103444 Investing in Universal Health Coverage Opportunities and Challenges for Health Financing in the Democratic Republic of Congo Barroy H., Andre F., Mayaka S., & Samaha H. 2014 Health Public Expenditure Review acknowledgments This report was prepared by Hélène Barroy (Health The team coordinated and benefited from the fol- Economist, GHNDR). Françoise André (Health lowing government representatives’ contributions, Economist and Consultant) and Serge Mayaka comments, and analyses: Marcel Mukengeshayi (Public Health Expert and Consultant) collected Kupa (MoPH—Secretary General), Alain Iyeti central government and provincial data and contrib- (MoPH—Research and Planning Division, ad interim uted to the analysis. Hui Wang (Junior Professional Director), Raphael Nunga (MoPH—Research and Associate, GHNDR) supported the statistical anal- Planning Division), Robin Miteo (MoPH—Research ysis work, focusing particularly on efficiency and and Planning Division), Gérard Eloko (MoPH— equity. The report was produced under the stra- National Health Accounts Program), Georges Minga tegic and operational support of Hadia Samaha (MoPH—National Health Accounts Program), (Sr Operations Officer) and overall guidance of Trina Epiphane Ngumbu (MoPH—D1), and Fidèle Haque (Practice Manager, Health, Nutrition and Mokute Mopolo (Ministry of Economic Planning— Population Global Practice, World Bank Group). AIMP). They are warmly thanked for their availabil- ity and interest in supporting the production of this Michel Muvudi (Health Specialist and Consultant), Review. Franck Adoho (Poverty Specialist), Luc Laviolette (Senior Nutrition Specialist), Nicolas de Borman The team also received input and support from (Health Economist and Consultant), Saidou Diop other development partners active in the health (Senior Financial Specialist), Chaid Bou Habib sector in the Democratic Republic of Congo. It is (Senior Country Economist), and Emmanuel Pinto especially grateful to UNICEF and Lluis Vinyals (Lead Economist) all provided inputs in their respec- Torres (UNICEF) for support with the budget data tive fields to help understand the macroeconomic collection and sharing process for three provinces. and health sector environment and all contributed Acknowledgments also go to Jean-Paul Mvogo to the general analysis. The final report also ben- (IMF), Dan Pike, and Hamish Colquhoun (DFID). efited from comments from Driss Zine-Eddine The team warmly thanks the Canadian International (Sr  Health Economist, GHNDR), Gyorgy Bela Development Agency for contributing to funding Fritsche (Sr Health Specialist, GHNDR), Patrick this Review. Mullen (Sr Health Specialist) and Chadi Bou Habib (Sr Economist, GMFDR).   iii Table of Contents Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii Section 3. Government Health Expenditure by Type . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Acronyms and Abbreviations . . . . . . . . . . . . . . . . . . . . . vii Section 4. Expenditure for Health Personnel . . . . . . 39 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii Chapter 6. Health Expenditure Introduction: Country Characteristics Performance . . . . . . . . . . . . . . . . . . . . . . . 42 and Study Presentation . . . . . . . . . . . . . . xiii Section 1. Financial Protection . . . . . . . . . . . . . . . . . . 42 Chapter 1. Macroeconomic and Fiscal Section 2. Equity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 44 Environment . . . . . . . . . . . . . . . . . . . . . . 1 Section 3. Technical and Allocative Efficiency . . . . . . . 48 Section 1. Macroeconomic Environment . . . . . . . . . . 1 Chapter 7. Decentralization: Impact Section 2. Fiscal Environment . . . . . . . . . . . . . . . . . . 2 on Deployment and Use of Health Section 3. Decentralization: A New Tax Order . . . . . 3 Resources . . . . . . . . . . . . . . . . . . . . . . . . . 53 Section 4. Public Finance Management . . . . . . . . . . . 4 Section 1. Financing Flows (excluding wages) . . . . . . 53 Chapter 2. Health System: Organization Section 2. Provincial Health Financing Resources . . . 54 and Resources . . . . . . . . . . . . . . . . . . . . . 7 Section 3. Provincial Government Health Expenditure 56 Section 1. Health System Objectives Chapter 8. Main Policy and Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7 Recommendations and Roadmap Section 2. Health Care Infrastructure for Implementation . . . . . . . . . . . . . . . . . 58 and Equipment Availability . . . . . . . . . . . . . . . . . . . . . 9 Section 3. Human Resources for Health . . . . . . . . . . 10 References . . . . . . . . . . . . . . . . . . . . . . . . 61 Section 4. Availability and Funding of Pharmaceuticals . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Figures Figure 1: Contribution of the Main Economic Sectors Chapter 3. Health System Performance 18 to Growth (% of GDP), 2010–2013 . . . . . . . . . . . . . 1 Section 1. Health Outcomes . . . . . . . . . . . . . . . . . . . 18 Figure 2: Share of National Revenues (excluding Section 2. Service Coverage . . . . . . . . . . . . . . . . . . . 23 grants) in GDP in DRC, Low-Income Countries Section 3. Service Quality . . . . . . . . . . . . . . . . . . . . . 24 and Sub-Saharan Africa . . . . . . . . . . . . . . . . . . . . . . . . 2 Figure 3: Distribution of National Revenue Collected Chapter 4. Health Financing . . . . . . . . . . 27 per Province and Distribution of National Revenue Section 1. Health Financing Sources . . . . . . . . . . . . . 27 Retrocessions per Province, 2010 . . . . . . . . . . . . . . . 5 Section 2. Adequacy of Government Financing Figure 4: Distribution of the Health Care Infrastructure for Health and Fiscal Space . . . . . . . . . . . . . . . . . . . . 31 by Province: Number of Hospitals per Province and Ratio per 100,000 Inhabitants . . . . . . . . . . . . . . . . . . 10 Chapter 5. Government Financing Figure 5: Registration and Compensation of Health for Health . . . . . . . . . . . . . . . . . . . . . . . . . 35 Personnel in the Public Sector, DRC . . . . . . . . . . . . . 13 Section 1. Health Budget . . . . . . . . . . . . . . . . . . . . . . 35 Figure 6: Ratios of Physicians to Nurses per 10,000 Section 2. Government Health Expenditure—Low Inhabitants in DRC and Sub-Saharan Africa and Volatile . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37 (all income levels) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 iv 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 7: Geographic Distribution of Physicians Figure 29: Economic Classification of Government per Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 Health Expenditure, 2007–2013 (% of total) . . . . . . . 39 Figure 8: Geographic Distribution of Nurses/Midwives Figure 30: Execution of Government Health per Province . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Expenditure by Type of Expenditure (% of allocations), Figure 9: Availability of Essential Drugs . . . . . . . . . . . 16 2007–2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Figure 10: Age Pyramid, 2013 . . . . . . . . . . . . . . . . . . 18 Figure 31: Concentration Curves and Lorenz Curves for Out-of-Pocket Payments, DRC . . . . . . . . . . . . . . 43 Figure 11: Fertility Rates in DRC and Other Countries in the Region . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20 Figure 32: Effects of Health Expenditure on the Pen’s Parade Curve of Household Consumption . . . . . . . . 44 Figure 12: Use of Modern Contraceptive Methods in DRC and Other Countries in the Region . . . . . . . 20 Figure 33: Inequalities in the Use of Services by Income and by Service Level, DRC, 2013 . . . . . . . 47 Figure 13: Change in Child Mortality from 1997 to 2013–2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 Figure 34: Distribution of Government Subsidies by Income Quintile, DRC . . . . . . . . . . . . . . . . . . . . . 48 Figure 14: Change in Child Malnutrition in DRC and Underweight Children in DRC and Peer Countries . . 22 Figure 35: Geographic Distribution of Government Expenditure Per Capita, 2013 (CDF) . . . . . . . . . . . . . 49 Figure 15: Change in Maternal Mortality, DRC, 1990–2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23 Figure 36: Rate of Skilled Birth Attendance (left) and Antenatal Care (right) vs. Government Health Figure 16: Coverage for Antenatal Care, Skilled Birth Expenditure Per Capita and Peer Countries (2013 Attendance and Basic Treatment for Children in DRC or most recent available data) . . . . . . . . . . . . . . . . . . 49 and Peer Countries, 2001–2014 . . . . . . . . . . . . . . . . 23 Figure 37: Maternal Mortality vs. Government Health Figure 17: Antenatal Care Coverage in DRC and Peer Expenditure Per Capita, 2005 and 2010 . . . . . . . . . . 50 Countries, 2007–2013 . . . . . . . . . . . . . . . . . . . . . . . . 24 Figure 38: Infant (upper figures) and Under-5 (lower Figure 18: Main Reasons for Nonuse of Health Care, figures) Mortality Rates vs. Government Health DRC, 2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Expenditure Per Capita, 2007 and 2012 . . . . . . . . . . 51 Figure 19: Operational Level of Hospitals by Province: Figure 39: Availability of Qualified Personnel vs. Use Running Water and Facilities for Consultations (% of of Services for Treating Fever and Infant Mortality . . 52 hospitals) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Figure 40: Investment Transfers Allocation Figure 20: Availability of Full Obstetric Services for and Execution, 2010–2013 Total by Province in CDF Mothers, Average Score per Province, 2012–2013 . . 26 Million . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Figure 21: Total Health Expenditure Per Capita in DRC Figure 41: External Assistance and Government and the Rest of Sub-Saharan Africa, 2012 (USD) . . . 28 Funds in Health Financing Per Capita, CDF 2013 . . . 55 Figure 22: Total Health Expenditure as percentage Figure 42: Central and Provincial Government Funds of GDP in DRC and the Rest of Sub-Saharan Africa, for Financing Health Care Per Capita, CDF, 2013 . . . 55 2008–2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 43: Health Expenditure Per Capita Figure 23: External Financing and Household by Financing Source, 2011–2013 . . . . . . . . . . . . . . . . 56 Expenditure by Disease, DRC, 2012, (USD thousands) . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29 Figure 44: External Assistance Disbursements by Province, CDF, 2010–2013 . . . . . . . . . . . . . . . . . . 57 Figure 24: Household Health Expenditure, DRC, 2012 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30 Tables Figure 25: Structure of Domestic Tax Revenue (% of total revenue), 2011 . . . . . . . . . . . . . . . . . . . . . 31 Table 1: Summary of Data Sources . . . . . . . . . . . . . . xvi Figure 26: GDC Elasticity of Government Health Table 2: Growth in National Revenue Collected Expenditure, DRC, 2007–2016 . . . . . . . . . . . . . . . . . 33 and Revenue Retrocessions, in CDF billions, 2007–2010 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4 Figure 27: DRC Health Budget, 2003–2013 . . . . . . . . 36 Table 3: Local Revenues Collected by a Sample of Four Figure 28: Execution of Health Expenditure Using Provinces (excluding revenue retrocessions), 2013 . 4 Internal Resources (MoPH and Earmarked Transfers), 2007–2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38 Table of Contents   v 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 4: Breakdown of Revenues Collected and Revenue Table 19: Government Compensation of Health Retrocessions per Province (total 2007–2010, in CDF Personnel, 2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41 billion) and actual revenue retrocession rate (%) . . . 5 Table 20: Share of Health in Household Expenditure Table 5: Analysis of Budget Execution Delays and Impact by Income Quintile, DRC . . . . . . . . . . . . . . . . . . . . . 42 on Sector Expenditure Effectiveness . . . . . . . . . . . . . 6 Table 21: Estimated Incidence of Catastrophic Health Table 6: Summary of Main Strategic Health Goals . . 8 Expenditure, DRC, 2013 . . . . . . . . . . . . . . . . . . . . . . 43 Table 7: New Governance of the Health System Table 22: Concentration Indicators, Infant and Under-5 in DRC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Mortality by Province (sample) and National Indicator, Table 8: Availability and Operational Status of 2007–2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 the Health Care Infrastructure in DRC, 2013 . . . . . . 9 Table 23: Health Allocations as Percentage of Provincial Table 9: Health Personnel by Professional Category Government Budgets, 2010–2014 (%) . . . . . . . . . . . . 56 in DRC (2013) and Sub-Saharan Africa (most recent available data) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 Boxes Table 10: Proportion of Health Facilities Meeting National Box 1: Role of Faith-Based Organizations Standards per Category, Type of Health Facility and in the Health System in DRC . . . . . . . . . . . . . . . . . . 9 Location, 2014 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Box 2: Demographic scenarios for 2050 . . . . . . . . . . 19 Table 11: Estimate of Drug Expenditure, 2013 . . . . . 17 Box 3: A unique initiative funded by domestic Table 12: Main Health Financing Indicators, resources: the PESS . . . . . . . . . . . . . . . . . . . . . . . . . . 26 2008–2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27 Box 4: Priority Recommendations for the Health Table 13: Expanding Fiscal Space, 2014–2019 . . . . . . 32 Sector from the First Public Expenditure Review, Table 14: Summary of Potential Increases DRC, 2008 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 in Fiscal Space for Health . . . . . . . . . . . . . . . . . . . . . . 33 Box 5: Using the Health MTEF to Improve Table 15: Budget allocations, MoPH Budget, the Budget Process . . . . . . . . . . . . . . . . . . . . . . . . . . 37 2007–2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35 Box 6: Performance-Based Financing: Findings Table 16: Government Health Expenditure of an Impact Study in Katanga . . . . . . . . . . . . . . . . . . 52 (Execution), 2007–2013 . . . . . . . . . . . . . . . . . . . . . . . 37 Table 17: Analysis of the Government Health Expenditure Execution Chain for All Resources, 2011–2013 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40 Table 18: Wage Expenditure, 2009–2013 . . . . . . . . . 41 vi  Investing in Universal Health Coverage Acronymns and Abbreviations AIMP Aid and Investment Management Platform HRH Human Resources for Health ARI Acute Respiratory Infection IMF International Monetary Fund CDF Democratic Republic of Congo francs MDG Millennium Development Goal CIDA Canadian International Development Agency MICS Multi-Indicator Cluster Survey DFID UK Department for International MoPH Ministry of Public Health Development MTEF Medium-Term Expenditure Framework DGCMP Public Procurement Audit Division NHA National Health Accounts (Direction Générale pour le Contrôle des NHDP National Health Development Plan Marchés Publics) NSI National Statistics Institute DHS Demographic and Health Survey PESS Health Facility Enhancement Program DPS Provincial Health Divisions (Divisions (Programme d’Equipement des Structures Provinciales de la Santé) Sanitaires) DRC Democratic Republic of Congo SNAME National System for Procurement of ETD Local Entities (Entités Territoriales Essential Medicines (Système National Décentralisées) d’Approvisionnement en Médicaments FEDECAME Federation of Central Medical Stores Essentiels) (Fédération des Centrales d’Achat de UHC Universal Health Coverage Médicaments Essentiels) UNDP United Nations Development Programme GAVI Global Alliance for Vaccines and UNICEF United Nations Children’s Fund Immunization USAID United States Agency for International GDP Gross Domestic Product Development GHNDR Health, Nutrition and Population Global VAT Value-Added Tax Practice WDI World Development Indicators HGR District Hospital WHO World Health Organization HIV/AIDS Human Immunodeficiency Virus vii Overview Background and Policy When a dollar is allocated to health, only 40 cents are Recommendations actually spent and 20 cents are, rightly, targeted at priority interventions. To progress toward UHC, it is The Democratic Republic of Congo (DRC) is pre- essential to make health spending more effective and paring to adopt a historic bill on Universal Health efficient—generating more “bang for the buck”—an Coverage (UHC). Yet the government is question- agenda item be advanced in parallel to the suggested ing how effectively health financing reforms will increase in budget. Inefficiencies have emerged improve coverage and financial protection of chil- within and outside the health sector. A set of Public dren and the rest of the population—three-fifths of Finance Management measures shall be accelerated children with respiratory infections, diarrhea, and as well as sector-specific actions be implemented fever are not treated for these conditions. to free up health resources. One key sector-specific International aid and households’ direct payments measure is to refocus efforts on strengthening the finance 80 percent of the health sector (half each), availability and performance of primary and com- the government less than the remaining 20 percent. munity care services, which are vital for efficiency This pattern must be reversed, as it is neither sustain- but also for coverage expansion, especially of the able (aid is likely to be reduced by 2020–2030) nor poorest. The Global Financing Facility (GFF) can equitable (direct payments affect the poorest more). help underpin this full overhaul of health sector pri- Unless the central government spends more than ority allocations and support the scaling up of key its current one dollar a year per capita on health, high-impact interventions. it will make no serious progress toward UHC. The Equity and financial protection are likely to be health budget envelope should move out of its cur- improved if the country moves from a user fee– rent, almost forgotten, state (4 percent of the overall based system that really only just finances the “bare budget); a doubling of the budget is a must to have bones” functioning of facilities while deterring use the necessary critical mass. of services and accelerating catastrophic spending. Such doubling is possible, as DRC has considerable As a way to transitioning toward more equitable and fiscal potential. Tapping into this potential could harmonized health financing, provincial resource allow the country to generate substantial additional pooling should be explored over the medium term revenues and invest more for UHC. Whereas other as decentralization is consolidated and, it is hoped, countries in sub-Saharan Africa can mobilize 1 or transfers from central government to provinces as 2 extra percentage points of gross domestic prod- well as local revenues are strengthened. For these uct (GDP), DRC should be able to manage more changes to be effective, retrocessions must acceler- than 8 percentage points of GDP. New taxes are not ate to near the constitutional objective of 40  per- needed, but a more effective tax collection system— cent, earmarked transfers to provinces must be high particularly in mining—is. By increasing revenues enough for these jurisdictions to exercise their core from 13 to 21  percent of GDP by 2020, the state mandate of implementing health financing, and fis- would then have the domestic resources to invest cal and budgetary capacities at decentralized level more in priority interventions for universal coverage. must be firmed up. viii 2014 HEALTH PUBLIC EXPENDITURE REVIEW Nationwide prepayment mechanisms such as man- Health System and Results—Some Encouraging datory health insurance may well be advisable over Outcomes Against a Backdrop of Poor Infrastructure the longer term. In the short and medium term, how- and a Fragmented System ever, defining and then implementing a quality care In the mid-2000s, the country launched a deep package alongside in-depth reform of provider pay- reform of the health sector as a pillar of socioeco- ment systems (based on experiences from the coun- nomic reconstruction. Dispensing with the emer- try’s performance-based financing mechanism since gency approach that prevailed in the late 1990s 2004) could be a more doable alternative to extend during the war, the government has paved the way coverage and move effectively toward UHC in the for longer-term development of the health system. medium term. Some progress has been made, but the organiza- tion of the health care and service delivery system remains relatively fragmented and fragile. Key Findings The health system is no exception to the country’s Macroeconomic and Fiscal Environment—Strong twin infrastructure deficit—quantity and qual- Growth, but for What? ity. Geographic coverage has improved at primary The DRC economy over 2009–2013 enjoyed strong health care level but the dearth of referral operations growth at an annual average of 7.4 percent, driven is still a major problem, while the quality of infra- by a booming mining sector which, over 2010–2013, structure and equipment is highly problematic—less accounted for more than 20 percent of GDP growth. than 30  percent of health facilities are considered Yet the central government has not managed to tap operational. into this upturn in national wealth by raising its revenues. Still, there are multiple signs for hope: service cov- erage rose for a number of interventions essential Although domestic revenues shot up from 2004 to maternal health. Over 2001–2014 for example, to 2012 they have stagnated since then—worse, at antenatal care coverage climbed from 68 percent to 13 percent of GDP they are actually sliding back due 88 percent and the skilled birth attendance rate from to low collection of indirect taxes and revenues from 61 percent to 80 percent. The use of insecticide-treated the natural resources sector. This revenue downturn nets soared from 6 percent in 2007 to 56 percent in has had a knock-on effect on expenditure, capping it 2014. Coverage of essential treatments for children at 12 percent of GDP since 2011–2012. This raises is, however, reported to have regressed slightly over the question of how to fund the investment needed the period, especially for fever and acute respiratory to achieve the government’s aim of becoming a infection (ARI). Similarly, the proportion of fully middle-income economy by 2030. ­ vaccinated children apparently fell from 53 percent One avenue will be tapping the strong potential to 46  percent from 2010 to 2013–2014. Cost is a for local revenue generation in several mineral-rich major barrier to access, though poor service quality provinces. Another will be decentralization, which also constricts demand. has already given the country a new governance and Other encouraging data are that, after a slow start tax order. In accord with the 2006 Constitution, in the early 2000s, the reduction in child mortality provinces over the last few years have received started gathering pace after 2010, and infant mortal- greater political, administrative, and fiscal pow- ity, estimated at over 90 per 1,000 over 1997–2007, ers. The interprovincial redistribution system being fell sharply by 2013–2014 (to 58). Less happily, set up, gradually, is smoothing revenue inequalities malnutrition still affects 43 percent of children, with across provinces and paving the way for a new sense nearly half of them suffering from severe malnutri- of national solidarity. tion. Finally, following a sharp upturn in maternal Overview  ix 2014 HEALTH PUBLIC EXPENDITURE REVIEW mortality due to conflicts in the late 1990s, the rate the staff needed to be fully operational, especially in fell steadily to 846 deaths per 100,000 births over rural areas. While Kinshasa has 1.3 physicians per 2007–2014, though the rate remains higher than the 10,000, most of the other provinces post half of this regional average of 510. ratio. Loose regulation of public sector employment in the health sector accelerates the imbalances and results in a growing stock of personnel in the capital Health Financing—Minimal Spending and Heavy and other urban areas. Reliance on External Aid and Out-of-Pocket Expenditure At USD 13 per capita, DRC spends less than one- The health labor market is relatively saturated. tenth the average of the rest of sub-Saharan Africa HRH supply is rising faster than demand, mainly on health. Health expenditure as a share of GDP due to increased production of medical personnel fell from 4.6 percent in 2008 to 3.8 percent in from private institutions on one side—over 2,000 2011–2012. trained physicians every year—and the public sec- tor’s limited absorption capacity on the other side. External assistance is the leading source of health Due to structural and cyclical rigidities the public financing, accounting for an average 40  percent sector can only enroll and pay a salary to 31 percent of total health financing sources over 2008–2013. of the total stock of personnel. Household funds are the second largest (averaging 39.3  percent over the period). Out-of-pocket pay- Against a background of a shrinking share for ments accounted for more than 90 percent of house- (nonpersonnel) operating and capital expenditure, hold health expenditure. personnel spending is worrisome, doubling from 42 percent in 2007 to more than 80 percent of total At less than 15 percent of total health expenditure, government health expenditure after 2010. The the government’s share of health financing is mini- increase in total payroll is due mainly to large pay mal. Health accounted for an average of 4 percent of raises in the civil service (more than 140  percent the central government budget, excluding debt and from 2009 to 2013). financial expenses, over 2007–2013. The increase in the payroll raises questions about the Still, actual government expenditure on health aver- medium-term financial sustainability of this expendi- aged less than USD  1 per capita over 2007–2013, ture. Allocations for personnel costs are to be dou- climbing to 0.7 percent of GDP only in 2013. Health bled to just cover wages and bonuses of current staff, as a share of total government spending averaged amounting to one and half times the MoPH budget 3.95 percent over 2007–2103. for 2013. In particular, sustainability of the “risk Since 2011, however, implementation of decentral- bonus”—providing 85  percent of government pay- ization has led to more government financing for checks for physicians and rose by 236 percent from health through earmarked transfers to the provinces. 2009 to 2013—is a valid question for the government. These new transfers amounted to 150 percent of the MoPH budget in 2013 and appear to point to a new Additional Fiscal Space for Health—Prospects Show model of resource allocation to decentralized level. Good Potential Recent estimates of the per capita health financing Payroll—A Heavy Burden for the Budget with gap—USD 18 to USD 33—highlight the vast needs Unsatisfactory Results for Personnel and Patients that must be met to achieve universal coverage of Despite a growing stock of human resources for health care services in the country. While the level of health (HRH), the population does not have adequate funding is a critical issue for DRC, these estimates access to qualified health professionals nationwide. shall not be taken as definite spending “targets” and Poor distribution of human resources is such that a be an impediment for pursuing expenditure and large majority of primary-level facilities do not have efficiency-oriented reforms. x  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW The country seems to have several possibilities for they are between provinces. Prevalence of malnutri- increasing the fiscal space for health. The central tion, however, as estimated by the number of under- government could, as said, increase its revenue by weight children, is increasing among the poorest and 8 percentage points of GDP over the long term, by in rural areas. collecting more from the natural resources sector Major income-related disparities are reported in use and by improving existing domestic revenue collec- of services, but some of these are narrowing, includ- tion. All other factors being equal, greater central ing use of antenatal care and skilled birth atten- government revenue collection could generate up to dance (though a quarter of the poorest women do 0.4 percentage points of GDP more for health. not receive any antenatal care, against just 8 percent Giving health a greater priority in the central gov- of the richest). The use of services for treating child- ernment budget (up to 8  percent of government hood diarrhea has improved for all quintiles, nota- expenditure) could increase the health budget by bly for the poorest children, using community-based 0.6 percentage points of GDP. Equally important, delivery mechanisms. Conversely, ARI and malaria improved execution of health expenditure, as well treatment coverage, which government facilities as efficiency, would be needed so as to raise health provide for free, has decreased. sector performance and free up resources for the Government health expenditure is highly focused on sector. All in all, it is expected that more than one hospitals (87 percent), which are used more by the additional point of GDP could be generated from richest quintile, pointing to inequitable access. All domestic resources over the medium term to finance operating expenditure, modest though it is, goes vir- progress toward UHC. tually exclusively to hospitals. As hospitals are more Prospects from external assistance are less clear. used by the richest, the poorest quintile receives only The latest outlook for external assistance is gloomy. 16 percent of total public subsidies. The distribution This source—equivalent to 5.6 percent of GDP in of health expenditure among provinces is also imbal- 2013—is forecast to shrink to 1.3 percent by 2030, anced, at a ratio of 5 (the capital) to 1 (the rest). for about USD 5 per capita in the short term (2015– For the same level of expenditure, DRC’s perfor- 2016), but declining over 2018–2020. Although no mance is much poorer than the rest of the region’s projections of external aid for health exist for DRC, infant and under-five mortality. Performance also it is probable that aid to the sector will follow the largely varies across provinces with the same level of same pattern as that of total external assistance, as inputs, suggesting high technical efficiencies. Raising the country transitions to having greater political the quality of pre- and on-the-job training for rural stability and becoming a middle-income economy. personnel seems essential to improve efficiency. Health Expenditure Performance—Poor Financial Impact of Decentralization on Health Financing— Protection; Equity and Efficiency Are Concerns Opportunities Must Be Seized More than 80  percent of household expenditure Decentralization has the potential to change the takes the form of out-of-pocket expenditure, such volume and execution of health financing. A larger that financial protection against illness is almost share of government funds for health is likely to now nonexistent. The poorest—spending more of their transit through the provinces. The government’s pri- income on health than the rich—are hit hardest, orities mean that budget allocations to the provinces as they are by catastrophic health expenditure, for these sectors have increased since 2010–2011, which overall affects more than 10  percent of the but execution of transfers remains poor. population. As with the central government, external assistance Disparities in child health outcomes are large, but is the biggest health financing source for the prov- appear to be narrowing among income groups, as inces. In 2013, amounts provided to the provinces by Overview  xi 2014 HEALTH PUBLIC EXPENDITURE REVIEW external partners were nearly three times as high as provincial public resources are primarily used in total government funds allocated for health. Central the provinces to cover personnel expenditure to government funds account for the majority of pro- top up or compensate extra personnel, not paid by vincial public funding. Except for Katanga, funds the central level. Only a few provinces use internal raised by the provinces from their own resources resources purely for capital spending. to finance health are extremely small (perhaps Finishing and aligning decentralization with UHC US 10 cents per capita). goals is central for the sector. Decentralization is a Provincial budgets allocate only a tiny share to funding and governance opportunity that must be health—an average of 4 percent of their all resources seized by health policy makers for the success of the to the health sector in 2010–2014. Central and UHC agenda. xii  Investing in Universal Health Coverage Introduction: Country Characteristics and Study Presentation Country Overview nomic performance is due mainly to the buoyancy of the mining sector on high world prices for minerals, Political and Economic Environment and of trade and services. In 2007–2008, following two decades of war, the Democratic Republic of Congo (DRC) entered a In 2010, the government undertook to conduct phase of political and institutional consolidation. extensive reforms to improve macroeconomic The signing of peace agreements in 2002 put an end and tax stability. The major reforms relate to to the second war in DRC that had contributed to public finance, economic governance, the business the deaths of over 4.5 million people. However, secu- climate, use of banking services, and the dedol- rity remains uncertain, especially in the east: despite larization of the economy. Although all of these a military victory won by the government army over reforms have been launched, outcomes are partial. the M23 rebellion in November 2013, some 40 armed Considerable progress has been made on fiscal groups remain at large in the Nord and Sud Kivu consolidation, enabling the central government to provinces. As the 2016 presidential elections near, bring down its budget deficit. Inflation has fallen tensions are rising in the capital again over whether significantly. the current president will run for a third term. The 2006 adoption of a new Constitution grant- Sociodemographic and Human Development Situation ing substantial powers to 26 new provinces DRC is the third most populous country in sub- stands as a strong symbol of national consoli- Saharan Africa with over 70 million inhabit- dation. The Constitution also targets five priority ants and a population growth rate estimated at sectors for decentralization: health, education, agri- 3.1 percent in 2014.2 With half the population under culture, rural development, and infrastructure. This 16 years old, there is a great deal of pressure on the process is gradually introducing earmarked fiscal labor market as over 33 million individuals would transfers to the decentralized levels. like to find work. The country remains mainly rural (61.2 percent), with the capital, Kinshasa, account- The country has posted strong economic perfor- ing for 11.7 percent of the population and standing mance for five years (2009–2013) with economic out for its high share of households with a female growth at 8.5 percent in 2013, higher than the head (25.6 percent). regional average of 5 percent. It also showed marked improvements in development policies and institutions (Country Policy and Institutional Assessment [CPIA], 2014).1 The good macroeco- 2 The absence of any recent population census (the latest dates back to 1984 and the next one is being prepared) and massive population 1 The CPIA score for DRC was raised from 2.7 to 2.9 in 2014 in mobility due to the conflicts make it hard to estimate the population recognition of the introduction of public finance reforms and in any detail. The latest estimates available place the population at rigorous macro-economic management. 77.4 million inhabitants (National Statistics Institute [NSI], 2014). xiii 2014 HEALTH PUBLIC EXPENDITURE REVIEW Poverty has been reduced. Over 2005–2012 growth of 1.64 percent, consistently higher than the poverty was rolled back 8 percentage points from regional average since 2000. Life expectancy at birth 71.3 percent to 63.4 percent (NSI, 2013). The drop remains among the lowest in the world at 50 years was even sharper in rural areas where the inci- in 2012. dence of monetary poverty fell 10 percentage points (65.2  percent in 2012). However, four provinces still have an incidence of over 70  percent: Kasaï Rationale and Study’s Objectives Oriental (78.6 percent), Equateur (77 percent), Kasaï DRC is at a turning point in developing its health Occidental (74.9 percent), and Bandundu (74.6 per- system. The completion of the second National cent). Access to basic infrastructure, especially water Health Development Plan (2011–2015) marks the and electricity, remains very low with just 1 percent end of a period of extensive health system recon- of rural households having electricity and running struction following two decades of war. The govern- water (NSI, 2013). ment has put real effort into improving coverage, DRC has a high rate of literacy with over 73 per- assisted by development partners, with virtual uni- cent of the population able to read and write in versal coverage of primary health care operations. 2012. A full 62  percent of the population reports A vast government initiative, the Health Facility having been to school. However, the net primary Enhancement Program (PESS), was launched in 2011 enrollment rate is 68 percent, pointing to late school (becoming effective in 2013) to upgrade health cen- enrollment and massive repetition. The net sec- ters, and district, provincial, and tertiary hospitals. ondary enrollment rate plunges to 36  percent for In 2015 a consolidation phase will be launched to the 12- to 17-year-old age bracket. Female school improve the quality of and financial access to health enrollment has improved hugely, and has caught up care. The strategic definition of the third National with the male rate. Health Development Plan (2016–2020) is expected to calibrate DRC’s new post-millennium health sys- Maternal and child health has improved, but tem targets. access to quality health care services remains problematic. Under-five mortality fell steadily Health is a pillar of decentralization. In 2010, over the decade to 104 per 1,000 in 2013/2014. Yet DRC launched a political, administrative, and tax 60 percent of children under five years old nation- decentralization process granting greater powers to wide are not covered by basic treatment services for the provinces and other local entities (ETDs). Health diarrhea, fever, and respiratory infections. Maternal is among the sectors top of the agenda for this health also remains a major concern. It is falling transfer of responsibilities. The process is expected more slowly than in other countries and is one of the to improve the response to needs by strengthening highest in the region with 846 deaths per 100,000 local governance. However, the real implications for births. Despite relatively satisfactory antenatal care health sector financing and governance are relatively coverage, there is a lack of emergency obstetric uncharted. care. At this rate, Millennium Development Goal The health financing system is also undergoing (MDG) 4 (60 deaths per 1,000 births) and MDG 5 a phase of transition and strategic redefinition. (332 deaths per 100,000 births) will not be achieved During the war and immediate postwar years, the by either 2015 (the MDG end date) or 2020 (the sector was financed largely by users and interna- national goal). tional aid in the form of emergency programs. Where DRC scores are second to last in the 2013 UNDP users had access to health facilities, they paid out of Human Development Report ranking with an their own pockets for their treatment under a cost- index of 0.338. Still, it appears to be catching up recovery mechanism set up in the 1990s. External with the region’s performances with annual index assistance financed most of the capital and operating xiv  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW costs. Through the mid-2000s, government funds Review also draws on the recently published health invested in health care barely reached USD 50 cents financing analysis guidance developed by WHO5 per capita a year. Since 2010, with action scaled up and the tools developed jointly by the World Bank in 2012–2013, the government’s commitment to and WHO to measure progress toward UHC.6 several priority sectors including health has seen a sharp increase in allocations. Government funding, in various forms, has focused on upgrading supply. Methodology Universal Health Coverage (UHC) is on DRC’s The study analyzes both the central and the provin- political agenda. Taking up this priority as a core cial level with a sample of six provinces selected non- objective of the current government’s roadmap,3 randomly: Bandundu, Equateur, Katanga, Maniema, authorities are set to progress toward UHC by Province Orientale, and Sud Kivu. The analysis is 2020. The administration is therefore preparing a based on primary and secondary data sources avail- new health financing strategy with a view to pass- able for DRC (Table 1) in the second quarter of 2014 ing a historic bill in 2015. Yet the absence of an up- with respect to: to-date diagnosis of health sector financing and the • The macroeconomic and fiscal environment; scarcity of data makes it a difficult exercise. • Health sector resources and expenditure from the The purpose of this study is to establish a diagno- government, households, and external sources; sis of the health sector’s national and provincial • Health system and health expenditure perfor- financing in order to assist DRC with its transi- mance—coverage, financial protection, efficiency, tion to UHC. The study’s specific objectives are to: and equity; and • Health outcomes. • Measure the availability of financial resources for health and analyze financing trends for 2007–2013; The provincial analysis is based on the primary and • Analyze the prospects for increasing available secondary data available for Katanga, Sud Kivu, and resources for health in the macroeconomic and fis- Province Orientale for the first quarter of 2014, and cal midterm environment; for Bandundu, Equateur, and Maniema for the sec- • Assess the nature and performance of health spend- ond quarter of that year. The main sources include ing in terms of coverage, financial protection, fiscal data for each province and the state of total equity, and efficiency; and and sector government expenditure, along with data • Investigate the impact of decentralization on the drawn from household surveys (Demographic and health sector’s financing. Health Surveys, Multi-Indicator Cluster Surveys [MICS], and 1-2-3 surveys on employment, the The main guiding tool used to conduct this study is informal sector, and household consumption). the World Bank Group’s analytical framework for public expenditure reviews for human development, The central data were collected and compiled and the health sector in particular.4 This framework by a World Bank team from April to September proposes an analysis of health financing based on 2014. Provincial data were collected in Bandundu, three main parameters: health finance sources; actual Maniema, and Equateur by a World Bank team sector expenditure by finance source, with a focus in September 2014. Data collection in Katanga, on government expenditure; and health expenditure Province Orientale, and Sud Kivu was conducted performance in terms of equity and efficiency. The by a UNICEF team from January to March 2014. 3 “Universal health coverage for mothers and children is a top national priority to be achieved by 2020,” H.E. the Minister of Health, 5 http://www.who.int/health_financing/tools/hf_for_uhc_situation_ National Health Steering Committee, Kinshasa, December 2014. analysis_v1.0.pdf?ua=1 4 http://siteresources.worldbank.org/EXTPERGUIDE/Resources/ 6 http://apps.who.int/iris/bitstream/10665/112824/1/WHO_HIS_ PER-Complete.pdf HIA_14.1_eng.pdf?ua=1 Introduction: Country Characteristics and Study Presentation   xv 2014 HEALTH PUBLIC EXPENDITURE REVIEW A World Bank team conducted the analysis from of the GIBS sub-financing commission and individ- September to December 2014. ual interviews. Broad-based consultations were held throughout The study’s preliminary findings were presented at a the study with all the central government authorities training workshop on UHC in Matadi (Bas Congo) (ministries of health, the budget, economic planning, in December 2014. Its main recommendations were economy and finance) and provincial authorities presented at the highest level to H.E. the Ministers (provincial ministries of health and governorate) of Finance, Public Service, and Health in Kinshasa in along with development partners (Global Fund December 2014 and April 2015. A final consultation to Fight AIDS, Tuberculosis and Malaria, GAVI, workshop was held in Kinshasa in April 2015 with UNICEF, IMF, DFID, USAID, and CIDA) in the form government and development partners. Table 1  Summary of Data Sources Type of data Source Origin Period covered Fiscal data Macroeconomic framework and Medium Term Ministries of Finance, of Budget, and 2009–2014 Expenditure Framework of Economic Planning General budget (Budget Act) Ministry of Budget 2007–2013 Expenditure chain Ministry of Economy and Finance 2007–2013 National Health Accounts (NHA) Ministry of Public Health 2008–2013 Budget execution statements Ministry of Budget 2010–2014 Data on the National Survey on the Situation of Women and National Statistics Institute, Ministry 1997 state of health, Children (ENSEF) of Economic Planning coverage Multiple Indicator Cluster Survey 2 (MICS 2) Ministry of Economic Planning 2001 and 2010 and financial Demographic and Health Survey (DHS) Ministry of Economic Planning 2007 and 2013/2014 protection 1-2-3 Household survey (poverty and service use) National Statistics Institute 2013 Annual reports (national and provincial) of the Ministry of Public Health 2007–2014 National Health Information System Data on 1-2-3 Household survey (household consumption National Statistics Institute, Ministry 2013 household and financial protection) of Economic Planning expenditure Data on Aid and Investment Management Platform (AIMP) Ministry of Economic Planning 2014 external NHA Ministry of Public Health 2008–2013 expenditure Provincial Macroeconomic framework per province and Ministry of Economic Planning, 2013–2014 ­fiscal data provincial Medium-Term Expenditure Framework Ministry of Public Health (MTEF) (where such exists) Provincial Governments Statement of provincial revenues Provincial Governments 2012–2014 Statement of revenue retrocessions (to the State Audit Office (Cour des 2014 provinces) Comptes) Provincial budget provisions Provincial Governments 2011–2014 (depending on the province) Provincial government expenditure monitoring Provincial Governments 2011–2014 (depending on province) xvi  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Report Structure actuals). Government resources for health are taken to mean all domestic financing sources allocated to The report contains eight chapters. Chapter 1 leads off the Ministry of Public Health and other health enti- with the main macroeconomic and fiscal determinants ties. Section 1 examines changes in the health budget to understand the general health financing situation in envelope over 2007–2013, while section 2 focuses DRC. Its sections 1 and 2 provide a brief overview of on actual spending. Section 3 analyzes govern- the country’s macroeconomic and fiscal environment. ment executed expenditure by nature and Section 4 Section 3 sheds new light on decentralization—a core reviews trends and types of personnel expenditure— health issue—and focuses on the fiscal implications of the largest share of government health expenditure. decentralization. Section 4 offers a brief analysis of the management of public finance, a key element for the Chapter 6 focuses on health expenditure perfor- effectiveness and quality of expenditure, especially in mance in the light of three main parameters: financial health. protection; equity; and efficiency. Section 1 analyzes financial protection using standard indicators (share Outlining the DRC health system, Chapter 2 starts of out-of-pocket payments, catastrophic expendi- with a brief overview of the main objectives and ture, and impoverishing expenditure). Section 2 scru- analyzes its organization and governance, with tinizes health outcomes and service use disparities/ a focus on health care delivery. It then presents a inequalities based on income, gender, and place of detailed analysis of the availability and distribution residence. The last section presents an overview of of inputs, especially infrastructure and equipment the efficiency of health expenditure in DRC, primar- (Section 2), health personnel (Section 3), and phar- ily in comparison to its peer countries. maceuticals (Section 4). Chapter 7 analyzes health financing from the Chapter 3 discusses the performance of the main point of view of the provinces, the new “entitled” health system’s outputs and outcomes. Section  1 authorities for the health sector, drawing on a sur- analyzes major changes in health outcomes based vey of financial and fiscal data from six provinces. on different household surveys. Section 2 addresses Section  1 presents an overview of health financing service coverage with a focus on mother and child flows after decentralization. Section 2 focuses on services. Section 3 explores service quality issues. financing sources for the health sector at the provin- In a detailed analysis of health financing sources, cial level, examining provincial government funds, Chapter 4 looks at changes in public, external, external assistance, and central government trans- and private sources over 2008–2013 (Section 1). fers. The chapter ends with an analysis of the vol- Section  2 analyzes the adequacy of government ume and type of decentralized government health funding for health financing needs, and assesses the expenditure. prospects for expanding fiscal space for health. The report concludes in a brief Chapter 8 with a series Chapter 5 examines government funds mobilized for of recommendations to improve short- and medium- health, both allocated in the budget and executed (or term health sector financing and performance. Introduction: Country Characteristics and Study Presentation   xvii Macroeconomic and Fiscal Environment 1 Section 1. Macroeconomic Figure 1  C  ontribution of the Main Economic Sectors to Growth (% of GDP), 2010–2013 Environment Other, Agriculture, DRC has enjoyed a period of strong, sustainable 8.6 7 economic growth over 2009–2013, averaging Services, 7.5 Extractive 7.4 percent a year. In 2013, DRC posted one of the Industries, 17.5 highest real growth rates in the region at 8.5 percent Transport and compared with a regional average of 5 percent. The Telecom, 9 growth rate is forecast to have reached 9 percent in 2014, posting an average of 8.7 percent over 2014– 2019 (World Bank, 2015). This good economic per- formance is due to an upturn in public and private investment in an environment of relative security and Other Trade, 24.6 political stability. However, DRC’s per capita GDP Industries, 25.8 (USD 445 in 2013) remains one of the region’s lowest. Source: Adapted from World Bank, 2015. The booming mining sector is driving growth up, accounting for around 20  percent of GDP growth over 2010–2013 (Figure 1). The mining sector’s value-added grew 10.5 percent on average over the subsector was the food industry. The tertiary period as investments made the previous decade sector—including trade, services, transport, telecom- ­ came onstream. Preliminary data point to a 28 per- munications and other services—grew 9 percent on cent increase in the mining sector’s value added in average over 2010–2013, although growth in 2013 2013, while the general mining production index was only 4.4 percent. The tertiary sector is estimated rose 17  percent. Copper production alone grew to have provided nearly 46 percent of total growth 52 percent, which largely offset a 7.4 percent price over 2010–2013. The tertiary sector’s strongest per- drop observed in 2013 (World Bank, 2015). formance was in trade, reflecting the general recov- ery bolstered by stabilized prices and exchange rates The manufacturing and tertiary sectors also in 2010–2011 (World Bank, 2014). showed buoyancy. Growth in manufacturing posted an estimated average 11.4 percent over 2010–2013, Half of DRC’s economic activity still comes from and 21.4  percent in 2013. Manufacturing report- the informal sector:  97  percent in agriculture, edly provided 25.8 percent of total growth over the 86 percent in manufacturing, and over 40 percent in four-year period. The most buoyant manufacturing mining and tertiary industries. 1 2014 HEALTH PUBLIC EXPENDITURE REVIEW Despite strong economic growth, the labor mar- (VAT)—which fell from 30  percent to 27  percent ket shows tensions. If the extended unemployment of domestic revenue from 2012 to 2013, and low rate is taken to include jobseekers who have stopped revenue collected from the natural resources sector looking for work due to a lack of vacancies, unem- (IMF, 2014; World Bank, 2014). This revenue ratio ployment could be as high as 17.7  percent of the places DRC far behind its peers posting revenues of population, with 31 percent of them in urban areas, 23 percent of GDP on average in sub-Saharan Africa and 39 percent in Kinshasa alone. Youth unemploy- over 2008–2012, and 15 percent in the least devel- ment is a huge concern, at 28 percent of the under- oped countries over 2009–2010 (excluding grants) 24s in the labor force. Nearly four in 10 of those (Figure 2). out of work hope to find a job soon. All in all, it The downturn in revenues has had a knock-on is estimated that over 33.6 million individuals want effect on the expenditure to bring the budget to find work. The informal sector outweighs all the deficit under control. Domestic government expen- other sectors with 88.6 percent of employed work- diture has stagnated at 12  percent of GDP since ers nationwide, ranging from 62.7  percent in the 2011–2012; its share in GDP rose from 6.4 percent capital to 94.8 percent in rural areas. in 2004 to 13.2 percent in 2011 before sliding back After peaking at over 40 percent in 2009, infla- to 12.1 percent in 2012 and 12.5 percent in 2013. tion has remained below 2 percent since 2012. The government has reduced the level of public In 2013, it came in at 0.8  percent following fiscal expenditure since 2011, mainly by introducing caps restraint, control of monetary aggregates, and an on fiscal commitments, in a move to sustain a posi- absence of major import price shocks. The decelera- tive domestic fiscal balance (0.5  percent in 2013, tion of inflation has coincided with a slowdown in excluding payment of arrears). the growth of the money supply, which has fallen from over 50 percent in 2008 and 2009 to around 20 percent since 2011. The reduction in the budget Figure 2  S  hare of National Revenues (excluding grants) deficit and the consequent restriction of monetary in GDP in DRC, Low-Income Countries and financing of central government expenditure have Sub-Saharan Africa helped contain inflation. 30.0 DRC’s external debt has fallen sharply since 2010 25% Minimum Recommended (IMF) 25.0 when the country reached the completion point for the Heavily Indebted Poor Countries Initiative 20.0 (HIPC). The public external debt ratio fell from 75 percent in 2009 to 18 percent in 2013, after sub- 15.0 stantial debt relief in 2010 (IMF, 2014). 10.0 Section 2. Fiscal Environment 5.0 Despite strong growth, the central govern- ment has not managed to tap into the upturn in 0.0 national wealth and raise its revenue. Domestic 2008 2009 2010 2011 2012 2013 revenue grew faster than GDP over 2004–2012, with Low Income Countries an elasticity of 1.45, but has stagnated since 2012. Sub-Saharan Africa In 2013, central government revenue was 13  per- DRC cent of GDP, down on the 14.9  percent of 2012, Source: Authors’ calculations, based on World Development Indicators mainly due to low indirect taxes—value-added tax (WDI, 2014) and IMF (2013). 2  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW The recent downturn in revenues had put the local entities authority to tax. Provincial resources burden of fiscal consolidation on expenditure comprise three main sources: local tax resources;7 and has accentuated the problems of financ- resources from national revenue allocated to the ing development. Fiscal consolidation has mainly provinces; and extraordinary resources. hit capital expenditure, transfers, and expenditure Decentralization implies in principle an increase on goods and services. Domestically financed capi- in allocations, especially in the social sectors for tal expenditure reached a low 1.2 percent of GDP, which responsibility has been transferred. The before recovering to around 2 percent in 2013. 2006 Constitution stipulates that 40  percent of Expenditure on goods and services has leveled off national tax revenue collected in the provinces is at 2.4 percent (World Bank, 2015). Payroll has con- to be passed on to the provincial authorities in the tinued to rise from 3.7 percent of GDP in 2009 to form of revenue retrocessions to increase their fiscal 5.0 percent of GDP in 2013, but remains moderate capacity. The Constitution also lays down the prin- relative to the average of sub-Saharan Africa (8.5 ciple of equalization across provinces in the form of percent). Extraordinary expenditure has increased an additional 10 percent allocation. Adding together sharply, from less than 0.5 percent of GDP in 2008 personnel compensation, operating subsidies, and to more than 1.6 percent in 2012, reflecting a sub- investment subsidies for the five priority sectors stantial rise in security expenditure over 2008–2012. combined, expenditure earmarked for the provinces Government expenditure is concentrated largely is estimated at approximately one-third (32.8  per- on current expenditure, to the detriment of cap- cent) of fiscal allocations in 2014 (as opposed to half ital expenditure. Current (personnel and operating) in 2009). expenditure accounted for an average 86 percent of Since 2007, the retrocession system (being set government expenditure over 2009–2013. At 41 per- up) has passed on only a small proportion of the cent of the total, the largest share of government revenue collected in the provinces.8 Tax revenue expenditure goes on government employee wages, a retrocessions form a relatively new source of pro- share that has increased at the rate of 30 percent a vincial budget financing, and more than CDF 100 year on average over the last 10 years (World Bank, billion has been transferred annually since 2008.9 2015). Expenditure on goods and services accounts However, retrocessions remain below the 40 percent for one-third of total expenditure, with a sharper threshold set by the Constitution: over 2007–2010, increase posted in transfers to the provinces. an average of 15.7 percent of national revenue col- lected was estimated to have been transferred back to the provinces (Table 2). Other sources estimate Section 3. Decentralization: A New even lower rates of actual retrocession, ranging from Tax Order 7.2 percent in 2010 to 11.4 percent in 2008 (IMF, The new Constitution gives the provinces their 2014). own fiscal and tax powers. The provinces and local entities (ETDs) have been granted autonomy 7 Own resources include provincial taxes and levies (including real and free administration of their resource manage- property taxes), public utility taxes and fees, and taxes specific to each province. ment (Article 3 of the Constitution). Act N.11/011 8 The Constitution provides for revenue to be “retained at the source,” of July  13, 2011 confirms the provinces’ fiscal but this does not happen in practice. Revenues from State-controlled autonomy and stipulates distinct fiscal authority: entities are paid into the Public Treasury. Central government then calculates budgets for the provinces (“retrocessions”) and retains a the central government budget is governed by the share to pay for the costs of the responsibilities not yet transferred to Budget Act, provincial budgets by provincial budget the provinces. 9 Over 2003–2006, the exchange rate was in a range of CDF 400–468 edicts, and local entities’ budgets by budget “deci- for USD 1. Since 2010, the rate has been higher than CDF 900 per sions.” The legislation also grants provinces and USD 1 (CDF 919 in 2013). Macroeconomic and Fiscal Environment    3 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 2  Growth in National Revenue Collected and Revenue panel), they only receive 40 percent of total revenue Retrocessions, in CDF billions, 2007–2010 retrocessions (Figure 3, left panel—although Katanga Actual and Bas Congo receive the largest shares of total Revenue* Revenue retrocession revenue retrocessions, at 17 percent and 13 percent   collected retrocessions rate (%) respectively). According to State Audit Office data, 2007 358.7 50.0 13.9 the adjustment would mean that these three high 2008 608.5 112.4 18.5 revenue-collecting provinces receive proportionally 2009 790.2 134.1 17.0 fewer retrocessions—6–8  percent of the sums col- 2010 1,034.8 143.2 13.8 lected. Conversely, the revenue retrocession rates Total 2007–2010 2,792.1 439.7 15.7 for Bandundu, Equateur, and Maniema—the three Source: Authors’ calculations based on State Audit Office data, 2013. provinces with the highest incidence of ­ poverty— * National revenue, except oil revenue. stand at around 600–1,250 percent (Table 4). Local revenues collected by the provinces are Section 4. Public Finance highly imbalanced across the provinces and Management are generally low despite the considerable tax Despite recent changes, shortcomings in public potential of certain areas. For example, Maniema finance management continue to undermine the collected CDF 534 million in 2013, or less than USD volume and quality of government expenditure 0.27 per inhabitant. Katanga, a mineral-rich prov- in DRC. The country has launched a far-reaching ince, collected roughly CDF 32 billion that year, public finance modernization campaign starting which is still only USD 2.90 per capita (Table 3). with the adoption of a vast program of public finance These small provincial revenues are such that some reforms via the establishment of COREF10 (2009), provincial governments today find themselves with the adoption of a Strategic Public Finance Reform very little fiscal leeway despite their greater sector Plan (2010), the revision of the Public Procurement responsibilities. Code (2010), and the passing of a new Public Finance Decentralization has introduced an interpro- Act (the LOFIP) in 2011.11 These reforms cover the vincial redistribution system that could tend to entire budgetary process as well as the administra- smooth inequalities among provinces. Although tion and accounts management of expenditure.12 three provinces (Katanga, Kinshasa, and Bas Congo) These reforms aim to deconcentrate the budget and collect 88 percent of national revenue (Figure 3, right finance ministry payment process to the line min- istries. The Administrative and Financial Divisions (AFDs), scheduled to be set up in each ministry in 2015, are the latest part of this reform. However, Table 3  Local Revenues Collected by a Sample of Four fiscal management performance remains suboptimal Provinces (excluding revenue retrocessions), 2013 at best. Local revenues Total local revenues per capita Provinces (CDF millions) (2013 current USD) 10 The Public Finance Reform Steering Committee (Comité Katanga 32,489 2.94 d’Orientation des Réformes des Finances Publiques) is the body in charge of steering the implementation of the public finance reform. P. Orientale 5,640 1.08 11 Act N.11/2011 of July 13, 2011. Sud Kivu 5,275 0.72 12 The pillars of the reform are to reform the tax system to modernize Maniema 534 0.27 the tax instruments and tax administration; streamline the management of government expenditure by improving budget Source: Authors’ calculations, based on Provincial Government data, 2014. procedures; improve the accounting system and cash-flow Note: The sample was defined based on data available in the last quarter of management; and improve the public finance auditing system via 2014. technical and institutional support to the audit bodies. 4  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 3   Distribution of National Revenue Collected per Province and Distribution of National Revenue Retrocessions per Province, 2010 5% 3% 6% 10% 0.5% 2% 5% 1% 5% Kinshasa Kinshasa Bas Congo 27% Bas Congo 13% Bandundu Bandundu 11% Equateur Equateur Katanga Katanga 35% Kasai Oriental Kasai Oriental 7% Kasai Occidental Kasai Occidental 9% Province Orientale Province Orientale Mainiema Mainiema 10% Nord Kivu Nord Kivu 7% 26% 17% 0.5% Source: Authors’ calculations, based on State Audit Office data, 2013. Table 4  Breakdown of Revenues Collected and Revenue Sector Expenditure Framework (MTSEF) for the sec- Retrocessions per Province (total 2007–2010, in CDF billion) tors was also expected to improve the predictability and Actual Revenue Retrocession Rate (%) of budget allocations in the priority sectors (health, Revenue Revenue Actual revenue education, agriculture, rural development, and collected retrocessions retrocession infrastructure). An MTSEF has been prepared for (% of total) (% of total) rate (%) the health sector since 2011. Although the tool has Kinshasa 27 10.1 5.9 been an effective training and information vehicle Bas Congo 26.4 13.3 7.9 on budget preparation technique at the Ministry of Bandundu 0.1 7 1,257.2 Public Health (MoPH), it has done little to improve Equateur 0.3 9.9 604.7 the predictability of budget allocations and choices. Katanga 34.6 17.4 7.9 Kasaï Oriental 0.6 7.4 200.2 Although expenditure channels have been sys- Kasaï Occidental 0.4 8.4 338.1 tematized, obstacles still hamper sound execu- Orientale 1.9 10.7 90 tion, especially for social allocations. The current Maniéma 0.1 4.4 583.7 execution procedure has been improved with four Nord Kivu 5.1 6.1 18.7 standard expenditure steps (Table 5). The Budget Sud Kivu 3.6 5.3 23.3 Commitment Plans—introduced to keep expen- diture in line with cash flow and restrict slippage Total 100 100 15.7 (multitude of emergency procedures,13 overruns, and Source: Authors’ calculations, based on State Audit Office data, 2013. initiation from outside the expenditure chain)—have limitations in practice, like delays, no connection to the original allocation, unsuitable timeframes, and Budget preparation still shows little standard- meager commitments. Although the actual transfers ization or systematization. The introduction of generally match these plans, expenditure validation general “steering” tools—the Medium-Term Fiscal and payment orders are more problematic. These Framework (MTFF) and budget preparation by program—has, to a certain extent, improved the ­ 13 The Extraordinary Expenditure Procedure (EEP) rolls the first three predictability of the expected volume of revenue and procedures into one (commitment, validation and payment order). It is used for emergency expenditure and extraordinary expenditure expenditure and improved the budget’s clarity and (ministerial orders N.291/CAB/Min/Finances/2008 and 039/Min/ credibility. The introduction of the Medium-Term Budget/2008). Macroeconomic and Fiscal Environment    5 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 5  Analysis of Budget Execution Delays and Impact on Sector Expenditure Effectiveness Expenditure Responsible steps administration Problems identified Commitment  Budget Delays issuing the quarterly Budget Commitment Plans. Disconnect between these plans and the sector budget. Limitation of sums available for commitment (liquidity). Technical ministry Delays/errors in the preparation (departments; sub-managers), signature (Minister) and submission (sub-manager) of the commitment voucher in accordance with the Budget Commitment Plan. Validation of Budget Delays in the signature (charged to the following quarter; loss of quarterly transfers for the MoPH). expenditure Long processing delays/standstills. Highly manual procedures. Payment Finance Payment orders can take two to three months to sign (loss of quarterly transfers). Carryover transfers. order Manual procedure. Payment Treasury Delays/errors in authorization for payment and the bank transfer. Significant lag between opening the dossier (requisition) and releasing the payment. Source: Authors; Ministry of Budget 2014; World Bank 2012; World Bank 2015. Note: A commitment is the act by which the state creates or recognizes an obligation on the state from which a charge will arise.Validation of expenditure checks that the debt exists and calculates the exact amount of the outlay. A payment order is an administrative act by which, in accordance with the results of the validation calculations, the order is given to the cashier to pay the state’s debt. Payment is the act by which the state is released from its debt. two steps remain highly manual and, although lead- Since 2010, the expenditure chain has been refor- times have been reduced, overdue payments are mulated. The aim of reforms to the computerized still the norm and carryover transfers are frequent expenditure monitoring base was to improve read- (World Bank, 2012). ability (reduce duplication), bring it into line with the budget classification and introduce new special The low level of government expenditure exe- modules (for payment in particular). The next step cution across all resources has fallen further in will be to set up an integrated information system recent years, whereas allocations have gener- for all ministries to consolidate monitoring of all ally been raised. Execution has fallen across the government expenditure. entire budget and for the social sectors in particular. The total expenditure execution rate tumbled from A reform of the public procurement system was 77  percent in 2008 to 52  percent in 2011 before launched in 2010 on account of its importance picking up to 56  percent in 2013 (World Bank, in expenditure. A Public Procurement Regulatory 2015). Expenditure execution of total resources fell Authority, Public Procurement Audit Division even more sharply for the social sectors and infra- (DGCMP), and management units in the depart- structure from 77 percent in 2008 to 35 percent in ments were set up. Operating resources remain low, 2012–2013. The low level of execution for the social however. Its 12-step process involves validation by sectors, especially health, is due mainly to the over- nonobjection at each step, but selection alone can estimation of allocations from external resources take three or four months. With external assistance, and nonexecution of transfers to the provinces several bodies have been set up to smooth purchas- (Chapter 5). ing and investment in health when done via public procurement (World Bank, 2015). 6  Investing in Universal Health Coverage Health System: Organization and Resources 2 Section 1. Health System Objectives national Operational Action Plan. (Provincial opera- and Organization tional action plans have also been written.) Health is one of the pillars of DRC’s socioeco- Consolidation of the health system calls for nomic reconstruction. The war has had a consider- improving the health districts (zones de santé) able impact on the health system and the population’s that form the mainstay of the system’s peripheral state of health. After two decades of conflict esti- governance. This focus should scale up access to qual- mated to have cost over 4.5 million lives, the govern- ity primary health care by having local management ment is aware of the need to build its human capital geared to needs. It should also reduce the system’s for future development. The second Growth and fragmentation by furthering the integration of vertical Poverty Reduction Strategy (GPRS  2, 2011–2015) health programs. The health district is the basic health adopted in 2011 prioritizes access to basic social ser- care planning and primary health care delivery unit, vices (Pillar III). The strategy is designed to contrib- and its activities are coordinated by a Management ute to poverty reduction mainly by improving the Team. A health district covers some 100,000–200,000 population’s health and nutrition. The government inhabitants, with 10  health centers and one district adopted the second National Health Development hospital (Hôpital Général de Référence, or HGR). Plan (NHDP II) in 2011 with its timeline aligned The Minimum Care Package provided by health cen- with that of the GPRS 2. This plan aims to consoli- ters covers first-referral curative, preventive, promo- date the health system to better meet the popula- tional, and assistance activities. The Supplementary tion’s health needs. Care Package covers the clinical services of internal medicine, surgery, ­ gynecology-obstetrics, and pediat- Since the mid-2000s, the country has engaged rics in district hospitals. in major reforms of health sector governance. Moving from the emergency approach of the late Over the last few years, more responsibility has 1990s, the government has paved the way for longer- been shifted to the provinces. In keeping with the term development, despite emergency operations in 2006 Constitution, governance of the health sec- the eastern part of the country especially. The strong tor, now the exclusive reserve of the provinces, has political will to set a long-term strategic vision for seen a redefinition of responsibilities between the the health sector can be seen from the multiple strat- health administration’s central and decentralized egies developed since 2010 (Table 6). Program tools levels (Table 7). The MoPH now has a regulatory have also been set up to improve monitoring of the and normative role (Article 202 of the Constitution) sector’s core outputs and performance, such as an with responsibility for policy implementation at outcomes framework for the NHDP (2011) and a the province level (Article 204). The 26 Provincial 7 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 6  Summary of Main Strategic Health Goals Date of adoption Strategy document Main objectives or strategic goals 2013 National malnutrition reduction strategy Reduce acute and severe malnutrition 2011 Growth and Poverty Reduction Strategy Improve access to basic social services to improve the population’s state of health (GPRS 2, 2011–2015) and nutrition 2011 National Health Development Plan II Develop the health district (NHDP II 2011–2015) Improve sector leadership and governance Improve intersector cooperation 2010 Health System Strengthening Strategy Correct distortions due to vertical programs (HSSS) Reorganize the central and intermediate level Develop human resources for health Reform the drugs sector Reform health financing Scale up intersector collaboration 2010 National Human Resources for Health Improve careers management (rational management and use of HRH) Development Plan (2011–2015) Improve HR retention Increase the output of health professionals Take forward in-service training 2005 Health Financing Strategy Increase sector financing Increase the predictability of internal and external government expenditure Scale up prepaid mechanisms (community-based insurance and other forms of insurance) Reduce the fragmentation of international aid for health Improve the populations’ financial accessibility to quality health care 2000 Adoption of national health standards Definition of quantitative and qualitative standards for the health care package Minimum Care Package Supplementary Care Package Table 7  New Governance of the Health System in DRC Health Divisions (Divisions Provinciales de la Santé, Administrative DPS) established in December 2014, reporting to organization Health organization Provincial Minister of Health, have been tasked with Central level Minister of Health’s Tertiary-level managing and overseeing provincial health admin- staff national hospital istration. Nevertheless, deconcentrated Provincial Secretariat General Health Inspectorates (Inspections Provinciales de for Health Central divisions and Santé, IPS) are to continue supervising and monitor- programs ing at the provincial level with central ministry over- Provincial Provincial MoPH Secondary-level sight. The old health care district has disappeared level Provincial Health provincial hospital and District Manager Offices (Bureaux chef de Divisions Zones, BCZs) oversee the primary-level health facil- Provincial Health Inspectorates ities (district hospitals, health centers, and health Operational Central Health District hospital posts). level District Office Health center Health post The nonprofit private sector has always played a dominant role in care delivery and health sys- Source: Authors, based on information from the MoPH, 2014. tem structure. Through 1991, 62 percent of health districts were managed by churches (Murru and Pavignani, 2013). It is estimated that the nonprofit private sector was managing over one-third of the health facilities in 2013, either on its own account 8  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Box 1  Role of Faith-Based Organizations in the Health System operations, due mainly to an increase in the output in DRC of medical personnel. This used to be a characteristic of the health districts in urban areas, but is now seen Some 35–40 percent of health districts are managed or increasingly in rural areas (NHDP, 2011). comanaged by networks of faith-based organizations; an estimated 40–50 percent of health facilities are managed by such bodies. Their delivery of care is well integrated into Section 2. Health Care Infrastructure the government system, without any real split or separate duties. They form part of service delivery considered public and Equipment Availability (with or without a government contract), rather than strictly The health sector is no exception to the coun- “private.” try’s infrastructure deficit. DRC had a recorded Health facilities managed by faith-based bodies may be 8,266 health centers, 393 district hospitals, and owned either by a church or the state. The Catholic Church, 5  provincial hospitals in 2013. The ratio of infra- for example, estimates that it works in 50 percent of health structure coverage to national targets was 95 percent districts and supports 27 percent of health facilities fully and for primary level, 76  percent for district hospitals, 67 percent partially. The MoPH pays the wages of some staff working in health facilities run by faith-based outfits. and 19  percent for secondary level. Although geo- An estimated 65 percent of staff working in health facilities graphic coverage has improved over the last decade, supported by the Catholic Church are registered as health especially at primary level where it has reached the personnel with the MoPH, and 30 percent are on MoPH standard of one center for every 10,000 inhabitants, payment registers. there is a dearth of referral operations with less than What makes the situation in DRC different from other one district hospital for every 200,000 inhabitants countries is that, beyond playing a major role in providing nationwide. Secondary referral hospitals (or provin- services in the community, health centers, and hospitals, cial hospitals) and tertiary (or national) hospitals faith-based networks also help manage or comanage health remain scarce in the public health system given a districts. Diocesan Offices are heavily involved in managing 30-year absence of any strategy or financial policy districts. for them. Source: Adapted from NHDP II, 2011 and DFID, 2010. Less than 30 percent of primary- and secondary- level facilities are operational (Table 8). Most or delegated on behalf of the central government. An health care operations/services offer patchy coverage estimated 35  percent of district hospitals are man- capacity due to failings tied to aging infrastructure aged by faith-based organizations (Box 1), 60  per- and lack of equipment. A 2013 MoPH study reports cent by central government, and 6  percent by the that only 31  percent of primary health centers are for-profit private sector (MoPH, 2010, NHDP). considered operational and capable of delivering the Recent years have seen the exponential growth of expected care package. Two-thirds are reported to the for-profit private sector, especially primary-level have been built using materials not made to last. A Table 8  Availability and Operational Status of the Health Care Infrastructure in DRC, 2013 Percentage operational Ratio (to DRC Target Actual Operational of total population) standard Health centers 8,628 8,266 2,588 0.31 1.08 per 10,000 1 per 10,000 District hospitals 516 393 31 0.08 0.49 per 100,000 1 per 100,000 Provincial hospitals 26 5 N/A N/A N/A N/A Source: MoPH, 2013; authors’ calculations. Health System: Organization and Resources   9 2014 HEALTH PUBLIC EXPENDITURE REVIEW mere 8  percent of district hospitals are considered hospitals ranges from 0.33 hospitals per 100,000 capable of delivering the supplementary care pack- inhabitants in Province Orientale to 0.71 in Kasaï age. Nationwide, 22  percent of the hospitals have Occidental. Although the province of Kinshasa has electricity and 32 percent have running water, drink- a population of over 10 million, it is not the best ing or otherwise. Just 1 percent have full laboratory covered province with just 33 district hospitals—or equipment. The majority of hospitals (59  percent) 0.31 hospitals per 100,000 inhabitants (or 1 hospi- were built and equipped before independence, and tal per 322,000 inhabitants) (Figure 4). little investment has been made since. Fewer than 10 hospitals were built over 1960–2010. Section 3. Human Resources Referral infrastructure is concentrated in urban areas. Most district hospitals are in or around pro- for Health vincial capitals, leaving the majority of the popula- Availability of Health Personnel: Understaffing tion without direct access to a referral establishment. versus Overstaffing For example, the three urban health districts in Sud The issue of availability of qualified health per- Kivu province each has three district hospitals, yet sonnel in DRC is thorny. By international stan- one-third of the rural health districts do not have dards, DRC has an acute health personnel crisis a single hospital. In Province Orientale, 29 rural (WHO, 2009). In 2013, the country was estimated health districts have no referral establishments to have just under 6,000 physicians and 72,000 whereas the city of Kisangani has three. Some prov- nurses, i.e., 11 qualified health personnel per 10,000 inces suffer from a greater lack of hospitals than inhabitants—according to WHO standards, the others: Maniema for example has just nine district country should more than double its medical staff to hospitals, Katanga 59. The distribution of district 23. With only 0.7 physicians per 10,000 inhabitants Figure 4  D  istribution of the Health Care Infrastructure by Province: Number of Hospitals per Province and Ratio per 100,000 Inhabitants 70 0.8 52 59 0.7 60 54 49 0.6 50 41 0.5 39 40 33 0.4 30 30 28 24 0.3 20 0.2 9 10 0.1 0 0 e a a vu go vu a l al u r a eu d al ng m as nt t Ki Ki en on un nt ie t h rie ta ua id rie an ns d C d nd Ka O or Su Eq cc s Ki O M Ba N Ba i O sa c e i Ka sa in ov Ka Pr Number of hospital (left) Ratio hospital for 100,000 (droite) Source: Authors’ calculations, based on MoPH data, 2011. 10  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 9  Health Personnel by Professional Category in DRC (2013) and Sub-Saharan Africa (most recent available data) Total Nurses and health Administrative   Physicians midwives personnel personnel DRC Staff (2013) 5,719 71,472 89,557 52,202 Ratio per 10,000 (2013) 0.7 8.9 11.1 6.5 International comparisons (per 10,000 inhabitants) Sub-Saharan Africa (2009) 0.3 1.7 2.0 N/A World (2009) 1.4 2.8 4.1 N/A Standards (per 10,000 inhabitants) DRC standard 0.3 4.8 1.3 N/A WHO standard N/A N/A 23 N/A Source: Authors’ calculations based on MoPH data (2013) for DRC and WHO for other countries. http:// apps.who.int/gho/data/view.main.92000 in 2013, DRC has one of the lowest physician ratios remain chronically understaffed (World Bank and in the world. University of Kinshasa, forthcoming). Yet the MoPH paints a picture of overstaffing in The health labor market is relatively saturated. Human Resources for Health (HRH). National HRH supply is rising faster than demand, mainly standards—below WHO standards—classify the due to increased production of medical personnel on country as overstaffed. In particular, nursing staff the one hand, and the public sector’s limited absorp- and midwives (8.9 practicing per 10,000 inhabit- tion capacity on the other. ants) are estimated at nearly double national stan- The country has seen surging production of dards (4.8 per 10,000) (Table 9). Although precise medical personnel. Such production across all numbers of practicing personnel are subject to dis- disciplines has boomed in the last decade with the cussion absent a comprehensive up-to-date census, expansion of universities and provincial schools. this estimate is corroborated by anecdotal evidence, The number of medical schools shot up from three especially in the hospitals and for certain categories in 1998 to 61 in 2013.14 The number of paramedi- of unskilled personnel, reportedly forcing some peo- cal training schools doubled over 1998–2013. ple to work on a monthly roster basis due to low Potentially over 2,000 physicians and 4,000 nurses workloads. can graduate from the country’s schools and fac- In fact, despite a growing stock of HRH, the ulties every year. The increase in graduate produc- population does not have adequate access to tion has had a positive effect on the total density qualified health professionals nationwide. Poor of working personnel, which soared from 0.1 to 0.7 distribution of human resources is such that a large physicians per 10,000 inhabitants over 2006–2013. majority of primary-level facilities do not have the The increase in the production of nursing graduates staff they need to be fully operational, especially in also appears to have driven up the number of nurses rural areas. Although some urban operations may well be overstaffed, an analysis by province suggests 14 The 2009 MoPH assessment and the 2013 HRH directory count 406 technical institutes of medicine as opposed to 255 in 1998, 100 that most of the health facilities, especially at lower higher technical institutes of medicine as opposed to 53 in 1998, and levels of the health care pyramid and in rural areas, 61 medical schools as opposed to 3 in 1998. Health System: Organization and Resources   11 2014 HEALTH PUBLIC EXPENDITURE REVIEW working in the country, with 8.9 nurses per 10,000 2012–2013, but seems not to have put any real brake inhabitants in 2013. on entry into service, especially at the decentralized level, due to flexible implementation of recruitment/ But growth in graduate production has come commissioning rules (estimated by the Civil Service at the expense of skill levels. The development of Ministry in 2014). Entry continued to rise sharply schools and training institutes in the provinces has over 2009–2013, especially for physicians. Today, become a lucrative business, with the emphasis on the sector has 20  percent unregistered “new units” the quantity rather than quality: two-thirds of new (35 percent on average across the entire civil service). training centers have no technical facilities for medi- cal students to practice what they learn (MoPH, It is hard to exit the civil service because it does 2013). Without a solid accreditation system, the not offer decent pensions. Thus many staff who increase in production of medical personnel does have reached retirement age remain in service and/ nothing to raise the availability of skilled personnel. or part of official staff numbers. Over 10  percent of staff registered by the MoPH are estimated due The supply of trained staff is larger than the for retirement, but stay on the payroll. In 2013, a public sector’s absorption capacity. DRC’s higher request was sent to the Civil Service Ministry for education system turns out over 2,000 trained approval to retire 9,182 health employees—4  per- physicians every year, more than one-third of the cent of all retirements to be brought in by end-2015 number of public sector physicians in 2013.15 This (among a total of 244,198 estimated by the Civil is too much for the civil service to absorb given its Service Ministry in 2014). structural (budget constraints) and cyclical (flow management) rigidities. Yet despite the public sec- Capacity for registration17 and compensation tor’s limited recruitment capacity, there is little regu- are constrained, resulting in nearly half the lation of personnel flows resulting in the majority workforce being effectively informal and unpaid. of the public workforce being unpaid by the central Just 31  percent of the workforce is registered and government. receives a formal monthly salary from the civil ser- vice. Of the personnel working, 49 percent are regis- There is little regulation of what used to be virtu- tered, but not on the payroll18 (Figure 5). ally automatic entry into the public sector. Before the early 2000s, entry into the public sector was Public sector constraints prompt a high brain drain almost automatic.16 Nearly all medical graduates were from the health labor market. Over 1995–2005, an assured of a position in the civil service, at central or estimated 9–13 percent of physicians and 12 percent of provincial level, from the moment they were registered nurses emigrated (Clemens and Pettersson, 2007). The with the Order of Physicians. Over the last decade, main destinations were other countries in the region entry has been regulated more, due to civil service caps (79 percent to South Africa, Zambia, Rwanda and, for on government employee numbers. A freeze on wage the eastern part of the country, Uganda and Angola), earners among medical personnel was introduced in followed by Europe (15 percent). On graduating, an estimated half of all graduates seek work in the pri- 15 Figures vary by source, but usually tally: 8,597 graduates from vate for-profit or nonprofit sector. Nongovernmental institutes of medical technology (ITMs) and institutes of medical organizations alone reportedly take on over 30  per- education (IEMs) over 2001–2005 and roughly 8,000 over 2007– cent of new medical graduates. 2009 (NHDP II); 3,424 graduates from all courses for the 403 ITMs/ IEMs (MoPH, Annual Report, January 2014). 16 At the central level, a medical graduate duly registered with the 17 Registration (immatriculation) is the official recognition of staff by Order of Physicians used to receive a letter of commission (right the Civil Service. It is equivalent to tenure. to practice in a health facility) following a central assignment 18 A recent study on the availability of HRH in four provinces committee meeting (MoPH Secretary General authority). The MoPH (Bandundu, Equateur, Katanga and Sud-Kivu) shows that the rate virtually never refused registration. The provinces used to be able to of paid staff is even lower at 21 percent of the sample (across 1,771 issue a letter of assignment/commission, which was then sent to the staff interviewed). Mechanization is lower at primary level due to the Secretary General for approval. absence of physicians on the staff. 12  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 5  R  egistration and Compensation of Health Personnel Distribution of Health Personnel in the Public Sector, DRC Nurses features disproportionately among 20% HRH. With 8.9 nurses working for every 10,000 inhabitants, paramedical tasks would appear to be 31% largely covered in DRC. The inverse physician/nurse ratio, with over 12 nurses to 1 physician (Figure 6), as opposed to 1 to 2.04 worldwide, confirms the shortage of physicians but also reflects a dispro- portionately high percentage of practicing nurses. This situation is due to the systematic delegation of clinical tasks to nurses at the primary level, met by a growing output of nursing graduates in the provinces. The geographic distribution of HRH shows large disparities between Kinshasa and the rest of the 49% country, especially for physicians. Kinshasa has 1.3 physicians per 10,000 inhabitants, while most Registered Remunerated Staff of the other provinces post half this ratio (Equateur, Registered Non-Remunerated Staff Bandundu, the two Kasaï, and Province Orientale). Non-Registered Non-Remunerated Source: Authors’ calculations, based on MoPH data 2013. Figure 6  R  atios of Physicians to Nurses per 10,000 Inhabitants in DRC and Sub-Saharan Africa (all income levels) 30 Namibia Botswana 25 Ratio nurses for 10,000 habitants 20 Sao Tome 15 Angola 10 DRC 5 Nigeria 0 0 1 2 3 4 5 Ratio physicians for 10,000 habitants Source: Authors, based on MoPH data for DRC and WHO, 2014 or latest available data. Health System: Organization and Resources   13 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 7   Geographic Distribution of Physicians per Province 7,000 1.4 6,000 1.2 5,000 1.0 4,000 0.8 3,000 0.6 2,000 0.4 1,000 0.2 0 0.0 go du r l al ga a vu e vu sa L ta u m al TA nt te en Ki Ki ha on un n nt ie rie ta ua TO id an rie ns C nd d d Ka iO or Su Eq cc s Ki M O Ba N Ba iO sa ce Ka sa in ov Ka Pr Number of physicians (left) Ratio physician per 100,000 (right) Source: Authors’ calculations, based on MoPH (2013) and Equateur DPS (2014) data. The only exception is Sud Kivu with a ratio of 1.1 Section 4. Availability and Funding (Figure 7). The reverse is true for nurses: Bandundu of Pharmaceuticals and Equateur seem better off than Kinshasa with over 13,000 nurses each against Kinshasa’s fewer Availability than 4,000 (4 per 10,000) (Figure 8). Poor availability of essential drugs in health establishments is one of the major bottlenecks The distribution of HRH is also problematic to quality health care provision. In 2010, just within provinces, with rural areas losing out. 15 percent of tracer drugs were reportedly available Across a sample of four provinces, the facilities lack- in district hospitals (MoPH, 2011). The SARA sur- ing staff are mainly in rural areas (Table 10). Rural vey conducted from 2012 to 2013 put average avail- health centers are mainly short of nurses (infirm- ability of tracer drugs at around 20 percent across iers) and administrators/managers.19 As an example, the more than 1,000 health facilities in its sample Equateur’s rural areas lack physicians (médecins), (Figure 9). None of the health facilities has all the administrators, and pharmacists. Acute shortages essential drugs. Oral rehydration salts—an essential are reported at district hospital level, while the stock treatment for diarrhea in children—are found in of personnel is closer to national norms in urban just 2 percent of health facilities. Access to essential areas. The main deficit overall is among physicians drugs is almost nonexistent in certain provinces: five and technicians, especially pharmacists. provinces have less than 20  percent availability of tracer drugs. Maniema, for example, has less than 13 percent of the drugs (1 in 17) on average. Lack of development and fragmentation of 19 National standards do not provide for physicians at health centers. the supply system is restricting access to 14  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 8   Geographic Distribution of Nurses/Midwives per Province 80,000 20.0 18.0 70,000 16.0 60,000 14.0 50,000 12.0 40,000 10.0 30,000 8.0 6.0 20,000 4.0 10,000 2.0 0 0.0 go du r l al ga a vu e vu sa L ta u m al TA nt te en Ki Ki ha on un n nt ie rie ta ua TO id an rie ns C nd d d Ka iO or Su Eq cc s Ki M O Ba N Ba iO sa ce Ka sa in ov Ka Pr Nurses (left) Ratio nurses per 10,000 hab. (right) Source: Authors’ calculations, based on MoPH (2013) and Equateur DPS (2014) data. Table 10  Proportion of Health Facilities Meeting National Standards per Category, Type of Health Facility and Location, 2014 Source: World Bank and University of Kinshasa, forthcoming. Note: The number in each cell is the proportion (maximum = 1) of health facilities meeting the health standards. A red cell means that the standards are met by a small number of health facilities and that there are not enough health personnel in this category, province, type of facility, and district. An orange or green cell is a sign of an increase in the proportion of health facilities meeting national standards. “Autre = other.” Health System: Organization and Resources   15 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 9   Availability of Essential Drugs Average availability of items % of facilities holding all the items Fluoxetine tablets Amitriptyline tablets Beclomethasone Calcium blocker Simvastatine tablets Metfomine capsules/tablets Enalapril or other ACE inhibitors Insulin injectable Glibenclamide capsules/tablets Omeprazole tablets Salbutamol inhaling Gentamicine injectable Ceftriax one injectable Zinc sulfate tablets/sirop Ibuprofene tablets Amoxicilline tablets/sirop/suspension Ampicilline injectable 0% 10% 20% 30% 40% 50% 60% Availability Index (%) Source: SARA survey, MoPH, forthcoming. quality drugs. A National System for Procurement the effect on the quality of the drugs circulated is of Essential Medicines (SNAME) set up in 2002 disastrous. Most private importers and wholesalers has done little to improve access to drugs. Drugs operate outside of the rules of best pharmaceutical storage and distribution is decentralized through a distribution practices. The private sector is not regu- network of 15 regional medical stores (Centrales lated and the quality of the drugs circulating is a de Distribution Régionales), but their performance problem. A small-sample study in Kinshasa province remains poor and many health districts do not have in 2011 found that over 60 percent of drugs in cir- one. Further, the quantification of needs is not very culation were unfit for consumption (expired, fake, reliable and is affected by vertical programs working or counterfeit) (MoPH, 2011). in “silos.” Logistical capacities remain poor, while When the drug required is available, it is not certain regions are relatively inaccessible (e.g. prov- always well managed and used. The Register on the ince of Equateur). Transport is extremely expensive, Use of Essential Drugs and Revenues is not in wide- weighing heavily on drugs budgets and making it spread use and is sometimes based on archaic tools. impossible to place small emergency orders. In terms of use, a recent analysis of medical prescrip- Frequent stockouts mean that health facilities tions reveals a strong tendency to overprescribe with and patients turn to the private sector for medi- an average of four drugs per prescription (antibiot- cal supplies. Yet this sector is poorly regulated, and ics in over 60 percent of cases). Overprescribing is 16  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW likely associated with the provider payment system Table 11  Estimate of Drug Expenditure, 2013 in all facilities, including primary care level, which Funding source USD (2013) typically relies on itemized (fee-for-service) billing. Households* 159,542,594 External assistance** 96,729,682 Funding Public sector*** 2,378,868 Drugs are funded largely by users. A cost recov- Source: Authors’ calculations, based on MoPH data. ery system is in place in DRC based on the Bamako * Estimate based on the NHA (2012 data). initiative. It was supposed to provide health facilities ** Amount disbursed in 2013, as reported by the development partners and with financial autonomy and guarantee financially published by the MoPH Research and Planning Division, 2014. sustainable drugs renewal. In practice, in DRC as in *** Estimate based on predicted expenditure for the PESS program drugs line in 2013, the main drug expenditure item using government funds. other countries in the region, cost recovery has not prevented stockouts. Yet the cost recovery system has reportedly made substantial proceeds, which have not all been systematically reinvested in purchasing and 99 distribution channels involving over 50 dif- drugs. A study conducted in Sud Kivu finds a small ferent partners. This fragmentation undermines the share of proceeds reinvested in health establishments system’s efficiency at all stages of the supply cycle and drugs availability in particular (DCC, 2014). (selection, quantification, procurement, and infor- For example, annual Federation of Central Medical mation management) and creates huge disparities in Stores (FEDECAME) orders are estimated to total drug accessibility between provinces (WHO, 2010). one-tenth of the amounts disbursed by households Government funds invested in drugs are pal- for drugs at the point of use of the services. try. Before the launch of the Health Facility External assistance is another important source Enhancement Program (PESS)—a vast national of drug funding. Donors are estimated to have spent initiative—government expenditure on drugs was over USD 96 million on drugs in 2013. These drugs virtually nonexistent (Table 11). This program now are distributed through two channels: centralized provides for the government to support the supply purchases by the national FEDECAME network of targeted health districts with essential drugs and and supplies (the majority of distribution) through other consumables from its own resources. Likewise, each partner’s own channels, which then handle the government has undertaken to finance part of distribution through to the health facilities. A 2010 the purchase of vaccines from its own resources (the WHO study finds more than 19 supply agencies 2015 allocation is USD 11 million). Health System: Organization and Resources   17 Health System Performance 3 Section 1. Health Outcomes DRC’s fertility rate is one of the highest in the region, at 6.6 children per woman (Figure 11). Demographic Issues Fertility reportedly rose slowly from 6.2 children DRC faces immense demographic challenges. per woman in 2007 to 6.3 in 2010 and 6.6 in With a natural increase rate estimated at 3.1 percent 2014. When this rate is disaggregated at provincial in 2014, forecasts suggest that the population will level, marked differences are found between urban double by 2050 and could approximate 150 million areas (5.4 children per woman) and rural areas (Box 2). Currently, nearly half of the population is (7.6). Adolescent pregnancies have also increased, under 16 years old and 20 million people are aged especially in rural areas, from 24  percent in 2007 15 to 30 (Figure 10). to 27  percent in 2013–2014 among women aged Figure 10  Age Pyramid, 2013 RDC RDC 80 ans et plus 70–74 ans 60–64 ans 50–54 ans FEMMES 40–44 ans HOMMES 30–34 ans Migrantes Migrants 20–24 ans Natifs Natives 10–14 ans 0–5 ans 8000 7000 6000 5000 4000 3000 2000 1 000 – – 1000 2000 3000 4000 5000 6000 7000 8000 Milliers Milliers Source: NSI, 2013. Note: Hommes = men; femmes = women; milliers = thousands; ans = years. 18 2014 HEALTH PUBLIC EXPENDITURE REVIEW Box 2  Demographic Scenarios for 2050 Demographic projections show that it is possible to intervene The essential difference between these assumptions lies in on the demographic variable, through accelerating contracep- the age structure of the DRC’s population in 2050. On the slow tive use.Three scenarios of dissemination of contraception have scenario, young people under 15 and those aged 15–24 are in been formulated: slow, with an increase in prevalence of mod- 2050 twice as numerous as in 2014. On the very fast scenario, ern contraception of 1.0 percentage point a year from 2014; the number of young people under the age of 15 are stabiliz- rapid, with an increase of 1.5 points a year; and very fast, with an ing and the 15–24 age group are less numerous than before. increase of 2.0 points a year. These scenarios lead to a decrease This scenario also sees a shift in the population growth rate in the fertility rate of about 6.5 children per woman today to from 3 percent to 1 percent in 2050 (2.4 percent in the slow respectively to 4.4, 3.1, and 2.2 children per woman in 2050. But scenario), which would lead to a more rapid increase of GDP irrespective of hypothesis, the population of the DRC will be per capita. In emerging countries, the elimination of the costs at least 100 million by 2030, and in 2050 could be 140 million attributable to a continued increase in numbers of pregnan- for the very fast scenario, double today’s, and 185 on the slow cies, births, and young people has generated a tax gain that has scenario. Current high fertility and the many young people of allowed them generally to better respond to social demands child-bearing age will therefore mean that the DRC’s popula- in health and education, and to implement policies that benefit tion will continue growing strongly in 2050 and beyond. from the demographic dividend. Box figure   Age Pyramids in 2014 and 2050 with Two Hypotheses of Slow and Rapid Dissemination of Contraception Slow Dissemination of Contraception Rapid Dissemination of Contraception 80–84 80–84 75–79 75–79 70–74 70–74 65–69 65–69 60–64 60–64 55–59 55–59 50–54 50–54 45–49 45–49 40–44 40–44 Age Age 35–39 35–39 30–34 30–34 25–29 25–29 20–24 20–24 15–19 15–19 10–14 10–14 '05–09 '05–09 0–04 0–04 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12 14 16 –16 –14 –12 –10 –8 –6 –4 –2 0 2 4 6 8 10 12 14 16 Effectifs de chaque groupe d'âge en millions Effectifs de chaque groupe d'âge en millions Hommes Femmes Hommes Femmes Source: Guengant J-P. et al. 2014 : Population, développement et dividende démographique en République Démocratique du Congo.  15–19. In rural areas, nearly one-third of women Main Causes of Death among Adults under 19 were mothers in 2013–2014. The main causes of illness and death have changed Use of modern contraceptives remains scant. little, with communicable diseases accounting This is despite a slight upturn from 6  percent to for a large proportion. (Malaria comes first fol- 8  percent among women aged 15–49 over 2007– lowed by respiratory infections, diarrhea, malnutri- 2013/2014 (as opposed to 15  percent in 1995), tion and HIV/AIDS). Adult mortality among those which places DRC behind peer countries (Figure 12). aged 15–49 is 5.4 per 1,000 (for women) and 5.2 Health System Performance   19 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 11  Fertility Rates in DRC and Other Countries in the Region Niger 2012 7.6 DRC 2013–2014 6.6 Burundi 2010 6.4 Uganda 2011 6.2 Nigeria 2013 5.5 Zambia 2013–2014 5.3 Equatorial Guinea 2011 5.1 Congo 2011–2012 5.1 Cameroon 2011 5.1 Sierra Leone 2013 4.9 Benin 2011–2012 4.9 Togo 2013 4.8 Liberia 2013 4.7 Rwanda 2010 4.6 Gabon 2012 4.1 Source: DHSs, date depending on national data availability. Figure 12  Use of Modern Contraceptive Methods in DRC and Other Countries in the Region Zambia 2013–2014 45 Rwanda 2010 45 Uganda 2011 26 Congo 2011–2012 20 Gabon 2012 19 Liberia 2013 19 Burundi 2010 18 Togo 2013 17 Sierra Leone 2013 16 Cameroon 2011 14 Niger 2012 12 Equatorial Guinea 2011 10 Nigeria 2013 10 DRC 2013–2014 8 Benin 2011–2012 8 Source: DHSs, date depending on national data availability. (for men). The trend rises with age for women from Maniema (34 percent). Malaria continues to account 4 per 1,000 at ages 15–19 to 7 at ages 45–49; for for over 40 percent of the causes of infant mortal- men the equivalent figures are just under 4 and 10. ity and is the number one reason for consultations nationwide. DRC has one of the highest incidences of malaria in the region. Malaria remains rampant The HIV/AIDS epidemic appears to have been with 23  percent of under-fives testing positive by brought under control with a prevalence of thick smear. Prevalence rates are higher in Province 1.2 percent among adults in 2013, down from Orientale (38 percent), Katanaga (32 percent), and 5 percent in 1990 (and 1.3 percent in 2007). Still, 20  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW HIV/AIDS is the fifth most common case of number addition, Schistosomiasis, hookworm infection, of years of life lost (YLLs), accounting for 5.4 per- Ascariasis, Trichuriasis, and Lymphatic Filariasis are cent of all YLLs. Seroprevalence among pregnant also prevalent, and likely to be some of the under- women fell from 4.3 percent in 2008 to 3.4 percent lying factors contributing to the burden of disease in 2011. However, the epidemic remains more wide- linked to malnutrition in DRC. The HAT program spread among women, as 1.6 percent of women aged is in transition, with changes such as a new rapid 15–49 are HIV positive against 0.6 percent of men. test and new oral treatment to be launched by end- HIV prevalence is very high among widows (7.9 per- 2015. There are regular outbreaks of cholera and the cent) and divorced/separated women (2.9  percent). Ebola virus in DRC. In 2014, the country recorded A large proportion of the total population (68 per- 8,700 cases of cholera with 249 deaths. The prov- cent) has never taken an HIV test, suggesting that ince of Katanga was the hardest hit. There were an national prevalence is underestimated. estimated 66 reported cases of Ebola causing 49 deaths in September 2014 (with the outbreak start- Tuberculosis prevalence started a downward ing in Jeera county). trend in 1990. The rate dropped from 160 per 100,000 in 2003 to 94 in 2012. However, annual incidence remains high at 384 cases per 100,000 Mother and Child Health due to the increase in coinfections. DRC still ranks After a slow start in the early 2000s, the reduc- among 22 countries the hardest hit by tuberculosis tion in child mortality started gathering pace in worldwide, and number five in Africa. 2010. Infant mortality (below 1 year old), estimated at over 90  per 1,000 over 1997–2007, fell steeply Neglected tropical diseases (NTDs) contrib- by 2013–2014 (to 58). Under-five mortality has also ute significantly to the burden of disease. Two seen distinct progress in recent years, falling from conditions in DRC—leprosy and Human African 172 per 1,000 in 1997 to 148 in 2007 and then 104 Trypanosomiasis (HAT)—are estimated to have the in 2013–2014 (Figure 13). highest prevalence of any NTDs globally. Seventy- five  percent of the global HAT cases are in DRC, Malnutrition continued to affect 43 percent of and 50 percent of all cases in DRC come from one children in 2013–2014, with nearly half of them district (Mai-Ndombe in Bandundu province). In suffering from severe malnutrition. The prevalence Figure 13  Change in Child Mortality from 1997 to 2013–2014 172 165 148 104 98 96 92 84 74 62 58 49 42 38 37 28 0–5 Years 1–4 Years 0–1 Year 0–1 Month ENSEF 1997 MICS 2001 DHS 2007 DHS 2013–2014 Source: ENSEF (1997), MICS (2001 and 2010), DHS (2007 and 2013/2014). Health System Performance   21 2014 HEALTH PUBLIC EXPENDITURE REVIEW of malnutrition has not changed much since 2010 year of estimation). However, maternal mortal- and there have even been significant upturns in the ity estimates vary by source. The 2007 DHS esti- eastern regions (> 40 percent). The rate of stunting mates maternal mortality at 543 deaths per 100,000 rose to 43 percent of children in 2014 from 38 per- births while the 2013–2014 DHS sets mortality at cent in 2001. No significant progress was made with roughly 800 deaths. The maternal mortality rate, wasting (8 percent) or underweight children (23 per- which stood at more than 1,000 maternal deaths cent) from 2007 to 2013–2014 (Figure 14). per 100,000 births prior to 2000, fell to some 900 around 2005 and then to 800 or so around 2010 Following a sharp upturn due to the conflicts (Figure 15). Continuation of this trend would place in the late 1990s, maternal mortality may have the maternal mortality rate at around 700 in 2014. fallen steadily to 846 deaths per 100,000 live Maternal deaths account for 35 percent of all deaths births over 2007–2014 (with 2010 the point of women aged 15–49. Figure 14  C  hange in Child Malnutrition in DRC and Underweight Children in DRC and Peer Countries 46 43 43 38 31 25 24 23 13 10 9 8 Stunted Emaciation Underweight MICS 2001 EDS 2007 MICS 2010 DHS 2013–2014 25% 2001 31% 31% 23% 2007 27% 25% 22% 2010 23% 24% 21% 2013 22% 23% 0% 5% 10% 15% 20% 25% 30% 35% Sub-Saharan Africa Low Income Countries DRC Source: ENSEF (1997), MICS (2001 and 2010), and DHS (2007 and 2013/2014) for DRC and WDI for other countries. 22  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 15  Change in Maternal Mortality, DRC, 1990–2014 Rapport de Mortalité Maternelle République Démocratique du Congo, 1990–2013 1400 1289 (Décès pour 100,000 naissances) Rapport de Mortalité Maternelle 1200 1100 1100 1000 930 1003 1000 846 800 710 699 730 600 MICS 2001 DHS 2007 549 400 DHS 2013 Estimations UN 549 Poly. (Tendance) 200 0 1985 1990 1995 2000 2005 2010 2015 Année Source: ENSEF (1997), MICS (2001 and 2010), DHS (2007 and 2013/2014). Section 2. Service Coverage of insecticide-treated nets represents an important advance for maternal health. The majority of preg- Service coverage rose for a number of inter- nant women sleep under such nets (56  percent in ventions essential to maternal health. Over 2014), but only 6 percent did so in 2007. 2001–2014, antenatal care coverage rose from 68  percent to 88  percent and skilled birth atten- Although the generic index of access to essen- dance from 61  percent to 80  percent (Figure 16). tial treatments for children (ARI, diarrhea, and This gave to DRC antenatal care coverage rates fever) has risen, an analysis by treatment reveals higher than those in peer countries for these two backsliding coverage. In particular, the share of interventions (Figure 17). A sharp upturn in the use children with respiratory infections who receive Figure 16  C  overage for Antenatal Care, Skilled Birth Attendance and Basic Treatment for Children in DRC and Peer Countries, 2001–2014  100% 90% 88% 80% 80% 70% 68% 61% 60% 50% 40% 40% 33% 30% 20% 10% 0% ANC Assisted Birth Delivery Access to Basic Treatments MICS 2001 DHS 2007 MICS 2010 DHS 2013–2014 Source: MICS 2001, DHS 2007, MICS 2010, DHS 2013/2014. Health System Performance   23 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 17  Antenatal Care Coverage in DRC and Peer Countries, 2007–2013 79% Sub-Saharan Africa 66% 75% Low Income Countries 56% 88% DRC 68% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% 2013 2010 2007 2001 Source: MICS 2001, DHS 2007, MICS 2010, DHS 2013/2014 and WDI for other countries. treatment is reported to have fallen from 42 percent Out-of-pocket payment by users acts as a bar- to 35  percent since 2007. The number of children rier to access to health care. The health care ser- treated with anti-malarial drugs is also estimated vices depend on cost recovery in the form of having to have fallen, from 45 percent to 40  percent over users pay at the point of use. Price is the number 2007–2014, when treatment for fever posted a one reported reason for nonuse of health care and downturn from 31 percent to 20 percent. Conversely, forms a barrier for 35 percent of people interviewed access to treatment for diarrhea is reported to have (Figure 18), and 40 percent among the poorest (NSI, improved from 32 percent to 39 percent of children 2013). Consequently, the national rate of service use with diarrhea. is very low, estimated at around 0.4 consultations per person a year. Immunization coverage has slipped over the last few years. Although coverage doubled over 2001– 2010 (from 23 percent to 45 percent fully vaccinated Section 3. Service Quality children), the proportion of fully vaccinated children is estimated to have fallen from 53 percent to 46 per- Service quality is largely affected by the poor cent from 2010 to 2013–2014. Immunization cover- availability of amenities and equipment required age remains behind peer and neighboring countries for good quality consultations. Only 64  per- (with Rwanda at 90  percent, Burundi 83  percent, cent of health facilities on average have a private and Cameroon 53 percent). In particular, little prog- room for consultations to respect patients’ privacy. ress has been made with BCG coverage over the last Availability of other amenities and conveniences is four years (down slightly from 84 percent to 83 per- also low at less than 50 percent for sanitary facilities cent) and Pentavalent-3 coverage (unchanged from (41  percent), improved water sources (37  percent), the 2010 coverage of 61 percent) despite encourag- communication equipment (28 percent), a vehicle in ing efforts from 2001 to 2007. working order with fuel (12 percent), power source 24  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 18  M  ain Reasons for Nonuse of Health Care, DRC, Figure 19  O  perational Level of Hospitals by Province: 2013 Running Water and Facilities for Consultations (% of hospitals) Others 12% Non-Necessity 100 Family Refusal 17% 1% 90 Time 80 1% 70 60 50 40 30 20 10 0 Automedication sa go du r vu vu a al l ga ta 31% u m nt te ha Ki Ki en on un n ie rie ta ua d- ns an id C nd d Ka iO or Eq Su s- cc Ki M Ba N Ba iO sa Ka sa Ka Price 35% % District hospitals with water % District hospitals with consultations equipment Distance Source: Authors’ calculations, based on MoPH data, 2011. 3% Source: Authors’ calculations based on 1-2-3 Survey data, NSI, 2013. of the health facilities have a urine pregnancy test. Availability of the other basic tests (blood sugar (9 percent), and computers and Internet connection type) is less than 20  percent. Only 21  percent of (2 percent) (MoPH, forthcoming a). Only less than health facilities have adequate means to diagnose 17 percent of the health facilities tested in Bandundu HIV/AIDS, mainly because HIV/AIDS prevention and 19  percent in Kasaï Occidental have the level and treatment services have not been incorporated of amenities and equipment required for curative into the health system. Services to prevent mother-to- consultations. Five provinces post an average avail- child transmission also remain thin on the ground, ability of 25 percent or less (MoPH, forthcoming a). at 15 percent nationwide. The quality of hospital infrastructure and equip- Although there is relatively good availability of ment is, in particular, very poor. In Bandundu antenatal care, the quality emergency obstetric and Equateur, just 5–7 percent of district hospitals care that is so important to a mother’s survival is have running water and less than 10  percent have lacking. Antenatal care is available in nearly 78 per- full consultation amenities (Figure 19). The health cent of health facilities, yet less than 5 percent of health facilities in the province of Kinshasa are somewhat facilities can offer quality emergency obstetric services. better equipped but are still not fully operational In Bandundu, Province Orientale, and the two Kasaï (just 62 percent of the district hospitals in that prov- provinces, over 99 percent of health facilities can pro- ince have running water). Only the province of Bas vide no quality emergency obstetric services for want Congo has over 40  percent of its district hospitals of equipment or trained personnel (Figure 20). fully equipped for consultations. In 2011 the central government launched the Diagnosis capabilities are also poor (person- PESS, which came into effect in 2013. The pro- nel, equipment, and analysis). Although there is gram has two major phases: to improve 1,000 health a relatively good nationwide availability of means centers and 200 district hospitals, and to target sec- to diagnose malaria (84  percent), just 40  percent ondary and tertiary levels (Box 3). Health System Performance   25 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 20  A  vailability of Full Obstetric Services for Mothers, Average Score per Province, 2012–2013 30% 25% 20% 15% 10% 5% 0% sa go du r e vu a vu ga l al ta u al m nt te ha Ki Ki en on un n nt ie rie ta ua d- ns rie an id C nd d Ka O or Eq Su s- cc Ki M O Ba N Ba i iO sa e nc Ka sa i ov Ka Pr Césarienne Transfusion sanguine Soins obstetriques d’urgence Source: Authors’ calculations, based on MoPH, forthcoming a. Box 3  A Unique Initiative Funded by Domestic Resources: The PESS In December 2011, the prime minister officially launched the At June 30, 2014, the Public Treasury had disbursed USD PESS, funded entirely by the government. This off-budget pro- 65 million, allocated to: (i) the purchase by UNICEF of equip- gram is designed to roll out in two major phases.The first, PESS- ment for all the targeted establishments (the equipment for the Primary Network (PESS-RP) phase, for a total of CDF 80 billion first two phases was distributed in September 2014 or is being (USD 85 million) is to renovate, equip, provide training for, and distributed); (ii) the design of standard plans for the health cen- supply essential drugs to 1,000 health centers and 200 district ters to be built, the renovation of two district hospitals and hospitals in the 11 provinces. The second, PESS-Secondary and eight health centers in Kinshasa, the launch of building work for Tertiary Network (PESS-RST) phase, being developed, might 34 health centers in keeping with the standard plans in the 11 well be able to mobilize a large amount (USD 90 million) to fit provinces (work is scheduled for completion in mid-2015), and out provincial and tertiary hospitals. the preparation of operations for the next phase (the funds ini- The PESS-RP was initially planned as a four-phase program tially earmarked for the first phase were redirected to building for a sum of CDF 20 billion (USD 21 million), with each phase fewer establishments following the standard plans); and (iii) the designed to renovate, equip, provide training for, and supply opening of financing facilities in the regional medical stores for essential drugs to 66 district hospitals and 330 health centers in 132 district hospitals and 660 health centers, as well as the the 11 provinces. UNICEF was placed in charge of purchasing purchase of contraceptives (through UNFPA). The financing the health care equipment (for an earmarked USD 50.8 mil- facility set up for capacity building had been entirely used up at lion), with essential drugs to be supplied by the FEDECAME this date, used, for example, on provincial training workshops, by means of financing facilities opened in the regional medical provision of management tools, a BCeCo study, and program stores for the establishments concerned (USD 6 million) and monitoring and monitoring assignments in the provinces. renovation/building work assigned to BCeCO (the National Source: Authors, based on information provided by the MoPH. Coordination Body for External Aid). In addition, USD 4.9 mil- lion was earmarked for capacity building. 26  Investing in Universal Health Coverage Health Financing 4 Section 1. Health Financing Sources Table 12  Main Health Financing Indicators, 2008–2013   2008 2009 2010 2011 2012 2013 Nominal GDP (from SNA 93) 10,741,806 14,788,935 18,592,321 21,912,978 25,249,857 27,596,000 (CDF millions) Nominal exchange rate (CDF/USD) 563.19 816.78 906.96 918.02 918.74 919.48 Population (NSI) (thousands) 68,076 70,991 72,784 75,259 77,817 80,462 Total health expenditure (USD thousands) 869,831 893,363 891,955 904,187 1,031,802 1,249,836 Total health expenditure (% of GDP) 4.6 4.9 4.4 3.8 3.8 4.5 Total health expenditure per capita (USD) 12.8 12.6 12.3 12.0 13.3 15.53 Government health expenditure (USD thousands)a 129,263 127,971 94,712 108,064 160,559 224,103 Government health expenditure (% of total health 14.9 14.3 10.6 12.0 15.6 17.9 expenditure) Government health expenditure (% of GDP) 0.7 0.7 0.5 0.5 0.6 0.8 Household health expenditure (USD thousands) 366,738 374,073 329,214 339,297 405,399 475,483 Household health expenditure (% of total health 42.2 41.9 36.9 37.5 39.3 38.0 expenditure) Household health expenditure (% of GDP) 1.9 2.1 1.6 1.4 1.5 1.7 Household per capita health expenditure (USD) 5.4 5.3 4.5 4.5 5.2 5.9 Out-of-pocket payments for healthcare (USD 366,408 272,974 308,432 327,446 343,053 459,418 thousands) Out-of-pocket payments (% of household spending) 99.9 72.9 93.7 96.5 84.6 96.6 Out-of-pocket payments for health care (% of total 42.1 30.6 34.6 36.2 33.2 36.7 health expenditure) Out-of-pocket payments for healthcare per capita 5.4 3.8 4.2 4.4 4.4 5.7 (USD) External health expenditure (USD thousands) 291,478 297,307 423,435 423,671 423,868 462,665 External health expenditure (% of total health 33.5 33.3 47.5 46.9 41.1 37.0 expenditure) External health expenditure (% of GDP) 1.5 1.6 2.1 1.8 1.5 1.6 External health expenditure per capita (USD) 4.3 4.2 5.8 5.6 5.4 5.7 Source: Authors’ calculations based on the 2013 NHA and the Ministry of the Economy. a. All expenditure by government entities active in the health field (MoPH and other ministries, Social Security), at every level using internal and external resources. Note: The findings based on the health financing indicators, as shown in this table, differ from those based on the 2013 NHA Program. The main reasons for this difference are that the updated GDP figure from the new SNA 93 classification is used and the population is greater per the latest (2013) estimates from the NSI. 27 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 21   Total Health Expenditure per Capita in DRC and the Rest of Sub-Saharan Africa, 2012 (USD) 1200 1000 DRC 800 600 400 200 13 0 ea ag a r i er n am a e Ta so ia Se a am al n da ire go e . ab dan Sw rde nd on ia a aw ep ca i da ne qu nd oo on on ric op an g ib g Fa itr an vo ila ab ne as ,R i Ve al am Af N Su ui bi ga er e Su hi nz Er az I w L G M G d’ a Et U h R o N in ra h ut e ad ut oz rk C er ot So So C M Bu Si M C C Source: Authors, based on the NHA. DRC’s health expenditure per capita is less than 2013, peaking at more than 47 percent in 2010. This one-tenth that in the rest of Africa. DRC spent share is equivalent to around 1.7 percent of GDP on USD 13 per capita on health in 2012 (Table 12), average over the period. Even though estimates vary compared with more than USD 140 on average in by source, all sources indicate a big jump in the last the rest of the region and USD 31 in low-income few years. The NHA show an increase from USD countries (Figure 21). Total health expenditure per 291 million in 2008 to USD 462 million in 2013; capita showed little growth over most of the period, the International Assistance Management Platform ranging between USD 12–13.30, but climbing in has external assistance doubling over 2007–2013 2013 to USD 15.53. from USD 255 million to USD 530 million; and the OECD Development Assistance Committee shows Health expenditure as a share of GDP has that health/population assistance climbed from USD shown a falling trend since 2008. Total nominal 169 million in 2006 to USD 363 million in 2009. health expenditure was stagnant over 2008–2011 at less than USD 900 million, rising to USD 1.2 bil- Despite this growth, per capita external assis- lion in 2013. Health expenditure as a share of GDP tance for health was still lower than in the rest declined after a big jump in GDP from 2008. It fell of the region. It stood at USD 5.7 in 2013 accord- from 4.6  percent of GDP in 2008 to 3.8  percent ing to the NHA (half the amount of other countries in 2011–2012, versus 6.5 percent in the rest of the in sub-Saharan Africa, which average USD 11) and region and 5.3  percent in low-income countries at USD 7 according to the International Assistance (Figure 22). Health expenditure rose to 4.5 percent Management Platform. Assistance for all sectors is of GDP in 2013. still low in DRC, at USD 23 per capita. External assistance is the leading source of health Even though the per capita amounts for health financing in DRC. External assistance accounted for are lower than for the rest of the region, health 40 percent of total health financing over 2008–2013, is one of the partners’ priority areas. In 2012, rising from 33  percent in 2008 to 37  percent in 26 percent of total assistance to DRC was for health, 28  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 22  T  otal Health Expenditure as Percentage of GDP in DRC and the Rest of Sub-Saharan Africa, 2008–2012 6.8 6.4 6.4 6.5 6.3 5.5 5.5 5.4 5.2 5.3 4.9 4.6 4.4 3.8 3.8 2008 2009 2010 2011 2012 Sub-Saharan Africa (developing countries) Low-income countries DRC Source: Authors, based on the NHA. up from 15  percent in 2007. External assistance Current expenditure, which accounted for nearly for health is tightly concentrated by source, with 95  percent of total expenditure over 2010–2012, five main donors accounting for more than 70 per- was covered in 2012 almost equally by external cent of it over 2007–2012. By contributions, these assistance (39 percent) and households (41 percent). five donors are the Global Fund (21.5  percent), External assistance in 2013 financed primarily drugs USAID (15.6  percent), World Bank (14.4  percent), and specific inputs (34 percent of total expenditure the Government of Belgium (10.3  percent), and reported to the MoPH, or CDF 96 million). Other GAVI (9.6 percent). Much support from partners is large expenditure items were services (15 percent of extrabudgetary, and the central government budget total expenditure) and infrastructure and equipment amounts to only a very small share of the funds from (12 percent). donors (roughly 10 percent). The execution rate var- Household funds are the second largest source of ies: 83 percent of 2012’s commitments and 42 per- health financing, just behind external assistance. cent of 2013’s (MoPH—Research and Planning Households accounted for an average of 39.3  per- Division, 2014). cent of total health expenditure over 2008–2013. External assistance is still focused primarily on Their share of expenditure declined slightly from financing for treatment of communicable dis- 42  percent in 2008 to 38  percent in 2013, reflect- eases (HIV/AIDS, malaria, and tuberculosis), ing the increase in external health assistance in real totaling nearly USD 150 million in 2012. Support terms. Household health expenditure averaged USD for vaccination accounts for the second-largest share. 5 per capita, in the range of USD  4.50–5.90 over At nearly USD 50 million, support for the health sys- the period. The private sector’s share of total health tem that is not related to specific diseases accounts expenditure stood at 4 percent in 2012, consisting of for the fourth largest share of external assistance direct payments of health costs by some employers expenditure (malaria is the third) (Figure 23). in the formal economy. External assistance finances the bulk of health Out-of-pocket payments account for more than investment (construction and facilities). In 90 percent of household health expenditure. In a 2012, it financed 89 percent of capital expenditure. fee-for-service system, users cover the total costs of Health Financing   29 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 23  E  xternal Financing and Household Expenditure by Disease, DRC, 2012, (USD thousands) 120,000 100,000 80,000 60,000 40,000 20,000 0 S s fe ia ns r i en l di ea s er asit s .. s g al pbs di on s a s .. si tio le d se m /pa ase se se a ition l d al. in m c. ID r io la h iti lo di ann ifi au a he am sea ea na ic / /A ot pre pic iarr ct ab utr to ma es cu ec se e d r IV is s t ed con as pl n er d sp i in H r b se r R ily le m a tu ry / es b to ta us s H ic ira tro a f at un no ise sp v at cc ted ec m rd re el nf m va lec tr ot he co e in g he ot ne n no External Households Source: Authors’ calculations, based on the 2012 NHA. Figure 24  Household Health Expenditure, DRC, 2012 government’s share of health financing is lim- Other, ited.20 The central government, via the MoPH Laboratory Tests, 5% and other ministries indirectly involved in health, 6% accounted for 15  percent of financing in 2008, a Curative share that decreased to 11 percent in 2010, but then Ambulatory Care, picked up to 17.9  percent in 2013. Government 13% funds cover most payroll costs (Chapter 5). Hospitalization, 44% There is no pooling system at the national level. A system of voluntary insurance plans has devel- oped sporadically, but coverage (1–2  percent of the population) and financial protection are weak. Drugs, The two most developed insurance networks are in 32% Bwamanda (since 1986) and Sud Kivu (since 1997). Source: Authors’ calculations, based on the 2011 NHA. The first has more than 130,000 voluntary members and the second 20,000–30,000 in 16 insurance plans. office visits, house calls, and drugs, except in a few Both networks offer identical coverage, including districts that have recently introduced a subsidized curative care in health centers and general hospi- flat-rate payment. Drugs (32  percent) and primary tals. They impose copayments of 20–50  percent of care (13  percent) account for nearly half the out- cost. The yearly individual premium is USD 1.50 in of-pocket payments by users. Forty-four percent is Bwamanda and USD 3.00–USD 6.00 in Sud Kivu. mainly to cover hospitalization costs (secondary and Coverage is delivered through a network of faith- tertiary) (Figure 24). Household expenditure is used based establishments directly linked to the insurance to cover the costs of maternal health (childbirth and postnatal care) and child health (mainly respiratory 20 This figure is based on the NHA definition and includes government infections and malaria). expenditure using internal and external resources. The breakdown provided in Chapter 4 is based on government expenditure using At less than 15 percent of total financing sources internal resources only, which explains the difference between the over 2008–2013 (average of 14.2 percent), the two estimates of “government health expenditure.” 30  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW organization. Some evidence shows an increase in Figure 25  S  tructure of Domestic Revenue (% of total the use of services by the insured, but limited finan- revenue), 2011 cial protection. The reduction in out-of-pocket pay- Oil Revenue ments for care is about 10  percent over five years among the enrolled population (Soglohoun 2012). VAT Revenue A compulsory insurance plan for teachers was recently introduced with mandatory membership and premiums. Mining Revenue Section 2. Adequacy of Government Financing for Health and Fiscal Space Achieving universal coverage of essential health services requires substantial funding. Closing the financial gap—estimates range from USD 18 Other Revenue to USD 33 per capita—would require a tripling of current health financing (MoPH 2011, Integrated Source: Adapted from World Bank, 2015. Healthcare Technology Package method).21 With government expenditure of about USD 1 per capita (see Table 16), financing needs will not be met unless 2009–2013 as copper production increased sharply fiscal space is expanded, especially as external assis- with high international prices, have not lifted related tance from donors may decline in the medium to revenue. Poor revenue collection, stemming from long term (IMF, 2014). organizational inefficiencies in the tax administra- tion and tax breaks for the mining industry, mean There are multiple possibilities for increasing fiscal that the effective tax rate for the mining industry is space. The medium-term macroeconomic projections around 13.9  percent,23 while tax revenue from the for DRC are reassuring. Real GDP growth is expected petroleum industry is estimated to be worth up to to reach an average of 8.9 percent growth from 2015 53.6 percent of the value of oil exports. Ultimately, to 2017. Inflation, which was less than 1  percent mining contributes to a quarter of total revenue in in 2013, is expected to rise again to an average of 2011 (Figure 25). 4.3 percent in the medium term. The fiscal deficit was small in 2013, but could increase in the medium term Tax revenue from other sectors also seems to to 3 percent of GDP, reflecting the expected stagnation fall short of its potential. The revenue share from in assistance from donors (IMF, 2014). other sectors (i.e. excluding extractive industries and donors) was 9.4  percent of GDP in 2012, against Mining industry revenue falls short of its poten- 22.9 percent in the region.24 The share of tax revenue tial.22 The growth of the mining industry over 2010– 2013, and the doubling of mining exports over 23 The mining tax system is under review. The government is exploring options for increasing royalties and for introducing a resource tax rent. Exemptions to the tax regime in the natural resource sector as a 21 A joint study by the World Bank and the Government in 2015 whole were also under review in 2015. should make accurate estimates of the financial requirements for 24 Note: Boundaries between tax, non-tax and mining revenue are universal coverage of maternal and child health care by 2030 as part blurry in the context of DRC. Part of the non-tax revenues are in of the Global Financing Facility. fact revenues from the mining and oil industry, while part of direct 22 DRC has identified deposits of some 50 minerals, including copper, and indirect taxes also originate from the mining sector. Double cobalt, uranium, diamonds, gold, and coltan. It has the second- counting and mis-classification frequently occur. For this reason, tax largest reserves of copper in the world and the largest reserves and non-tax mining revenue was not distinguished in the present of cobalt—an essential component of electronic circuits. It also analysis. Further classification and analytical work is ongoing by the produces oil. World Bank and the Government. Health Financing   31 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 13  Expanding Fiscal Space, 2014–2019 2014–2015, 2016–2017, 2018–2019, Percentage of GDP 2013 short-term medium-term long-term Domestic revenue 13 14.9 14.9 21 Expenditure using 12.5 14.6 15.4 20.5 domestic revenue Source: Adapted from IMF, 2014 and World Bank, 2015. in total revenue has improved in recent years to There is room for expansion in the medium 8.7 percent of GDP in 2012. The VAT amounted to term through improved macroeconomic per- 4.5 percent and is a key indirect tax and source of formance. Government funds for the health sector country revenue. It has though posted a poor perfor- from domestic sources could be increased by at least mance since 2013, with a drop in revenue equivalent 1 percentage point of GDP. By linking health expendi- to 1 percentage point of GDP, following an increase in ture to national income growth, the resources raised tax credits granted to the private sector and changes can be used to meet some of the health sector needs in accounting methods relating to VAT refunds. in the medium term. Government health expendi- ture started tracking growth with greater elasticity The central government should be able to in 2011–2012 (e = 2.14 over 2011–2013, up from increase its revenue by 8 percentage points of e = 0.55 over 2007–2010). If this trend (e > 2) were GDP in the long term. The IMF estimates that sustained until 2016, government health financing all categories of revenue fall short of their poten- could reach CDF 300 billion, up from CDF 186 bil- tial, sales tax aside. This means that revenue could lion in 2013 at current prices (Figure 26). Similarly, increase from 13 percent of GDP in 2013 to 21 per- overall domestic financing including “the investment cent in 2019 and government expenditure could grants to provinces and local entities” could increase increase from 12 percent in 2013 to 20.5 percent in from CDF 463 billion to CDF 830 billion. 2019 (Table 13). The IMF argues that full tax poten- tial could be achieved by collecting more revenue Greater revenue collection would generate up from the natural resources sector and by strength- to 0.4 percentage points of GDP more for health, ening administration and collection of existing (and all else equal (Scenario 1—Table 14). This sce- future) domestic revenue. nario, which is backed up by the IMF Article IV pro- jections and World Bank PEMFAR, seems plausible. There are different options for expanding a In the medium term, central government revenue country’s fiscal space for health. The literature collection could increase by 8 percentage points of and empirical evidence reveal four main types of GDP from 13 percent to 21 percent. Achieving this mechanisms that can be used to increase resources increase depends on improving collection capacities for health: (i) economic growth, if it leads to more for existing taxes and a reform of the collection sys- central government revenue, can automatically tem. An improved business climate and better con- increase the size of the health budget; (ii) prioritiza- trol of natural resources are also needed to lead to tion of the overall budget can give a larger share of effective increase in fiscal space. the central government budget to the health sector; (iii) raising new resources earmarked for the health Health could be given greater priority in the sector (assistance from donors, excises) can increase central government budget (up to 8 percent the resources available; and (iv) efficiency gains of government expenditure) and increase by made through improved expenditure execution and 0.6 percentage points of GDP, all else equal utilization can free up resources for the health sector (Scenario 2). However, it remains to be seen (Tandon and Cashin 2010; Mathonnat 2010). whether this scenario is politically and technically 32  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 26   GDP Elasticity of Government Health Expenditure, DRC, 2007–2016 National Income vs. Government Health Spending Democratic Republic of the Congo (2007–2016) Scenario 1 300 (e = 2.14) Government Health Spending (executed) 2016 Scenario 2 (e = 1.33) 250 (in billions, local currency) 2015 2014 200 2013 Scenario 3 Elasticity = 2.14 (e = 0.55) (2011–2013) 150 2012 2011 100 Elasticity = 0.55 2010 (2007–2010) 2007 2008 2009 50 10000 20000 30000 40000 GDP in Billions, Local Currency Source: Authors, based on government payment data and World Bank projections (2014). Table 14  Summary of Potential Long-Term Increases in Fiscal Space for Health Gov’t. health Total gov’t. exp. exp. Gov’t. health exp. Conditions (% of GDP) (% of total gov’t. exp.) Baseline Scenario 0.7 12.5  5.3 (2013) Scenario 1 1.1 20.5  5.3 Increase in central government revenue from 13% to 21% of GDP and subsequent increase in gov’t expenditure Scenario 2 1.0 12.5  8.0 Increase in health expenditure from 5.3% to 8.0% of central government budget Scenario 3 1.3 12.5 10.3 Improved spending effectiveness through better execution of health budget (to 80% of allocations) Combined 1.7 17.0 10.0 Increase in central government expenditure to 17% of Scenario GDP; increase in health expenditure as a percentage of central government budget (2 points) and in the execution of budget allocations (2 points) Source: Authors. feasible. The current cuts in government expenditure low and leaves little room for reallocation between stemming from lower revenue do not augur well for sectors. The 2015 budget outlook seems to support increased budgets for line ministries (Chapter 5). The this hypothesis, with a 5 percent cut expected in the budget surplus of 0.5 percent (excluding payments health budget from 2014. This cut would mainly in arrears) results from fiscal consolidation relying affect the matching funds from the government, the on compression in priority expenditures, including PESS, and the operating budgets of provinces and capital and social spending. Similarly, the share of health facilities—the core functioning of the health expenditure using domestic sources (12.5 percent) is system beyond wages. Health Financing   33 2014 HEALTH PUBLIC EXPENDITURE REVIEW Improved execution of health expenditure different local taxes that produce little and are inef- would be an effective way to free up additional ficient. Local mining revenue, which is categorized resources (Scenario 3). By increasing execution as “extraordinary revenue,” accounts for a small up to 80  percent, government expenditure could share of total local revenue at only 8 percent of the reach 1.3  percent of GDP, all else equal. However, total reported. The effective tax rate from the min- execution of government expenditure using domes- ing sector is less than 2 percent: for a combined out- tic sources (excluding current expenditure) is closely put of coltan, copper, and zinc of USD 2.17 billion linked to the ineffectiveness of transfers to the prov- in Katanga, revenue collected was only USD 52.3 inces. The improved execution of these transfers lies million (2012) (Provincial Government data, 2013). outside the MoPH’s remit (Chapter 5). External assistance will be unable to sustain The potential gains from more efficient expenditure a lasting increase in fiscal space for health. It cannot be quantified exactly, but could also be quite accounted for 4.4 percent of GDP in 2013 (exclud- large (Chapter 6). ing 1.4  percent as loans). In the medium term, the share is forecast to stabilize at 4.5 percent in 2017, Decentralization could also be an opportunity but to decline to 1.3 percent by 2030 (IMF 2014). for health financing. In addition to central govern- There are no reliable data on external assistance ment budget transfers to the provinces, three types for health in DRC, which makes it impossible to of domestic financing sources can now be used: rev- assess the sustainability of this financing accurately, enue retrocessions, investment subsidies, and local or its future potential. However, estimates call for resources. The introduction of revenue retrocessions external assistance—the main financing source for has been very valuable politically and is expected to health—to stabilize at about USD 5 per capita in the lead to an overall increase in transfers for provinces. medium term (2015), then decline over 2018–2020. Investment grants earmarked for such priorities as Off-budget assistance is expected to decrease as health may be a new source of health financing at funds for humanitarian assistance shrink. In-budget the local level, as long as the transfers are executed assistance could, however, rise, particularly with the (Chapter 7). On the strength of their broader man- creation of the Global Financing Facility (GFF) for date for implementing health policy, provinces could maternal and child health. HIPC funds were a major also raise more local revenue, especially from natu- source of government revenue for health, accounting ral resources industries. for 83.6 percent of the MoPH’s internal resources in For provincial fiscal space to develop in the long 2007. The residual funds are now used to finance the term, the heavy reliance transfers from central health sector, accounting for 8.9 percent of financ- government must give way to revenue raised ing in 2013. These funds are expected to dry up in locally. An analysis of the types of revenue in prov- the medium term as the country pays off its debt. inces with strong economic capital, such as Katanga, Finally, a “harmonized contract” at provincial level reveals the low levels of revenue raised locally, given would increase fiscal space for health by enhancing their potential for raising tax revenue. Almost half the efficiency of external assistance, through better (48.8 percent) the revenue in Katanga’s 2013 budget coordination and harmonization. (Such a contract came from revenue retrocessions. Local tax revenue was adopted in December 2014 to replace some of was tiny and totaled only 0.11  percent of the rev- the multiple contracts in use, improve coordination enue that the province collected in 2013. The local of external assistance, and allow for more predict- tax system is highly fragmented, with more than 50 able financing.) 34  Investing in Universal Health Coverage Government Financing for Health 5 Box 4  Priority Recommendations for the Health Sector from over 2007–2013, from CDF 39 billion to CDF 186 the First Public Expenditure Review, DRC, 2008 billion. However, given inflation and the devalua- tion of the currency over 2007–2013,25 the MoPH’s Government Budget and Disbursement Process: budget allocation barely doubled at constant 2007 1) Improve the credibility and usefulness of the budget by prices, rising from CDF 39 billion to CDF 74 billion. introducing a transparent prioritization process within a real- istic total envelop and involving the provinces in the budget The central government devoted 4.2 percent of process. 2) Establish a system to collect and compile data on its budget to health, equivalent to 0.7 percent external assistance for health. 3) Improve budget execution of GDP in 2013 (Figure 27). Health’s share of the of nonsalary allocations. 4) Curtail the use of exceptional central government budget is tiny and much smaller payment procedures for MoPH expenditures. than the average for the region. The MoPH averaged Human resources: Establish an accounting of civil 3.7  percent of the central government budget over servants and nonregularized personnel in each province. 2007–2013, on a downward trend since 2003.26 The Decentralization: Lead a consultative and policy education budget had the opposite trend, rising over development process for decentralization in the health sector. 10 years to 9 percent in 2013. Source: World Bank, 2008. The health budget closely tracks the total cen- tral government budget, which is pro-cyclical. Section 1. Health Budget 25 Over 2003–2006, the exchange rate was in a range of CDF 400–468 for USD 1. Since 2010, the rate has been higher than CDF 900 per In real terms, allocations for health, excluding USD 1 (CDF 919 in 2013). external funds, increased fourfold over 2007– 26 During the period covered by the first Public Expenditure Review, 2013 and increased 10-fold in nominal terms health accounted for an average of 4.6 percent of the central government budget from 2003 to 2006. However, this figure over 10 years. The MoPH’s budget from domes- included external financing and is not comparable to the 2007–2013 tic resources increased 4.8 times in nominal terms average, which excluded external funds and was based exclusively on internal resources. Table 15  Budget Allocations, MoPH Budget, 2007–2013 2007 2008 2009 2010 2011 2012 2013 MoPH budget (current CDF millions) 38,663 44,287 77,597 91,354 124,649 175,472 186,145 MoPH budget (% of total budget)* 2.8 3.8 4.5 3.1 3.4 4.3 4.2 MoPH budget (% of GDP) 0.5 0.4 0.5 0.5 0.6 0.7 0.7 Source: Authors’ calculations based on disbursement data and 2013 GDP under SNA 93, IMF 2014. * Excluding debt and financial expenses. 35 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 27  DRC Health Budget, 2003–2013 5,000,000 5.5% 6.0% 4,500,000 5.0% 4,000,000 3,500,000 4.0% 3,000,000 4.2% 2,500,000 3.0% 2,000,000 2.0% 1,500,000 1,000,000 1.0% 500,000 0 0.0% 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Total state budget (current CDF, millions) MoPH budget (current CDF, millions) MoPH budget (% total state budget) Source: Authors’ calculations based on MoPH data (2008–2013) and the 2008 Public Health Expenditure Review (2003–2007). The health budget was never a priority over 2003– Since 2011, decentralization as implemented 2013. A fall in revenue led to slower growth of the has led to more government financing for health central government budget and automatically to through earmarked transfers to the provinces slower growth of the health budget. The govern- (called “investment subsidies to provinces and ment’s policy largely consisted of adjusting expen- local entities”).27 These new transfers amounted diture to match the revenue available to curb the to 125  percent of the MoPH budget in 2012 and fiscal deficit. The main burden of adjustment fell on 150 percent in 2013. They came to CDF 43 million financing for social and infrastructure expenditures in 2011 and exceeded CDF 277 million in 2013. (IMF, 2014). They increased the health budget as a share of the The health budget is volatile from one year to 27 At first, in 2011, these transfers were recorded in a specific section the next and is strongly linked to the total bud- (Section 88) by province, but without specifying the field. Starting get. The budget posted large annual variations, with in 2012, these transfers were specified in Section 88 “investment steep nominal increases in 2009 (75 percent), 2011 subsidies to provinces and local entities” under the expenditure heading “Provinces and Local Entities/Health—Chapter 88100” (36 percent), and 2012 (41 percent), and more mod- and then broken down by province and nature of expenditure. erate increases in 2008, 2010, and 2013 (6–18 per- These transfers are for areas where the provinces have sole jurisdiction. These are projects that are defined by the provinces and/ cent). Hence the budget was unpredictable. The or coordinated with the provinces, and then executed by central recent introduction of health budget guidelines— government on behalf of the provinces. This will continue as long Health Medium-Term Expenditure Framework as the provinces’ public procurement systems are not operating in compliance with the public procurement code (see Memorandum (MTEF)—has not been effective at stabilizing the of Understanding on procedures for using investment transfers in health budget or making it more predictable (Box 5). sectors that are the sole jurisdiction of the provinces, March 2013). 36  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Box 5  Using the Health MTEF to Improve the Budget Process budget by 6 percentage points in 2013, from 4.2 per- cent (excluding earmarked transfers) to 10.4 percent From 2011 the Research and Planning Division at the MoPH (including earmarked transfers). The new budget has run a program to improve the budget process via a transfers seem to be leading to a refined model of results-oriented management concept that uses the MTEF resource allocation for the sector. as a tool. Since 2012, the national MoPH and provincial min- istries have compiled a national and provincial MTEF each year. This tool is featured in the roadmap for government Section 2. Government Health expenditure reform initiated by the Ministry of Economy and Expenditure—Low and Volatile Finance, making the health sector a trailblazer for a reform to be extended to all other sectors. Central government expenditure using domes- The benefits are twofold. First, results-based management tic resources for health averaged less than practices are picked up by provincial planning and budgeting 1 dollar (USD 0.84) per capita over 2007–2013 teams. These teams will play a central role in future alloca- —among the lowest in the world. (Table 16) tions of resources for health. Second, the tool makes it easier Even though government health expenditure to develop arguments in defense of the health budget when rose by around half in real terms over 2007–2013, choices are being made for the annual budget. In 2014, sound it showed a declining trend as a share of total arguments helped the MoPH obtain a 20 percent increase in the budget initially announced for nonwage expenditure. This government expenditure (see Table 16). Health represents an additional USD 10 million in the health allocation. expenditure as a share of total government expen- However, the unpredictability of external resources diture averaged 3.95  percent over 2007–2013, but and uncertainty surrounding decentralization makes the was on a declining trend, especially over 2007–2009. medium-term budget process an especially delicate exer- Over the period, government health expenditure as a cise that often has little link to macroeconomic realities. The share of GDP ranged between 0.3 percent (2009) and MTEFs are developed using incomplete and patchy data: the 0.7 percent (2013). provinces have no clear idea of the domestic and external resources that they will receive the following year. There- The earmarked transfers to the provinces fore, MTEFs are hardly ever used to manage resources and increased the health budget, but had little effect are more of a theoretical exercise. The MoPH’s efforts to on actual health expenditure. Despite the sharply improve the budget process are hampered by the uncer- higher budget allocations for transfers to the provinces tainty surrounding decentralization and the fragmentation (especially in 2012 and 2013—see previous section), of external financing. The MoPH’s financing strategy should the transfers increased the MoPH’s actual expenditure enable the government to set out its official vision of the by only 2 percent in 2012 and 0.1 percent in 2013. health financing and decentralization architecture, which will improve the budget process.  Execution of government health expenditure is volatile and has declined since 2011. The average Source: World Bank/CIDA Project, 2014. execution rate for MoPH expenditure and transfers Table 16  Government Health Expenditure (Execution), 2007–2013   2007 2008 2009 2010 2011 2012 2013 MoPH expenditure (constant 2007 CDF millions) 49,415 39,025 30,123 34,177 46,916 50,253 76,522 MoPH expenditure + transfers (constant 2007 49,415 39,025 30,123 34,177 46,916 51,119 76,614 CDF millions) Central government health expenditure (% of 6.8 3.9 3.3 3.3 4.2 4.3 5.3 total government expenditure) Central government health expenditure (% of 0.6 0.4 0.3 0.4 0.5 0.5 0.7 GDP) Central government health expenditure per 1.5 1.0 0.5 0.5 0.7 0.7 1.0 capita (constant 2007 USD) Source: Authors’ calculations based on disbursement data (2014). Government Financing for Health   37 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 28  E  xecution of Health Expenditure Using Internal Resources (MoPH and Earmarked Transfers), 2007–2013 140 127.8 120 103.4 103.8 100 92.1 79.2 80 71.5 66.6 68.2 60 41.7 40 32.3 20 0 2007 2008 2009 2010 2011 2012 2013 Execution rate MoH budget (% of budget allocations) Execution rate MoH + transfers (% of budget allocation) Source: Authors, based on MoPH data. Note: Earmarked transfers were only made from 2011. stood at 74  percent over 2007–2013, better than Section 3. Government Health in 2004–200728 but with sharp annual variations. Expenditure by Type Execution slipped sharply after 2011, including transfers to 32  percent in 2012 and 41  percent in The economic classification of government 2013. In contrast, expenditure exceeded the budget health expenditure shows changes in recent transfers by 127 percent in 2007 and 103 percent in years, with a shrinking share for operating and 2008. Expenditure was erratic, with a decline until capital expenditure. Current expenditure accounts 2010 followed by annual increases ranging from for an average of nearly 90 percent of government 7 percent to more than 50 percent. health expenditure using internal resources. As a share of the total, personnel spending doubled from The declining execution rate for government 42  percent in 2007 to more than 80  percent over health expenditure since 2011 stems primarily 2009–2012 (against around 90  percent for educa- from nonexecution of transfers to the provinces. tion). The share of operating expenditure steadily These transfers undermine the effectiveness of the declined from around 26 percent in 2007 to 8 per- health budget and cut the execution rate in half— cent in 2013. The share of capital expenditure also in 2013 from 103  percent (excluding transfers) to declined from around 32 percent in 2007 to 3 per- 41  percent (including transfers). In 2011, no pay- cent in 2012, but it then jumped to 27  percent in ments were made on these budget items. In 2012, 2013 after the start of the PESS and the recognition only 1 percent of the budgeted amount was executed of related expenditure in the accounts of the MoPH and in 2013 only 0.1 percent (Figure 28). (Figure 29).29 28 The 2008 Government Health Expenditure Review reported that 29 All of the 2013 PESS expenditure (CDF 20 billion) was attributed to the execution rate for government health expenditure ranged from the MoPH as “investment using internal resources,” even though the 55 percent to 65 percent (World Bank, 2008). MoPH did not set aside provisions in its budget for these operations. 38  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 29  E  conomic Classification of Government Health Expenditure, 2007–2013 (% of total) 100% 3.2% 3.2% 7.9% 90% 17.4% 20.9% 8.0% 26.8% 31.9% 80% 70% 60% 25.8% 50% 85.8% 84.0% 84.2% 40% 80.7% 75.9% 64.6% 30% 20% 42.3% 10% 0% 2007 2008 2009 2010 2011 2012 2013 Personnel expenditure Operating expenditure Capital expenditure Source: Disbursement data and World Bank calculations, 2014. Analysis of expenditure by function was difficult construction, and renovation seems to follow an because of inaccurate reporting of some expen- implementation and execution pattern that has no diture items. Government expenditure is most relation to the budget allocations for these items. directed to health promotion activities, public health A detailed analysis of the budget execution campaigns, and programs that deal with major dis- process shows that the main problems lie with eases. Together these accounted for 42  percent of the MoPH (commitments) and the Ministry total government health expenditure, or CDF 91.5 of Economy and Finance (payments). Over billion, in 2013. Expenditure on operating health 2011–2013, expenditure commitments on equip- facilities, pharmaceutical products, and vaccines ment, services, and transfers by the MoPH came was small at only 5 percent of the total that year. to 14  percent, 21  percent, and 59  percent of the Personnel expenditure has maintained a steady allocations received. Expenditure commitments share while execution of other types of expen- are more in line with the allocations received for diture varied widely (Figure 30). When personnel personnel expenditure (94 percent), goods and ser- expenditure is stripped out, no clear trends emerge vices (116  percent), but less so for construction in expenditure execution. Personnel expenditure (67  percent). The second problem lies with pay- execution tended to be in line with allocations, with ments. Payments come to only 55  percent of the execution rates ranging from 84 percent in 2012 to transfers for goods and equipment and 40 percent 102  percent in 2008. Conversely, other operating of the transfers for construction (Table 17). On the expenditure was far over budget from 2007 to 2009, whole, it seems that much less use was made of but with average execution rates of less than 50 per- exceptional procedures, which represent less than cent after that. Capital expenditure for equipment, 10 percent of total allocations. Government Financing for Health   39 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 30  E  xecution of Government Health Expenditure by Type of Expenditure (% of allocations), 2007–2013 700 600 500 400 300 200 100 0 2007 2008 2009 2010 2011 2012 2013 Personnel Goods and supplies Equipment Construction, renovation Source: Authors, based on disbursement data (2014). Table 17  Analysis of the Government Health Expenditure Execution Chain for All Resources, 2011–2013 Total allocations Expenditure in current CDF commitments Validations Payment Payments billions (2011–2013) (% of allocations) (%) orders (%) (%) Personnel 351.5 94.1 94.1 94.0 93.6 Goods and equipment 45.8 116.0 115.0 63.8 54.5 Services 5.5 20.9 20.2 18.3 18.1 Transfers 35.7 58.7 58.5 45.1 41.2 Equipment 597.5 14.4 14.4 14.1 13.5 Construction, rebuilding, renovations 59.9 67.0 49.2 59.5 39.5 Source: Authors, based on consolidated data from the MoPH Research and Planning Division, Ministry of Budget and Ministry of Economy and Finance 2014. Section 4. Expenditure for Health 2009–2013 was around 200  percent (Table 18). Personnel Most of the increase stems from large pay rises in the civil service, since the number of employees on Health personnel expenditure has shot up since the payroll increased by only 25  percent. Average 2007. The total payroll using internal resources civil service pay increased by more than 140  per- increased nearly sixfold over 2007–2013, from CDF cent, from CDF 340,457 in 2009 to CDF 816,192 21 billion to CDF 125 billion. The increase over (USD 380) in 2013. Bonuses rose by 236  percent, 40  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 18  Wage Expenditure, 2009–2013 Table 19  Government Compensation of Health Personnel, 2013 Increase   2009 2013 (%) Average by rank, official pay scale Total compensation of 41.6 124.9 200 (constant 2013 US dollars) government employees Maximum Minimum Risk bonus (%) (current CDF billions) Medical director 806 727 86 Wages 14.54 33.80 132 Junior doctor 660 85 Bonuses 27.14 91.18 236 Head nurse 190 155 39 Number of wage-earning 32,450 40,626 25 Nurse 138 125 24 employees Manager 117 101 21 Number of bonus-earning 90,000 112,507 25 employees Source: MoPH 2013 and authors’ calculations. Source: Authors’ calculations based on MoPH data, 2013. in compensation for health employees possible, but the average rising from CDF 301,633 in 2009 to systematic use of bonuses, which now account for CDF 810,474. The larger bonuses primarily ben- around 85  percent of physicians’ compensation efited physicians more than nurses, averaging USD from the government (Table 19), raises problems for 660–806 (USD 138 for nurses). The “hazard bonus” the medium and long term: bonuses are not included started as top-up pay and an incentive; it is now in base pay; there are no financial incentive schemes; the main component of civil service pay. The bonus and bonuses are not taken into consideration when accounted for more than 85 percent of government calculating retirement pensions. paychecks for physicians and 73 percent of person- Wages account for only a minor share of total nel expenditure in 2013, up from 65 percent in 2009. compensation of health personnel, as seen. A The increase in personnel expenditure stems recent survey30 shows that compensation from gov- from better coverage of health personnel work- ernment sources (wages and bonuses) account for ing in or for the civil service. Government efforts a small part of total compensation for most health have provided some form of compensation, includ- personnel, even though it accounts for a larger share ing bonuses, for nearly 80 percent of health person- of physicians’ compensation (42  percent). In con- nel (112,507 out of a total of 141,759 in 2014), even trast, users’ out-of-pocket payments based on cover- if a salary is the norm for only 31  percent of the ing costs are the main source of compensation for whole staff. Further, wage increases have brought all personnel in health facilities (except for central physicians’ pay into line with that in neighboring and provincial administration staff). External assis- countries. The system for depositing employees’ pay tance, in the form of wage supplements (per diems, directly into their bank accounts has been in place bonuses, and incentives) is the main source of income since end-2013, making payments more secure. for the executives supervising the health district, but it accounts for a smaller share for other personnel, The increase in payrolls raises questions about from 27 percent for nurses to 11 percent for physi- the medium-term financial sustainability of cians. Income from activities outside the health field this expenditure. It is estimated that the alloca- account for around 20  percent of total income of tions for personnel costs will need to be more than health personnel and this share is proportionately doubled (CDF 250 billion) to only cover wages and higher for unskilled personnel (28 percent) and for bonuses for all of the employees recognized by the pharmacists (27 percent). MoPH in 2014. This amount represents one and half times the MoPH’s total budget in 2013 (CDF 186 billion). Paying bonuses made a large increase 30 World Bank and University of Kinshasa, forthcoming. Government Financing for Health   41 Health Expenditure Performance 6 Section 1. Financial Protection The concentration curve of out-of-pocket expen- diture lies below the line of equality, which shows That 90 percent of total health expenditure is that the richest spend less on health than the household out-of-pocket expenditure is a pri- poorest (Gini=0.432). Furthermore, out-of-pocket mary indicator of the poor financial protection payments exceed households’ ability to pay, since provided by the health financing system in DRC. the concentration curve of payments lies below the Prepaid systems from voluntary insurance plans Lorenz curve. The poor spend proportionally more cover only a tiny proportion of the population. than their ability to pay than the richest (Figure 31). Furthermore, access to insurance plans is linked to income: 0.7 percent of men and 1 percent of women Equity in household expenditure varies by prov- in the poorest quintile reported that they had insur- ince and by place of residence. The difference ance, whereas 12 percent of men and 15.3 percent of between urban areas (11.0 percent) and rural areas women in the richest quintile have insurance (DHS, (4.7  percent) stems from the lower use of services 2013/2014). in rural areas, especially costlier hospital services. The provinces of Maniema, Nord Kivu, and Sud Financing health through households’ out-of- Kivu are where private health expenditure is pro- pocket expenditure disadvantages the poorest, portionally the highest (11  percent of nonfood who spend a larger proportion of their income expenditure). Bandundu and Katanga are the two for their health. The share of health in household provinces where the distribution favors the rich the expenditure increases with poverty. Health absorbs most, whereas Sud Kivu has the least inequitable 6.48  percent of household nonfood expenditure distribution of out-of-pocket health payments. In for the population as a whole, from 11.44 percent four provinces (Bandundu, Equateur, Maniema, and (poorest quintile) to 4.48  percent (richest quintile) Province Orientale) health payments are equivalent (Table 20). Table 20  Share of Health in Household Expenditure by Income Quintile, DRC Q1 Q2 Q3 Q4 Q5 Mean Total private health expenditure per 20,341 28,265 35,185 40,344 59,188 39,270 household (current CDF) Share of health in household total 3.39 2.97 2.80 2.48 2.02 2.41 expenditure (%) Share of health in household nonfood 11.44 10.97 9.83 7.78 4.48 6.48 expenditure (%) Source: Authors, based on 1-2-3 Survey data, NSI 2013. 42 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 31  Concentration Curves and Lorenz Curves for Out-of-Pocket Payments, DRC DRC 1 .8 payment/expenditure (%) Cumulative proportion of .6 .4 .2 Gini(exp.)=0.432*** C.I.(OOPE)=0.331*** 0 0 .1 .2 .3 .4 .5 .6 .7 .8 .9 1 Cumulative proportion of population (%) Line of equality Lorenz curve Out-of-pocket health payment Source: Authors, based on 1-2-3 Survey data, NSI 2013. ***P < 0.001 to the ability to pay (the concentration curves and total expenditure, probably affected nearly the Lorenz curves meet). In the other provinces, pay- 13 percent of the population in 2013 (Table 21). ments are regressive. However, another commonly used measure for esti- mating catastrophic expenditure—health expendi- With no reliable data from the last decade, it is ture of at least 40  percent of household nonfood hard to estimate real changes in financial pro- expenditure—gives a lower estimate of incidence tection. According to WHO, the household share of catastrophic expenditure, affecting 3.84  percent of health expenditure fell from 90 percent in 2002 of the population. Catastrophic expenditure affects to 50  percent in 2012. There are questions about the poor more: using an intermediate measure of the reliability of these figures, however. Failure to 20 percent of nonfood expenditure, 16.3 percent of include data on external assistance could result in the poorest spend more than this share on health, overestimation of households’ expenditure as a but only 10.5 percent of the nonpoor. share of health financing in the early 2000s. The high estimate of household expenditure in 2002, at Table 21  Estimated Incidence of Catastrophic Health 90  percent of total health expenditure, could lead Expenditure, DRC, 2013 us to think that the share of household expendi- ture decreased substantially from 2002 to 2008, Household health expenditure (share of total household >40% nonfood expenditure) whereas this expenditure seems to have stagnated at Share in the total population (%) 3.84 a level equivalent to the regional average of around Household health expenditure (share of total household >10% 50 percent. expenditure) The incidence of catastrophic health expen- Share in the total population (%) 12.91 diture, estimated at 10 percent of household Source: Authors, based on NSI data, 2013. Health Expenditure Performance   43 2014 HEALTH PUBLIC EXPENDITURE REVIEW Impoverishing expenditure is estimated to affect Section 2. Equity roughly 1 percent of the national population and 2 percent of the rural population. In Pen’s Parade, Poverty and Place of Residence all the households in the first to third income quin- Despite the fall seen in incidence of poverty over tiles have impoverishing health expenditure. Despite 2005–2012, disparities between provinces and their low health expenditure, it keeps them below the between the provinces and Kinshasa are still poverty line. The situation is different for the richer very stark. Nationally, the incidence of poverty fell households in the fourth and fifth income quintiles. by 8 percentage points, from 71.3 percent in 2005 Still, it seems that a large proportion of households to 63.4  percent in 2012, but it is still greater than are pushed below the poverty line by major health 70 percent in four provinces (Kasaï Occidental and expenditure (Figure 32). For some households, Oriental, Equateur, and Bandundu). The regional impoverishing health expenditure amounts to half poverty incidence is lowest in Kinshasa at 36.8 per- their disposable income. cent, or equal to half the Kasaï Occidental. Over Figure 32  Effects of Health Expenditure on the Pen’s Parade Curve of Household Consumption 8 6 Consumption as Multiple of PL 4 2 0 0 .2 .4 .6 .8 1 Cumulative Proportion of Population, Ranked from Poorest to Richest Pre-OOP Consumption Post-OOP Consumption Source: Adept Survey, MoPH 2014b, based on data from MICS and NHA 2010. Note: The figure shows Pen’s Parade for household unadjusted consumption and out-of-pocket health expenditure. Household consumption is expressed as a curve starting below the extreme poverty line based on minimum food requirements set at a threshold of USD 1.08. The “paint drip” shows how much subtracting out-of-pocket health payments reduces consumption for each household. If the drip falls below the poverty line, the household is not classified as poor on the basis of its unadjusted consumption, but is poor based on its net consumption. 44  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW 2005–2012, half the provinces saw declines in pov- significantly more likely to die before their first erty incidence of more than 13 percentage points, birthday in Bandundu or their fifth birthday in while Kasaï Occidental and Oriental as well as Equateur. However, the degree of inequality declined Maniema recorded higher incidence. The decline in over 2007–2013. incidence was greatest in rural areas, where it fell by There has also been notable progress on infant 10.6 percentage points, against only 1.4 points in and child mortality in rural areas over 2007– urban areas (NSI, 2013). 2013. Infant mortality fell from 12 percent to 7 per- cent in rural areas and from 8 percent to 6 percent Inequalities in Health Outcomes in urban areas. Infant and child mortality rates Infant and child mortality is concentrated in the fell from 18  percent to 13 percent in rural areas poorest population groups, but the disparities and from 12  percent to 10 percent in urban areas narrowed over 2007–2014. In 2007, nearly 12 per- (MoPH, forthcoming b). cent of children born to the poorest households died The prevalence of childhood diseases is higher in before their first birthday, against 6 percent in richer the poorest population groups. The prevalence of households. Seven years later, infant mortality rates acute respiratory infection (ARI) is much lower for were more uniform between the income quintiles at children of mothers with a higher education, 3 per- 6  percent (richest) and 7  percent (poorest). Infant cent, versus 6 percent for mothers with a secondary and child mortality rates also improved faster for the education, 7 percent for mothers with a primary edu- poorest, falling from 18 percent to 13 percent, against cation, and 8 percent for mothers with no education. a decline from 11 percent to 9 percent for the rich- The variations according to the index of economic est. Concentration indicators show that the gaps for well-being are minor, even though children from infant mortality narrowed from –0.11 to –0.03, while households in the highest quintile suffer less from those for infant and child mortality narrowed from ARI, with a prevalence of 5 percent, versus an aver- –0.09 to –0.06 (MoPH, forthcoming b; Table 22). age of 7 percent for the other quintiles. In addition Child health disparities exist between the to social and demographic factors, the prevalence provinces, but they have narrowed over time. of childhood diarrhea also varies by province, with Bandundu is the only province that posted infant one out of five children suffering from diarrhea in and under-five mortality rates that were always Sud Kivu and Kasaï Oriental, versus less than one in below the national average for the last 15 years, 10 in Maniema. It appears that access to improved whereas Sud Kivu, Province Orientale, and Katanga latrines does not have a significant impact on the have always had rates higher than the national aver- prevalence of diarrhea among the children surveyed age. Children born into the poorest households are (DHS, 2013/2014). Table 22  Concentration Indicators, Infant and Under-5 Mortality by Province (sample) and National Indicator, 2007–2013 Bandundu Equateur Katanga Maniema K. Oriental Sud–Kivu National Under-5 mortality 2007 –0.0642 –0.0789 –0.1226** –0.0728 –0.0270 –0.1653** –0.1206** 2013 –0.0322 0.0237 –0.1028** 0.0068 0.0328 –0.0915** –0.0265 Infant mortality 2007 –0.0972** –0.0607** –0.1730 –0.0175 0.0575 –0.1935 –0.1104 2013 0.0004 0.0352 –0.0615 0.0110 0.0569 –0.0286 –0.0174 Source: Authors, based on 2007 and 2013–2014 DHS. Health Expenditure Performance   45 2014 HEALTH PUBLIC EXPENDITURE REVIEW The prevalence of malnutrition, as estimated by The distribution of mothers’ use of services is sig- the number of underweight children, is increas- nificantly more favorable to the rich in Bandundu, ing among the poorest. It rose from 27  percent Equateur, and Province Orientale. to 31 percent for children from the poorest house- The increase in the use of antenatal care and holds but declined from 15 percent to 8 percent for skilled birth attendance favored the poorest children from the richest households. All provinces households. The use of these services is still concen- are seriously affected by malnutrition, with preva- trated among the richest, but the scale of inequality lence of more than 20 percent, apart from Kinshasa, narrowed gradually over 2007–2014, with concen- where prevalence was 5  percent in 2013–2014. tration indexes falling from 0.087 to 0.058 for four Maniema, where 33 percent of children are under- antenatal visits. The disparities are still stark between weight, is by far the most seriously affected. The the provinces and Kinshasa, where 76  percent of decrease in malnutrition was smaller in rural areas, women receive full antenatal care; in eight out of 11 where prevalence fell from 29 percent to 27 percent provinces, fewer than half receive such care. over 2007–2014, than in urban areas, where it fell from 20 percent to 14 percent. The same pattern is The use of modern contraception is still strongly found for stunting, which is decreasing less rapidly an urban phenomenon slanted toward the in urban areas. ­ richest—17 percent of women—versus 6 per- cent in rural areas. In Kinshasa and Bas Congo, The figures broken down by province show 21  percent of married women report that they use that childhood malnutrition increased in Bas a modern contraception method, whereas most of Congo and Maniema, whereas it decreased or the other provinces fall below the national average remained the same in the rest of the country. of 8 percent. The differences between rich and poor In Maniema, the prevalence of underweight chil- are wide: only 4 percent of the poorest women use dren is estimated to have increased by 11 percentage a modern contraceptive method, against 20 percent points over 2007–2014. In Bas Congo it increased of the richest. from 26  percent to 28 percent. In contrast, preva- lence in Kinshasa fell by two-thirds to 5 percent in The use of services for treating childhood diar- 2013. The same pattern is found for stunting, which rhea improved for all quintiles, including the increased by some 2 percentage points in Bas Congo, poorest. A significant increase of 10 percentage Maniema, and Kasaï Occidental, but decreased in all points, from 21 percent to 31 percent, was seen in other provinces. access to treatment for diarrhea for the poorest chil- dren. The increase was sharper in the provinces than in Kinshasa. Provinces such as Kasaï Occidental and Inequalities in Use of Services31 Oriental as well as Province Orientale saw increases The most recent data show major inequalities of more than 10 percentage points. With 31 percent in coverage of maternal health services based and 32  percent of children having access to treat- on income and place of residence. One-quarter ment for diarrhea, Equateur and Province Orientale of the poorest women do not receive any antenatal fall below the national average of 39 percent. care, whereas only 8 percent of the richest women do not receive any. Nine percent of the women liv- Access to treatment for ARI and for malaria, ing in Kinshasa do not receive any antenatal care, which government facilities provide for free, whereas the figures are more than 20  percent in decreased, especially among the richest. Access to Katanga, Kasaï Occidental, and Kasaï Oriental. ARI treatment fell for all children, but the decrease was sharper for the richest households at 15 per- 31 This subsection draws on analysis carried out for this report and a centage points than among the poorest households separate Equity Analysis conducted using Adept software. at 6 percentage points. Access is still very restricted 46  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW for all quintiles, at 39  percent for the richest and Oriental saw their coverage rates fall below 40 per- 33 percent for the poorest. In Province Orientale and cent and below the national average of 46 percent. Nord Kivu, access to this treatment fell by half to Even though coverage rates increased in Equateur less than 25 percent in 2013–2014. Access to treat- and Province Orientale, these two provinces still ment for ARI also fell sharply in urban areas from have the lowest vaccination coverage rates, with 46 percent to 35 percent of children infected, match- 33  percent and 31  percent of their children fully ing the level of treatment for rural children. Access vaccinated. to malaria drugs also fell, particularly for the rich- The use of secondary level health facilities shows est households, from 63 percent to 43 percent. This a fairly equal distribution between the income decline may stem from less use by the richest quin- quintiles, but access to tertiary health facilities tiles of government health centers that distribute is still slanted toward the richest quintile. In con- free malaria treatments. On the other hand, access trast, only 24 percent of the top quintile use primary to means of prevention and, more specifically, access health facilities, compared with 45  percent of the to insecticide-treated nets that are distributed free poorest (Figure 33). The use of private health ser- in local communities is fairly equitable, with little vices is concentrated in urban areas, accounting for differences in use between the income quintiles and 46 percent of the care consumed, against 13 percent places of residence. in rural areas. The decrease in vaccination coverage is esti- mated to have affected the poorest more Benefits of Government Financing severely,with a decline of 6 percentage points over Government health expenditure benefits hospi- 2010–2014, whereas the coverage rate increased tals more. Operating expenditure, modest though by 9 percentage points for the richest households. it is, goes virtually exclusively to hospitals. Further, Provinces such as Bandundu, Katanga, and Kasaï Figure 33  Inequalities in the Use of Services by Income and by Service Level, DRC, 2013 Use of primary care facilities Use of secondary level facilities Use of tertiary level facilities 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0 50.0 Use of secondary level Use of tertiary level facilities Use of primary care facilities facilities Q5 10.3 11.7 23.9 Q4 5.1 10.6 39.6 Q3 4.7 7.9 44.2 Q2 2.7 8.4 43.8 Q1 3.0 9.0 45.4 Source: Authors, based on 1-2-3 Survey data, NSI 2013. Health Expenditure Performance   47 2014 HEALTH PUBLIC EXPENDITURE REVIEW as of 2013, new fixed grants for district hospi- subsidies. Subsidies for ambulatory care in govern- tals’ operating costs were introduced. Government ment hospitals are used more by the richest quintile. expenditure on health centers, in the form of operat- These are the most inequitable subsidies, bringing ing subsidies, is virtually nonexistent. Yet if we con- very little benefit to the poorest quintile, which sider all the current expenditure, including wages, receives 5.8 percent, as opposed to the top quintile, the distribution may be slightly more balanced, which receives 40.7 percent (MoPH, forthcoming b). even though the majority of qualified personnel are The geographic distribution of government employed by hospitals. However, the breakdown of expenditure is inequitable. The majority of gov- expenditure by level of health facility is difficult to ernment expenditure is inversely related to local estimate because of the lack of detail in the classifica- income. Kinshasa receives CDF 2,431 per capita tion. In addition, the primary level of health facilities a year, whereas the other provinces, except Bas encompasses health centers and district hospitals, Congo, receive less than half that. The ratio between which makes the cut-off between levels even vaguer. the capital and the province receiving the lowest Government health expenditure appears rather amount per capita is 5 to 1 (Figure 35). inequitable. The richest quintiles use hospital ser- vices more than the poorest (Figure 34), but it seems that current expenditure benefits higher levels of Section 3.Technical and Allocative health facilities more than primary health facilities. Efficiency While 87  percent of government expenditure goes For the same level of health expenditure, health to government hospitals—with 70 percent for hos- outcomes in DRC fall short of those of peer pitalization services and 17 percent for ambulatory countries. Government expenditure on maternal care—only 13  percent goes to ambulatory care in health in DRC seems to be more efficient and pro- health centers. Under the proportional cost assump- vides antenatal care and skilled birth attendance tion, government expenditure favors the top quin- coverage that is higher than the regional average tile, which accounts for 22 percent of the total. The (Figure 36). However, the maternal mortality rate is poorest quintile receives 16  percent of total public much higher than the regional average, which hints Figure 34  Distribution of Government Subsidies by Income Quintile, DRC 45 40 35 30 25 20 15 10 5 0 Subsidies for Subsidies for Subsidies for Total public subsidies inpatient care in ambulatory care ambulatory care to health facilities hospitals provided by hospitals provided by health centers Lowest quintile Q2 Q3 Q4 Highest quintile Source: Authors, based on NHA and MICS data, 2010. Note: Calculations under the proportional cost assumption. 48  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 35  G  eographic Distribution of Government Expenditure per Capita, 2013 (CDF) 3000 2500 2431 2000 1500 1317 958 1000 722 731 813 824 848 944 700 546 661 500 0 ga u al e u Eq tal ur Ba ma du Ba nne go sa iv al i O Kiv nt te en ha n un on -K nt ie rie ta e ua rie id an ns oy d nd C d Ka O Ka Nor Su cc s- Ki M M O i sa ce Ka sa in ov Pr Source: Authors, based on disbursement data (2013). Note: For the sake of accuracy, the geographic distribution of government expenditure is calculated on the basis of personnel expenditure in each province. Estimated on the basis of personnel expenditure. Figure 36  R  ate of Skilled Birth Attendance (left) and Antenatal Care (right) vs. Government Health Expenditure per Capita and Peer Countries (2013 or most recent available data) Accouchements assistes par du personnel qualifie 100 100 90 90 Femmes enceintes recevant des soins DRC 80 DRC 80 70 70 Sub-Saharan Average prenataux (%) 60 60 (% du total) 50 Sub-Saharan Africa 50 40 Average 40 30 30 20 20 10 10 0 0 0 2 3 4 5 6 7 8 0 2 3 4 5 6 7 8 9 Log depenses publiques de sante par tete (US$ courants) Log of depenses publiques de sante par tete (US$ courants) Source: Authors, based on WDI data. Note: Skilled birth attendance rate is percentage of total births). Antenatal care is percentage of pregnant women. Log of government health expenditure per capita in current USD. at major problems with the quality and effectiveness Government health expenditure by prov- of services (Figure 37). For the same level of expen- ince does not seem to be linked to priorities diture, DRC’s performance is also much poorer than for action or to health outcomes. For example, the rest of the region for infant mortality and under- Maniema has higher government health expendi- five mortality (Figure 38). ture per capita than the other provinces, but it has Health Expenditure Performance   49 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 37  Maternal Mortality vs. Government Health Expenditure per Capita, 2005 and 2010 2005 2010 1500 1500 (modeled estimate, per 100,000 live births) (modeled estimate, per 100,000 live births) 1000 C Congo, Dem. Rep. 1000 Maternal mortality ratio Maternal mortality ratio Congo, Dem. Rep. Sub-Saharan African Average Sub-Saharan African Average 500 500 0 0 0 5 50 100 250 1000 2500 10000 0 15 50 100 250 1000 2500 10000 Public Health Expenditure per Capita (current US$) Public Health Expenditure per Capita (current US$) Source: Authors, based on WDI data. Note: x-axix log scale. Gray area indicated 95% confidence interval for the fitted line. not achieved better coverage for health services. The mortality (Figure 39). Coverage of antenatal care allocations for Bandundu, Equateur, and Province does not seem to vary according to the level of per- Orientale are fairly similar, but Bandundu shows sonnel available in each province, but skilled birth better outcomes on vaccination coverage and skilled attendance does seem to be inversely linked to per- birth attendance. The priority on allocations for sonnel density. Equateur, which has the most health hospitals also makes it harder to improve the per- personnel of the six provinces in the sample, seems formance of primary care and the health system as a to have the lowest rate of skilled birth attendance, at whole (Chapter 6, Section 2). less than 20 percent. The performance of health personnel varies by DRC introduced a performance-based financing province and could be greatly improved. The lack mechanism (PBF) in 2004 (Box 6). The mechanism of reliable data makes it impossible to assess pro- was scaled up quickly to purchase services based on ductivity by output per person.32 However, dispari- the performance of health facilities. In DRC, PBF ties in coverage indicators and health outcomes by is used to link financing to specific outcomes, and province, compared with health human resources, payment is only made once the expected results suggest that the performance of health person- have been achieved and verified. In 2014, PBF was nel could be improved appreciably. Bandundu and implemented in more than 120 health districts in the Province Orientale have the same level of qualified country, covering some 20 million people. Purchases personnel, but their use rates vary by a factor of two of services included both the minimum care pack- in terms of coverage for child health care and infant ages provided by health centers and the supplemen- tary package provided by district hospitals and other 32 A performance assessment of the health care delivery system will be carried out in 2015. It could shed light on personnel performance facilities. An independent agency performs the care and service quality. purchase function. 50  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 38  I nfant (upper figures) and Under-5 (lower figures) Mortality Rates vs. Government Health Expenditure per Capita, 2007 and 2012 2007 2012 150 100 Infant Mortality Rate (1,000 live births) Infant Mortality Rate (1,000 live births) Congo, Dem. Rep. 100 Congo, Dem. Rep. Sub-Saharan African Average Sub-Saharan African Average 50 50 0 0 0 5 50 100 250 1000 2500 10000 0 15 50 100 250 1000 2500 10000 Public Health Expenditure per Capita (current US$) Public Health Expenditure per Capita (current US$) 9435_FG06_08.eps 9435_FG06_09.eps 2007 2012 200 200 Under-5 Mortality Rate (1,000 live births) Under-5 Mortality Rate (1,000 live births) 150 150 Congo, Dem. Rep. Sub-Saharan African Average Congo, Dem. Rep. 100 100 Sub-Saharan African Average 50 50 0 0 0 5 50 100 250 1000 2500 10000 0 15 50 100 250 1000 2500 10000 Public Health Expenditure per Capita (current US$) Public Health Expenditure per Capita (current US$) Source: Authors, based on WDI data. Note: x-axix log data. Gray area indicated 95% confidence interval for the fitted line. Health Expenditure Performance   51 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 39  Availability of Qualified Personnel vs. Use of Services for Treating Fever and Infant Mortality 100 80 Prov. Orientale 90 70 80 60 70 50 60 50 40 40 30 Bandundu 30 20 20 10 10 0 0 0 2 4 6 8 10 12 14 16 Infant mortality (left scale) Fever treatment (right scale) Source: Authors, based on MoPH data for personnel and DHS 2013/2014 for the other data. Note: Each dot represents a province. The y-axis shows access to treatment of fever for children under five in light blue. Infant mortality is shown in dark blue. The x-axis shows the ratio of qualified personnel per 10,000 population. Box 6  Performance-Based Financing: Findings of an Impact Study in Katanga Haut-Katanga (population: 1.26 million) was part of the World activities, indicating more motivated personnel. The mechanism Bank Health Sector Rehabilitation Support Program (HSRSP), did not lead to a reduction in the supply of nontargeted ser- which covered 83 health districts in five DRC provinces. An vices and it did not have a negative impact on service quality (or impact assessment was carried out in 98 health areas, which a positive impact either). However, the study also shows that were assigned randomly to 2 groups: 48 districts received the test project did not lead to a significant increase in the use performance-based financing and the others received fixed of services, despite lower costs for users. The use of services payments. was similar between the PBF districts and the other districts. The impact assessment in Haut-Katanga found that the PBF The main obstacle that households reported to account for mechanism was effective in increasing the supply of targeted nonuse of services was a lack of confidence about the benefits services. Personnel attendance was 14 percent higher in the of the care provided. PBF areas. The number of preventive health sessions was also Source: Huillery and Seban, 2014. greater, along with the number of community-based outreach 52  Investing in Universal Health Coverage 7 Decentralization: Impact on Deployment and Use of Health Resources Section 1. Financing Flows terms of execution. Bandundu and Equateur receive (excluding wages) more than twice the transfers budgeted, thus match- ing the amounts actually transferred to Katanga. Provinces receive three types of earmarked transfers: investment (capital spending); operating; and fixed Despite increased transfers, government expen- grants for deconcentrated services. However, a large diture is still mainly controlled and executed by share of government expenditure (excluding wages) the central government, which is a sign of incom- is still executed at the central level on behalf of the plete decentralization. Investment in the provinces provinces. is still largely controlled by the central government, in the absence of effective public procurement man- Investment is the priority. Total budgeted invest- agement in most provinces. The central government ment allocations more than doubled from 2010 executes the PESS directly; no provisions are made to 2013, accounting for more than 75  percent of for the program in the provincial budgets. The cen- total allocations. However, the amounts executed tral government also continues to execute personnel are small, averaging for the six sampled provinces expenditure and no guidance has been provided to 26 percent over 2010–2013. date about transferring the payroll function to the Operating allocations are smaller,at roughly provinces. one-quarter the size of investment allocations, and The provincial budget-making process is still in its show little growth, though the central government infancy. A lack of reliable information and prob- is improving execution of transfers. The execution lems tracing revenue and expenditure hamper rate rose from 75 percent in 2010 to 94 percent in budget planning. Planning processes do not have 2013, taking operating grants to half the transfers adequate macroeconomic guidelines; the expected executed. Fixed grants for deconcentrated services, revenue and expenditure amounts are often out of the third type of transfer, have minimal allocations, step with the availability of local and central govern- though on this tiny base they have quite high execu- ment resources, on the one hand, and disbursement tion. They ultimately account for only 1 percent of capacities, on the other. Technical staff have little actual transfers. say in budget planning, which, despite a vote by the The geographic distribution of transfers shows Provincial Assembly, is primarily a political process. no discernible pattern and their execution is The recent introduction of guidelines such as MTEF unpredictable. Katanga receives the largest sub- in the provinces have helped provide guidance for sidy, which is 10 times as great than the subsidy for forecasting revenue and expenditure, but budget Maniema (Figure 40). However, the ratio is 5 to 1 in elaboration process remains perfectible. Further, the 53 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 40  I nvestment Transfers Allocation and Execution, 2010–2013 Total by Province in CDF Million 800000 700000 600000 500000 400000 300000 200000 100000 0 Maniema Sud Kivu Kasai Oriental Bandundu Equateur Katanga Allocations Executions Source: Authors, based on compilation of various central and provincial sources. provincial budgets are presented using nonstandard in health financing at provincial level: in 2013, the formats that differ between provinces. Budgets are amounts provided to provinces by external partners not broken down by administration; areas of juris- were nearly three times as great as total govern- diction are still vague and vary from one province to ment funds allocated for health (and executed rates the next. This increases the lack of transparency in were of course far lower). In some provinces, such sector budgets.33 Some budgets are also artificially as Maniema or Sud Kivu, external assistance is up inflated by incorrect recognition of resources and to six times as high as central and provincial gov- expenditure executed by the central government, ernment funds (Figure 41). The majority partner in such as wages. Sud Kivu, Katanga, and Kasaï Oriental is USAID. In the three other provinces in the sample, the World Bank is the main donor through the Health Sector Section 2. Provincial Health Financing Rehabilitation Support Program (HSRSP). Resources Central government funds account for the major- As with the central government, external assis- ity of provincial government funding. Katanga tance is the leading health financing source for aside, funds raised by the provinces to finance health provinces. Estimates of external assistance vary by are small (Figure 42). In Kasaï Oriental, Sud Kivu, source, and few provincial authorities are aware of Bandundu, and Equateur, provincial resources for them. Our analysis based on questionnaires filled in health come to CDF 53 and CDF 97 per capita— by development partners at national and provincial less than USD 0.10. In Maniema, provincial funds level show that external assistance is predominant are greater, but are still only one quarter that of cen- tral government health transfers. In Katanga, how- 33 Bandundu: Ministries of Public Health, Social Affairs, Humanitarian ever, central government and provincial government Affairs, and of Youth and Sports; Equateur: Ministries of Public Health, of Social Affairs, and relations with Parliament; Maniema: funds seem to be in balance at about USD 0.60 per Ministries of Public Health and of Social and Humanitarian Affairs. capita each. 54  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 41  External Assistance and Government Funds in Health Financing per Capita, CDF 2013 Bandundu Equateur Maniema Sud Kivu Katanga Kasai Oriental Central External Source: Authors, based on data compiled from provincial authorities and development partners using questionnaires. Figure 42  C  entral and Provincial Government Funds for Financing Health Care per Capita, CDF, 2013 1373 658 1336 1043 946 736 608 366 97 53 74 53 Bandundu Equateur Maniema Sud Kivu Katanga Kasai Oriental Provincial Central Source: Authors, based on national and provincial budget data. Decentralization: Impact on Deployment and Use of Health Resources   55 2014 HEALTH PUBLIC EXPENDITURE REVIEW Table 23  Health Allocations as Percentage of Provincial External assistance, however, declined by 13 percent Government Budgets, 2010–2014 (%) (Figure 43). Starting in 2010 there was a down- Health allocation wards trend in Bandundu, Equateur, Maniema and Province (% of provincial budgets)a Katanga. Sud Kivu, which suffered from a succession Bandundu 5.0 of military and humanitarian crises, posted a signifi- Equateur 5.0 cant increase in external assistance up until 2012, Maniema 7.3 followed by a slight decline in 2013 (Figure 44). Sud Kivu 2.0 Katanga 3.0 Kasaï Oriental 2.7 Section 3. Provincial Government Health Expenditure Source: Authors, based on provincial budget documents. a. From internal resources and unearmarked transfers.  Government health spending per capita is very low. Among the six provinces, it varied little among Provincial budgets devote a tiny share of expen- them (except for Maniema), from CDF 809 to CDF diture to health. Excluding earmarked transfers, 1,364 (USD 0.90 to USD 1.50) in 2013. provinces allocate an average of 4  percent of their Central and provincial government resources resources to health (Table 23). are primarily used in the provinces to cover per- Decentralization efforts meant that expenditure sonnel expenditure.34 Personnel expenditure is financed by government funds increased more primarily paid by the central government, but some than expenditure funded by other financing sources. Earmarked transfers from central govern- 34 Apart from Katanga and Maniema, personnel expenditure accounted ment for health nearly doubled from 2011 to 2013 in for more than 90 percent of total government expenditure executed the provinces in the sample, rising from CDF 3,476 in 2013: 92 percent in Bandundu, 97 percent in Kasaï Oriental, 99 percent in Sud Kivu, and 99 percent in Equateur. Maniema and to CDF 6,091 per capita. Provincial funds for health Katanga were the closest to the national average, at 61 percent and increased less rapidly, but still rose by around half. 62 percent. Figure 43  Health Expenditure per Capita by Financing Source, 2011–2013 28,940 22,035 19,254 5,793 6,091 3,476 841 590 1,250 2011 2012 2013 Provincial Central External assistance Source: Authors, based on data compiled from provincial authorities and development partners using questionnaires. 56  Investing in Universal Health Coverage 2014 HEALTH PUBLIC EXPENDITURE REVIEW Figure 44  E  xternal Assistance Disbursements by Province, CDF, 2010–2013 45,000,000,000 40,000,000,000 35,000,000,000 30,000,000,000 25,000,000,000 20,000,000,000 15,000,000,000 10,000,000,000 5,000,000,000 0 2010 2011 2012 2013 Bandundu Equateur Manierma Kasai Oriental Katanga Sud Kivu Source: Authors, based on data compiled from development partners using questionnaires (disbursements). provinces top up wages with their own resources, (office of the District Manager, district executives, such as Bandundu, Kasaï Oriental, and Sud Kivu. provincial authorities). In a sample of three provinces (Katanga, Sud Kivu, and Kasaï Oriental), operating A few provinces use internal resources purely expenditure accounted for half of the disbursements on investment. This is true for Katanga (100%, or by external partners. The other main expenditure more than 37 percent of total government resources items are services (15 percent), equipment, goods, spent on health) and Maniema (100% and 18 per- and supplies (18  percent), and personnel (13 per- cent). Investment in other provinces was financed cent) for bonuses and other wage supplements. by the central government, but the amounts were very small at less than 1 percent Kasaï Oriental as Expenditure execution using provinces’ internal an example. resources is fraught with problems. The average execution rate for such expenditure in 2013 stood Operating expenditure is the poor cousin in at 20  percent (Bandundu, 25 percent; Maniema, provincial government expenditure. An average 24 percent and Sud Kivu, 19 percent). No significant of less than 5  percent of expenditure of all central improvement was seen from 2011 to 2013. and provincial resources combined goes to operat- ing health facilities and the health system in gen- There is a lack of monitoring and supervision of eral in the provinces in the sample. However, larger provincial expenditure. A simplified expenditure “new” subsidies were awarded to district hospitals chain is being introduced in the provinces, but it was to cover operating costs, especially in Bandundu and not yet operational in most provinces in the sample Maniema, since 2013. These subsidies came to CDF in 2014. Expenditure monitoring is not computer- 242 million in Maniema and CDF 309 million in ized, or only partially computerized, as is the case in Bandundu. Katanga. Monitoring by the Ministries responsible for the budget is not comprehensive, or even nonex- External assistance was apparently used mainly istent in Sud Kivu, at the commitment and/or valida- to finance the operating expenditure of health tion stages. facilities and the associated management bodies Decentralization: Impact on Deployment and Use of Health Resources   57 8 Main Policy Recommendations and Roadmap for Implementation Timeline Short Medium Long Lead Topics Diagnosis Recommendations term term term Authority 1.  Make natural-resources wealth more “healthy” to invest for UHC Fiscal Space Government resources for 1a) Increase the mobilization of domestic X Ministry of the health (USD 1 per capita) are revenues from natural resources and Economy and inadequate to meet needs. improve the system for collecting and the Budget, There are feasible options for administering the existing revenues for MoPH increasing fiscal space for health. a gain of up to 8 percentage points of GDP. X 1b) Revise the priority-setting exercise within the government budget and aim to double the health envelope (to 8 percent of the general budget). X 1c) Use the sectoral MTEF as an advocacy tool rather than a constraint to build up health budget credibility and gain power in budget negotiations. X 1d) Enhance the predictability and traceability of external health assistance through systematic collection of financial information for allocation, execution, and forecasts. 2.  Free up resources for priority interventions for UHC through better effectiveness and efficiency of health spending Effectiveness The amount, nature, and 2a) Define a joint roadmap between the X MoPH, Ministry and distribution of government Ministry of Economy and Finance and of Finance Efficiency health expenditure are MoPH to boost effectiveness of health inequitable, and lack efficiency expenditure and smooth rigidities and and effectiveness. delays. The predictability, effectiveness, equity, and efficiency of 2b) Enhance the expenditure- X government health expenditure commitment procedure through capacity can be improved. building at the MoPH level to avoid loss of budgeted allocations and improve the execution rate (up to 90 percent). 58 2014 HEALTH PUBLIC EXPENDITURE REVIEW Timeline Short Medium Long Lead Topics Diagnosis Recommendations term term term Authority 2c) Conduct a complete overhaul of X allocations within the health sector with the aim of refocusing efforts on priority preventive and primary care as well as priority, cost-effective interventions, with the support of the GFF. 2d) Speed up development of purchasing X agents, ensuring a functioning purchaser/ provider spilt and therefore greater efficiency of health spending. 3.  Revise and enforce public sector employment, deployment and remuneration rules Personnel Government personnel 3a) Step up the census of active X MoPH, Ministry Expenditure expenditure tripled over 2009– personnel (short term). of the Civil 2013 and only covers one-third Service, College of personnel. 3b) Stabilize the pay scale and staffing X of Physicians Full compensation of registered levels through stricter enforcement of personnel would help improve civil service hiring and employment rules motivation and service quality. It (short term). would also stabilize staffing levels in the short term. 3c) Finalize the system of direct bank X deposit of wages and bonuses (short term). 3d) Run a cost simulation of the total X payroll, including salaries and bonuses for all active personnel. 3e) Initiate rethinking of national X health human resources standards and incentives policy in order to curb geographic and skills imbalances. 3f) Regulate quality of training through X accreditation of training institutes. Decentral- Decentralization has not been 4a) Improve execution of revenue X MoPH, ization optimized or harmonized with retrocessions to bring it up to Provincial sector funding and governance. constitutional commitment (40 percent Governments Health resources for coverage of national revenue collected). expansion can be better mobilized and applied at 4b) Clarify the roles and the division of X decentralized levels. labor for implementing health financing reforms between the central, provincial, and deconcentrated entities (in particular for expenditure and personnel management). 4c) Increase the collection of local X revenues by consolidating and simplifying the provincial tax system and increasing revenues from natural resources. Main Policy Recommendations and Roadmap for Implementation   59 2014 HEALTH PUBLIC EXPENDITURE REVIEW Timeline Short Medium Long Lead Topics Diagnosis Recommendations term term term Authority 4d) Develop a budget tool kit to X enhance budget preparation at provincial level and sensitize on health and UHC priorities to align allocations with goals. 4e) Pilot resource-pooling mechanisms X at provincial level as a first step toward better alignment across different sources of funds (domestic, external, and private). 60  Investing in Universal Health Coverage References Clemens, M.A., and G. Pettersson. 2007. «  New Data ———. 2010b. Stratégie de Renforcement du Système de on African Health Professionals Abroad. » Working Santé: deuxième édition, Kinshasa. Paper N°95, Center for Global Development, ———. 2011a. Cadre de suivi et évaluation du PNDS Washington, DC. 2011–2015, Kinshasa. Huillery E., and J. Seban. 2014: Pay for Performance, ———. 2011b. Plan National de Développement des Motivation and Final Output in the Health Sector. Ressources Humaines pour la Santé 2011–2015, Empirical Evidence from the Democratic Republic of Kinshasa. Congo. Working Paper, JPAL. ———. 2013. Annuaire des ressources humaines de la IMF (International Monetary Fund). Democratic Republic santé 2013, Edition 2013, Kinshasa. of Congo. 2013. Poverty Reduction Strategy Paper (PRSP2), Kinshasa. ———. Various years. Rapports annuels d’activités 2011, 2012, and 2013, Kinshasa. ———. 2014. Article IV Consultations, Washington, DC. ———. Forthcoming a: Indice de disponibilité et de Mathonnat, J. 2010 : Disponibilité des ressources finan- capacité opérationnelle des services de santé. Enquête cières pour la sante dans les pays d’Afrique sub- SARA 2014, Kinshasa. saharienne, Notes et documents. Agence Française de Développement, Paris. ———. Forthcoming b  : Etude sur l’équité en santé, Adept, Kinshasa. Ministry of Planning and Macro International. 2008. Enquête Démographique et de Santé, République Ministry responsible for planning and monitoring imple- Démocratique du Congo 2007. (Demographic and mentation of the modernity revolution/Ministry of Health Survey 2007.) Calverton, Maryland, United Public Health. 2014. Enquête Démographique et de States. la Santé 2013–2014: rapport préliminaire, Kinshasa. Ministry of Planning and Reconstruction/UNICEF. 2002. Ministry responsible for planning and monitoring imple- Enquête Nationale sur la Situation des Enfants et des mentation of the modernity revolution/Department Femmes (MICS2/2001): rapport d’analyse. Volume II, of macroeconomic studies. 2013. Rapport de la revue Kinshasa. du DSCRP 2 et du PAG. Kinshasa. Ministry of Planning/National Statistics Institute/ Murru, M. and E. Pavignani. 2013. La prestation des UNICEF. 2010. Enquête par grappes à indicateurs soins de santé en situations de crise: République multiples/MICS-RDC 2010, Kinshasa. Démocratique du Congo, Etude multi-pays de l’Université de Queensland, Queensland. MoPH (Ministry of Public Health  ; Ministère de la Santé Publique). 2001. Normes de la zone de santé, NSI (National Statistics Institute). 2012. Projections de Kinshasa. la population de la RDC et des provinces de 2011 à 2015: document interne de travail, Kinshasa. ———. 2005. Politique et stratégies de financement du secteur de santé, Kinshasa. ———. 2013. Résultats de l’Enquête 1-2-3, République Démocratique du Congo, Kinshasa. ———. 2006. Stratégie de Renforcement du Système de Santé, Kinshasa. Soglohoun, P. 2012. Contribution des Mutuelles de Santé à L’Organisation de la Demande des Services et Soins ———. 2010a. Plan National de Développement Sanitaire/ de Santé en RDC. Bethesda, MD: Health Systems PNDS 2011–2015 (National Health Development 20/20 project, Abt Associates Inc. Plan), secrétariat général de la santé, Kinshasa. 61 2014 HEALTH PUBLIC EXPENDITURE REVIEW State Audit Office (Cour des Comptes). 2013. Audit des ———. 2012. Évaluation de la Gestion des Finances recettes à caractère national des régies financières, Publiques de la République Démocratique du Congo: République Démocratique du Congo, Kinshasa. Rapport PEFA, Kinshasa. Tandon A., and C. Cashin. 2010. Assessing Public Health ———. 2014. Rapport de Suivi de la Situation Économique Expenditure from a Fiscal Space Perspective, HNP et Financière de la République Démocratique du Discussion Paper, World Bank, Washington, DC. Congo, 1ère Édition, Kinshasa. UNDP. 2014. Rapport national sur les OMD en 2013: ———. 2015. Public Expenditure Management and évaluation de la mise en œuvre des Objectifs du Financial Accountability Review (PEMFAR), Millénaire pour le Développement, Kinshasa. Washington, DC. WHO (World Health Organization). 2006. World Health World Bank and University of Kinshasa. Forthcoming: Report: Moving Together for Health. Geneva. Survey of Health Personnel Availability and Compensation in DRC, Provinces of Bandundu, ———. 2014: “Guidance on Conducting a Situation Equateur, Katanga and Sud Kivu (Disponibilité Analysis of Health Financing for Universal Coverage.” et rémunération du personnel de santé en RDC), Draft for circulation. Geneva. Kinshasa. World Bank. 2008. Public Expenditure Review, Washington, DC. 62  Investing in Universal Health Coverage