indigenous Harm Reduction Programme for tion and Other Sexually Transmitted Diseases in IDUs in a Developing Country, Nepal." Abstract Developing Countries: Public Health Importance and PD0508. Paper presented at the Tenth International Priorities for Resource Allocation." Journal of Infec- Conference on AIDS, Yokohama, Japan. August tious Diseases 174 (suppl. 2): S162-75. 14-17. Lifesaving and Lifegiving Society, Kath- mandu, Nepal. Over, Mead, Phare Mujinja, Daniel Dorsainvil, and Indrani Gupta. Forthcoming. "Impact of Adult Peng, Xizhe. 1994. "Recent Trends in China's Popula- Death on Household Expenditures in Kagera, Tan- tion and Their Implications." CP 30. London School zania." Working paper. World Bank, Policy Research of Economics, Suntory-Toyota International Centre Department, Washington, D.C. for Economics and Related Disciplines. London. Over, Mead, M. Randall, P. Ellis, Joyce H. Huber, and Perriens, Joseph, Kenneth Hill, Nicholas Prescott, and Orville Solon. 1992. "The Consequences of Adult Ill- Chaiyos Kunanusont. 1997. "Health Gains from Health." In Richard A. Feachem, Tord Kjellstrom, Antiretroviral Therapy." Geneva: UNAIDS. Christopher J. L. Murray, Mead Over, and Margaret A. Phillips, eds., The Health of Adults in the Devel- Philipson, Tomas J., and Richard A. Posner. 1993. Pri- oping World. New York: Oxford University Press. vate Choices and Public Health: The AIDS Epidemic in 346 SELECTED BIBLIOGRAPHY an Economic Perspective. Cambridge, Mass.: Harvard Jaime Sepulveda Amor, and Sheldon Segal, eds., University Press. AIDS and Women s Reproductive Health. New York: Plenum Press. Phoolcharoen, Wiput, and Seri Phongphit. 1996. "HIV Prevention Works Report: Case Study Thailand." Pokrovsky, V. V., I. V. Savchenko, N. N. Ladnaya, and Paper presented at the Eleventh International Confer- others. 1996. "HIV Infection Surveillance in Russia ence on AIDS, Vancouver, B.C., Canada. July 7-12. in 1987-1995 (Statistics)." Russia AIDS Centre, Moscow. Pickering, Helen, and H. A. Wilkins. 1993. "Do Un- married Women in African Towns Have To Sell Sex, Porapakkham, Yaowarat, Somjai Pramarnpol, Supatra or Is It a Matter of Choice?" Health Transition Review Athibhoddhi, and Richard Bernhard. 1996. "The 3 (suppl.): 17-27. Evolution of HIV/AIDS Policy in Thailand: 1984- 1994." AIDSCAP Policy Working Paper Series WP5. Pickering, Helen, M. Quigley, J. Pepin, and others. Family Health International, Washington, D.C. 1993. "The Effects of Post-Test Counseling on Condom Use among Prostitutes in The Gambia." Poshyachinda, V. 1993. "Drug Injecting and HIV Infec- AIDS 7: 271-73. tion among the Population of Drug Users in Asia." Bulletin on Narcotics 45 (1): 877-90. Pinkerton, Steven D., and Paul R. Abramson. 1996. "Implications of Increased Infectivity in Early-Stage Prescott, Nicholas. 1997. "Setting Priorities for Govern- HIV Infection." Evaluation Review 20 (5): 516-40. ment Involvement with Antiretrovirals." Paper pre- sented at the Informal Consultation on The Implica- 1997. "Effectiveness of Condoms in Preventing tions of Antiretroviral Treatments, World Health HIV Transmission." Social Science and Medicine 44 Organization, Geneva. April 29-30. World Bank, (9): 1303-12. East Asia and Pacific Department, Human Resources Operations Division. Piot, Peter. 1994. "Differences between African and Western Patterns of Heterosexual Transmission." In Prescott, Nicholas, Chaiyos Kunanusont, W. Phoolcha- A. Nicolosi, ed., HIV Epidemiology: Models and roen, W. Rojanapitayakorn, Joseph Perriens, and D. Methods. New York: Rowen Press. Boonyuen. 1996. "Formulating Rational Use of Anti- retrovirals in Thailand." Abstract Mo.B. 533. Paper Pitayanon, Sumalee, Sukontha Kongsin, and Wattana presented at the Eleventh International Conference Janjaroen. 1997. "The Economic Impact of HIV/ on AIDS, Vancouver, B.C., Canada. July 7-12. Min- AIDS Mortality on Households in Thailand." In istry of Public Health, Department of CDC, AIDS David Bloom and Peter Godwin, eds., The Economies Division, Bangkok, Thailand. of HI V and AIDS: The Case of South and South East Asia. Delhi: Oxford University Press. Putnam, Robert. 1993. "The Prosperous Community- Social Capital and Public Life." American Prospect 13: Plange, Nii-K. 1990. "Report on Prostitution in Fiji." 35-42. University of the South Pacific, Department of Soci- ology, Suva, Fiji. Putnam, Robert, R. Leonardi, and R. Nanetti. 1993. Making Democracy Work: Civic Traditions in Modern Plummer, Francis A., Stephen Moses, and Jackoniah 0. Italy. Princeton: Princeton University Press. Ndinya-Achola. 1991. "Factors Affecting Female-to- Male Transmission of HIV-1: Implications of Trans- Quinn, T. C., R. 0. Gannon, D. Glasser, S. L. mission Dynamics for Prevention." In Lincoln Chen, Groseclose, W. S. Brathwaite, A. S. Fauci, and E. W. 347 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Hook 3d. 1990. "The Association of Syphilis with the mitted Diseases in a Rural Ugandan Population: Risk of HIV Infection in Patients Attending STD Simulation Model Estimates." InternationalJournal of Clinics." Archives of Internal Medicine 150 (6): Epidemiology 26 (1): 180-89. 1297-1302. Rojanapithayakorn, Wiwat, and Robert Hanenberg. Quinn, T. C., A. Ruff, and N. Halsey. 1994. "Special 1996. "The 100% Condom Program in Thailand." Considerations for Developing Nations." In P. A. AIDS 10 (1): 1-7. Pizzo and C. M. Wilfert, eds., Pediatric AIDS. The Challenge of HIV Infection in Infants, Children and Rowley, Janet, Roy Anderson, and T. W. Ng. 1990. Adolescents. Baltimore, Md.: Williams and Wilkins. "Reducing the Spread of HIV Infection in Sub- Saharan Africa: Some Demographic and Economic Radhakrishna, R., K. Subbarao, with S. Indrakant and Implications." AIDS 4 (1): 47-56. C. Ravi. 1997. "India's Public Distribution System: A Ryder, Robert, V. Batter, M. Nsuami, and others. 1991. National and International Perspective." World Bank, "Fertility Rates in 238 HIV-1 Seropositive Women in South Asia Country Department 2 (Agriculture and Zaire Followed for 3 Years Post-Partum." AIDS 5: Water Operations), Washington, D.C. 1521-27. Reeler, Anne V. 1990. "Injections: A Fatal Attraction?" Ryder, Robert W., Mibandumba Ndilu, Susan E. Social Science andMedicine31(10): 1119-25. Hassig, Munkolenkole Kamenga, D. Sequeira, M. Kashamuka, H. Francis, F. Behets, R. L. Cole- Refeno, Germain, Victor Rabeza, Gora Mboup, and bunders, and A. Dopagne. 1990. "Heterosexual Juan Schoemaker. 1994. Madagascar: Enqu&te Na- Transmission of HIV-1 among Employees and Their tionale Demographique et Sanitaire 1992. Antana- Spouses at Two Large Businesses in Zaire." AIDS 4 narivo, Madagascar: Centre National de Recherches (8): 725-32. sur l'Environnement, Ministere de la Recherche Appliquee au Developpement, and Calverton, Md.: Sarkar, Swarup, Anindya Chatterjee, Clyde B. McCoy, Macro International, Inc. Abu S. Adbul-Quader, Lisa R. Metsch, and Robert S. Anwyl. 1996. "Drug Use and HIV among Youth in Reggy, Atieno, R. J. Simonds, and Martha Rogers. 1997. Manipur, India." In Clyde B. McCoy, Lisa R. "Preventing Perinatal HIV Transmission." AIDS 11 Metsch, and James A. Inciardi, eds. Intervening with (suppl. A): S61-67. Drug-Involved Youth. Thousand Oaks, Calif.: Sage. Sarkar, Swarup, N. Das, Panda, T. N. Naik, K. Sarkar, Rezza, G.,B. C. Singh, J. M. Ralte, S. M. Aier, and S. P. Tri- AIDS among Italian Drug Addicts: Evaluation of B.hC.1993. 'Rapid S. of HiV ang InjecTi- Treatment Progra s adIfrmte Stratge.") pathy. 1993. "Rapid Spread of HIV among Injecting Treatment Programs ana Informative n trategles. Drug Users in North-Eastern States of India." Paper presented at the Fourth International Con- Bu-tnoNacis45()9105 Paper ~~~~~~~~~~~~Bulletin on Narcotics 45 (1): 91-105. ference on AIDS, Stockholm, Sweden. June 12-16. Sato, Paul. 1996. "Sentinel H1V Surveillance." In Jona- Robbins, A., and P. Freeman. 1988. "Obstacles to than Mann, and Daniel Tarantola, eds., AIDS in the Developing Vaccines for the Third World." Scientific World II: Global Dimensions, Social Roots, and Re- American 259 (5): 126-33. sponses. The Global AIDS Policy Coalition. New York: Oxford University Press. Robinson, Noah Jamie, Daan W. Mulder, Bertran Auvert, and Richard J. Hayes. 1997. "Proportion of Sauerborn, R., P. Berman, and A. Nougtara. 1996. "Age HIV Infections Attributable to Other Sexually Trans- Bias, but No Gender Bias, in the Intra-Household 348 SELECTED BIBLIOGRAPHY Resource Allocation for Health Care in Rural Burkina Ronald, P. Piot, and F. A. Plummer. 1988. "Human Faso." Health Transition Review: The Cultural, Social Immunodeficiency Virus Infection among Men with and Behavioural Determinants ofHealth 6 (2): 131-45. Sexually Transmitted Diseases." New EnglandJournal ofMedicine. 319: 274-78. Scheper-Hughes, Nancy. 1993. "AIDS, Public Health, and Human Rights in Cuba." Lancet 342 (8877): Singh, Sujata. 1995. "Three Year Stint at Sonagachi." All 965-67. India Institute of Hygiene and Public Health, Calcutta. August. Sepulveda, Jaime. 1992. "Prevention through Informa- tion and Education: Experience from Mexico." In Sittitrai, Werasit. 1994. "Nongovernmental Organiza- Jaime Sepulveda, Harvey Fineberg, and Jonathan tion and Community Responses to HIV/AIDS in Mann, eds. AIDS Prevention through Education: A Asia and the Pacific." AIDS 8 (suppl. 2): S199-206. Workld View. New York: Oxford University Press. Smith, James, and Alan Whiteside. 1995. "The Socio- Serwadda, David, Maria J. Wawer, Stanley D. Musgrave Economic Impact of HIV/AIDS on Zambian Busi- Nelson K. Sewankambo, Jonathan E. Kaplan, and nesses: Report for the BEAD [Business Exchange on Ronald H. Gray. 1992. "HIV Risk Factors in Three AIDS and Development Group] and CDC." Con- Geographic Strata of Rural Rakai District, Uganda." monwealth Development Corporation, London. AIDS 6 (9): 983-89. SOMARC (Social Marketing for Change), Office of Serwadda, David, M. Wawer, N. Sewankambo, R. H. Public Relations. 1996. "Crisis Communications Les- Gray, C-J. Li, R Kelly, and T. Lutalo. 1995. "Trends sons Learned: Case Study of Niger and the Philip- in'HV IcidnceandPrevalence in Rakai District, pines." SOMARC Occasional Paper 20. Futures in HIV Incidence and Peaec nRkiDsrc Uganda." Abstract MoCO85. Ninth International Group International, Washington, D.C. Conference on AIDS and STDs in Africa, Kampala, Span, Paula. 1996. "Needle Exchanges Inject Contro- Uganda. December 10-14. Institute of Public Health, versy in AIDS Prevention." Washington Post, July 16, Makerere University, Kampala, Uganda. p. Al. Shaeffer, Sheldon. 1995. "Impact on Education." In The "Sri Lankan Condom Sales Higher but Not High Impact of HIV on Children in Thailand. Program on Enough, Say AIDS Workers." 1996. AIDS Analysis AIDS, Research Report 16. Bangkok: Thai Red Cross Asia 2 (6): 15. Society. Ssengonzi, Robert, Martina Morris, Nelson Sewan- Siegel, Joanna E., Milton C. Weinstein, and Harvey V. dambo, Maria Wawer, and David Serwadda. 1995. Fineberg. 1991. "Bleach Programs for Preventing "Economic Status and Sexual Networks in Rakai Dis- AIDS among LV Drug Users: Modeling the Impact of trict, Uganda." Abstract WeC259. Ninth Inter- HIV Prevalence." American Journal of Public Health national Conference on AIDS and STDs in Africa, 81 (10): 1273-79. Kampala, Uganda. December 10-14. Uganda Virus Research, Entebbe, Uganda. "Signs of Change as Taboo Subjects Get Air Time in Information Campaign. Country Profile: Pakistan." Stanecki, Karen A., and Peter 0. Way. 1997. "The 1996. AIDSAnalysisAsia2 (3): 14-15. Demographic Impacts of AIDS-Perspectives from the World Population Profile: 1996." IPC Staff Paper Simonsen, J. N., D. Cameron, M. Gakinya, Jo Ndinya- 86. International Programs Center, Population Divi- Achola, L. J. D'Costa, P. Karasira, M. Cheang, A. R. sion, U.S. Bureau of the Census. Washington, D.C. 349 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Stewart, Graeme. 1997. "Compliance with Antiretroviral Taha, T. E., J. K. Canner, J. D. Chiphangwi, and others. Therapy." Paper presented at the Informal Consulta- 1996. "Reported Condom Use Is Not Associated tion on the Implications of Antiretroviral Treatments, with Incidence of Sexually Transmitted Diseases in World Health Organization, Geneva. April 20-30. Malawi." AIDS 10: 207-12. Stimson, Gerry V. 1993. "The Global Diffusion of In- Tan, Michael L., and Manuel M. Dayrit. 1994. "HIV/ jecting Drug Use: Implications for Human Immuno- AIDS in the Philippines." AIDS 8 (suppl. 2): S 125-30. deficiency Virus Infection." Bulletin on Narcotics 45 (1): 3-17. Tchupo, J. P., J. Timyan, C. Miken, J. G. Ouango, and S. Watts. 1996. "PSI Research Report Summary: Bur- 1994. "Reconstruction of Subregional Diffusion kina Faso Condom KAP Survey, April. 1995." Popu- of HIV Infection among Injecting Drug Users in lation Services International, Washington, D.C. Southeast Asia: Implications for Early Intervention." AIDS 8 (11): 1630-32. Tembo, George, Hany Friesan, Godwil Asiimwe-Okiror, Rita Moser, Warren Naamara, Nathan Bakyaita, and 1996. "Drug Injecting and the Spread of HIV Joshua Musinguzi. 1994. "Bed Occupancy Due to Infection in South-East Asia." In L. Sherr, J. Catalan, HIV/AIDS in an Urban Hospital Medical Ward in and B. Hedge, eds. The Impacts ofAIDS: Psychological Uganda." AIDS 8: 1169-71. and Social Aspects of HIV Infection. Reading, U.K.: Harwood Academic. Temmerman, M., S. Moses, D. Kiragu, and others. 1990. "Impact of Single Session Post-Partum Coun- Stoneburner, Rand L., and Carballo, Manuel. 1997. "An seling of HIV Infected Women on Their Subsequent Assessment of Emerging Patterns of HIV Incidence in Reproductive Behavior." AIDS Care 2: 247-52. Uganda and Other East African Countries." Final report of consultation. Family Health International, Thomas, James C., and Myra J. Tucker. 1996. "The AIDS Control and Prevention Project. International Development and Use of the Concept of a Sexually Centre for Migration and Health, Geneva. Transmitted Disease Core." Journal of Infectious Diseases 174 (suppl. 2): S 134-43. Stover, John, and Peter 0. Way. 1995. "The Impact of Interventions on Reducing the Spread of HIV in Thongthai, Varachi, and Philip Guest. 1995. "Thai Sex- Africa: Computer Simulation Applications." African ual Attitudes and Behaviour: Results from a Recent JournalofMedicalPractice 2 (4): 110-20. National Survey." Paper presented at the conference Gender and Sexuality in Modern Thailand. Austra- Strauss, John, and Duncan Thomas. 1995. "Human lian National University, Canberra. July. Resources: Empirical Modeling of Household and Family." In Jere Behrman and T. N. Srinivasan, eds., Titmuss, Richard. 1972. The Gift Relationship: From Handbook of Development Economics. Vol. 3A. New Human Blood to Social Policy. New York: Vintage York: North-Holland. Books. Subbarao, Kalanidhi, Aniruddha Bonnerjee, Jeanine Tokars, J. I., R. Markus, D. H. Culver, S. A. Schable, P. Braithwaite, Soniya Carvalho, Rene Ezemenari, Carol S. McKibben, and C. I. Bandea. 1993. "Surveillance Graham, and Alan Thompson. 1996. "Social Assist- of HIV Infection and Zidovudine Use among Health ance and Poverty-targeted Programs: Lessons from Care Workers after Occupational Exposure to HIV- Cross-Country Experience." World Bank, Poverty infected Blood." Annals of Internal Medicine 118 (12): and Social Policy Department, Washington, D.C. 913-19. 350 SELECTED BIBLIOGRAPHY Townsend, R. 1994. "Risk and Insurance in Village ulation Division. U.S. Bureau of the Census, Wash- India." Econometrica 62: 539-91. ington, D.C. UNAIDS. 1996a. "HIV and Infant Feeding: An Interim U.S. Bureau of the Census (database). 1997. "HIV/ Statement." Geneva. AIDS Surveillance Database." Version 1.1, Release 0. Population Division, International Programs Center, 1996b. "HIV/AIDS: The Global Epidemic." Washington, D.C. Fact Sheet. Geneva. December. U.S. GAO (General Accounting Office). 1993. Needle 1996c. "The HIV/AIDS Situation in Mid- Exchange Programs: Research Suggests Promise as an 1996: Global and Regional Highlights." Paper pre- AIDS Prevention Strategy (GAP/HRD-93-60). Wash- sented at the Eleventh International Conference on ington, D.C.: U.S. Government Printing Office. AIDS, Vancouver, B.C., Canada. July 7-12. Valdiserri, Ronald O., David W. Lyter, Laura C. Levi- 1996d. "UNAIDS Fact Sheet, Mid-1996" ton, Catherine M. Callahan, Lawrence A. Kingsley, Geneva. and Charles R. Rinaldo. 1989. "AIDS Prevention in Homosexual and Bisexual Men: Results of a 1997. The Impact of HIV and Sexual Health Randomized Trial Evaluating Two Risk Reduction Education on the Sexual Behaviour of Young People: A Interventions." AIDS 3 (1): 21-26. Review (Updated). Joint United Nations Programme on AIDS. Geneva. Van Dam, C. J., G. A. Dallabetta, and P. Piot. 1997. "Prevention and Control of Sexually Transmitted UNAIDS/Country Support. 1996. "Reported AIDS Diseases in Developing Countries." In K. K. Holmes, Cases as of 31 December 1995." World Health P. F. Sparling, W. E. Stamm, P. Piot, J. Wasserheit, Organization/Global Programme on AIDS, February and Per-Anders Mardh, eds., Sexually Transmitted 1. Geneva. Diseases. 3d ed. New York: McGraw-Hill. United Nations. 1993. Urban and RuralAreas by Sex and van Praag, Eric, E. Katabira, S. Anderson, E. Ngugi, and Age: The 1992 Revision. New York: United Nations, D. Koy. 1996. "Can HIV/AIDS Care Initiatives Be Department for Economic and Social Information Part of Integrated Care? Lessons from Developing and Policy Analysis, Population Division. Countries." Abstract Th.B.400. Paper presented at the Eleventh International Conference on AIDS, 1995. "International Migration Policies, 1995." Vancouver, B.C., Canada. July 7-12. World Health Poster. New York: United Nations Department for Organization, Geneva. Economic and Social Information and Policy Analy- sis, Population Division. Van der Gaag, J. 1995. Private and Public Initiatives: Working Together for Health and Education. Direc- U.S. Bureau of the Census. 1996. "World Population tions in Development series. Washington, D.C.: Profile 1996, with a Special Chapter Focusing on World Bank. Adolescent Fertility in the Developing World." U.S. Department of Commerce. U.S. Government van de Walle, Dominique, and Kimberly Nead, eds. Printing Office, Washington, D.C. 1995. Public Spending and the Poor: Theory and Evidence. Baltimore: Johns Hopkins Press. 1997. "Recent HIV Seroprevalence Levels by Country: January, 1997." Research Note 23. Health Van der Ploeg, Catharina P. B., Carina Van Vliet, Sake Studies Branch, International Programs Center, Pop- J. De Vlas, Jackoniah 0. Ndinya-Achola, Lieve 351 CONFRONTING AIDS: PUBLIC PRIORITIES IN A GLOBAL EPIDEMIC Fransen, Gerrit Van Oortmarssen, and J. Dik F. Committee on Health Research Relating to Future Inter- Habbema. 1997. "STDSIM: A Microsimulation vention Options. Geneva. Model for Decision Support in STD Control." Erasmus University, Centre for Decision Sciences in WHO/EC Collaborating Centre on AIDS. 1996a. Tropical Disease Control, Rotterdam, the Nether- "HIV/AIDS Surveillance in Europe," First Quarterly lands. Forthcoming in Interfaces. Report 49 (March). Saint-Maurice, France. Vlahov, David. 1997. "Role of Needle Exchange Pro- _ . 1996b. "HIV/AIDS Surveillance in Europe," grams in AIDS Prevention." In NIH Consensus Con- Second Quarterly Report 50 June), Saint-Maurice, ference, Interventions to Prevent HIV Risk Behaviors: France. Program andAbstracts. Bethesda, Md.: National Insti- tutes of Health. WHO/GPA. 1995. "Global Prevalence and Incidence Wawer, M. J., C. Podhisita, U. Kanugungsukkasem, A. of Selected Curable Sexually Transmitted Diseases: Pramualratana, and R McNamara. 1996a. "Origins Overview and Estimates." WHO/GPA/STD/95.1. and Working Conditions of Female Sex Workers in Geneva. Urban Thailand: Consequences of Social Context for HIV Transmission." Social Science and Medicine 42 WHO, Program on Substance Abuse. 1994. "Multi-City (3): 453-62. Study on Drug Injecting and Risk of HIV Infection." A report prepared on behalf of the WHO Interna- Wawer, M., N. K. Sewankambo, R. H. Gray, and oth- tional Collaborative Group. Geneva. ers. 1996b. "Community-based Trial of Mass STD Treatment for HIV Control, Rakai, Uganda: Prelimi- Wirawan, D. N., P. Fajans, and K. Ford. 1993. "AIDS nary Data on STD Declines." Abstract Mo.C.443. and STDs: Risk Behaviour Patterns among Female Eleventh International Conference on AIDS, Van- Sex Workers in Bali, Indonesia." AIDS Care 5 (3): couver, Canada. July 7-12. Columbia University, 289-303. Center for Population and Family Health, New York. Wong, K. H., S. S. Lee, and W. L. Lim. 1993. "HIV Weisbrod, Burton. 1977. The Voluntary Non-Profit Surveillance among Drug Users in Hong Kong." Sector. Lexington, Mass.: Lexington Books. Paper presented at the Ninth International Confer- Weniger, Bruce G., and Seth Berkley. 1996. "The Evolv- ence on AIDS, Berlin. June 611. ing HIV/AIDS Pandemic." In Jonathan Mann and World Bank. 1993a. "Staff Appraisal Report: Brazil Daniel Tarantola, eds., AIDS in the World IL Global AIDS and STD Control Project." Report 11734 BR. Dimensions, Social Roots, and Response. The Global World Bank, Country Department I, Human Re- AIDS Policy Coalition. New York: Oxford University s Press. sources Division, Washington, D.C. Weniger, Bruce G., K. Limpakarnjanarat, K. Ungchusak . 1993b. Tanzania AIDS Assessment and Planning S. Thanprasertsuk, K. Choopanya, S. Vanichseni, T. Study. World Bank Country Study. Washington, Uneklabh, P. Thongcharoen, and C. Wasi. 1991. D.C. "The Epidemiology of HIV Infection and AIDS in Thailand." AIDS 5 (suppl. 2): S71-85; erratum in . 1993c. World Development Report 1993: Invest- AIDS 1993 7 (1): 147. ing in Health. New York: Oxford University Press. WHO (World Health Organization). 1996. Investing in . 1996a. "AIDS Prevention and Mitigation in Health Research and Development, Report of the Ad Hoc Sub-Saharan Africa: A Strategy for Africa." Report 352 SELECTED BIBLIOGRAPHY 15569. World Bank, Africa Region, Technical De- Injections on Daily Medical Practice. Amsterdam: Free partment, Human Resources and Poverty Division, University Press. Washington, D.C. Yeager, Rodger, and Craig Hendrix. 1997. "Global Sur- 1996b. "Tanzania-The Challenge of Reforms: vey of Military HIV/AIDS Policies and Programs." Growth, Incomes, and Welfare." 3 vols. Vol. 1: Main Civil-Military Alliance Newsletter 3 (1): 1-8. Report. Africa Region, Eastern Africa Department, Country Operations Division. Washington, D.C. Yu, Elena S. H., Qiyi Xie, Konglai Zhang, Ping Lu, and Lillian L. Chan. 1996. "HIV Infection and AIDS in . 1997a. World Development Report 1997. New China, 1985 through 1994." American Journal of York: Oxford University Press. Public Health 86 (8): 1116-22. . 1997b. World Development Indicators 1997. Washington, D.C. Zheng, Xiwan. 1996. "HIV/AIDS Epidemic in China." In HIV/AIDS in China, no. 4: 11-12. World Bank, Wyatt, H. V. 1993. "Injections, Infections, and Steril- Asia Technical Department, AIDS in Asia Unit, ity." In M. Bloem and I. Wolffers, eds., The Impact of Washington, D.C. 353 THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433 USA Telephone: 202-477-1234 Facsimile: 202-477-6391 Telex: MCI 64145 WORLDBANK MCI 248423 WORLDBANK World Wide Web: http://vww.worldbank.org/ E-mail: books@worldbank.org * * * * EUROPEAN COMMISSION * * Directorate-General for Development Health, Family Planning and AIDS Unit (VIII/G/1) Rue de la Loi, 200 B-1049 Brussels, Belgium Telephone: (32.2) 296.36.98 Facsimile: (32.2) 296.36.97 World Wide Web: http://europa.eu.int/ UNAIDS UNICEF * UNDP * UNFPA UNESCO *WHO *WORLD BANK Joint United Nations Programme on HIV/AIDS 20, avenue Appia CH-1211 Geneva 27, Switzerland 90 00 0 Telephone: (41.22) 791.3666 Fac,irniile: (41.22) 791.4187 World Wide 'eb: http://www.unaids/org/ 9 '780195 211177 ISBN 0-19-521117-0 DERINRD BY EDWARD TOWLAS