Health Financing System Assessment Papua New Guinea DEPARTMENT OF HEALTH Health Financing System Assessment Papua New Guinea Disclaimer This publication has been funded by the Australian Government through the Department of Foreign Affairs and Trade. The views expressed in this publication are the author’s alone and are not necessarily the views of the Australian Government. The Australian Government neither endorses the views in this publication, nor vouches for the accuracy or completeness of the information contained within the publication. The Australian Government, its officers, employees and agents, accept no liability for any loss, damage or expense arising out of, or in connection with, any reliance on any omissions or inaccura- cies in the material contained in this publication. This publication is intended to provide general information only and before entering into any particular transaction users should: rely on their own enquiries, skill and care in using the information; check with primary sources; and seek indepen- dent advice. Contents Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ix Abbreviations and Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xiii Section One: Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Section Two: Background . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.1 Economic Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3 2.2  Demographic Health Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 2.3 The Disease Burden . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 2.4 Health System Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Human Resources for Health . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.5  Health Care Access and Utilization Context . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 2.6  2.7 Health Financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Section Three: Health Expenditure Trend Analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 Section Four: PNG Reliance on Donor Financing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Key PNG Donors: GFATM and Gavi . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 4.1  4.2 Gavi: Immunization in PNG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 GFATM: HIV and AIDS Response, TB, and Malaria Treatment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 4.3  Section Five: Financial and Institutional Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 5.1 Financial Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Institutional Sustainability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 5.2  Section Six: Summary and Policy Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 Appendix One: Data Challenges in PNG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Budget and Expenditure Data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Health System Outputs and Outcomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 49 Boxes Box 1-1: What is a Health Financing System Analysis? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Box 2–1: Data Challenges in PNG . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Box 3–1: Service Improvement Programme Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Box 5–1: Examples of Partnerships to Deliver Health Services . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 iii iv Health Financing System Assessment Figures Figure 2-1: PNG’s GDP per Capita and Lower-Middle-Income Status Threshold (1993–2017) . . . . . . . . . . . . . 4 Figure 2–2: Annual GDP Growth (%) (1993–2017) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Figure 2–3: General Government Revenue and Expenditure (1983–2022) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5 Figure 2–4: Fertility Rate and Population Growth (1967–2015) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 Figure 2–5: Population Health Outcomes (1967–2015) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Figure 2–6: International Life Expectancy Comparison (1960–2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Figure 2–7: Country Comparison of Population Health Outcomes (2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Figure 2–8:  Disease Burden Attributable to Communicable Diseases, NCDs, and Injuries (1990–2015) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9 Figure 2–9: Governance Structure of the Health System . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Figure 2–10: Selected Health Access Indicators (2005–15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 2–11: Selected Service Delivery Indicators (2006–15) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figure 2–12: THE as a Proportion of GDP (1995–2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Figure 2–13: THE and Public Health Expenditure (PHE) as % of GDP in East Asia Countries . . . . . . . . . . . . . . . 15 Figure 2–14: Health Expenditure per capita (1995–2015) (current US$ and constant LCU) . . . . . . . . . . . . . . . . 16 Figure 2–15: Trends in the Structure of THE (1996–2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Figure 2–16: Health Financing Mix (2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16 Figure 2–17: External Resources on Health as % of THE (1995–2014) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 Figure 3–1:  Government Health Sector Operational Budgets, Warrants, and Expenditure (2010–17) (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 3–2: Health Sector Fund Flows . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 3–3:  Health Sector Fund Flows at the Subnational Level : Comparison Between Traditional and PHA Models . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 3–4:  Share of Health Sector Budget by Agency (2011–17) (revised budget figures, domestic financing only) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 Figure 3–5:  Payroll Overruns in NDoH and HMS (2012–16): Personnel Emoluments Revised Budget and Actual Expenditure, Domestic Financing Only (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . 23 Figure 3–6:  Total Government Health Sector Expenditure by Fund Source (2011–17) (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 3–7:  Cash Release for HFGs by February (Province and Region) (2008–12) . . . . . . . . . . . . . . . . . . . . . . 25 Figure 3–8:  Health Sector 2015 Budget, Warrants, and Expenditure, Domestic Financing (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Figure 3–9:  Health Sector 2016 Budget, Warrants, and Expenditure, Domestic Financing (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Figure 3–10: Comparison of Revised Appropriations by Agency and Economic Item Categories, Domestic Financing (2015–17) (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 Figure 4–1: Official Development Assistance (ODA) for Health by Channel (US$ millions) . . . . . . . . . . . . . . . 28 Figure 4–2: ODA for Health by Donor (US$ millions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 28 Figure 4–3: Gavi’s Transition Policy and Country Preparedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 4–4: Gavi Disbursements by Year and Program (US$) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Figure 4–5: Gavi Cofinancing Requirements (US$ millions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Figure 4–6: GFATM Disbursements by Disease (US$ millions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31 Figure 4–7: GoPNG Agreed Spending for GFATM Funds (US$ millions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Figure 4–8: Number of LLINs Distributed (millions) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Figure 5–1: GoPNG Health Sector Projections (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36 Figure 1A–1: Budget Data Variance (2014 and 2015) (millions of Kina) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 46 Contents v Tables Table 2–1: Top Ten Causes of Morbidity and Premature Mortality (various years) . . . . . . . . . . . . . . . . . . . . . . 9 Table 2–2: Health Workers per 1,000 Population (2009 and 2016) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12 Table 2–3: Comparison of Access to Health Services and Infrastructure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 Acknowledgments The World Bank, with assistance from the Papua New Economist), Neesha Harnam (Health Specialist), and Chan- Guinea National Department of Health (NDoH), undertook dana Kularatne (Senior Economist). this study. This effort was one of the subtasks under the Papua New Guinea Programmatic Health Advisory Ser- Toomas Palu (Practice Manager of East Asia and Pacific vices and Analytics. Region, Health, Nutrition, and Population Global Practice) and Christoph Kurowski (Global Solution Lead in Health This report was produced and written by a task team Financing, Health, Nutrition, and Population Global Prac- consisting of Xiaohui Hou (Senior Economist and Task tice) provided technical comments and overall supervi- Team Leader), Laurin Janes (Health Economist), Sophie sion on this report. The team would like to thank Michel Brown (Consultant), Katie Barker (Consultant), and Pra- Kerf (Country Director for Papua New Guinea and Pacific nita Sharma (Health Economist). Kerry Main Pagau, Islands, East Asia and Pacific Region) and Steffi Stallmeis- Quenelda Clegg, and Amanda Sookun provided techni- ter (Former Country Manager for Papua New Guinea), and cal and administrative support. Maude Ruest Archam- Patricia Veevers-Carter (Country Manager for Papua New bault (Health Economist) coordinated the peer review Guinea) for their overall guidance and support. The report process. The team would like to thank Ajay Tandon (Lead was edited by Chris Stewart. Health Economist), Aneesa Arur (Senior Economist), Jewel- wayne Salcedo Cain (Consultant), Nicolas Rosemberg The authors would also like to sincerely thank the Secre- (Economist), and Sarah Alkenbrack (Senior Health Econo- tary and staff of the NDoH for providing information for mist) for providing the analyses and useful comments this study, as well as staff members from the Global Fund for various drafts. The team would also like to thank col- to Fight AIDS, Tuberculosis, and Malaria (GFATM) and Gavi, leagues from the Australian Department of Foreign Affairs The Vaccine Alliance. and Trade (DFaT) for providing thorough comments and Financial support for this work was received from DFaT peer reviewers Netsanet Walelign Workie (Senior Health and Gavi. vii Executive Summary Objective areas. The rural terrain impedes access to health services and presents a challenge in attracting and retaining health This Health Financing System Assessment (HFSA) eval- personnel. Many access and utilization indicators have uates the financial and institutional sustainability of stagnated or declined in the last few years. A regional com- health financing in Papua New Guinea (PNG), at a time parison highlighted poor maternal health care access; out- when PNG is preparing to graduate from development patient visits per person also declined from 2010 to 2014. partner (DP) support in key areas. The assessment pro- Quality of care, staff presence, and out-of-pocket (OOP) vides an overview of PNG’s economic situation, including payments—even though they are rarely ­ catastrophic— DP assistance, and examines the strength of the system to also impede access to health services. absorb responsibility for four key program areas (human immunodeficiency virus (HIV) and AIDS response; tuber- Human Resources in the Health System culosis; malaria; and immunization) that have been largely supported by DPs. Examination of the governance and PNG is facing a health worker shortage that could financial structure of the health system provides insight worsen in the medium to long term. The service-­ into areas for improved efficiency and capacity building delivery worker ratio of 15.4 per 10,000 population (2016) to meet upcoming cofinancing agreements and improve is low. The existing health workforce is rapidly aging, com- health outcomes in PNG. pounding the health worker shortage issue. The workforce is inequitably distributed between rural and urban areas, with a substantial predisposition toward urban areas. Key Findings Another compounding factor is PNG’s lack of significant The HFSA reviewed PNG’s health system to determine progress in medicine and health-training capacity, which the institutional and financial sustainability of health has resulted in a shortage of qualified staff to replenish the resourcing and financing and established the follow- aging workforce. ing key findings: The Composition and Level of Health Sector Economic and Demographic Context Expenditures Economic growth from 2002 to 2014 was impressive, Total Health Expenditure (THE) as a percentage of averaging approximately 5.7 percent per year. Growth GDP relies predominantly on government spending, rates have subsequently slowed due to subdued global with low OOP spending. THE per capita in real terms has commodity prices, a major domestic drought, and weak been declining since 2004, in part because the absolute nonresource sector growth. Strong economic growth in increase has been offset by population growth. THE per earlier years led to an increase in health expenditure, how- capita is low by regional standards and is expected to fur- ever, infant and maternal mortality rates (MMR) remain ther decline due to decreasing government expenditure high for a country of PNG’s development status and within and a growing population. This challenge underscores the the region. The recent decline in economic growth threat- need to prioritize financing within the health sector and ens to impact the health sector at a critical time when improve efficiencies to make the most of the allocated DP support is expected to decline. funding. Health Care Access and Utilization External Financing on Health There is a significant disparity in access to health care On average, 20 percent of total annual health spend- throughout PNG, especially between rural and urban ing in PNG is from DPs and funding levels, sources, and ix x Health Financing System Assessment recipients are volatile. Australia is the largest bilateral Health Plan (NHP) to the four disease program plans. donor, while GFATM and Gavi have become increasingly First and foremost, PNG needs to determine which important partners in health service delivery. In recent components of each key disease program to integrate years, implementation of donor-funded programs has and provide cost and resource requirements (such as increasingly shifted towards nongovernmental organiza- with Gavi support and the immunization program). tions (NGOs) and civil society. As the PNG Government These program activities need to be linked to Provin- moves toward fulfilling their cofinancing commitments cial Health Authority (PHA) implementation and inte- connected to donor support transition, it is imperative grated into provincial service plans down to the health that it demonstrates the ability to strategically plan for a facility level. This will establish a clear demarcation of successful transition to continue critical health programs. roles and responsibilities to effectively coordinate and implement the national strategies already in place for The Financial Sustainability of GFATM the four priority disease programs. and Gavi-supported Programs • Increase health system efficiency to increase the The increasing reliance on GFATM and the current fiscal space. It is particularly important for the govern- transition from Gavi support raise issues of financial ment to increase the efficiency of current spending. sustainability. PNG is currently preparing to meet the Higher spending on health can contribute to better cofinancing agreements connected to Gavi transition. The outcomes, but so can improvements in health spend- government has yet to cost and appropriate the funding ing efficiency. There appear to be significant inefficien- necessary to meet the steadily increasing share of health cies in health spending, although the extent of these spending that Gavi has supplied, or to integrate the donor- inefficiencies is not entirely clear. The systematic docu- funded activities into the budget. Given the persistently mentation of the budget, expenditure, personnel, and high disease burden, it is unlikely that GFATM support health utilization information will help track expendi- will transition soon. In preparation for the eventual tran- tures and expenditure outcomes. Improved efficiency sition and to ensure program sustainability, however, the will ensure that resources are available in a timely government should begin to integrate the donor-funded manner and will reduce cost pressures. Improving the activities in the budget. efficiency of spending will create a stronger health sys- tem, and better prepare it for the Gavi transition and The Institutional Sustainability of GFATM decreased vertical support from GFATM. and Gavi-supported Programs • Strengthen fiscal analysis and public financial man- Gavi and GFATM programs currently rely on separate agement (PFM). Poor financial management systems systems and human resources, which presents chal- within the National Department of Health (NDoH) have lenges for future support and integration. Silo pro- prevented donors from increasing on-system support grams represent inefficiency in health staffing, that is through the Health Sector Improvement Program (HSIP) already in high demand, and do not lead to the best out- trust account. Improving this capacity would serve as a comes as evidenced by PNG’s low immunization rates. precondition and allow donors to increasingly rely on government systems. Short-term estimates of funding needs and cofinancing requirements must be included Policy Recommendations in the Medium-term Expenditure Framework (MTEF), • Prepare for medium- to longer-term donor transi- to better understand future resource requirements tion. PNG is not ready to graduate from GFATM and and communicate them effectively to central agencies. other donor support in the medium term. To improve PFM needs to be continually strengthened overall to preparedness for graduation, the government should improve the cost-effectiveness of service delivery and unbundle the support it is currently receiving and free up resources from poorly performing areas. determine the key components to integrate and sus- tain in the national budget. The government should • Improve the timely release of warrants. Warrant also use this period to improve the financial and institu- releases need to be timely to improve health sys- tional capacities in preparation for future transition and tem performance. While untimely disbursement is a graduation. Department of Treasury (DoT) issue, the NDoH needs to improve its own budget execution and disburse- • Manage Gavi transition and clarify types and levels ment of operational funds for hospitals and Christian of interventions from the major disease programs to Health Services (CHS). Although the issuing of warrants be integrated. Integrating the major disease programs is unpredictable and ad hoc, NDoH has the potential to as part of the overall health system strengthening (HSS) create further bottlenecks if funds to hospitals and CHS effort will require further linkages between the National are not disbursed immediately upon receipt from DoT. Executive Summary xi • Protect government financing to the frontlines to key disease areas and the health strategy implemen- ensure the successful implementation of the Free tation challenges as DP support declines (particularly Healthcare Policy. The government has committed with Gavi). The challenges of the health sector, in terms to eliminating user fees that are a barrier to health care of integration and transition, need to be more clearly access, especially for the poorest segments of the pop- articulated. Furthermore, policy options and strategies ulation. To date, the related funds are not disbursed in a should be established in close consultation and part- timely manner, which results in delays at the provincial nership with the DoT and the Department of National level and forces facilities to charge fees. Warrants for Planning and Monitoring, as well as with other sectors health function grants (HFG) and the Free Healthcare like transportation and education. Policy funding must be released reliably to eliminate user fees. A majority of this funding must be utilized • Improve data accessibility, comprehensiveness, reliably to finance facility operations (in-kind or cash) and quality. Improve data sources for domestic fund- rather than nonhealth activities at the provincial and ing (national and provincial), external funding, disease district level. burdens, health access, health workforce, and health infrastructure so they can be effective planning and • Strengthen reporting and information sharing. monitoring tools. The National Health Information Increased engagement and membership at the Min- System (NHIS) requires improvements and there is an isterial Economic Committee, as well as the Provincial urgent need to assess the country’s disparate data col- and Local Level Service Monitoring Authority (PLLSMA) lection systems across different service-delivery levels could help the NDoH, as the lead agency for the health and programs. sector, to improve communication and report on the Abbreviations and Acronyms ADB Asian Development Bank GFATM Global Fund to Fight Aids, Tuberculosis and Malaria ANC Antenatal Care GNI Gross National Income APEC Asia Pacific Economic Cooperation GoPNG Government of Papua New Guinea BMU Basic Management Unit HFG Health Function Grant CCM Country Coordinating Mechanism HFSA Health Financing System Assessment CHS Christian Health Services HIV Human immunodeficiency virus CHW Community Health Worker HMS Hospital Management Services CPIA Country Policy and Institutional Assessment HR Human resources CRS Creditor Reporting System HSIP Health Sector Improvement Program CSO Civil Society Organization HSS Health System Strengthening DALY Disability-adjusted Life Year IBBS Integrated bio-behavioral study DFaT Australian Department of Foreign Affairs and Trade IFMS Integrated Financial Management System DIRD Department of Implementation and Rural IHME Institute for Health Metrics and Evaluation Development IMF International Monetary Fund DoT Department of Treasury IMR Institute of Medical Research DP Development Partner IPV Inactivated Polio Vaccine DPLLGA Department for Provincial and Local-Level Government Affairs LLG Local-level Government DPT3 Diphtheria, pertussis, tetanus LLIN Long-lasting insecticidal nets DSIP District Services Improvement Programme MDG Millennium Development Goal EAP East Asia Pacific MDR Multidrug resistant Gavi The Vaccine Alliance MMR Maternal Mortality Rate GDP Gross Domestic Product MR Measles-rubella MSM Men who have sex with men xiii xiv Health Financing System Assessment MTEF Medium-term Expenditure Framework PHC Primary Health Care NAC National AIDS Council PHE Public Health Expenditure NCD National Capital District PLLSMA Provincial and Local Level Service Monitoring Authority NCD Noncommunicable Disease PNG Papua New Guinea NDoH National Department of Health PR Principal Recipient NEFC National Economic and Fiscal Commission PSI Population Services International NFM New Funding Model PSIP Provincial Services Improvement Programme NGO Nongovernmental organization RAM Rotarians Against Malaria NHIS National Health Information System STI Sexually Transmitted Infection NHP National Health Plan TB Tuberculosis NHSS National Health Service Standards THE Total Health Expenditure ODA Official Development Assistance UHC Universal health coverage OECD Organisation for Economic Co-operation and Development UNICEF United Nations Children’s Fund OOP Out-of-pocket WEO World Economic Outlook OSF Oil Search Foundation WHO World Health Organization PFM Public Financial Management WVI World Vision International PHA Provincial Health Authority XDR Extensively drug resistant section one Introduction The health sector of Papua New Guinea (PNG) has seen Subsidized Secondary Care.3 These policies focus primarily significant positive developments in recent years. on financial protection for the most vulnerable popula- These include investments in critical infrastructure (espe- tion groups. The confluence of these events, together with cially medical stores, training facilities, and hospitals), the other demographic and the epidemiological transition,4 is establishment of key policies, the steady rollout of the Pro- placing, and will continue to place, enormous strain on the vincial Health Authority (PHA) model, and a reduction in health sector of PNG. the incidence of malaria. Given this backdrop, the Health Financing System Key indicators of health access and quality have, Assessment (HFSA), which evaluates the financing however, barely improved or have even declined (for system and institutional sustainability, comes at an example, maternal and child health) from 2006 to opportune time. The report begins by providing a com- 2015 (National Department of Health—NDoH 2016). prehensive background, including an overview of PNG’s In addition, the availability of resources for health financ- economic situation, health demographics, health financ- ing will be limited in the medium term as the government ing, human resources for health and the health system. grapples with constraints such as domestic revenue mobi- It then analyses two areas critical to sustainable health lization, achieving a balanced budget, and restoration of financing: (i) PNG expenditure, with a focus on levels and macroeconomic stability. Overall, the pace of long-term sources of health expenditure, as well as resource alloca- progress in key health indicators is slowing down and, tion; and (ii) PNG reliance on donor resources, in particular, given the reduced expenditure projections, the health the GFATM and Gavi. sector is in a critical phase. In 2016, PNG entered the accelerated transition phase from Gavi support.1 During this time, the government will be expected to increase its share of cofinancing for vaccines, while Gavi gradually decreases their contribu- tion. By 2021, the Government of PNG (GoPNG) will gradu- ate from Gavi support, while maintaining access to Gavi prices. The government also needs to continue to meet its cofinancing obligations with GFATM and prepare for the eventual transition from the program. The government is simultaneously moving towards universal health coverage (UHC),2 along with many other developing countries, and has recently established Fee Free Primary Health Care and 3.  The free health policy came into effect on February 24, 2014; however, there are concerns about implementation of the policy in terms of its efficacy, implementation schedule, and lack of indicators to determine both the extent of financial protection through the policy and the coverage of health care services. Despite the name of the new 1.  While GoPNG’s vaccine cofinancing is expected to increase, there policy, primary health care services have officially been free in PNG since are discussions around extending Gavi’s support beyond 2021 in the 1975. Nevertheless, due to a lack of reliable financing, health facilities form of HSS. The idea is to ensure the effective delivery of immunization do, in practice, charge user fees as these are often the only source of services. This has been discussed, but not fully agreed yet. operational revenue. 2.  The goal is to ensure that all people obtain the health services they 4.  For example, aging and the higher burden of noncommunicable need without suffering financial hardship. diseases (NCDs). 1 2 Health Financing System Assessment Box 1–1: What Is a Health Financing System Analysis? The HFSA asks a set of key questions about the nature of a health system, the prioritization of health expenditure within the government budget, the sources of financing and how they change over time, the magnitude of resources available and how they compare to other countries, trends in economic growth and their impact on health financing, and opportunities for increased efficiency in spending. The World Bank conducts HFSAs to systematically assess a country’s health financing system to identify critical con- straints and opportunities to accelerate and sustain progress towards universal health coverage (UHC). It represents a systematic and comprehensive approach, with a common core and flexible and modular overall design. The assess- ment focuses on institutions and tries to highlight linkages to service delivery issues. The HFSA focuses on efficiency, equity, and the overall macrofiscal context. This document represents the core HFSA, but additional in-depth modules can be considered to further explore topics of interest including: • Transition from development assistance; • Fiscal space; • PFM; • Purchasing/payment systems; • Hospital financing; and • Human resources. This HFSA is part of a series of reports across the East Asia and Pacific (EAP) Region, including Indonesia, the Solomon Islands, Kiribati, and Vanuatu. Given the different characteristics of health financing in these countries, various focus areas receive more attention, while maintaining a degree of consistency around core topics. section two Background between 1996 and 2010. Estimates from a 2009–10 house- Key messages: hold survey indicated that the PNG national poverty rate was still approximately 39.9 percent (PNG National Statisti- • Despite impressive recent economic growth, PNG cal Office 2013). is now facing a challenging economic forecast that will make safeguarding the health sector difficult. Economic growth in recent years has been impres- • PNG failed to reach the Millennium Development sive, albeit volatile, reflecting the impact of fluctuat- Goals (MDGs) for under-five mortality and mater- ing international commodity prices (Figure 2–2). PNG nal health, and is far from reaching Sustainable experienced high growth from 2010 to 2014. The recent Development Goals (SDGs). decline in commodity prices and drought in 2015 and early 2016 has, however, weakened the external and fis- • The government-funded health system is sup- cal positions and lowered economic growth. Lower real plemented by the government-subsidized CHS. GDP growth rates in 2016 also reflect the base effects of Governance is fragmented and ambiguity around the country’s first liquefied natural gas project (PNG LNG), roles and responsibilities persists. which commenced in 2014 and reached full capacity in • An insufficient and aging health workforce cou- 2015. GoPNG undertook two consecutive supplementary pled with challenging terrain present further chal- budgets in 2015 and 2016 to reduce expenditure due to lenges to access and provision of health services. the revenue shortfalls (International Monetary Fund—IMF 2017). Adopting a prudent 2017 budget, the authorities • Overall, PNG has experienced a decline in the uti- responded to the recent commodity price shock through lization of outpatient services between 2010 and fiscal consolidation and tax measures. 2014. There are significant disparities in access to, and utilization of, health services. Furthermore, The short-term economic outlook is not positive. A key rural outreach has continued to decrease. factor dampening nonresource growth is the required fis- • Health spending, as a share of general govern- cal consolidation recently initiated by the government. ment spending, decreased from over 9 percent The fiscal deficit has been increasing—a budget scenario in 2010 to 6.8 percent in 2014. External resources that is expected to persist from 2018 to 2021 (Figure 2–3). for health play an important role in total health Over the medium term, additional resource projects may financing. lead to increased government revenue. PNG is in a situ- ation of severe capacity constraint because of economic fragility. The Country Policy and Institutional Assessment 2.1 Economic Context (CPIA) rating in the Harmonized List of Fragile Situation for PNG is a lower-middle-income country with a GDP PNG is 3.13 in FY 2017. A country is in a “fragile situation” per capita of US$2,688 (2016). Previously classified as when its harmonized average CPIA rating is 3.2 or less. a lower-middle-income country in the 1990s, PNG was PNG is one of the most resource-dependent economies reclassified as low income in 2001.5 In 2008, PNG returned in EAP, with resources (including natural gas, oil, gold, to lower-middle-income status (Figure 2–1). Growth in nickel, and copper) accounting for 19 percent of GDP in GDP per capita did not result in a decline in poverty rates 2015, compared with a regional average of 4.9 percent (IMF 2016). This indicator is also high in comparison with other lower-middle-income countries where the average 5.  This was due to a sharp depreciation of the exchange rate, starting in 1994. The exchange rate has also caused GDP per capita to decline since is 7.5 percent. The IMF is supporting the government’s 2012 (see Figure 2–1). 3 4 Health Financing System Assessment Figure 2–1: PNG’s GDP per Capita and Lower-Middle-Income Status Threshold (1993–2017) 3,000 10,000 GDP per capita, current US$ (left axis) 2,500 8,000 GDP per capita (current LCU) GDP per capita (current US$) 2,000 6,000 GDP per capita, current LCU (right axis) 1,500 4,000 1,000 2,000 Lower-middle income Low income Lower-middle income 500 0 1993 1997 2001 2005 2009 2013 2017 Source: World Development Indicators; IMF World Economic Outlook. Note: LCU refers to local currency unit; 2017 data projected. Figure 2–2: Annual GDP Growth (%) (1993–2017) 18.2 20 GDP growth (annual %) 15 11.6 8.7 10 7.3 7.4 6.6 6.6 5.9 6.1 4.7 4.7 4.4 4.5 3.9 3.6 3.7 5 2.9 3.0 2.8 2.6 2.5 2.3 2.0 2.2 2.1 1.9 0.6 0 –0.0 –2.5 –3.4 –5 1997 –6.3 –10 1993 1994 1995 1996 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Year GDP growth annual (%) Non-resource GDP growth (annual%) Source: IMF 2016a. Note: 2017 data projected. efforts to establish and operate a Sovereign Wealth Fund. is likely to persist, at least in the short term. Govern- This was legislated in 2015 but had no inflow of funds up ment revenues and grants, as a percentage of GDP, have until the end of 2016 (DoT 2016). fallen slowly since 2012 and were projected to decline fur- ther in 2016 and 2017. In 2015, government revenues fell The government’s fiscal status directly limits the avail- substantially due to falling commodity prices that reduced ability of resources to the health sector, a situation that the country’s fiscal space and increased government debt. Background 5 Figure 2–3: General Government Revenue and Expenditure (1983–2022) 20 30 18 25 16 14 20 Billions (current LCU) 12 (% of GDP) 10 15 8 10 6 4 5 2 0 0 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 2015 2017 Proj. 2019 Proj. 2021 Proj. Revenue Expenditure Revenue as a percentage of GDP Source: IMF 2016a. Note: The PNG Government committed to gradually eliminating the deficit just after 2021, so expenditure will likely be lower than the IMF projection. In 2016, the outcome for tax revenues and dividends was and education sectors will be challenging in the medium lower than budgeted (even though commodity prices term. had been factored in), indicating significant constraints in the government’s ability to raise domestic revenues (DoT 2017). Continuous shortfalls of revenue in recent years Demographic Health 2.2  reflect an inability to translate growth in the resource sec- Context tor into increased government revenue as well as issues PNG had a total population of 7.6 million in 2015 with domestic tax collection. (World Bank 2016). The total fertility rate is relatively high Investing in health and education will be challenging at 3.8 (2014), as is population growth at 2.1 percent (2015) in the short term. The IMF (2015) has stressed “the need (Figure 2–4). UN population projections suggest that PNG’s to maintain prudent macroeconomic policies to ensure population will increase by one-quarter to over 10 million debt sustainability.” The IMF noted the importance of by 2030 (United Nations 2015). The age distribution of the safeguarding social sectors within the context of “needed population is an important factor impacting the utilization strong fiscal consolidation.” With an economic backdrop of health services; younger and older subgroups gener- of “tight fiscal policies,” and competing needs such as the ally have much higher utilization rates. PNG’s population 2017 general election and preparations for the Asia Pacific is relatively young: about 38 percent of the population is Economic Cooperation (APEC) conference 2018 (a K 3 bil- below 15 years of age, and only 3 percent is above 65 years lion commitment), prioritizing investment in the health of age. The median age is only 21. 6 Health Financing System Assessment Box 2–1: Data Challenges in PNG During the compilation of this report, many inconsistencies and challenges in economic, fiscal, and health-related data were apparent. The magnitude of these challenges makes answering fundamental questions surprisingly difficult. The answers can significantly change the narrative around particular issues and therefore affect the ability to adequately plan and implement policy and make effective policy recommendations. This in itself represents a constraint to effec- tive health service delivery in the country. The challenges include: • Data fragmentation. Data relevant to the health sector is held by many agencies, without open access. Budget data is maintained in different central agencies and provincial governments, and health system data on infrastruc- ture, staff, and outcomes is held in various units within NDoH and by implementing partners. • Quality of data. Health sector budget data captured in the Integrated Financial Management System (IFMS) and in budget reports varies significantly from report to report and actual expenditure data is recorded with significant (up to 50 percent) and inexplicable deviations from budgets. Data on health outputs and outcomes is derived from provincial government reports, which are not verified. • Data sharing. Access to data often requires formal requests between department heads, and these are sometimes not endorsed. No agencies provide online data repositories in an accessible format, even for data that is in the pub- lic domain (for example, detailed budget outcome figures after completion of the budget year). • Inconsistency between national and international data. Information available in both national and international repositories is often difficult to reconcile or differs significantly (for example, health workforce ratios, health access statistics, budget figures, and GDP). More detail of key data challenges is provided in Appendix One so that readers and future researchers can understand how this affects analysis and successful planning and service delivery in the sector. Figure 2–4: Fertility Rate and Population Growth (1967–2015) 6.5 3 Population growth 6 (right axis) Average number of children 5.5 2.5 Percentage (%) 5 4.5 2 Total fertility rate (left axis) 4 3.5 1.5 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 Source: World Bank 2016. Access to health services in PNG is particularly chal- condition and many villages can only be reached on foot lenging as most of the population (86 percent) live in (Department of National Planning and Monitoring 2015). rural villages which are often difficult to access. The In coastal areas, the main mode of transport is by boat. In challenging terrain of PNG limits the population’s access addition to rural challenges, an increasing population in to health services and health providers’ ability to reach unplanned urban settlements is presenting a new set of remote populations for even basic programs such as health challenges. immunization. Less than 39 percent of roads are in good Background 7 Figure 2–5: Population Health Outcomes (1967–2015) 180 65 Life expectancy (right axis) 160 60 Mortality rate (per 1,000 live births) 140 120 55 Under-5 mortality rate 100 Years (left axis) 50 80 60 45 Infant mortality rate 40 (left axis) 40 20 0 35 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 Source: World Bank 2016. Figure 2–6: International Life Expectancy Comparison (1960–2014) 75 70 65 60 Years 55 50 45 40 35 1960 1963 1966 1969 1972 1975 1978 1981 1984 1987 1990 1993 1996 1999 2002 2005 2008 2011 2014 East Asia & Pacific (IDA & IBRD) World Pacific Island Small States Lower Middle Income Papua New Guinea Low Income Source: World Bank 2016. PNG still lags other countries in the region and those of declined from 159 per 1,000 live births in 1967, to 89 in similar income status in a number of health indicators. 1990, and 57 in 2015, however, PNG failed to reach the Mil- Life expectancy rose from 43 years in 1967 to 56 years in lennium Development Goal (MDG) of an under-five mor- 1990, and 63 years in 2014 (Figure 2–5). The pace of increase tality rate of 30 per 1,000 by 2015, and is far from the new in life expectancy has declined notably since 1980, which SDG of 25 per 1,000 by 2030 (UNICEF6 et al. 2014). is uncharacteristic for a country at PNG’s level of economic development (Figure 2–6). The under-five mortality rate 6.  UNICEF: United Nations Children’s Fund. 8 Health Financing System Assessment Figure 2–7: Country Comparison of Population Health Outcomes (2014) Life expectancy Infant mortality 85 125 Nigeria 75 Ghana Vietnam Sri Lanka Lao PDR Papua New Guinea 75 50 Timor-Leste Samoa Tonga Kiribati Cambodia Vanuatu Cambodia FSM Fiji FSM Kiribati 25 Solomon Islands Rate per 1,000 live births Lao PDR Timor-Leste Solomon Islands Vietnam Fiji 65 Samoa Vanuatu Years Papua New Guinea 10 Tonga Ghana Sri Lanka 5 55 Nigeria 2 LOWER UPPER LOWER UPPER LOW MIDDLE MIDDLE HIGH LOW MIDDLE MIDDLE HIGH 45 INCOME INCOME INCOME INCOME 1 INCOME INCOME INCOME INCOME 250 500 1000 2500 10000 35000 100000 250 500 1000 2500 10000 35000 100000 GNI per capita, US$ GNI per capita, US$ Source: World Development Indicators database. Note: Both y- and x-axes logged. Although infant mortality has declined by one-third to continue (Figure 2–8). Lower respiratory infections are since 1960 (to 45 per 1,000 live births in 2015), life responsible for the largest share of the overall disease bur- expectancy and infant mortality rates are worse than den, causing 8.3 percent of DALYs lost due to morbidity and expected for PNG’s income level (Figure 2–7). Further- premature mortality in 2015 (Table 2–1). NCDs such as dia- more, PNG’s outcomes compare especially unfavorably to betes have rapidly increased in PNG between ­ 1990–2015. better-performing countries such as Vietnam, Sri Lanka, High fasting plasma glucose, smoking, and household air Cambodia, and other Pacific countries. PNG is also off- pollution are prominent risk factors (IHME 2016). track for the maternal health MDG. The maternal mortality rate (MMR) remains high at 220 per 100,000 live births.7 Malaria, HIV and AIDS remain significant contributors Malnutrition rates are high and persistent: almost one- to the disease burden, although malaria’s share is on half of children under five years of age are stunted and 28 the decline due to decreasing incidence and preva- percent are underweight. Malnutrition indicators are rela- lence. The World Health Organization (WHO) estimates tively high, even among richer income groups. that 94 percent of the population lives in high-­transmission areas (>1 case per 1,000 population). The prevalence of HIV is estimated at 0.7 percent. HIV is a mixed epidemic as key 2.3 The Disease Burden subpopulation categories (sex workers, MSM,9 and trans- gender people) living in urban areas have a dispropor- PNG is undergoing an epidemiological transition, like tionately high incidence. Certain rural areas are, however, other countries in the region. The NCD share in PNG is characterized as geographic “hotspots” due to high levels rising rapidly. In 1990, 38 percent of disability-adjusted of (unprotected) sexual partner turnover and concurrency. life years (DALYs)8 were attributable to NCDs; by 2015 this number had risen to 54 percent and this trend is expected Tuberculosis (TB) is a serious public health issue and the leading cause of death in PNG. According to the 7.  Probably higher. Other studies suggested 541/100,000 based on WHO Global TB Report 2017, the estimated TB incidence death audits and hospital data; or 730/100,000 (Vince 2017). rate was 432 per 100,000 population (35,000 cases) 8.  DALYs refer to aggregated healthy years of time lost at the population level because of disease-related morbidity and premature mortality. 9.  MSM: Men who have sex with men. Background 9 Figure 2–8: Disease Burden Attributable to Communicable Diseases, NCDs, and Injuries (1990–2015) 100% 10% 11% 11% 11% 11% 11% Proportion % of disease burden 80% 38% 40% 43% 46% 50% 54% 60% 40% 52% 49% 46% 20% 43% 39% 35% 0% 1990 1995 2000 2005 2010 2015 Communicable diseases Noncommunicable diseases Injuries Source: Institute of Health Metrics and Evaluation 2016. Note: The burden of disease data is drawn from hospital records of death that are limited in PNG, so the data may not be fully accurate. Table 2–1: Top Ten Causes of Morbidity and Premature Mortality (Various Years) Rank in DALYs 2015 Condition 1990 2000 2015  1 Lower respiratory infections (including TB) 14.8% 12.9% 8.3%  2 Ischemic heart disease 3.5% 4.3% 6.3%  3 Cerebrovascular disease 4.6% 4.9% 5.5%  4 Chronic obstructive pulmonary disease 4.1% 4.3% 4.7%  5 Diabetes mellitus 1.7% 2.6% 4.2%  6 Asthma 2.8% 3.0% 3.4%  7 Road injuries 2.5% 2.6% 2.9%  8 Iron-deficiency anemia 2.1% 2.3% 2.6%  9 Diarrheal diseases 5.6% 4.3% 2.6% 10 Neonatal preterm birth complications 3.4% 3.4% 2.6% Source: Institute of Health Metrics and Evaluation 2016. and the prevalence rate was 529 per 100,000 popula- lack of an active case-finding program. The NCD has been tion (39,000 cases). The estimated mortality rate, exclud- declared the most important TB hotspot, with five times ing deaths associated with HIV, was 44 per 100,000 the national average TB case notification and contains 25 (3,600 deaths). percent of the total TB caseload (NDoH 2015). Multidrug-resistant (MDR) and extensively drug-resis- tant (XDR) TB occurrences have risen to unprecedented 2.4 Health System Structure levels in hotspots across PNG, particularly in Western PNG has a government-funded health system through- Province, Central Province, Gulf Province, and the National out much of the country. It is supplemented by gov- Capital District (NCD). Western Province has the highest ernment-subsidized health services provided by various number of drug-resistant cases in PNG. All three districts Christian missions. At the time of independence, a central- in the Western Province have experienced a doubling of TB ized NDoH managed the whole health system. Since inde- notification rates since 2011. In Daru Island, Western Prov- pendence, the government has made successive attempts ince, an unprecedented outbreak of MDR-TB is occurring to decentralize the provision of services to provincial and with a prevalence of nearly 1 percent. Patched data also district governments, including the introduction of two indicated a high number of MDR-TB cases in the NCD and significant Organic (Constitutional) Laws (1977 and 1995). Gulf provinces. The actual number of cases may, however, This legislation attempts to give provinces and local-level be significantly higher than current estimates due to the 10 Health Financing System Assessment governments (LLGs) increased control over health service Services have each negotiated a separate funding agree- delivery and resources. The 1995 Organic Law devolved ment with NDoH. It is estimated that churches provide primary health care (PHC) services to the provincial level. 47 percent of primary health services, and a significant Hospitals (including their budgets, human resources, and share of secondary services, particularly in rural areas. The payroll) are, however, managed by an autonomous board, role of church health services varies significantly across which is independent of the NDoH. This structure contrib- the country. In rural areas, church health services provide utes to a diffuse and fragmented health delivery system. up to 60 percent of primary health services and, in some provinces such as West Sepik (Sandaun), CHS accounts for To address health system fragmentation, the 2007 Pro- 80 percent of all health services provided (Christian Health vincial Health Authority Act enabled provincial gov- Services 2013).11 ernments to establish a PHA to be responsible for both primary and secondary health care in their province. The National Health Administration Act 1997 required This legislation is supposed to streamline the provision of the NDoH to develop a set of national health stan- health services at the provincial level and bring together dards to operationalize the NHP. In 2011, the Medical the provincial health departments, hospitals, and district Standards Division, through extensive consultation with health services under one management board, however, clinicians, public health, and health service managers, under this legislation, it is not mandatory for provincial developed the National Health Service Standards (NHSS) governments to establish a PHA. for PNG. These standards contribute to the government’s Vision 2050 and were endorsed by the National Health By 2016, 10 provinces have fully established PHAs, Board at their March 2011 meeting. and an additional five provinces plan to establish a PHA by the end of 2017 (NDoH 2017). Initial implemen- The NHSS outlined a structure for health service pro- tation of the PHA was slow in the three pilot provinces, as vision in PNG for the following 10 years. The standards many staffing, financial management, and institutional are: (i) an important tool for national, provincial, and dis- hurdles had to be overcome. Since then, implementation trict planning and delivery of health services; and (ii) a has accelerated and an independent review found early blueprint for providing safe, quality care consistent with success in the financial management and service delivery the NHP objectives. The minimum standards envisage a support PHAs are providing. hierarchical structure for health services across the nation: Level One village aid posts; Level Two community health The model suffers from an inconsistent legal and regu- posts; Level Three health centers and/or urban clinics; latory framework. A key example is that provincial gov- Level Four district hospitals; Level Five provincial general ernments should immediately pass the funding for rural hospitals; Level Six regional referral hospitals; and Level health service delivery to PHAs after it has been trans- Seven national referral hospital(s) offering secondary and ferred from the national level (see Section Three: Health complex tertiary-level clinical services. Expenditure Trend Analysis for a mapping of these flows). This is only a voluntary arrangement, however, that allows The NHSS specify: some provinces to retain a share of funding for their own activities. (i) core clinical and PHC services such as medical, sex- ually transmitted infections (STIs), and maternal and Future strengthening of the PHA model requires com- child health; mitments in terms of adequate levels of resourcing, timely release of funds, recruitment of skilled staff, (ii) support services, such as diagnostics (medical imag- improved communication, coordination and effective ing and pathology, pharmacy, drugs, and medical reporting between all stakeholders, and sufficient lev- supplies), infection control, and public health; els of support from national departments. If such issues (iii) management support, including leadership, teacher are resolved and capacity constraints are removed, the training, and health information; PHAs may prove to be a successful model that unifies all service-delivery functions within the province under one (iv) health workforce productivity and capability; management structure. (v) drugs, equipment, and medical supplies; and Church health services are the key partners in deliver- ing health services in PNG.10 CHS and Catholic Church (vi) health infrastructure (that is, buildings that meet minimum standards) (Government of PNG 2011). 10.  CHS only administer funding to various church organizations for the delivery of health services and liaise centrally with NDoH. They are 11.  The Catholic Church is not part of CHS and has negotiated a not technically service providers, but this report will refer to CHS as such separate MoU for funding. The Catholic Church is the largest church throughout the report. provider. Background 11 Figure 2–9: Governance Structure of the Health System DPLLGA/ Development DIRD PLLSMA Treasury Planning partners NEFC Finance Governor on Church health PSIP Province/PHA NDoH & HMS services Open member District on DSIP Level 2–4 Level 5–7 Church facilities (hospitals) facilities Planning Budgeting Reporting Any dotted capital investment Level 1 (aid posts) Source: Based on a consultative process between World Bank and NDoH staff. Note: (i) Not all DPs provide funding directly to Church Health Services; (ii) HMS: Hospital Management Services. Governance mechanisms in the sector remain convo- service delivery; (iii) the Departments of Finance and Trea- luted and, in many ways, ineffective. While the PNG sury, which budget for, and release operational expendi- Government has goals to streamline the health system ture, and receive financial reports on budget expenditure structure and service delivery, in practice, governance and performance (but very inconsistently) from provinces is fragmented in many provinces due to the persistent and NDoH; (iv) the Department of Planning, which bud- ambiguity surrounding functional roles and responsibili- gets for the capital investment component and receives ties. Figure 2–9 shows the governance mechanisms of the performance reporting on infrastructure projects; and health sector, with respect to planning, budgeting, and (v) the National Economic and Fiscal Commission (NEFC), reporting channels. which budgets for rural facilities in provinces, but does not receive financial or performance reports from provinces An extensive list of central agencies receives siloed directly. reporting data. These include: (i) the Department of Implementation and Rural Development (DIRD), which This reporting structure creates a clear mismatch receives reports specifically on provincial and district infra- between those agencies that budget, plan, and set structure out of PSIP and DSIP12 funding; (ii) the Depart- standards; the agencies that monitor adherence to ment for Provincial and Local-Level Government Affairs standards and receive financial and performance (DPLLGA), which receives reports on provincial health reports; and the agencies that can impose sanctions (for example, by withholding funding). For example, 12.  PSIP: Provincial Services Improvement Programme; DSIP: District NEFC and NDoH set provincial operational budgets and Services Improvement Programme. service delivery requirements, but PSIP reporting data 12 Health Financing System Assessment is received by DIRD, DPLLGA, Treasury, and Finance (all to replenish the aging workforce. The School of Medi- of whom rarely share information with each other). This cine and Health Sciences is the only accredited medical creates poor incentive and accountability structures that school in PNG, and it graduated 316 doctors during the severely limit the departments’ ability to pressure prov- period 2010–16. Divine Word University is in the process inces to pursue beneficial reforms. of establishing the second medical school in the country. Recognizing the lack of training capacity, the Government of PNG and donors have invested in more training schools 2.5 Human Resources during 2012–16. This will not resolve the urgent shortage for Health of health workers, however, as it takes years to produce a qualified health worker. PNG is facing a health worker shortage that could worsen in the medium to long term. The doctor ratio The quality of the existing health workforce is incon- per 1,000 population was 0.06 in 2009 and 0.07 in 2016; sistent. This is partly due to the regulatory bodies poor while the ratio of Community Health Workers (CHWs) to accreditation and audit processes and the inability of the 1,000 population dropped significantly from 0.66 to 0.49. Nursing Council and Medical Board to update the cur- The aggregate number of health workers is consider- riculum for nurses and health workers in a timely, routine ably lower than the WHO minimum threshold of 2.28 per manner. The NDoH and provincial authorities continue to 1,000 population that is required to achieve the health- implement strategies to combat shortages of health work- related MDGs. This critical shortage is impacting the PNG ers in rural and remote areas, with very limited success. health system’s ability to reduce maternal mortality rates. The government has endorsed a Health Workforce The workforce is inequitably distributed between rural Enhancement Plan, which lays out priorities for work- and urban areas, with a substantial predisposition force strengthening. The eight nursing schools in PNG toward urban areas. The NCD has the highest number of trained 1,364 nurses during the period of 2010–15, and employed health workers (15 percent) with the presence of the 13 CHW training schools trained 2,326 CHWs during Port Moresby General Hospital, followed by Central Prov- the period 2010–16. The number of nurse and CHW gradu- ince (9 percent), while Hela has the least (1 percent). West ates will increase hereafter with the addition of five nurs- New Britain has the highest number of vacant positions ing schools and four CHW training schools (NDoH 2016). (89 percent) and Milne Bay has the least (8 percent). Rea- Despite an increase in the number of nurses being trained, sons for the higher numbers of vacancies include recruit- only 56 percent of available nursing positions were filled ment funding constraints, remoteness of health facilities, in 2016. rural-urban migration, and an increased shift from clinical support to administration (NDoH 2016). Timely and more accurate information on human resources in the health profession is needed. The last The existing health workforce is rapidly aging, com- comprehensive survey of the health workforce was con- pounding the health worker shortage issue. In 2009, ducted in 2009. This renders assessing progress and short- roughly 16 percent of the workforce was 55 years of age or falls in human resources challenging and represents a key older, which qualified them for retirement. An additional constraint to effectively planning reforms. Undertaking 38 percent was aged 45–54, meaning they will be eligible a comprehensive assessment and maintaining consis- for retirement inside the following decade. Just 12 percent tent records of the workforce is a priority for the depart- of the workforce was under 35 years old. Loss of health ment. Coordination between the National Department workers to retirement decreases service capacity. of Personnel, NDoH, and central agencies needs to be strengthened. For example, positions on the payroll need Another compounding factor is PNG’s lack of signifi- to be reconciled with NDoH records to avoid possible mis- cant progress in medical and health-training capac- matches between positions and payments.13 ity, which has resulted in a shortage of qualified staff 2.6 Health Care Access Table 2–2: Health Workers per 1,000 and Utilization Context Population (2009 and 2016) There are significant disparities in access to, and uti- Category Baseline (2009) 2016 lization of, health services, particularly between rural Doctors 0.06 0.07 and urban areas. As discussed earlier, the terrain of PNG presents a challenge to access and provision of health Nurses 0.49 0.44 CHWs 0.66 0.49 13.  It should be noted that some progress has been made on the Source: NDoH 2016. Nursing Council’s registration data. Background 13 care services. The poorest quintile of the population cites initiatives, preventive child health programs (for example, staff absenteeism, the cost of health care, distance, lack of immunization), and community health education. In geo- transport, transport costs, and the quality of health care as graphically challenging settings, immunization coverage the main barriers to visiting a health facility in the case of directly correlates with the prevalence of outreach patrols. illness (Irava et al. 2015). Several aid posts have been closed, partly due to lack Maternal health is another area that highlights chal- of human resources, and rural outreach has continued lenges in access and utilization of health services in to fall. There were 42 outreach clinics per 1,000 children PNG. Only one-third of women have access to modern under-five in 2010 but this had fallen to a ratio of 37 per contraceptive methods. Only 66 percent of pregnant 1,000 children under-five in 2015 (Figure 2–11). Anec- women attend four or more antenatal care (ANC) vis- dotal reasons for this reduction, cited by health manag- its and less than 53 percent of women deliver with the ers, include shortages and delays in operational funding assistance of a skilled attendant (Table 2–3). These indica- and shortages in available personnel. This is due to a slow tors have decreased since 2014 in the government’s own release of funds from the central level and delays in fund National Health Information System (NHIS) (NDoH 2016) distribution and support provision at the provincial level (Figure 2–10). (see also Figure 3–2). In 2013, 55 percent of facilities did not receive external funding for outreach clinics; opting to Rural outreach plays a critical role in ensuring access rely on user fees, own salaries or not providing the service to health services for rural and remote populations. at all (Howes et al. 2014). Outreach provides a key platform for maternal health Table 2–3: Comparison of Access to Health Services and Infrastructure Women Receiving ANC Four DPT3 Population Family or More Immunization Using Population Planning Times Skilled Birth Coverage Improved Using Needs from Any Attendance among Drinking- Improved Country/ Satisfied Provider at Delivery 1-year-olds water Sanitation Region (%) (%) (%) (%) Sources (%) Facilities (%) Cambodia 56 89 89 89  76 42 Fiji 44 98 100 99  96 91 Micronesia, 55 80 100 72  89 57 Fed. Sts. Kiribati 22 88 80 87  67 40 Lao PDR 50 53 42 89  76 71 PNG 32 66 [63] 53 [37] 62  40 19 Samoa 27 93 83 66  99 91 Solomon 35 91 86 98  81 30 Islands Timor-Leste 22 84 30 76  72 41 Tonga 34 99 98 82 100 91 Tuvalu 31 93 98 96  98 n.a. Vanuatu 49 76 89 64  95 58 Vietnam 76 96 94 97  98 78 EAP 48 89 85 85  86 69 Low-income 30 80 58 80  72 36 Lower- 48 86 78 84  85 61 middle- income Source: World Health Organization and World Bank 2017. Note: (i) DPT3: Diphtheria, pertussis, tetanus. (ii) 2015 data from PNG Government’s NHIS in square brackets [ ] if available. (iii) n.a. not available. 14 Health Financing System Assessment Figure 2–10: Selected Health Access Indicators (2005–15) 100% 100 90% 90 80% 80 Skilled birth attendance 70% 70 Measles immunization (% of children under 1) 60% 60 Pentavalent (% of children 50% 50 under 1) 40% 40 Antenatal care 30% 30 Family planning (couple years protection per 20% 20 1,000 WRA) (right axis) 10% 10 0% 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: NDoH National Health Information System 2016. Note: WRA: women of reproductive age (15–44). Figure 2–11: Selected Service Delivery Indicators (2006–15) 100% 200 90% 180 80% 160 % of facilities with at least one 70% 140 supervisory visit per annum % of months that facilities do not 60% 120 have stock-outs of selected supplies 50% 100 Outpatient visits per 100 people 40% 80 (right axis) 30% 60 Rural outreach clinics per 1,000 children (right axis) 20% 40 10% 20 0% 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: NHIS 2016. PNG has experienced an overall decline in the utiliza- was comparable to the low-middle-income country aver- tion of outpatient services between 2010 and 2015. age of 4.5 percent and to other countries in the region: Fiji The average number of outpatient visits to a health facil- (4.5 percent), the Solomon Islands (5.1 percent), and Vanu- ity (health centers and hospitals) per person, per year atu (5.0 percent) (Figure 2–13) (World Bank 2016). dropped from 1.62 (2010) to 1.23 (2015) (Figure 2–11). Rates varied across provinces but the decline in utiliza- The budget allocated to the health sector as a share tion was consistent across all provinces except two—­ of GDP represents the fiscal costs of health policy Bougainville and Oro. relative to the size of the overall economy. Govern- ment allocation to health as a share of the total budget, in turn, reflects the fiscal costs of government health policy. 2.7 Health Financing Health spending, as a share of general government spend- ing, decreased from 13.2 percent in 2013 to 9.5 percent in Despite steady growth, Total health expenditure (THE), 2014 (World Bank 2016).14 as a proportion of GDP, has been stagnating between 4 and 5 percent since 2007. As Figure 2–12 illustrates, THE rose rapidly between 2000–01, a 2.7 percent increase in one 14.  It should be noted that 2014 is the latest year available for this year. THE peaked in 2004 at 8.4 percent, before falling by 4.4 information in the World Development Indicators (World Bank 2016) percent between 2004 and 2007. In 2014, THE (4.4 percent) database. Background 15 Figure 2–12: THE as a Proportion of GDP (1995–2014) 10.0% Health expenditure (% of GDP) 8.0% 6.0% 4.0% 2.0% 0.0% 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 Health expenditure, total (% of GDP) Health expenditure, public (% of GDP) Source: World Bank 2016. Figure 2–13: THE and Public Health Expenditure (PHE) as % of GDP in East Asia Countries Total health expenditure Public health expenditure 20 20 15 15 FSM FSM 10 Kiribati 10 Cambodia Samoa Kiribati Samoa Vietnam Ghana 5 Solomon Islands 5 Share of GDP (%) Share of GDP (%) Papua New Guinea Tonga Solomon Islands Nigeria Fiji Papua New Guinea Vanuatu Tonga Sri Lanka Ghana Vanuatu Fiji Vietnam 2 Lao PDR 2 Cambodia Sri Lanka Timor-Leste Timor-Leste Nigeria 1 1 Lao PDR LOWER UPPER LOWER UPPER LOW MIDDLE MIDDLE HIGH LOW MIDDLE MIDDLE HIGH INCOME INCOME INCOME INCOME INCOME INCOME INCOME INCOME .5 .5 250 500 1000 2500 10000 35000 100000 250 500 1000 2500 10000 35000 100000 GNI per capita, US$ GNI per capita, US$ Source: World Development Indicators database. Note: Both y- and x-axes logged. Real THE (private and general government expendi- Overall health expenditure is exposed to volatility ture on health) per capita was K 86 (using constant from a variety of sources. In the absence of an effective 1998 prices) in 2014 (Figure 2–14). This marks a signifi- sovereign wealth fund, government expenditure is highly cant rise from K 60 in 2007, but a decrease in expenditure dependent on the international commodity price cycle from 2012 (K 88) and 2013 (K 95) levels. When compared and the performance of resource projects in the country. to 2014 international and regional standards, PNG’s health Donor financing represents a large share of PHE and has expenditure per capita (US$92) is low in comparison with been historically volatile. In addition, within a given bud- the low-middle-income countries average (US$265), EAP get year, cash flow tends to be inconsistent as the gov- (US$643), Fiji (US$204), the Solomon Islands (US$102), and ernment receives revenue lumped around corporate tax Vanuatu (US$158). collections in May, August, and September. This reduces 16 Health Financing System Assessment Figure 2–14: Health Expenditure per capita (1995–2015) (Current US$ and Constant LCU) 120 140 Real THE per capita (right axis) 120 100 100 Constant 1998 LCU 80 Current US$ 80 60 60 40 Nominal THE (left axis) 40 20 20 0 – 1995 1997 1999 2001 2003 2005 2007 2009 2011 2013 Source: World Bank 2016. Figure 2–15: Trends in the Structure of THE (1996–2014) 100 80 60 40 20 0 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Private Health Expenditure (minus OOPs) as % of THE Out of Pocket Expenditure (OOPs) as % of THE General Government Health Expenditure as % of THE Source: World Health Organization 2016. cash flow in the first quarter of the calendar year, as there Figure 2–16: Health Financing Mix (2014) currently is no effective mechanism to smooth expendi- ture through the usage of short-term debt instruments, General Government Health 10 such as treasury bills. Since public expenditure represents Expenditure (GGHE) as % of the main source of health sector funding, fluctuations in Total Health Expenditure 19 this component have a strong, direct effect on THE. Private Health Expenditure (PvtHE) as % of Total Health Expenditure (THE) Health financing is predominantly centralized. In 2014, 81 Out of Pocket Expenditure government spending—including government spending (OOPs) as % of Total Health financed by external sources—accounted for over 80 per- Expenditure (THE) cent of total health spending; the remaining 20 percent was attributed to private expenditure (Figure 2–16). In Source: World Health Organization 2016. comparison, the percentage of government expenditure as a share of THE for Fiji was 66 percent, the Solomon Islands (92 percent), and Vanuatu (90 percent) indicating Background 17 Figure 2–17: External Resources on Health as % of THE (1995–2014) External resources on health as % of THE 35 30 25 20 15 10 5 0 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 External resources on health as % of Total Health Expenditure (THE) Source: World Bank 2016. that PNG Government spending is in the mid-range for the the poor do not visit health facilities in the case of illness immediate region (World Bank 2016). (Irava et al. 2015). Only 10.46 percent of total health spend- ing is attributed to OOP payments. OOP expenditure was External resources on health,15 as a percentage of less than 30 percent of total consumption in every house- THE, continue to play a significant role in PNG’s hold in PNG. Only 0.02 percent of households spent more health financing. In 2014, external resources accounted than 20 percent and less than 1 percent of households for 21 percent (Figure 2–17) and have stabilized around spent more than 10 percent of their budgets on OOP pay- 20 percent in the preceding years. This share is high when ments. An estimated 0.3 percent of households were, how- compared to the low-middle-income average of 3.3 per- ever, impoverished because of OOP health spending, and cent and some Pacific countries, for example, Fiji (9 per- 3 percent of the poor were pushed further into poverty. cent). The percentage is, however, lower than for many others in the region: Timor-Leste (31 percent), the Solo- The government has renewed its intention to reduce mon Islands (56 percent), and Vanuatu (48 percent). OOP spending by introducing a fee-free basic and sub- sidized specialized health care policy that requires all PNG has relatively low out-of-pocket (OOP) expendi- facilities to stop charging user fees for primary care. ture on health by international standards. Given com- K 20 million has been allocated to compensate facilities paratively low access rates, there may be a significant for the loss of user fees—in addition to the existing health degree of foregone care that reduces OOP spending, and function grant (HFG) that funds facility operations. The low OOP does not necessarily mean financial protection. In persistent slow release of funding by central agencies and fact, health care costs are cited as one of the main reasons delays in channeling funds at the provincial level (see also Section Three for further discussion) means that facilities 15.  External resources for health are funds or in-kind services that are do not receive the financial support required to function provided by entities not part of the country in question. The resources may come from international organizations, other countries through without user fees. This has prevented the successful elimi- bilateral arrangements, or foreign NGOs. These resources are part of THE. nation of user fees at primary public health facilities to date. section three Health Expenditure Trend Analysis budget books.16 It is also assumed that all PHA allocations Key messages: and HFGs are spent on service delivery at the subnational level. • A snapshot of THE from 2010–15 reveals signifi- cant volatility in government spending and mis- General government expenditure almost doubled alignment between appropriations and actual from K 8.1 billion in 2010, to K 15.5 billion in 2014, expenditure—both under and over budget. before falling back to K 13.8 billion in 2016. The health • Although the majority of the population resides in sector budget increased significantly from 2010 to 2016 rural areas, the budget for frontline service expen- (Figure 3–1), but this likely overstates the actual avail- diture is quite small. NDoH, followed by hospitals, ability of resources to the sector. In 2015 and 2016, both account for the largest expenditure shares. warrants (an authorization from DoT which allows agen- cies to access budgeted funds during budget execution) • The untimely release of warrants results in det- and expenditure were below the budget allocation as cash rimental delays and inefficient use of resources. rationing prevented the full release of budgeted funds. Timely disbursement is critical to improved health outcomes. Expenditure likely gives a poor indication of actual • The 2015 and 2016 health budgets were adjusted spending, due to issues with the government’s to meet the government goal of a balanced bud- accounting system (see also discussion on data issues get by 2021. in Appendix One), so warrants should serve as the main indicator of the actual availability of funding to the • The 2017 health sector budget represented the sector. Warrants increased by about one-third from 2011 first budgeted reduction of K 314 million or 19 per- to 2014, and have remained constant since then. During cent, from K 1,647 million in 2016 to K 1,333 mil- the period from 2011 to 2016, total government expendi- lion in 2017. ture increased by 10 percent per annum, while the health sector budget increased by an average of 12 percent per annum, annual health sector warrants increased by 7 per- This section looks at current and historical budget allo- cent per annum, and health sector expenditure increased cation patterns within the PNG health sector. The focus by 9 percent per annum. Taking either expenditure or war- of any analysis of health financing is to gain as much infor- rants as an indication, the health sector, therefore, grew mation about budget allocations to improve efficiency more slowly than total government expenditure during and, in turn, service delivery. It relies primarily on data this period. from the government’s accounting system. Health sector spending referred to in this section does not, therefore, In nominal terms, government health expenditure include any sector spending by provincial governments started to increase significantly in 2013, when the (apart from PHAs or HFGs and so excludes internal rev- enue), salaries of health workers at subnational facilities, and donor spending that is not included in the national 16.  Most significantly, this excludes funding from Gavi and GFATM. 19 20 Health Financing System Assessment Figure 3–1: Government Health Sector Operational Budgets, Warrants, and Expenditure (2010–17) (millions of Kina) 1,800 1,600 1,400 1,200 Kina millions 1,000 800 600 400 200 2010 2011 2012 2013 2014 2015 2016 2017 Original budget Revised budget Warrants Expenditure Trend warrants Source: DoT from IFMS data 2017. incoming government changed from the previous managed by the governor (in a province) and open mem- path of fiscal consolidation to a path of rapid fiscal ber (in a district), a part of which is available for invest- expansion with commensurate increases in the deficit. ments in health infrastructure. The flow of funds from the The increase was not, however, distributed evenly across central level to provincial governments and health facili- agencies. From 2011 to 2016, NDoH funding decreased by ties is mapped in Figure 3–2. The budget also allocates almost 3 percent, as the budget was reallocated to PHAs funding to the Institute of Medical Research (IMR) and the and HFGs. HMS expenditure increased by about 4 percent National AIDS Council (NAC) (omitted in Figure 3–2). NAC or 1.4 percent on an annual basis. is responsible for HIV policy and coordination of the HIV response. In real terms the picture is less positive, as the health sector budget increased by 9 percent per annum from Fund flows in PHA provinces differ from nonPHA prov- 2011 to 2016, while warrants increased by 4 percent, inces. It is mainly because funding for the provincial hos- and expenditure increased by 6 percent. Given popu- pital (Level 5) and rural service facilities (Level 1–4) both lation growth of approximately 2.8 percent, real expen- flow through the PHA, rather than NDoH/HMS and the diture (measured through warrants) just kept pace with provincial government (see Figure 3–3 for a comparison of population growth, even during a time of significant fis- the two fund flow models). cal expansion. Overall, the fiscal inconsistencies of the government’s accounting system and large in-year devia- Between 2011 and 2017, hospitals share of the total tions from budgets due to the revenue shortfalls in the domestic health appropriations ranged between 30 resource sector make it difficult to provide a clear and and 45 percent (Figure 3–4). This includes salaries of accurate assessment of the trend in resources available to hospital staff, other operational expenditure, and funding the sector. for rehabilitation and upgrading of facilities. NDoH expen- diture declined from 40 percent of the domestic health The health sector budget finances the NDoH hospitals, sector budget in 2011 to 26 percent in 2017.18 In addition CHS, Rural Health Services (HFGs),17 and PHAs. Prov- to wages and salaries, the NDoH budget also includes inces receive two sources of funding for frontline facilities, procurement of medicines and other consumables. The the HFG, for facility operations and discretionary funds budget for frontline service expenditure (through HFG 17.  HFGs are not intended to fully fund operational budgets. The HFG is 18.  A large share of external funding is traditionally allocated under calculated based on provinces using their own internal revenue towards NDoH in budget books, out of a requirement of the accounting system health. It finances operational budget transfers for frontline health to display donor funds within the budget of a government agency, but facilities. the majority of this is managed outside the department. Health Expenditure Trend Analysis 21 Figure 3–2: Health Sector Fund Flows Fund sources Agents Service providers National & regional Treasury NDoH & HMS hospitals Church health Finance Church facilities services Internal revenue Province Provincial hospitals Households District Level 2–4 facilities Development Level 1 facilities LLG partners (aid posts) Cash flows In-kind flows Warrants Dotted—only a minority of facilities show this flow or it is not a formal arrangement Source: Based on discussions between a World Bank consultant and NDoH staff. and CHS) is quite small although the majority of the pop- roll costs in NDoH and HMS (including all hospital staff ) ulation lives in rural areas. It has, however, been steadily continuously run over budget appropriations (Figure 3–5). growing from 16 percent in 2011 to 21.6 percent in 2017. Salaries in the civil service are paid fortnightly directly into worker’s accounts through a separate payroll system At an aggregate level, wage and nonwage expenses which does not require warrants. This means that the gov- are relatively balanced. Some 54 percent of recurrent ernment payroll costs are based on the actual number of expenditure in the health sector is allocated for wages staff on the payroll, rather than budget appropriations (an increase from 47 percent in 2011), while goods and which are not reconciled with the payroll. This leads to services make up 46 percent in 2017.19 This is shifting extensive annual payroll overruns across whole-of-gov- increasingly towards salaries, given their rigidity. The pay- ernment (K 471 million or 11 percent in 2016) (DoT 2017), which also affects the health sector and diverts funding from budgeted goods and services expenditure. 19.  Note that the salaries for rural health workers (Level 1–4 facilities) are lumped into the general provincial salaries budget and are, therefore, never shown as part of the health sector budget. It is The capital investment component (previously the estimated that approximately K 140 million are allocated to provincial separate development budget) of the health sector health staff (10 percent of the health sector budget). 22 Health Financing System Assessment Figure 3–3: Health Sector Fund Flows at the Subnational Level : Comparison Between Traditional and PHA Models Traditional Model Central government NDoH & HMS Provincial hospitals (treasury & finance) Provincial government Level 1–4 facilities & treasury District Provincial Health Authority Model Central government NDoH & HMS Provincial hospitals (treasury & finance) Provincial treasury Provincial health Level 1–4 facilities authority District Cash flows In-kind flows Warrants Source: Based on discussions between a World Bank consultant and NDoH staff. Figure 3–4: Share of Health Sector Budget by Agency (2011–17) (revised budget figures, domestic financing only) 100% 90% 80% % of revised health sector budget 70% 60% 50% 40% 30% 20% 10% 0% 2011 2012 2013 2014 2015 2016 2017 NDoH Hospitals IMR NAC CHS HFG PHAs Source: DoT IFMS data 2017. Health Expenditure Trend Analysis 23 Box 3–1: Service Improvement Programme Funding In 2013, the PNG Government introduced a new directive to target resources to provincial, district, and local levels, through direct discretionary development grants, with specific guidelines for sector allocation. The new directive of subnational financing encompassed the PSIP, DSIP, and the Local Level Government Services Improvement Programme. Based on DIRD guidelines, these funds (totalling about K 1.492 billion annually since 2013) were to be distributed according to the formula: 30 percent of all direct subnational financing to be allocated to infrastructure projects, 20 percent to health (totalling approximately K 1 billion from 2011–16), 20 percent to education, 10 percent to the law and justice sector, 10 percent to administration, and 10 percent to the economic sector. The DSIP funds were given their own set of DIRD spending guidelines approved by the National Executive Council in 2012, however, these guidelines have since been abolished. The NDoH has no oversight of these funds, as reporting is provided to the DIRD, which does not release reports to central agencies or NDoH. In addition, fund usage reporting is poor, with DIRD reporting that only 31 percent of provinces and 32 percent of districts acquitted their usage of funds in 2016. Changes in facility levels captured in the NHIS from 2011–15 do not indicate a significant increase in open Level 1–4 facilities, for which DSIP and PSIP funding is primarily responsible. In 2012, only 6 percent of facilities surveyed in a representative study reported receiving funding through DSIP in the four years prior to the survey (Howes et al., 2014). Staff in NDoH also raise concerns that infrastructure built through these funds does not comply with the NHSS, and operational sustainability is often not guaranteed. Overall, while no systematic evidence is available, it appears that PSIP and DSIP do not contribute meaningfully to the estab- lishment of health infrastructure at the subnational level. This is significant, as national investments ceased to be tar- geted to subnational infrastructure, under the assumption that they would be covered through PSIP and DSIP, leaving a clear funding gap in a critical area. Figure 3–5: Payroll Overruns in NDoH and HMS (2012–16): Personnel Emoluments Revised Budget and Actual Expenditure, Domestic Financing Only (millions of Kina) 600 70% 500 60% 60% Actual (% of budget) 400 50% Kina millions 40% 300 39% 30% 30% 200 20% 100 12% 13% 10% 0% 2012 2013 2014 2015 2016 Revised budget Expenditure Overspend (%) Source: Chart prepared by World Bank consultant based on data provided by NDoH. Note: Some overspend may be due to reporting errors in the IFMS. budget is relatively small (Figure 3–6).20 Capital invest- Capital investment component appropriations include ment is allocated mostly to nonsalary expenditure such as DP support through project support grants and con- other operational expenses and capital formation expen- cessional financing (shown separately in Figure 3–6). diture such as construction, renovation, and improvement. While no consistent breakup is available, a large share of Capital expenditure also captures spending that is opera- external funding, captured under the capital investment tional in nature, such as operational materials, supplies, component of the budget, is used for programs that are travel, and subsistence spending for particular projects. operational in nature. The capital investment component of the budget fell significantly from 2016 (K 214 million) to 2017 (K 98 million), even though actual expenditure 20.  This does not include potential funding for health infrastructure in 2016 was significantly lower than the budget (more sourced from PSIP and DSIP, as consistent data, showing how much is allocated by provinces and districts, is not available. discussion below). The Public Expenditure and Financial 24 Health Financing System Assessment Figure 3–6: Total Government Health Sector Expenditure by Fund Source (2011–17) (millions of Kina) 1,800 1,600 1,400 1,200 Kina millions 1,000 800 600 400 200 2011 2012 2013 2014 2015 2016 2017 GoPNG operational GoPNG capital DP grants DP loans Source: DoT from IFMS data 2017. Accountability 2015 acknowledged that the capital invest- fees at primary facilities. The year-to-year unpredictability ment component of the budget faces considerable under- of funding to the HFG can be seen in Figure 3–7. spends, particularly due to capacity limitations and weak planning systems. This problem worsened in 2015 and 2016 despite sev- eral commitments from central agencies to improve There is no fixed rule on the budget allocation ratio of fund flows. In 2016, only 11 percent of the HFG was maintenance to capital spending, which is reflected in released by the end of the second quarter.21 At the same the physical condition of health facilities across PNG. time, health sector agencies—which receive their full Maintenance allocation depends on the initial capital remaining balances for their operational and capital base. Given the lack of investment in health infrastructure, investment budgets toward the end of the year—face dif- it is likely that the share of capital spending for new invest- ficulties in spending their funds in time. It takes provincial ments would exceed that for maintenance. In addition to administrations about two to three months to spend or the ambiguity around the budget allocation ratio, there is transfer the money received to districts, LLGs or facilities. a general lack of consistency around the responsibility for Once again, these delays result in the inefficient use of administration and execution of the maintenance budget. resources. Continued progress in the timely disbursement of funds is critical to improved health outcomes in PNG. There are many small and microsized health facili- ties (aid posts, health centers and subcenters) that In line with the Medium Term Fiscal Strategy require substantial maintenance. A survey conducted 2013–2017, the government is gradually working ­ in 2010 (Cairns and Xiaohui 2015) found that only four of toward a balanced budget by 2021. To reduce the 2015 the 35 facilities visited were in “good” condition and most budget deficit to 3.5–4 percent of GDP, the government required substantial repair. Funding for routine mainte- approved a 2015 Supplementary Budget that required nance is reflected in the calculation of the HFG (and is, a K 28.1 million or 2 percent reduction across the total therefore, the responsibility of provinces), but provinces health sector budget. The savings were realized from capi- do not allocate sufficient funding for this activity. Facility tal investment projects across hospitals (K 25.5 million) upgrades and new facility construction should be funded and from the operational budget of NDoH (K 2.6 million). out of PSIP and DSIP, but the available evidence suggests All other health sector agencies were expected to receive that this is not taking place in a reliable, systematic way. the full amount of 2015 appropriations.22 NDoH expendi- ture figures show a significant underspend, which is likely Warrant Release caused by errors in the government financial manage- ment system. The release of funds to rural health services is often untimely and unpredictable. This results in disruptions to health service delivery, hinders health managers’ capac- 21.  Based on reports from the IFMS. 22.  Actual expenditure is most likely reported with errors in the 2015 ity to implement planned activities, contributes substan- budget, as NDoH shows a large underspend and HMS a large overspend tially to inefficiencies, and prevents the elimination of user (see also separate discussion on data limitations in PNG in the appendix). Health Expenditure Trend Analysis 25 Figure 3–7: Cash Release for HFGs by February (Province and Region) (2008–12) Year Highlands Islands Momase Southern 100% to Feb 2012 50% In 2012, everyone got zero 0% 100% In 2011, every except Southern Highlands province got 50%, which was good to Feb 2011 50% 0% 100% to Feb 2010 50% 0% 100% In 2009, all Highlands and 4 Southern provinces get zero to Feb 2009 50% 0% 100% In 2008, almost every province got 50% which was good to Feb 2008 50% 0% EHP Enga SHP Simbu WHP ENB Manus NIP WNB ESP Madang Morobe Sandaun Central Gulf MBP Oro Western Source: Cairns and Xiaohui 2015. Note: The shaded area represents the 40% target funding that provinces need to commence their service delivery responsibilities. The 2016 health sector budget was increased from the 2016 health sector budget, and not all agencies received revised 2015 Supplementary Budget by 13 percent, their full funding by the end of the year. Only K 23 million from K 1.388 billion to K 1.567 billion. One of the key of the K 225 million that was newly budgeted for hospital drivers of the increase was the transfer of K 225 million infrastructure was released by the end of the year. from the PSIP (direct development grants to members of parliament of K 5 million per district, per province) funding The health sector budget was reduced to K 1.333 mil- to support hospital infrastructure rehabilitation and rede- lion in 2017. The main reductions are found in the velopment. The government has not, however, clarified operational budget of NDoH, (which faces a K 59 million usage and distribution of these funds. Additional fund- or 18.4 percent operational budget cut, including drug ing was approved for three new PHAs in 2016. Despite an procurement and salaries) and Hospital Management overall increase to the health sector in the 2016 budget, Services (mostly due to the termination of the K 225 mil- in comparison to the 2015 Supplementary Budget, some lion appropriation for provincial hospitals). CHS received agencies were negatively impacted. The distribution of the similar funding, or a mild reduction when considering budget increases has been varied due to significant reduc- a reallocation from hospital funding to CHS late in 2016. tions in some agency operational budgets, such as CHS A positive development is the maintenance of the HFG, and the NDoH (see Figures 3–8 and 3–9). Ongoing cash which protects frontline service delivery from reductions flow constraints negatively affected the execution of the across the sector. Figure 3–8: Health Sector 2015 Budget, Warrants, and Expenditure, Domestic Financing (millions of Kina) 600 160% 500 144% 140% % of revised budget 120% 400 Kina millions 99% 100% 90% 88% 86% 300 80% 200 43% 60% 25% 40% 100 20% 0% NDoH Hospitals IMR NAC CHS HFG PHAs Original budget Revised budget Warrants Expenditure Expenditure (% of rev. budget) Source: DoT from IFMS data 2016. 26 Health Financing System Assessment Figure 3–9: Health Sector 2016 Budget, Warrants, and Expenditure, Domestic Financing (millions of Kina) 700 120% 100% 100% 98% 600 93% 92% 92% 100% 500 % of revised budget 74% 80% Kina millions 400 60% 300 40% 200 100 20% 0% NDoH Hospitals IMR NAC CHS HFG PHAs Original budget Revised budget Warrants Expenditure Expenditure (% of rev. budget) Source: DoT from IFMS data 2017. Figure 3–10: Comparison of Revised Appropriations by Agency and Economic Item Categories, Domestic Financing (2015–17) (millions of Kina) 600 500 400 Kina millions 300 200 100 0 2015 2016 2017 2015 2016 2017 2015 2016 2017 2015 2016 2017 2015 2016 2017 2015 2016 2017 2015 2016 2017 NDoH Hospitals IMR NAC CHS HFG PHAs PE G&S Capital formation Source: DoT from IFMS data 2017. Note: G&S: Goods and Services; PE: Personnel Emoluments. section four PNG Reliance on Donor Financing funding has declined significantly to US$99.1 million in Key messages: 2015 (Figure 4–1). Donor health funding, from partners such as DFaT and Gavi, is projected to decline further in • On average, 20 percent of total annual health the near future. spending in PNG is from DPs and funding is vola- tile with regards to levels, sources, and recipients. Australia remains the largest bilateral donor, but its Australia is the largest bilateral donor. GFATM and share of external resources fell from 86 percent in 2006 Gavi have become increasingly important part- ners in health service delivery in PNG. to 71 percent in 2015 (Figure 4–2).23 In 2006, Australia and New Zealand together contributed over 96 percent of • PNG has entered the accelerated transition phase external resources to the PNG health sector. By 2015, this from Gavi support and must prepare to meet strict share had fallen to 72 percent. During this period, multi- cofinancing agreements and assume responsi- lateral resources from the GFATM and Gavi came to play a bility for key functions currently provided by the substantial and important role (20.1 percent of all external donor. resources in 2015). • PNG will benefit greatly from early preparation for Donors are increasingly supporting health initiatives the eventual graduation from GFATM by unbun- through nongovernmental organizations (NGOs) and dling and gradually integrating the support into civil society rather than through the PNG Government. government–funded activities. In several cases, this was due to concerns about financial • GFATM has invested over US$190 million to date. management practices within NDoH. This raises concerns Support is targeted to the HIV and AIDS response, regarding financial and institutional sustainability. GFATM malaria, TB, and HSS. Grants are now imple- resources are directed to NGOs and civil society rather mented through NGOs and nonprofits rather than directly to the public sector. After an audit revealed than the NDoH. The designation of NGOs and discrepancies in the use of funds, NDoH resigned as a PR nonprofit organizations as Principal Recipients and GFATM was classified as off-budget support. GFATM (PRs) increases the capacity of nongovernmental drugs and vaccines are procured directly from providers of service providers but limits the ability of NDoH choice. Gavi support is provided in the form of immuniza- to understand service-delivery functions pro- tion services support, health system strengthening (HSS), vided by donors. Whether NDoH can utilize the increased capacity of NGOs depends on the ability and vaccines. Gavi vaccines are procured through the UNI- to enter into successful partnerships in the even- CEF Supply Division only.24 tual case of donor transition. External resources, particularly the share contrib- • GFATM’s multiple PRs and subgrantees illustrate uted by Gavi and GFATM, comprise a much larger PNG’s fragmented financing and service delivery. share of health expenditure in vaccines, and the three GFATM diseases. This raises concerns in light of decreas- ing external funds, and/or more general budget support Donor funding was a significant source of health to the health sector. There is likely to be an expansion spending in PNG between 2006 and 2015, averag- ing 20 percent of total health spending per year (Fig- ure 2–17). Total external funding to the health sector in 23.  The reduction in recent years is both due to a decline in the grant allocation and a reduction of the exchange rate. PNG increased substantially, from US$38.4 million in 2006, 24. http://www.gavi.org/country/all-countries-commitments-and- to a peak of US$164.4 million in 2012. Since then, donor disbursements/ 27 28 Health Financing System Assessment Figure 4–1: Official Development Assistance (ODA) for Health by Channel (US$ millions) 180 70% 160 60% 60% 140 US$ millions (current) 50% 120 46% 100 40% 34% 80 31% 30% 60 20% 40 18% 20 10% 11% 10% 8% 6% 3% 0% 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Total health ODA Health ODA channeled through public sector Public (% of total) (right axis) Source: Organisation for Economic Co-operation and Development (OECD) Creditor Reporting System (CRS) 2017. Note: OECD-CRS include Gavi resources in multilateral arrangements, not to the public sector. Gavi procures supplies via UNICEF Supply Division. ODA channeled through public sector includes both donor and recipient governments. Figure 4–2: ODA for Health by Donor (US$ millions) 180 160 140 120 Others US$ millions (current) 100 WHO USAid 80 UN 60 New Zealand Japan 40 Global Fund 20 Gavi EU 0 Australia –20 ADB 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: OECD-CRS 2017. Note: The graph presents the figures based on OECD data. DFaT actual funding might be higher than presented. DFaT communicated that its 2015 spending on health in PNG was A$98.5 million, including joint understanding funding on health activities. of concessional lending in the sector by the World Bank Key PNG Donors: GFATM 4.1  and Asian Development Bank (ADB) in 2017 and 2018. The World Bank board recently approved a US$15 million and Gavi credit for the emergency TB project to target an unprec- Over the past decade to 2016, the GFATM and Gavi have edented increase in MDR and XDR TB in several hotspots played an increasingly important role in PNG’s health in PNG. The ADB is preparing a credit-financed program to service delivery. In 2015, GFATM funds comprised 15 per- provide budget support for existing health operations and cent and Gavi funds comprised 5 percent of external fund- an HSS component.25 ing to PNG’s health sector. GFATM resources have been directed to malaria treatment (43 percent), HIV and AIDS 25.  Negotiations for the ADB credit are yet to be finalised. response (32 percent), and TB (25 percent). Gavi resources PNG Reliance on Donor Financing 29 Figure 4–3: Gavi’s Transition Policy and Country Preparedness TRANSITION Low income Phase 1 Phase 2 Phase 3 Variable duration Variable duration 5 years 5 years Initial Preparatory Accelerated Full self-financing transition transition self-financing Linear co-financing— Co-financing share Co-financing increases Country finances 100% $0.20 per dose increases by 15% gradually to reach full of vaccines with access a year financing in 5 years to UNICEF tenders for vaccines Low-income country Eligibility threshold: End of Gavi financing threshold: $1045 GNI $1580 GNI capita per capita Source: Gavi 2017a. Note: These thresholds are updated annually. have been directed to vaccines and immunization- concerns regarding the “willingness to pay” provision as specific activities. Both donors have also directed some the CCM has limited influence on government expendi- funding to HSS. ture, and the government is no longer a PR.27 GFATM and Gavi resources have conditions (commit- Gavi requires recipient countries to contribute toward ments) and counterpart-financing requirements to the cost of the vaccines to ensure the immunization encourage financially sustainable programs. PNG’s program is sustained after Gavi’s financial support cofinancing share is expected to increase over the short to ends (Gavi 2017b). Support is based on Gross National medium term. It is important that these requirements and Income (GNI) per capita, and when the three-year aver- commitments are included in the Medium-term Expen- age exceeds the threshold (currently set at US$1,580), diture Framework (MTEF) as well as the annual health countries enter phase two or accelerated transition. Coun- budget. tries gradually transition from donor support to domestic resources when they are 100 percent self-financing (see Under the GFATM new funding model (NFM), PNG Figures 4–3 and 4–4). PNG began phase two of the Gavi must contribute 20 percent of the grant amount in program in 2017. counterpart funding as a lower-middle-income coun- try. In addition, 15 percent of grant funds are contingent These programs operate on the theory that, as coun- on domestic investment. This is known as “increasing tries become richer, their capacity to self-finance future commitments” and formerly called “willingness to improves and donor funding can be transferred to pay.” In practical terms, if a government does not make an countries with greater needs, thus linking donor eli- additional investment, year-three funds are cut (from the gibility to GNI per capita. In reality, economic growth is three-year 2015–17 grant) to offset the 15 percent coun- almost never synonymous with equal access to health care try requirement. For PNG, additional domestic investment across a population, nor does it guarantee a government’s includes the budget for HIV drugs. capacity to increase funding to the health sector. Gavi is phasing out its PNG support at a time when the broader For 2015–17, the government committed to spend health system still lacks capacity and faces many con- US$119.7 million, an additional investment of straints, including human resources (HR) shortages and US$48 million. The additional investment equates to a PFM problems, which pose a risk to a successful transition. 67 percent increase in government commitment on the three programs from 2012–14.26 In 2014, the chair of the GFATM Country Coordinating Mechanism (CCM) noted 26.  Note the actual spending is less than the commitment between 2012–14. 27.  30 Health Financing System Assessment 4.2 Gavi: Immunization in PNG Neither initiative required cofinancing. HSS support was committed for 2016 and 2017, but cash for 2014 was only Gavi provides cash support for HSS, operational costs disbursed in 2015 due to low utilization. Gavi is providing for campaigns, and vaccine introduction grants. All US$3 million from 2013 through 2017 to assist HSS. other Gavi support is in the form of vaccines; Gavi cur- rently supports five vaccine initiatives in PNG: pentavalent, As explained in a previous section, Gavi is entering pneumococcal, measles, measles-rubella (MR), and inacti- the accelerated transition phase and PNG must fully vated polio vaccine (IPV). finance vaccines by 2021. At that point PNG will be able to access Gavi prices from manufacturers for five years. It In 2016, Gavi disbursed US$8.87 million; including is highly recommended that PNG develop a costed transi- US$5.2 million worth of pentavalent and pneumo- tion plan with Gavi as a basis for targeted support from coccal vaccines (Figure 4–4). Gavi support for IPV was partners and potential time-limited investment from Gavi. in the form of one-time cash assistance. Gavi also pro- vided operational costs for the MR campaign on a one- GoPNG’s financial commitment to immunization is evi- time basis to facilitate vaccine rollout across the country. dent from its funding for routine vaccines, however, in Figure 4–4: Gavi Disbursements by Year and Program (US$) 10,000,000 9,000,000 8,000,000 7,000,000 Vaccine introduction grant 6,000,000 MR—operational costs IPV US$ 5,000,000 Tetra DTP-Hib 4,000,000 Pneumo 3,000,000 Penta Measles-Rubella 2,000,000 Measles 1,000,000 ISS 0 HSS 2012 2013 2014 2015 2016 2017 Source: Gavi 2017c. Figure 4–5: Gavi Cofinancing Requirements (US$ millions) 4.0 3.5 3.5 3.5 3.0 2.9 2.5 2.5 US$ million 2.0 2.0 2.0 1.5 1.1 1.0 1.0 0.7 0.6 0.5 0.5 0.5 0.5 0.3 2009 2010 2011 2012 2013 2014 2015 2016 2017 Proj. 2018 Proj. 2019 Proj. 2020 Proj. 2021 Proj. 2022 Proj. Source: Gavi 2017c. PNG Reliance on Donor Financing 31 past years, PNG health centers at all levels have experi- quarterly installments—this should prevent unnecessary enced vaccine stock-outs due to delayed or inaccurate stock-outs. reporting of vaccine supply and demand. A key reason for such stock-outs is that the annual vaccine demand fore- casts, that include central-level stock information, are often GFATM: HIV and AIDS 4.3  not submitted within the given timelines to the NDoH pro- Response, TB, and Malaria curement unit and UNICEF’s Supply Division. In 2013, the Treatment Hepatitis B vaccine was unavailable at the national level for two months because stock levels at the district level The Global Fund has been active in PNG since 2004, were not documented. The following year, stock-outs of with total disbursements of more than US$190 million the BCG, OPV, and measles vaccines were also reported to (Figure 4–6). Until 2008, NDoH was the single principal have lasted for two months at the district level, interrupt- recipient (PR). The Global Fund expressed concern over the ing the delivery of vaccinations.28 As these vaccines are not capacity of NDoH to perform the PR role after the Office of supported by Gavi, and given the upcoming Gavi transi- the Inspector General raised concerns (The Global Fund tion, stock-outs of various government-funded vaccines, 2014). Numerous donors identified the relevant NDoH combined with a default of 2014 cofinancing obligations, trust account as lacking rigor in terms of effectiveness and have raised questions regarding the government’s ability administration systems. Since 2010, GFATM grants have, to self-finance immunization programs. therefore, been implemented through NGOs and non- profit organizations. NFM grants were structured for the Gavi funding was channeled through a trust account years 2015–17. Four development partners implement held within NDoH, however, an audit in 2016 revealed GFATM grants under the NFM—Oil Search Foundation discrepancies in the management of this account. This (OSF), Population Services International (PSI), Rotarians led Gavi to request a refund of US$750,000 and to channel Against Malaria (RAM), and World Vision International its support outside the government system. This highlights (WVI). This creates fragmentation and appears to be mis- constraints in the effective management and accounting aligned with the principles of the Paris Declaration. of the NDoH trust account as a key issue preventing greater reliance of donor programs on government systems. Due to broad budget categories, it is difficult to esti- mate specific interventions and how much the GoPNG Many funding and implementation decisions are made spends on the three diseases that are supported by at subnational levels. While vaccines and supplies are GFATM funding. For example, TB is combined with lep- funded nationally and distributed to the provinces, little rosy, and HIV is combined with other STIs. Funding for funding is available for nonvaccine service delivery costs the three diseases, which is considered as counterpart by (that is, operational costs) at subnational levels. Funds for GFATM includes: (i) drugs purchased out of the general vaccine procurement are disbursed relatively regularly in drugs budget (even though this is not publicly visible in budget books); (ii) budget allocations to the respective units in NDoH; (iii) a share of the operational budget (both 28.  Regular stock-outs are the result of a series of interrelated issues salaries and goods and services) for provincial health ser- across the medical supply chain, including operating procedures, facilities, transport, the cold chain, and reporting (see Brown and Gilbert vice delivery (for example, 10 percent of rural health work- 2014). ers’ salaries, which are not visible in the budget books, Figure 4–6: GFATM Disbursements by Disease (US$ millions) 40 35 30 US$ millions 25 20 15 10 5 0 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 Malaria HIV TB Source: The Global Fund 2017. 32 Health Financing System Assessment Figure 4–7: GoPNG Agreed Spending for vention program goals, the grant aims to more than GFATM Funds (US$ millions) double the number of sex workers tested for HIV and pro- vide them with condoms and anti-retroviral treatment. A Earmarked GoPNG spending 140 similar program of testing, condom distribution, and risk reduction discussion, and STI treatment and partner noti- 120 fication is outlined for vulnerable populations. The grant 100 includes US$1.55 million for an integrated bio-behavioral 80 study (IBBS) amongst key populations.30 60 The three-year grant is for US$14.2 million, however, 40 it is not known how much the government spends on 20 HIV counseling, testing or treatment in NDoH facilities, nor what activities are undertaken by either CHS or 0 2012–14 Spending 2015–17 Commitment NDoH hospitals.31 In contrast, over one-half of the grant is spent on HR costs (55 percent); 16 percent is spent on HIV TB Malaria travel-related costs and an additional 16 percent on exter- nal professional services. In terms of program sustainabil- Source: Based on discussions between a World Bank consultant ity, it is highly unlikely that the government will be able to and NDoH staff. finance similar HR costs in the future. It could be possible to provide similar HIV prevention outreach through training are assumed to contribute towards malaria); and (iv) any and capacity building for aid post staff, and by improving relevant components of loan-funded donor projects— referral systems at facilities.32 GoPNG identified outreach currently the ADB’s Rural Primary Health Services Project. as a minimum priority area under its HFGs. In other words, Overall, GFATM rates counterpart funding as well as sys- a minimum outreach-spending threshold is required for all tems to track it as sufficient. It is clear, however, that the provinces, but funding for outreach activities continues to government is spending far less than GFATM. Figure 4–7 be constrained. presents budget data related to the three diseases. The actual spending is much less than the budgeted amount DFaT is transitioning from targeted HIV and AIDS due to a variety of reasons. programs to an integrated sexual and reproductive health-funding approach. As such, GFATM funding for Current GFATM Grants targeted HIV and AIDS services is becoming increasingly important. It will be important for the GoPNG to under- Health services to the target populations are partly stand the changes in the different donor programming so delivered using NDoH’s systems, in combination with that they can identify and address gaps that arise. Even service provider networks operated by CHS. For exam- though the delivery model runs in parallel to the govern- ple, LLIN distribution is managed by RAM, and implemen- ment system, NDoH should carry more responsibilities in tation is partly supported by government agencies and coordinating and planning to maximize value for money. CHS. As a result, NDoH remains involved despite no longer receiving GFATM funds directly. In addition, NDoH receives Malaria funds through the TB/HSS grant and the PSI malaria grant. PSI and RAM are the two NGOs selected as PRs to sup- The PSI grant also provides support to the IMR. Funding to port the implementation of the malaria grants. The the NDoH and IMR are on a no-cash basis. CHS providers malaria grants aim to provide coverage (60 percent) of also receive subgrants from PRs (for example, Anglicare long-lasting insecticidal nets (LLINs) (Figure 4–8 highlights receives a subgrant from OSF for HIV). the increase in distribution of LLINs); increase the use of HIV and AIDS appropriate malaria prevention measures; and maximize access to early diagnosis and treatment for malaria. In The GFATM grant for the HIV and AIDS response is addition, HSS components focus on strengthening malaria managed by OSF and focuses on reducing the risk of epidemic preparedness and response capacities at all HIV and STI transmission by improving access to pre- vention programs and improving links to HIV testing, care, treatment, and support services.29 In terms of pre- 30.  Under the terms of the grant, the size estimation and bio-behavioral data generated by the IBBS related to the size, location, and prevalence of key populations must be reported to the GFATM by December 3, 2017. 29.  OSF was established and is supported by Oil Search. Oil Search is Papua New Guinea’s largest company, employer, and investor. As a key 31.  In the first 18 months there have been disbursement issues of the player in PNG’s oil and gas industry, Oil Search believes that proactive GFATM HIV grant. participation in the development of the country by the corporate sector 32.  For example, aid posts could also provide sex workers with referrals is not only needed, but an obligation. to facilities that provide testing. PNG Reliance on Donor Financing 33 Figure 4–8: Number of LLINs Distributed (millions) 9 8.5 8 7.2 7 Nets distributed (millions) 6 5.2 5 4.2 4 3.1 3 2 1.7 1.4 1 0.73 0.23 0.23 0 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Source: The Global Fund 2017. Note: Figures are cumulative. levels; and program management, prioritizing the district funding for the replacement of nets, and an understand- level. The goal is to reduce annual parasite incidence to 41 ing of replacement costs is unknown, thus risking the sus- per 1,000 by 2018. tainability of outcomes. The RAM grant focuses on vector control, with 31 per- The PSI component of the three-year malaria grant cent of the grant total allocated to nonpharmaceuti- focuses on NHIS case management and improving the cal products, 30 percent to procurement and supply efficacy of LLINs through educational outreach pro- chain management costs, 23 percent for HR costs, and grams, facilitating consistent messaging, and produc- 10 percent for travel. RAM targets include 454,429 LLINs ing low literacy packaging. As is the case with the grant distributed via the mass campaign, and 75,000 distrib- for the HIV and AIDS response, the majority of funds are allo- uted to targeted risk groups through continuous distribu- cated for HR (50 percent) and travel-related costs (24 per- tion. Provincial health offices and PHAs will be involved in cent). Little is directed to health products (only 1.0 percent the logistical planning for distribution, and government of the grant), in the form of nonpharmaceuticals, health health facilities will serve as the main centers of distribu- product equipment, procurement, and supply-chain man- tion to households. Mothers can also receive bed nets agement costs. NDoH and IMR are recipients under the PSI while visiting antenatal clinics. The total cost requested grant. The NDoH component of the grant supports routine for LLIN distribution is US$29.5 million between 2014 and reporting, regional meetings, program salaries, adminis- 2017 (The Global Fund 2017). tration costs, and the purchase of some IT equipment. WHO recommends a three-year LLIN serviceable TB/HSS lifespan; as a result, despite distribution campaigns, The TB grant aims to reduce the estimated preva- GoPNG will need to assume an LLIN serviceable life lence and death rate of TB to one-half the 1990 lev- of three years and budget for replacements. A study els by 2020. This means reducing the prevalence rate in Rwanda that considered survivorship/attrition (a mea- to 339/100,000 population (or lower), and reducing the sure of the number of nets remaining), and fabric integrity, death rate to 30/100,000 population by 2020. The grant found that after two years LLIN survivorship decreased to also supports the availability of drugs and laboratory diag- 42 percent (Hakizimana et al 2014).33 The study estimated nostics for TB, HIV, and malaria under HSS, and laboratory that almost 60 percent of LLINs needed replacement after services with an emphasis on PHC. two years, and were of little-to-no benefit to a user. The findings are instructive for PNG as they create a replace- The WVI activities are centered on TB care and pre- ment and redistribution plan. NDoH has yet to allocate vention in 12 provinces where 28 of the 30 targeted underperforming basic management units (BMUs) are located. These 28 BMUs comprise only 11 percent 33.  In Rwanda, the National Malaria Control Programme in charge of LLIN distribution and replacement predicts the proportion of nets of total BMUs, however, they account for 53 percent of remaining at any given time by a three-year NetCALC net loss model. the national TB burden, 75 percent of “defaulters” and 34 Health Financing System Assessment 65 percent of “smear not done” cases nationally. Other The designation of NGOs and nonprofit organizations components focus on program management of drug- as PRs increases the capacity of nongovernmental ser- resistant TB in three provinces (NCD, Western Province, vice providers but limits the ability of NDoH to under- and Gulf Province) with high drug-resistant TB prevalence, stand service-delivery functions provided by donors. and increased HIV testing of TB patients in eight high HIV- The multiple PRs and subgrantees are another example of burden provinces. The grant also supports HSS, focusing PNG’s fragmented financing and service delivery. Whether on the national pharmaceutical supply-chain manage- NDoH can utilize the increased capacity of NGOs depends ment system, health information systems, monitoring and on their ability to enter into successful partnerships in the evaluation, and service delivery. NDoH is a recipient of the eventual case of donor transition. grant primarily under HSS components. section five Financial and Institutional Sustainability Key messages: 5.1 Financial Sustainability Given that both donors contribute significant amounts • Financial resources available to the health sec- to specific areas of PNG’s health sector, the financial tor are expected to decline, and it is unlikely that sustainability of GFATM and Gavi-financed programs GoPNG will be able to fully finance current DP sup- needs to be carefully assessed. Figure 5–1 provides a port as they graduate from assistance programs. summary of future financing to the health sector over the It is, therefore, important to identify interventions (including LLINs and TB care) that are affordable medium term to 2021, as projected in 2017 budget docu- within the projected resource availability and can ments. The summary shows a decline in financial resources be integrated into government budgets. available to the sector—in both monetary terms and as a proportion of the national budget. In addition, the GoPNG • In addition, to increase financial sustainability national budget figures do not capture GFATM and Gavi and prevent the loss of health gains, it is recom- support and, therefore, risk overlooking any financial mended that the NDoH: requirements for the transition and graduation phases of both DPs. Central agencies also do not capture GFATM • adequately budget for vaccines, delivery costs, cold chain, and operational expenses; and disbursements or contributions as part of their DP expen- diture reports or in either the Mid-Year Economic Fiscal • clarify functional responsibilities for medical Outlook or the Final Budget Outcome. supplies and drug distribution at every level. Upcoming program funding responsibilities should • To ensure institutional sustainability, it is critical be integrated into the government’s budget plan- for PNG to develop replacement strategies for the ning cycle. Given the complexity of the situation, an human and technical resources that exit as donor early integration of funding into the budget will enhance support ends. Local organizations and NGOs may the chances of a successful post-graduation transition. serve as resources as grants close. Recommendations for smaller-scale integration efforts • To continue Gavi’s gains and improve PNG immu- include financing LLINs and TB care and prevention efforts nization rates and overall institutional sustainabil- in underperforming BMUs. ity, GoPNG will need to: Similarly, to smooth the transition, NDoH should • align vaccine and vaccine consumables pur- update vaccine and related resource requirements in chase and delivery; the budget and MTEF. Like GFATM, Gavi is a substantial donor in immunization. To prevent the loss of immuniza- • improve the National Regulatory Authority to tion gains when Gavi vaccine support ends, it is important meet WHO-recommended functions; and for the NDoH to understand, communicate, and incorpo- • investigate taking advantage of available rate the required financial resources in the budgets. The Gavi vaccine prices through UNICEF in a more government has initiated the process of budgeting for the coordinated national effort to save financial vaccines previously provided by Gavi, with a goal to gradu- resources. ate by 2021 and fully finance the procurement, distribu- tion, and administration of the vaccines. Further analysis 35 36 Health Financing System Assessment Figure 5–1: GoPNG Health Sector Projections (millions of Kina) 16,000 16.0% Health share of national budget (%) 14,000 GoPNG budget (Kina million) 14.0% 12,000 11.9% 12.0% 11.5% 10,000 10.2% 9.8% 10.0% 10.0% 8.9% 9.3% 8,000 8.6% 8.2% 8.0% 8.0% 6,000 6.0% 4,000 4.0% 2,000 1,114 1,117 1,421 1,583 1,647 1,333 1,208 1,140 1,107 1,093 2.0% 0 0.0% 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Proj. Proj. Proj. Proj. National budget Health sector budget Health share of national budget WHO/Abuja recommendation of 15% health share of national budget Source: Government of PNG 2017. Note: (i) Health sector excludes provincial funding on health, as there are no projections included for this in the National Budget Book. (ii) WHO/Abuja Declaration: In April 2001, the African Union countries met in Abuja and pledged to set a target of allocating at least 15 percent of their annual budget to improve the health sector and urged donor countries to scale up support. is recommended to determine the effectiveness of the lack of confidence in NDoH capacity. PRs use NDoH sys- national immunization program and establish a strategy to tems; grants are implemented through CHS or NDoH, and fully integrate Gavi support as part of the transition phase. through a myriad of NGO networks.34 The grants increase the technical capacity of NGOs to deliver services and NDoH should adequately budget for the cost of vac- finance some NDoH staff. In addition, CHS providers may cine deliveries (including cold chain and operational also become recipients of subgrants. GFATM programs expenses) and factor in high distribution costs. Estab- have been aligned to support the disease programs within lishing clear responsibility for the distribution of medical NDoH and the national strategies for TB, HIV and AIDS, and supplies and drugs from medical stores to facilities will malaria treatment. It is important to recognize that GFATM increase reliable coverage for the population. NEFC rec- and other donor partner support is not limited to financial ommends that provinces take responsibility for drug dis- support. As donors transition out, they also take human tribution from the transit stores to facilities. NDoH has, and technical resources with them. It is, therefore, critical however, assumed responsibility for this function in many that strategies for successful transitions are put into place provinces, and many mixed and informal funding modali- to ensure institutional sustainability as well as financial ties exist. Unfortunately, stock-outs of essential drugs sustainability. remain a persistent problem. The use of multiple organizations (NGOs, nonprofits) Ultimately, financial sustainability of these programs, to implement GFATM grants has multiple advantages, which have contributed to health gains, will depend on but also presents risks to the system. The parallel deliv- PNG’s ability to improve stewardship of existing fund- ery builds capacity and knowledge across a wide range of ing, leverage development partnerships to increase local organizations and can potentially build long-term capacity and strengthen existing health systems or institutional capacity given the decentralized and tiered identify key changes in current practices, and improve health system in PNG. It is likely that some of the many financial planning and integration efforts in the short organizations will continue to be active in the health sec- to medium term. tor when grants end, whether at the central, provincial, or district level. 5.2 Institutional Sustainability Institutional capacity is also an important issue for 34.  Despite this, funding from GFATM is not captured in the PNG budget books (a process that is managed by the Department of PNG—as discussed, GFATM support is channeled National Planning without the input of NDoH) even if some of it is through nonprofit organizations and NGOs due to a eventually utilized by NDoH. Financial and Institutional Sustainability 37 As funding is unlikely to continue at the same rate, it is management present significant constraints if vaccine difficult to predict actual frontline delivery impacts. It consumable deliveries are not aligned with vaccine deliv- is unlikely that the activities will continue in as many dis- eries. To establish a sustainable immunization system, tricts over the long term unless they are integrated into the NDoH should align vaccine consumable deliveries with services provided by facilities and aid posts. GoPNG could, vaccine deliveries. in future, work with RAM and their distribution network to distribute LLINs. As discussed earlier, analysis is required Medicine stock-outs, reflecting the poor stock man- for a LLIN distribution and replacement plan over the next agement and procurement system in PNG, remain one five years, particularly given WHO guidelines regarding of the big challenges. DPs have provided assistance to a three-year LLIN life span. In addition, it is important to help project the demand and manage the overall stock maintain improvements in TB care in the underperform- for the three major diseases and immunization. Even with ing BMUs and to continue HIV and TB testing. Initially, this that level of support, stock-outs happened, partly due to will require improvements in government facility-referral poor communication between central, regional, and pro- systems. vincial medical stores and, therefore, the supply does not match the real demand in the areas. The capacity of the Similar challenges also exist in immunization pro- National Regulatory Authority also needs to be strength- grams. Vaccines are currently ordered and distributed ened, as it does not meet WHO-recommended functions without vaccine consumables. Geography and cold-chain for a self-procuring country. section six Summary and Policy Recommendations PNG can successfully transition from key areas of Next steps: donor support by improving budgeting, planning, prioritization of strategies, procurement systems, and • Improve reporting process of DPs (requiring their HR. The sustainability of externally supported programs increased cooperation) to receive better projections, will depend upon the government’s ability to unbundle with a clear breakup of program components, chan- the support being provided and determine the compo- nels, and cofinancing requirements. nents that need to be sustained and integrated into gov- • Undertake a donor mapping exercise. ernment systems in the medium term. • Improve the enabling environment to bring donor sup- PNG’s current weak fiscal position has implications for port back on system, including more reliable financial the health sector. During 2012–16, central-level health management of trust accounts. expenditure has been volatile. Fiscal policies continued to tighten in 2016 to meet debt targets. Given APEC prepara- • Review the feasibility of using current NGO program tions and electoral spending in 2017 and 2018, it is unlikely implementation partners, currently used by donors, to the health budget will increase significantly. Budget con- maintain important health services once PNG gradu- straints represent an imperative for the NDoH and CHS to ates from donor support. assess current spending and costing models. • Use a donor transition plan to prepare the sector and The recommendations below span short-term, urgent articulate the requirement for additional support to priorities, to longer-term and deeper reforms for the central agencies. health sector. The recommendations are based on the analyses in this report, and on recent reports produced 2. Continue Partnerships in Service Delivery  by the World Bank and other DPs. Some issues can be Technical addressed by the NDoH, while others will benefit from the Capacity support of DPs. Urgency Requirement Cost Timeframe Medium Medium Medium Medium- to long- 1. Prepare for Medium- to Longer-term term Donor Transition  Technical The NDoH should continue partnerships and dialogue Capacity with NGOs and CHS to establish alternative modalities Urgency Requirement Cost Timeframe for financing and service delivery across the key dis- Medium Medium-high Medium-high Medium- to ease programs. Opportunities exist to mobilize resources long-term through the private sector. PNG is not ready for eventual graduation away from Next steps: GFATM and other donor support in the medium term. To improve preparedness for graduation, the government • Map NGOs and private sector providers and their com- should unbundle the support it is currently receiving and parative advantages, so that partnership agreements determine the key components to integrate and sustain in can be considered when funding constraints ease over the national budget. The government should also use this the medium term. period to improve the financial and institutional capacities in preparation for future transition and graduation. 39 40 Health Financing System Assessment Box 6–1: Examples of Partnerships to Deliver Health Services As a way of strengthening the health system and using limited resources efficiently, alternative service-delivery and financing modalities through Civil Society Organizations (CSO)s, churches, and NGOs could be considered. Some examples of current partnerships include: Between NDoH and CHS: In 2014, a partnership agreement between the CHS and NDoH was signed to more formally reaffirm the already existing partnership. Under the 2014 Partnerships Policy, government entities and partners in health service delivery are required to adhere to more formal agreements. The partnership agreement formally established a health service delivery and training partnership with CHS by allocating and distributing government grants to members. Between provincial governments, hospitals, and the Oil Search Health Foundation: a) Gulf Provincial Government and OSF The OSF partnership with the Gulf provincial government, working with the Kikori district government, focuses on improving and running “integrated primary and preventative health programs.” These contribute to improved and measurable health outcomes for agreed public health authorities,” through the implementation of primary and preventative health care services using a health care model focused on communities and district levels. One of their objectives is to move away from “disease specific, resource inefficient” approaches through a more integrated primary approach and to include broader public health priorities beyond HIV and AIDs, TB, malaria, and maternal-child services. The partnership also includes Oil Search’s contribution to HSS, by rebuilding district health systems and acting as a “systems enabler,” given their private sector comparative advantage to build partnerships with stakeholders, other DPs, and with local authorities and leaders. b) Hela Provincial Hospital Board and Oil Search The Oil Search partnership agreement provides several support services to the Hela provincial hospital. In par- ticular, the partnership aims to support the hospital CEO and management by strengthening their leadership roles. The Oil Search Health Foundation provides supplemental funding for hospital staff (both clinical and support staff ), fund- ing support for the recruitment of 200 priority positions in the hospital, general financial management improvements, and the role of liaison with DPs for additional funding and other support. 3. Integrate Major Disease Programs and facility level. This will establish a clear demarcation of roles Clarify Types and Levels of Interventions and responsibilities to effectively coordinate and imple- ment the national strategies already in place for the key Technical Capacity disease programs. Urgency Requirement Cost Timeframe Next steps: Medium Medium High Medium- to long- term • Establish which donor responsibilities require integra- tion into service-delivery functions at all facility levels. Integrating the major disease programs as part of the overall HSS effort will require further linkages between • Identify nongovernment sector partnerships that the National Health Plan to the key disease program provide a competitive advantage over government plans. First and foremost, PNG needs to determine which providers. components of each key disease program to integrate and provide cost and resource requirements (such as with Gavi • Enact reforms across medical supply and pharmaceu- support and the immunization program). These program tical procurement and distribution services to support activities need to be linked to PHA implementation and health system preparation for the transition phase. integrated into provincial service plans down to the health Summary and Policy Recommendations 41 4. Manage Gavi Transition  Next steps: Technical • Identify the areas of inefficiency, including both alloca- Capacity tive inefficiency and technical inefficiency. For example, Urgency Requirement Cost Timeframe in the case of salary overruns, which divert funding from High Medium Medium Short- to medium- goods and services and prevent recruitment of critical term new staff, efficiency could be improved through: (i) a The transition from Gavi support is already under- payroll cleansing exercise; (ii) updating staff numbers way, as PNG’s cofinancing requirements are increasing across the health sector; and (iii) improving the effi- by 20 percent each year from 2017 onwards. To suc- ciency of drug procurement and distribution to prevent cessfully transition, the following actions are necessary: stock-outs and expiry. (i) identify a funding source for the increasing cofinancing • Improve budget coordination with provinces through requirements; (ii) identify the components of support that service plans and the Annual Implementation Plan must be integrated over the coming years; and (iii) maxi- process. Enhance alignment with plans and encourage mize the benefit of the remaining years of HSS support. funding to frontline facilities. Next steps: • Improve the impact of programs that have experienced • Communicate increasing cofinancing requirements— funding increases but still perform below potential, and their inevitability—clearly to central agencies, so especially the HFGs, PSIP, and DSIP. Consider direct facil- they can be factored into the budget. ity funding and improve coordination with provinces and districts. • Identify components of Gavi support that must be inte- grated over the short to medium term. • Implementing agencies and health facilities need to identify areas for waste reduction (such as unnecessary • Improve support mechanisms for vaccine delivery by travel, purchase of stationery, printing, and procure- improving communication between medical stores ment of supplies) that will create more fiscal space in and NDoH, aligning vaccine and vaccine consumables the short term for the key program areas. deliveries, and strengthen the capacity of frontline facilities to deliver immunization services. 6. Fiscal Analysis and PFM Technical 5. Increase Health System Efficiency Capacity to Increase the Fiscal Space  Urgency Requirement Cost Timeframe Technical High Medium Low Short- to medium- Capacity term Urgency Requirement Cost Timeframe High Medium-high Medium Short- to long- Short-term estimates of funding needs and cofinanc- term ing requirements must be included in the MTEF, to better understand future resource requirements and It is particularly important for the government to communicate them effectively to central agencies. increase the efficiency of current spending. Higher Overall, PFM needs to be continually strengthened to spending on health can contribute to better outcomes, improve the cost-effectiveness of service delivery and free but so can improvements in health spending efficiency. up resources from poorly performing areas. There appear to be significant inefficiencies in health spending, although the extent of these inefficiencies is Poor financial management systems within NDoH have not entirely clear. The systematic documentation of the prevented donors from increasing on-system support budget, expenditure, personnel, and health utilization through the Health Sector Improvement Program information will help track expenditures and expenditure (HSIP) trust account. Improving this capacity would serve outcomes. Improved efficiency will ensure that resources as a precondition and allow donors to increasingly rely on are available in a timely manner and will reduce cost pres- government systems. sures. Improving the efficiency of spending will create a Next steps: stronger health system—one that is better prepared for the Gavi transition and decreased vertical support from • Improve reporting from donors to furnish the MTEF GFATM. with projections of disbursements and counterpart requirements. 42 Health Financing System Assessment • Identify key PFM constraints at the subnational level Next steps: and identify innovative ways of delivering training. • Improve reporting from donors to furnish the MTEF • Clarify necessary capacity and processes required by with projections of disbursements and counterpart donors to use government systems at a DP summit. requirements. • Consider using technical assistance to build capacity in • Identify key PFM constraints at the subnational level the HSIP account and place a donor-funded accountant and identify innovative ways of delivering training. in the unit, as is the case in other Pacific countries. • NEFC and NDoH should continue to analyze provincial 7. Ensure Timely Release of Warrants  expenditure and identify areas that impede fund flows to facility operations. They should continue to work Technical Capacity Urgency Requirement Cost Timeframe with provinces to improve the chart of accounts to accurately track allocations to facilities. High Low Low Short-term • NDoH should use the Annual Implementation Plan pro- Warrant releases (which authorize an agency to access cess to oversee provincial budgets and encourage suf- budgeted funds throughout the budget year) need to ficient allocations to facility operations. be timely to improve health system performance. While untimely disbursement is a DoT issue, the NDoH needs to • Consider the feasibility of funding facilities directly (as improve its own budget execution and disbursement of is currently the practice with schools) to reduce cash operational funds for the hospitals and for CHS. Although flow bottlenecks at the provincial level. the issuing of warrants is unpredictable and ad hoc, NDoH has the potential to create further bottlenecks if funds to • Improve the provision of medical supplies so facilities hospitals and CHS are not disbursed immediately upon are not forced to collect user fees during stock-outs. receipt from the treasury. 9. Reporting and Information Sharing  Next steps: Technical Capacity • Continue to inform central agencies of priority HFGs for Urgency Requirement Cost Timeframe uninterrupted frontline service delivery. Medium Low Low Short- to medium- term • Track the timing of fund release to hospitals and CHS as a PFM performance indicator. Increased engagement and membership at the Min- isterial Economic Committee, as well as the Provincial 8. Protect Government Financing to and Local Level Service Monitoring Authority (PLLSMA) the Frontlines to Ensure the Successful could help the NDoH, as the lead agency for the health Implementation of the Free Health Care Policy sector, improve communication and report on the key Technical disease areas and the health strategy implementa- Capacity Urgency Requirement Cost Timeframe tion challenges as donor partner support decreases (particularly with Gavi). The challenges of the health High Medium Low Short- to medium- term sector, in terms of integration and transition, need to be more clearly articulated. Furthermore, policy options and The government has committed to eliminating user strategies should be established in close consultation and fees, which are a barrier to health care access—­ partnership with the DoT and the Department of National especially for the poorest segments of the population. Planning and Monitoring, as well as with other sectors like To date, the related funds are not disbursed in a timely transportation and education. manner, resulting in delays at the provincial level and forc- Next steps: ing facilities to charge fees. Warrants for HFG and the Free Health Care Policy funding must be released reliably to • Prepare an accessible summary of a donor transition for eliminate user fees. A majority of this funding must be uti- central budget agencies; include clear transition sched- lized reliably to finance facility operations (in-kind or cash) ules and predictable funding requirements for integra- rather than activities at the provincial and district level. tion into the national budget. Summary and Policy Recommendations 43 10. Improve Data Accessibility, • Improve the NHIS and strengthen staff capacity to use Comprehensiveness, and Quality  various tools for reporting and information sharing. Continue progress on the eNHIS system to improve Technical Capacity timeliness and accuracy of data. Urgency Requirement Cost Timeframe High High Medium Short- to medium- • Work with DoT to ensure budget figures are accurate term and can be used to estimate sector funding trends. • Consolidate data sources for health workers and com- There is an urgent need to assess the country’s dispa- plete payroll reconciliation with DoT, Department of rate data collection systems (across different service Finance, and Department of Personnel Management to delivery levels and programs, both routine NHIS and reduce salary overruns. eHealth pilots) in order to improve the NHIS. Pilot pro- grams are being implemented to strengthen the routine • Liaise with the WHO to identify reasons for inconsisten- NHIS system, but are not coordinated across the health cies in donor funding levels reported in the CRS. sector, which means that current, accurate data for sur- veillance of all diseases and services is not available. The • Consolidate and strengthen the disease-reporting sys- Global Fund grant included K 2.4 million to support NHIS tem to provide accurate data, which will improve finan- strengthening, and to support NDoH in reviewing and cial planning and targeted health staff and supplies assessing options to establish proper electronic systems. deployment. Next steps: • Improve data sources for domestic funding (national and provincial), external funding, disease burdens, health access, health workforce and health infrastructure so they can be effective planning and monitoring tools. appendix one Data Challenges in PNG Budget and Expenditure Data IFMS This report aimed to provide and use PNG Government The national budget is managed through the IFMS, data where possible because it gives a more granular, which provides detailed reports on agency budgets, up-to-date view on budgets and expenditure than data the release of warrants, and actual spending by proj- from international repositories. NDoH has access to two ect and activity, fund source, and economic item. This sources of budget data for the health sector—national system, in principle, allows very detailed analysis of expen- budget books and data from the IFMS. Both data sources diture trends and this report benefited greatly from this have significant limitations that are often not apparent to system. the uninformed viewer. Several issues significantly affected the ability to ana- lyze trends in the sector, including: Budget Books Budget books only show health sector totals since • Significant expenditure deviations from both bud- 2015—before that, the health sector was included in gets and warrants, and large variations in figures the social sector (together with education). Sector fig- depending on the time at which a report was run. In ures (in the expenditure section of Budget Volume 1 and 2014, health sector expenditure was 146 percent of the the Final Budget Outcome) had several errors that signifi- revised budget and NDoH expenditure was 158 per- cantly distort analysis. The errors included the omission of cent of the budget. In 2015, sector expenditure was external financing in sector figures in the 2015 Final Bud- 79 percent of the revised budget, while NDoH expen- get Outcome (indicating a large underspend) without a diture was 25 percent of the revised budget. Officers in note to clarify the omission. In addition, sector budgets do NDoH, Treasury, and Finance could not explain these not add to total expenditure—leaving a large unexplained deviations on request. component that reduces the reliability of sector forward • Variations were especially apparent in the 2014 and projections in the 2015 and 2017 budget books. 2015 budgets, where figures from reports a few Macroeconomic data contains recurrent errors. One months after the completion of the budget year are example was the inclusion of two different Gross Domes- significantly different from reports one or two years tic Product (GDP) series in the 2017 budget books, without after. In the case of the 2015 budget, changes in the any explanation as to which one is correct and a miscalcu- system were apparently only made more than a year lation in the deficit (which does not equal the difference after the completion of the Final Budget Outcome for between revenue and expenditure). This makes calculat- that year (so sometime mid-2016). These changes were ing ratios and trends uncertain (for example, health as a significant enough that they affected the analysis and percentage of GDP in 2017 is either 2.3 percent or 1.7 per- conclusions in this report. cent depending on which GDP series is used). Figure 1A–1 shows a comparison of 2014 and 2015 Some important health expenditures are not visible health sector budgets from a variety of data sources, either in budget books or the IFMS. This includes poten- including preceding budget books (that is, projec- tially significant health infrastructure spending out of PSIP tions), current budget books, the Final Budget Out- and DSIP, for which there is no reliable data, and salaries of come document, and IFMS reports run at a variety of provincial health workers (Level 1–4 facilities), which are times. This clearly demonstrates the large range of esti- incorporated in a broad provincial staffing grant. mates of what should be very similar figures. 45 46 Health Financing System Assessment Figure 1A–1: Budget Data Variance (2014 and 2015) (millions of Kina) 2,000 1,800 1,600 1,400 1,200 1,000 2014 budget book 2015 budget book 800 2016 budget book 2015 final budget outcome 600 Official IFMS report (three months after completion of 400 budget) IFMS report—early 2017 200 Current IFMS revised budget Current IFMS warrants 2014 2015 Source: Based on discussions between a World Bank consultant and NDoH staff. Subnational Expenditure commitments and disbursements in this system deviate—sometimes significantly—from figures The problem of data variation is exacerbated at the in donor’s own reports. For example, DFaT disburse- subnational level as the IFMS only captures a lump ments in the CRS are only about 45 percent of DFaT’s sum transfer of the HFG as a single line to each prov- own reported figures in 2014 and 2015. GFATM funding ince. At the next level, provincial governments run a in the PNG CRS data misses the HIV and AIDS compo- legacy accounting system. NEFC regularly tries to review nent, even though this is shown in GFATM’s own reports. provincial expenditure (funded through HFGs), but large CRS estimates funding by whether funding is chan- variations in the chart of accounts and budget classifica- neled through donor or recipient governments, but has tions used makes this task challenging. The result is a com- a (large) category where these two are mixed, making it plete lack of consistent and reliable information on how difficult to estimate precisely how much development much provinces spend on frontline service delivery vs. assistance is on PNG Government systems. provincial administration and particular disease programs (for example, malaria, TB, HIV, as this is often not reported • IHME Development Assistance for Health: This is in provincial budgets). another well-prepared international repository that also includes funding for international NGOs. This lack of data presents a challenge in reviewing This source highly overstates GFATM and DFaT funding. expenditure against standards and policy priorities. It IHME does not have information on whether funding is also makes it impossible to estimate with any certainty how on government systems. much the government allocates as counterpart financing against the GFATM diseases and vaccination. GFATM cur- • PNG Government Budget Books and Accounting rently uses estimates to calculate cofinancing actuals of System: ODA donor figures (both budgets and actu- the government (for example, 10 percent of salaries of als) are captured in budget books each year by the health workers are used in malaria-related activities). Department of National Planning and Monitor- ing. The break-up of these is, however, too coarse for External Health Funding detailed analysis (for example, no break-up into dis- ease areas or cash and in-kind support) and forward There are a variety of funding sources for DP expen- estimates are generally too inaccurate to inform policy diture available in PNG and challenges were found in making (donors generally state that their projections each source: are quite imprecise, and a large donor simply assumes • OECD CRS: The OECD CRS is an international that funding will continue at precisely the current level repository that is generally well audited, however, over the forward estimates). The information is unable Data Challenges in PNG 47 to be used as a tool to understand how and where health system and publishes an annual update through donor funding in the sector is allocated and will change the NDoH website. The NHIS also tracks all facilities in the over time. The reporting of actuals in the system is country, including their operational status. The NHIS relies inconsistent, and there are several years where actual on information from provincial governments. Reporting funding (which should be reported by donors after the rates generally fluctuate between 85 and 90 percent. NDoH budget year to the Department of National Planning also cites concerns that provinces overstate achievements and Monitoring) is missing for some donors. For this and the quality and status of facilities. Anecdotally, NDoH reason, the report excluded donor funding in charts, staff have been told on many occasions by clients or staff unless otherwise indicated. of referral facilities that a facility has been closed for a long time, even if it is reported as open by the province. • Direct Donor Projections to NDoH: NDoH asks donors to prepare and provide projections to NDoH NDoH, the government more generally, and CHS have directly, with a break-up of in-kind support and dis- several databases that contain important informa- ease areas. The projections are not consistently pro- tion on health sector employees. Many of these systems vided, however, and forward estimates are not reliable duplicate information contained in other systems and, in enough for policy decisions. Without a clear break-up of other cases, the same information is collected but differ- program areas, funding channels, and implementation ent codes are used. These data systems include: (i) the gov- partners, NDoH cannot understand the current sup- ernment payroll system (Department of Finance); (ii) the port it will eventually have to absorb into government CHS payroll; (iii) the Health Care Practitioner’s Professional systems. Registration System (within NDoH); (iv) Health Human Resource Management System (within NDoH); (v) the National Headcount Survey 2009 (within NDoH); and Health System Outputs (vi) the Health Management Information System (within and Outcomes NDoH). All these systems have significant constraints and data gaps which have degraded over recent years. This report used information from international repos- This means that the only reliable audit of health workers itories as well as the NDoH NHIS. The NHIS captures was in 2009, and since then only imprecise estimates are input, process, output, and outcome indicators of the available. Bibliography Brown, A. N., and B. Gilbert. 2014. “The Papua New Guinea Government of Papua New Guinea. 2011. “National Health medical supply system—documenting opportunities and Service Standards for Papua New Guinea 2011–2020.” challenges to meet the Millenium Development Goals.” Waigani: Government of Papua New Guinea. Journal of Pharmaceutical Policy and Practice 7(5). ———. 2017. “2017 National Budget Book: Economic and Cairns, A., and X. Hou. 2015. “Financing the frontline: an Development Policies Vol. 1.” analytical review of provincial administrations’ rural health expenditure 2006–2012.” Washington, DC: World Bank. Hakizimana, E., B. Cyubahiro, A. Rukundo, A. Kabayiza, A. Mutabazi, R. Beach, R. Patel, J. E. Tongren and C. Karema. Christian Health Services. 2013. “Christian Health Services 2014. “Monitoring long-lasting insecticidal net (LLIN) Technical Assistance Mission Report.” durability to validate net serviceable life assumptions, in Rwanda.” Malaria Journal, 13:344. Department of National Planning and Monitoring (PNG). 2015. “Medium Term Development Plan 2: 2016–2017.” Howes, S., A. A. Mako, A. Swan, G. Walton, T. Webster, and Waigani: Department of National Planning and Monitoring. C. Wiltshire. 2014. “A lost decade? Service delivery and reforms in Papua New Guinea (2002-2012).” Canberra: The Department of Treasury (PNG). 2016. “2016 Mid-Year National Research Institute and the Development Policy Economic and Fiscal Outlook (MYEFO) Report.” Waigani: Centre. Department of Treasury. http://www.treasury.gov.pg/ html/national_budget/files/2013/budget_documents/ Institute for Health Metrics and Evaluation (IHME). Related%20Budget%20Documents/2016%20Mid%20 2016. “Global Burden of Disease database.” http://www Year%20Economic%20and%20Fiscal%20Outlook%20 .healthdata.org/gbd. Report.pdf International Monetary Fund. 2015. “IMF Executive Board ———. 2017. “Final Budget Outcome 2016.” Waigani: Concludes Article IV Consultation with Papua New Guinea.” Department of Treasury. http://www.treasury.gov.pg/ Press Release. Washington, DC: International Monetary html/national_budget/files/2013/budget_documents/ Fund. http://www.imf.org/external/pubs/ft/scr/2015/ Related%20Budget%20Documents/2016%20FBO%20 cr15318.pdf. -%20FINAL.pdf ———. 2016. “IMF Executive Board Concludes 2016 Arti- Gavi. 2017a. “Gavi’s Transition Policy and Country cle IV Consultation with Papua New Guinea.” Press Release. Preparedness.” http://www.results.org.au/gavis-transition- Washington, DC: International Monetary Fund. https:// policy-and-country-preparedness/ www.imf.org/en/Publications/CR/Issues/2017/01/30/ Papua-New-Guinea-2016-Article-IV-Consultation-Press- ———. Gavi. 2017b. “Gavi’s business model.” http:// Release-and-Staff-Report-44596 www.gavi.org/about/gavis-business-model/country- commitment-to-co-financing/ ———. 2016a. “World Economic Outlook (WEO) Database.” https://www.imf.org/external/pubs/ft/weo/2017/01/ ———. Gavi. 2017c. “Papua New Guinea: Key information weodata/index.aspx on co-financing.” http://www.gavi.org/country/papua- new-guinea/documents/cofiss/co-financing-information- ———. 2017. “2016 Article IV Country Report.” Washing- sheet-papua-new-guinea/ ton, DC: International Monetary Fund. 49 50 Health Financing System Assessment Irava, W., K. Barker, A. Somanathan, and X. Hou. 2015. “A ———. 2017.“Papua New Guinea Country Overview.”http:// snapshot of health equity in Papua New Guinea: an analy- www.theglobalfund.org/en/portfolio/country/?loc=PNG sis of the 2010 household income and expenditure survey.” Working Paper. Washington, DC: World Bank. UNICEF/WHO/World Bank/United Nations. 2014. “Levels & Trends in Child Mortality: Report 2014.” New York: UNICEF. National Department of Health (PNG). 2009. National Head Count Survey.Waigani: National Department of Health. United Nations. 2015. “World Population Prospects The 2015 Revision: Key findings & advance tables.” New York: ———. 2015. “MDR/XDR-TB Emergency Response Team United Nations. Annual Report.” Waigani: National Department of Health. Vince, J. D. 2017. “Millennium Development Goals: prog- ———. 2016. “2015 Sector Performance Annual Review.” ress in Oceania.” http://adc.bmj.com/content/archdischild/ Waigani: National Department of Health. 100/Suppl_1/S63.full.pdf ———. 2017. “2016 Sector Performance Annual Review.” World Bank. 2011. “PNG Health Workforce Crisis: A Call Waigani: National Department of Health. to Action.” http://documents.worldbank.org/curated/ en/216511468332461651/pdf/NonAsciiFileName0.pdf National Statistical Office (PNG). 2013. “Papua New Guinea—Household Income and Expenditure Survey ———. 2016. “World Development Indicators Data- 2009–2010: Summary Tables.” http://catalog.ihsn.org/ base.” https://data.worldbank.org/data-catalog/world- index.php/catalog/3243. development-indicators Organisation for Economic Co-operation and Develop- ———. 2017. “Global Tuberculosis Report.” Geneva: World ment (OECD). 2017. “Creditor Reporting System (CRS).” Health Organization. https://stats.oecd.org/Index.aspx?DataSetCode=CRS1 World Health Organization. 2016. “Global Health Observa- The Global Fund. 2014. “Malaria control in Papua tory (GHO) data.” http://www.who.int/gho/en/ New Guinea: Scaling up for impact.” https://www .theglobalfund.org/en/portfolio/country/grant/?k= World Health Organization and World Bank. 2017. e4617a46-d656-46ab-9fc7-d0600c7d2687&grant= “Tracking universal health coverage: First global PNG-812-G08-M monitoring report.” http://www.who.int/healthinfo/ universal_health_coverage/report/2015/en/