32818 Africa Region Human Development Working Paper Series Resource Allocation and Purchasing in Africa: What is effective in improving the health of the poor? Tonia Marek Rena Eichler Philip Schnabl Africa Region The World Bank Washington, D.C. Copyright © August 2004 Human Development Sector Africa Region The World Bank The findings, interpretations, and conclusions expressed herein are entirely those of the authors They do not necessarily represent the views of the World Bank Group, its Executive Directors, or the countries that they represent and should not be attributed to them. Cover design by Word Express Typography by Word Design, Inc. Cover photo: Antelope, Burkina Faso ­ Musée de l'Homme, Paris in Afrique Noire by Laure Meyer, Terrail Editions, Paris, 1991 Contact: Tonia Marek Lead Public Health Specialist The World Bank tmarek@worldbank.org ii Contents Acknowledgement . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .v List of Acronyms . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .vi Executive Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1 Ten key findings emerge from this review: . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2 Nine main messages to guide future interventions: . . . . . . . . . . . . . . . . . . . . . . . . . .3 I Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .4 II Current Practices and Trends in Resource Generation . . . . . . . . . . . . . . . . . . . . . . . . . .5 Emergence of untapped opportunities upon closer study of who finances health care .6 Funding health services through tax and social health insurance schemes . . . . . . . . . .7 Charging user fees . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7 Implementing employer-based insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .8 The rise of community-based health insurance (CBHI) . . . . . . . . . . . . . . . . . . . . . . .9 The role of external development assistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9 III Current Practices and Trends in Resource Pooling . . . . . . . . . . . . . . . . . . . . . . . . . . .12 Establishing Social Health Insurance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12 Setting up Medical Aid Societies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .15 Promoting risk pooling schemes for the informal sector . . . . . . . . . . . . . . . . . . . . .15 Building on the Bamako Initiative: A pooling mechanism with local ownership . . . .20 IV Current Practices and Trends in Resource Allocation . . . . . . . . . . . . . . . . . . . . . . . . .24 Utilizing budget mechanisms to ensure better resource allocation . . . . . . . . . . . . . .24 Decentralization: an ongoing trend . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24 Better allocation through primary health care . . . . . . . . . . . . . . . . . . . . . . . . . . . . .27 Policy measures which have considerable public health impact . . . . . . . . . . . . . . . .28 Technicians overlook the importance of local politicians in allocating resources . . . .30 Some Trends to Better use external aid for resource allocation . . . . . . . . . . . . . . . . .30 V Who is Covered? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33 Addressing inequities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .33 Where else can the poor go for services? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .34 iii VI Current Trends and Practices in Purchasing Health Services . . . . . . . . . . . . . . . . . . . .38 Public­public arrangements: Local level purchasing within a decentralized public model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 Public­private arrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .38 Private­private arrangements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .41 Demand-side strategies to enable purchasing among target groups . . . . . . . . . . . . . .41 VII Conclusion: Misconceptions, Constraints, and Future Opportunities . . . . . . . . . . . . . .44 Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .49 iv Acknowledgement Financial support from a CIDA (Canadian International Development Agency) Grant and from a Dutch Trust Fund were used to carry out this study and is gratefully acknowledged. The authors would also like to thank Ok Pannenborg, Oscar Picazo, Alex Preker, and Agnes Soucat for their constructive comments, and Isabelle Dupond for her help in finalizing the document. v List of Acronyms AfDB African Development Bank AIDS Acquired Immuno-Deficiency Syndrome ARI Acute Respiratory Infection CBHI Community-based Health Insurance CFW Cry for the World Foundation CHF Community Health Funds CIMAS Commercial/Industrial Medical Aid Societies ESA Eastern and Southern Africa GDP Gross Domestic Product HIPC Heavily Indebted Poor Country Initiative HIV Human Immuno-Deficiency Virus ITN Insecticide Treated nets KHHS The Kisiizi Hospital Health Society MBB Marginal Budgeting for Bottlenecks MCH Maternal and Child Health MOH Ministry of Health NGO Non-governmental Organization NHA National Health Accounts NHIF National Hospital Insurance Fund (Kenya) OECD Organization for Economic Cooperation and Development ORT Oral Rehydration Therapy PCP Primary Care Physician RAP Resource Allocation and Purchasing SSA Sub-Saharan Africa SWAP Sector Wide Approach Program UN United Nations USAID United States Agency for International Development WDR World Development Report published by the World Bank WHO World Health Organization vi Executive Summary T he primary objective of this paper is to comes. This review is not exhaustive, but gives assess how Resource Allocation and a clear idea of the current situation in Africa, Purchasing (RAP) is being implemented putting forth information and insight into in Africa, and to consolidate some of what approaches have worked and what cur- the lessons that have been learned. This objec- rent trends in this area are today. Further, this tive is accomplished by reviewing trends and paper highlights a number of promising efforts specific cases in Africa where RAP strategies that have been implemented or are currently have increased the effectiveness of a health sys- underway that could provide direction useful tem, in terms of efficiency, equity, quality, own- to decision makers who should take action on ership or a combination of these factors. The these issues. Although these findings are not selection of case studies is solely based on the new -- their organization and presentation is availability of empirical evidence to show that -- and they are worth repeating. the arrangement had improved health out- 1 2 Resource Allocation and Purchasing in Africa Ten key findings emerge from this review: 1) The public sector finances less than half of total expenditure on health. The remaining costs are financed with out-of-pocket funds, that go primarily to purchasing private sector services. The private sector herein comprises formal or informal service providers as well as Non-gov- ernmental Organizations (NGOs). Since money spent on private providers is unpooled, it pro- vides little bargaining power for individual purchasers of commodities, resulting in inefficien- cy and inequity. Those issues must be addressed by governments. 2) The private sector plays a more significant role than the government in most African coun- tries compared to the Organization for Economic Cooperation and Development (OECD) countries. 3) Social health insurance is an existing trend for the formal as well as for the informal sec- tors. Community-based health insurance (CBHI) mutuals are created in a rapidly increasing rate in Africa. The problem of re-insurance for the sake of sustainability needs to be addressed. 4) Some governments depend primarily on donors to cover health costs, and some even rely on donors to cover recurrent costs as well, which is of serious concern. 5) Although there is a need to increase funding for health in Africa, there is also a need to improve the utilization of existing funding as a large portion of available funding is not being disbursed. 6) Large funding allocations earmarked for addressing and treating specific diseases have the potential to skew key health sector priorities if attention is not paid to the long-term strength- ening of health systems in Africa. This is a growing problem as donors have adopted a trend of increasingly financing vertical initiatives, rather than focusing on the health system as a whole. 7) Most governments in Africa have not yet assumed their stewardship function and are still struggling with service delivery issues rather than focusing on outcomes and performance. 8) Donors are part of the existing problem, adding to the fragmentation and the lack of own- ership. 9) Decentralization in the health sector is often undertaken in parallel with others moves toward decentralization in and by the government. To date, the health sector has only achieved minimal decentralization, and health expenditure tracking surveys show that funds have had difficulties reaching the base. 10) Primary health care has permitted great strides in improving health outcomes, but more targeting is needed now to ensure that resources actually reach the poor and to improve health indicators. Executive Summary 3 Nine main messages to guide future interventions: On the policy-side of RAP arrangements, there is need for African governments to: (i) pay much more attention to equity of health services and systems. This can be achieved through risk pooling mechanisms, targeted subsidies to the poor, and marginal budget- ing for bottlenecks, in addition to other targeted interventions. (ii) decrease dependency on donors for selected activities; (iii)develop policies to work with the private sector in order a) to alleviate the burden of financing health services by the poor who already purchase services from the private sector; and b) to harness those resources to achieve public health goals; (iv) implement public health policies that have a large impact and are cost-effective, for example in nutrition, immunizations, smoking. In the same line, an effort must be made to build on successes and go to scale. On the organizational-side of RAP arrangements, the primary messages are: (v) develop a culture of results rather than process, by using performance-based contracts with lower level authorities as well as with the private sector. Governments might inte- grate this approach into the governments' overall decentralization process; (vi) less fragmentation in the financing and delivery of health services must be ensured, wherein the Sector Wide Approach Programs (SWAPS) might be a good start; (vii)provide subsidies to the poor so that they are able to pool resources, increased attention to the potential capacity and contribution of health mutuals is one way to achieve this. On the institutional-side of RAP arrangements, African Ministries of Health need to: (viii) assume a stronger role as stewards, that is provide regulation, supervision, monitoring and competitive arrangements. Develop mechanisms, guidelines, and ensure adequate training to work with the private sector; (ix)separate financing from provision of services in order to help with governance problems which might be associated to RAP arrangements in public bureaucracies. CHAPTER I Introduction S trategic Resource Allocation and Pur- In order to assess whether health systems are chasing (RAP) is a health system func- adequately serving the poor, let's begin by tion and process where pooled examining the flow of funds. Generally, the resources are allocated to health serv- flow of funds undergoes three phases: 1) col- ice providers, and the providers, whether pub- lection, 2) pooling of revenues, and 3) use of lic or private, receive a coherent set of incen- these revenues to compensate public and pri- tives to encourage them to deliver priority vate providers. (Preker et al., 2000). In all health services efficiently. This function African countries these three functions co-exist involves a continuous search for the best ways under different organizational configurations to maximize health system performance by and not all funding is required to pass through deciding which interventions should be pur- all three phases. For example, public health chased, how, and from whom (WHO, 2000). expenditure is typically pooled and then allo- Figure 1 below illustrates how RAP arrange- cated through RAP arrangements, whereas ments complement the stewardship function out-of-pocket payments bypass the pooling of government. phase and go directly to providers. Figure 1: Sharing Decision Rights Between Stewardship Function and RAP Arrangement Decisions Stewardship Function RAP sub-function How much? Defining resource level Collecting and managing money Who to buy for? Defining beneficiaries Identifying and targeting patients What to buy? Defining strategic benefit coverage Deciding which services to purchase from providers Who to buy from? Setting rules of the game Selecting providers How to pay? Choosing payment mechanisms Source: Preker et al., 2002 4 CHAPTER II Current Practices and Trends in Resource Generation I n a worldwide comparison based on an but is less than the 6.2% of upper-middle index of performance which reports income countries (World Bank, 2002a). Out of how efficiently health systems translate 44 countries with data, only 10 spent more expenditures into health as measured than 5 percent of GDP on health services by disability-adjusted life expectancy (DALE), (WHO, 2000). Several countries have been The World Health Organization (WHO, 2000) able to significantly increase their public health ranks most African health systems within the spending per capita in the last decade. For bottom quarter. There are several possible example between 1990 and 1998, Senegal explanations for this state of affairs; one could increased it by 19 percent, Ethiopia by 10 per- be the fact that most African countries do not cent, Burkina Faso by 6.4 percent. On the have adequate resources to establish any func- other hand, a number of countries saw a tioning health systems. For example, Sub-Saha- decrease in those percentages, the worst being ran African (SSA) countries have total per capi- Burundi with ­11 percent, Gambia and Tanza- ta health expenditures ranging from US$ 4 in nia followed with ­3 percent. (Wagstaff, 2002) Ethiopia to US$ 230 in South Africa, with a A report by the WHO Commission on median situated at US$ 18, an average at US$ Macroeconomics and Health (2002) therefore 30 if South Africa is not included, and US$ 40 concludes that significant increases in donor if it is included (World Bank, 2002). In 1997, funding are needed to provide basic health in absolute per capita terms South Africa, services in Africa. (WHO, 2002a) However, Namibia, Gabon, Botswana and Mauritius care is required when pretending to increase spent the largest amounts on health care (more expenditures at the country level on the basis than US$ 100 per capita per year) (WHO, of donor financing. Increased financing may 2000). All the other countries spend less, with not translate into larger expenditures at the 19 countries spending less than US$ 25 per country level as will be seen later in this paper. capita. (McLaughlin, 2004). Moreover, increases in expenditures may not Total health expenditure in SSA, as a per- be efficient or directed to those determinants centage of Gross Domestic Product (GDP), is that have the largest impact or outcome. (Got- on average 4.9 percent, which compares to the tret & Preker, 2003) Other possible explana- 5 percent of all lower-middle income countries, tions and solutions can be found in this paper. 5 6 Resource Allocation and Purchasing in Africa The following is a review of existing prac- pay for service fees and for mostly unregulated tices and issues with accompanying suggestions over the counter medicines (WHO, World that provide guidance in addressing these limi- Health Report 2000) Since those out-of-pock- tations. et expenditures are unpooled, they tend to be inefficiently spent and highly inequitable (ESA NHA, 2000). Emergence of untapped opportunities In Guinea, total health expenditures were upon closer study of who finances US$ 34 per capita in 2001, of which only US$ health care 3 came from the public sector (including donors), and the rest was provided by mainly Analyzing public expenditures gives only half out-of-pocket. (Schwabe et al., 2003) the picture since at least one-half of all health Residents of the more industrialized OECD expenditure is financed by the private sector. countries spend relatively much less out-of- This figure is 58 percent for the whole SSA pocket on health, 16 percent versus 36 percent according to the 2003 World Development in Eastern and Southern African countries, as Indicators (World Bank, 2003) and 43 percent private insurance companies' contribution is according to data from 9 National Health higher in industrialized countries. Since in Accounts (NHA) in Eastern and Southern Africa, most of out-of-pocket for health is Africa (ESA) (see Graph 1 below). Graph 1 spent buying services from the private sector, also shows that 84 percent of private expendi- one can say that the private sector plays a tures come from out-of-pocket in those nine more significant role than the government in countries of ESA. Out-of-pocket spending rep- less industrialized countries than in OECD resents more than 65 percent of total health countries. expenditures in seven other countries of West- Although health is still underfunded by gov- ern Africa: Burkina (69 percent), Mauritania ernments in Africa, the potential for generating (70 percent), Nigeria (72 percent), Sudan (79 large resources is there, if out-of-pocket expen- percent), Cameroon (80 percent), Sierra Leone ditures could be channeled to more efficient and Democratic Republic of Congo (90 per- and equitable health services. This could be cent). Most of the out-of-pocket expenditures done in part by working with the service Graph 1: Sources of Total Health Expenditures in 24 OECD and 10 ESA Countries 100% 90% 27% 80% Donors 70% 59% 60% Public 30% 50% Other Private 40% 7% 30% 25% Private (Out 20% of Pocket) 36% 10% 16% 0% OECD AFRICA Sources: for OECD: WHO, 2002c Annex 5; World Bank 2002a; ESA NHA 2000. Current Practices and Trends in Resource Generation 7 providers used by the population, namely the recurrent expenditures (Shaw and Griffin, private sector. 1995). Experiences from Ghana (Nyonator and Kutzin, 2000), Uganda (Konde-Lule and Okel- lo, 1998), Mauritania (Audibert and Mathon- Funding health services through tax nat, 2000) and Zambia (Bossert et al., 2000) and social health insurance schemes have shown that user fees have a positive impact on drug availability and service quality. Tax-funded health expenditure is around 14 A study in Nigeria showed that although percent of GDP for 16 African countries with Bamako Initiative facilities had a better avail- available data (World Bank, 2003). In most ability of essential drugs than non-Bamako Ini- African countries, adequate revenue collection tiative centers (35 drugs available versus 15), through public taxes has proven difficult the Bamako Initiative gave rise to higher rates because a large proportion of the population of drug prescribing, which, in turn, calls for works in the informal sector and tax collection promoting the rational use of drugs as part of mechanisms are weak. Since most public strengthening the Bamako Initiative scheme health expenditure is tax-funded, overall pub- (Uzochukwu et al. 2002). Thus, cost recovery lic spending on health services is very low. The has, in many cases, improved access to and the prospects for increased public funding in the quality of health care. This has benefited the medium term are thus slim. population at large, including the poor. Several African countries complement tax However, since the introduction of cost revenues by mandating participation by govern- recovery schemes, there has been a debate as to ment workers and/or formal sector workers in their impact on very poor people. For the social health insurance. In 1993, a World Bank poorest of the poor even small fees are too high Survey identified 14 African countries which to afford. Exemption schemes designed to pro- have some form of social health insurance. The tect poor people from user fees have usually population covered ranged between 0.1 percent failed to do so. In Ghana in 1999, only one out (Ethiopia) and 25 percent (Kenya) (Shaw and of 1,000 patients was granted an exemption Griffin 1995). However, WHO estimates that (Nyonator and Kutzin, 2000) while 45% of Kenya was the only country where social securi- the population lives with less than US$ 1 per ty contributions exceeded one percent of public day (World Bank, WDR 2003, p. 58). In rural health expenditure. (WHO, 2000) Social insur- Ethiopia, 52 percent of all patients were ance has not been a major contributor to the exempted from paying user fees but a survey generation of resources in SSA. Details of social found no relation between exemption status health insurance and lessons for SSA are pre- and income (Engida and Mariam, 2000). sented in section IV that covers risk pooling, These results render it difficult to estimate the drawing on lessons from the Kenya National real impact of cost recovery on poor people. Hospital Insurance Fund (NHIF). Essentially, we do not know the counterfactual -- How accessible would health services be for poor people in the absence of cost recovery Charging user fees schemes? While there might be no fees for poor people, at the same time, there might not be It is generally believed that cost-recovery any services either. schemes administered locally (i) ensure a steady For example, Mauritania's public health cash flow at the local level; (ii) promote more services charge very low user fees. However, efficient drug use; and (iii) provide the commu- there are significant shortages in essential nity with more flexibility in financing necessary drugs and equipment. A study showed that the 8 Resource Allocation and Purchasing in Africa availability of syringes in primary care infra- change and the introduction of new measures structures was between 1.4 percent and 74.2 to improve services. More research is needed in percent with the majority of regions being at this area. the lower end (Hahmed and Soucat, 2004). Another source (Audibert & Mathonnat, 2000) suggests that the introduction of user Implementing employer-based fees in Mauritania, with its accompanying insurance strategies (better running of facilities) might have positive effects on the effective ability of According to NHAs, in 1998 five African the poor to access health care. Another countries (Botswana, Cote d'Ivoire, Namibia, instance is found in the Malawi context -- South Africa, and Zimbabwe) had forms of Malawi's public sector does not charge for private insurance that accounted for more than services, while the NGO sector, which provides 5 percent of total private health expenditure about 35% of health services does charge for (WHO, 2000). services and 32 % of the NGOs' income comes The term "employer-based" insurance is a from user fees (Picazo, 2002). Uganda abol- scheme that is financed by payroll-deductions ished user fees in March 2001, and recent evi- and sometimes by employer contributions -- dence suggests that this policy change and is exclusive to the formal sector. The increased utilization of health services by the resources mobilized through employer-based different income groups. At the same time the insurance and thus the contributions to gener- Government improved health sector funding ating resources for the health systems are rela- with increased amounts for pharmaceuticals tively small (Shaw and Griffin, 1995). An and health workers' emoluments (Tashobya et exception to this is Zimbabwe, where medical al., 2003), so it is difficult to know if the aid societies provide private insurance to peo- increased utilization should be attributed to ple employed in the public and the private sec- the abolition of user fees or to the increased tor that represents roughly 8 percent of the funding. total population. Medical aid societies are non- In response to these problems, the Govern- profit organizations that collect premiums ment of Ghana, for example, intends to phase from businesses and government and use these out the system of requiring front payments for resources to pay health care providers for serv- drugs and supplies (the "cash and carry" sys- ices provided to beneficiaries. Details of med- tem) and it will improve and extend the ical aid societies and lessons for SSA are pre- exemption scheme while also further develop- sented in Section IV which covers risk pooling. ing health insurance schemes (World Bank African countries planning to promote Ghana PAD 2002). This policy stems from the employer-based health insurance must be care- widely growing consensus that out-of-pocket ful to assess the equity impact of such schemes. payments reduce equity since they impose a Private insurance can mobilize additional burden on those least able to pay (Chawla et resources for health needs and has the poten- al., 1996). tial to improve health access by stimulating the Thus, there seems to be evidence that any health industry. However, in countries with change in the existing user fees policies (either severe human resources shortages, such adding user fees or getting rid of them), brings schemes might actually pull scarce clinicians about some positive changes in terms of uti- away from serving the poor in favor of serving lization of health services. This is probably due the insured population. At least in the short to several factors, among which the most run, this would have negative equity implica- prominent are a political commitment to tions for the system. Current Practices and Trends in Resource Generation 9 Table 1: Informal Sector Employment as a Percentage tain. Moreover, these schemes have to deal of Employment with adverse selection, moral hazard, small Informal sector employment risk pools, and the need for additional Country (year) as % of employment resources to protect against catastrophic risk. Uganda (1993) 84 Given these challenges and the important role Zambia (1993) 81 that such schemes might play, more opera- Ghana (1997) 79 tional research in Africa is needed to under- Gambia (1993) 72 stand the design features that best achieve Mali (1996) 71 Tanzania (1995) 67 social priorities including the equity impact of Kenya (1995) 58 CBHIs. Madagascar (1995) 58 Cote d'Ivoire (1996) 53 Benin (1992) 48 The role of external development Mauritius (1992) 24 assistance Botswana (1996) 19 South Africa (1995) 17 External aid Source: ILO, 2002a Although the real value of development assis- Considering that many people in Africa are tance to Africa has declined since the employed in the informal sector (see Table 1), mid­1990s, the importance of external aid for the adoption of employer-based insurance the health sector is remarkable. About US$ 1.2 needs to be complemented with other means billion or 10 percent of total health expendi- and schemes to reach informal workers. ture, in SSA is externally financed. Thus, this region receives the largest proportion (38.5 percent) and the highest per capita share (US$ The rise of community-based health 2.45) of global development assistance for insurance (CBHI) health (McLaughlin, 2000). Donor assistance for health care varies considerably from coun- Community-Based Health Insurance provides try to country. In 1998, in 23 African countries financial protection through local level risk more than 20 percent of public health expen- pooling to the informal sector. Households diture was financed by donors. In six countries contribute premiums to a risk pool that pro- foreign aid provides more than 40 percent of vides coverage for a package of benefits that total health expenditure (WHO, 2000). can range from primarily ambulatory care to a For example, 84 percent of Chad's public comprehensive package that includes hospital- expenditure on health in 2000 was financed by ization. In Western and Central Africa the external grants and loans, 12 percent by the number of CBHIs has risen from 67 to approx- Government's own resources, and 4 percent by imately 827 from 1997 to 2000 (Abt Associ- communities (Hahmed and Soucat, 2004). In ates, 2000a). Details of CBHI schemes and les- Mauritania in 1999, donors contributed 53 sons for SSA are presented in Section III that percent to the overall health budget (Soucat, covers risk pooling. 2004). In ESA, donors on average contributed The impact of the aforementioned schemes 27 percent of total health expenditures. on equity is not yet fully understood. CBHIs In some countries, even the recurrent budg- have the potential to improve welfare and et is mainly covered by external money. This health access for the poor. However, the extent was the case in Chad in 2000 where 81 percent to which they have done so in the past is uncer- of recurrent expenditures were funded by 10 Resource Allocation and Purchasing in Africa external sources and only 19 percent by the bursement of available funds. For example, in Government's own funds (Hahmed and Sou- the period from 1994 to 1999, the World Bank cat, 2004). In Guinea in 2001, external financ- approved 27 International Development Asso- ing paid for 32 percent of recurrent costs, ciation (IDA) loans in Africa with a total com- which included 93 percent of vaccines, 91 per- mitment of US$ 751 million. However, by June cent of training, and 84 percent of vehicles 2000 only 53 percent had been actually dis- maintenance paid by the outside (Schwabe et bursed. The undisbursed balance amounted to al., 2003). In Mozambique in 1997, 47% of US$ 353 million. (World Bank, 2002d) Other recurrent costs were financed by donors, and donors like the United Nations (UN), the donors also covered 92% of drugs and medical African Development Bank (AfDB) and bilat- equipment (S. Chao and K. Kostermans, eral donors experience slow disbursement as 2002). well. In Malawi, between 1994 and 1998 only The World Bank and other multilateral 61 percent of the donors' health allocation was donors have supported health expenditures disbursed (Picazo, 2002). with the Debt Initiative for Heavily Indebted African countries experience implementa- Poor Countries (HIPC). The HIPC initiative tion constraints which limit the ability to uti- was launched in 1996 with the goal of reduc- lize funds effectively. Among those, severe ing unsustainable levels of debt. As of July shortages of human resources represent a 2003, 27 of the 38 countries that potentially major challenge to health systems in Africa. qualify for assistance under HIPC had used Another reason for slow disbursement is the this initiative which resulted in the freeing of complexity involved in coordinating different about US$1.0 billion in annual debt-service donors' priorities with local health strategies. savings. As a result of the HIPC debt relief, Administrating and coordinating donors con- debt stocks for those 27 countries are expected sumes scarce local government administrative to decline by two-thirds in net present value resources. A recent study in Rwanda lists 27 terms. Part of the debt relief initiative is a international donors contributing to the health poverty reduction strategy, which includes pro- budget, each donating amounts ranging from grams to improve health care services for poor US$ 100,000 to US$ 4 million (Schneider et al., people. In the 27 countries, poverty-reducing 2000). Another problem with donor assistance expenditures, including those for health, have is that it is often not tracked in a consistent increased from about US$6.1 billion in 1999 to manner, and is often maintained "off budget", US$8.4 billion in 2002, and are projected to which makes proper planning and manage- increase to US$11.9 billion in 2005 (IMF, Sept. ment by the government more difficult. 12, 2003. HIPC Initative: Status of Implemen- tation). Multiplicity of initiatives The high dependency on donor funding in a number of countries is a serious concern. Any An additional challenge facing SSA is the pres- country's health system should aim at having sure placed on African health systems by the enough reliable funding within the country's spread of the HIV/AIDS epidemic. It is esti- own resources to maintain basic health services mated that in SSA alone, more than 55 million for a growing population (ESA NHA, 2000). people will die from AIDS within the next twenty years. In Botswana, life expectancy has The paradigm of undisbursed aid dropped below 40 years, a level not seen since 1950 (Joint UN Program on HIV/AIDS, 2002). Despite the growing need for additional Therefore, several initiatives to increase resources, there is the problem of non-dis- resources to fund programs that address the Current Practices and Trends in Resource Generation 11 challenges of prevention, and care of the infect- Foreign aid and ownership ed and affected people have been undertaken to react to this new threat in Africa. The most Another important constraint faced by coun- prominent initiative is The Global Fund to tries in the context of external development Fight AIDS, Malaria and Tuberculosis which assistance is the need for ownership by local has collected US$ 2.1 billion from industrial- governments in order to ensure a program's ized countries prior to May 2002. The World success. Recent evidence suggests that external Bank also launched the Multi-Sectoral AIDS aid does not help to improve health indicators Program (MAP) which up until May 2003 had if a government is not committed to the pro- funded 21 HIV/AIDS Projects with a net com- grams. Even the conditionality of loans or mitment of US$ 722 million. (World Bank, grants will have little effect on a country's 2002b) Other recent initiatives also aim to health outcomes if the local government does increase funds for HIV/AIDS as well as for not fully support the project(s) (Devarajan et other health conditions. While a strong com- al., 2001). mitment from the international community for Most African governments have not yet funding programs to address, treat and prevent assumed their stewardship function and are HIV/AIDS and other health problems are need- still struggling with delivering services them- ed, limited attention has been given to ensuring selves, rather than embarking on strategic that local governments have the adequate reforms of the existing service delivery system. capacity to effectively manage these new funds. Africa could benefit from Latin America's les- Large funding allocations targeted specifically sons on health reform strategies, where it was for specific diseases have the potential to skew found that a major factor in the success of priorities if attention is not paid to the long- reforms was that a relatively stable and coher- term strengthening of African health systems. ent "change team" was formed, supported by Some of the problems that plague govern- the President and other major political actors, ments at the central level also find their way to and included members drawn from the Min- local levels of government. This is the case in istries of Planning and of Finance. This change Tanzania, where the health office in Dar Es team was isolated from the broader political Salaam, the country's capital city, has to man- process until it had developed a significant, age 12 different financing sources (Dar Es technically defined package of reforms (HSPH, Salaam City Medical Office of Health, 2003), 2000). thus stretching its management capacity. CHAPTER III Current Practices and Trends in Resource Pooling Establishing Social Health Insurance that a large portion of the population, in both the formal and informal sectors, is receiving T hrough social health insurance some financial protection from high cost health schemes, governments can provide events. The NHIF has increased revenue for financial protection to households by both private and public providers raising the aggregating contributions and pooling total amount of resources available for health. risk. Social insurance agencies have the poten- This availability of funds seems to have also tial to drive improvements in the performance stimulated the development of the private sec- of the delivery system by active and strategic tor which has enabled consumers to have choic- purchasing. There are few examples of govern- es (See Box 2 below). Weaknesses of the NHIF ment-managed social insurance in SSA. Tanza- include its limited capacity that extends to most nia is in the early stages of implementation of aspects of insurance design, implementation, a program for civil servants, and Mozambique management and ongoing monitoring. More- has begun withholding wage contributions for over, poor design features, such as perverse a future social insurance program (Conversa- incentives in the reimbursement system and tion with Daniel Kraushaar, 2002). Other weak monitoring of fraud and abuse, have led countries in SSA have considered introducing to some unintended negative consequences. social health insurance for civil servants and Individuals with the technical skills needed to employees of formal sector companies (Ghana, effectively run health insurance schemes are in Nigeria, Uganda, South Africa, Zimbabwe) but short supply in Africa. An additional factor that little progress has been made. Only Kenya has must be well thought out when considering a long history of experience with social insur- social insurance is the impact of the AIDS epi- ance through its NHIF. demic which is a catastrophic risk that can ren- When evaluating the performance of the der broke nascent insurance schemes in popula- NHIF in Kenya and aiming to determine tions with high prevalence of the disease. whether this is a model that can be adapted to If other African countries plan to implement other countries in the region, it is important to or expand similar schemes, they would have to look at both the system's successes, failures and analyze the supply side of health services to problems. The biggest strength of the system is adequately assess its impact on equity. Social 12 Current Practices and Trends in Resource Pooling 13 Box 1: The Kenya National Hospital Insurance Fund (NHIF): a model for Africa? The Kenya National Hospital Insurance Fund (NHIF) was established in 1967 to finance improved access to private hospitals, nursing and maternity homes to middle and upper class Kenyans (Kraushaar 1997). By law, all Kenyans earning over the equivalent of US$ 19 per month must contribute to the NHIF. In the early phases of the program, this included prima- rily formal sector workers and civil servants. In recent years, because of inflation, informal sec- tor workers and agricultural workers have also been included. Those formally employed con- tribute through payroll deductions and this is considered "standard". Others can voluntarily enroll for a flat amount per month. Membership has grown considerably from roughly 60,000 in 1967 to roughly 1.4 million members in 1985/1986. Assuming an average of five depend- ents per member, this translates to coverage of roughly seven million people. NHIF members and dependents receive coverage for inpatient care and the cost of drugs at certified inpatient facilities but do not receive coverage for hotel services. Outpatient and pre- ventive services are not included in the benefits package. Only "hospitals", which includes government, mission and private hospitals, nursing homes and maternity homes, are eligible to receive reimbursement. The NHIF pays a fixed per diem fee. Substantial balance billing occurs in private higher cost hospitals while the NHIF fee is accepted to fully cover costs in other hospitals. In 1992, the law on contributions was changed from requiring a fixed sum contribution to requiring one equal to 2 percent of income for standard members. Voluntary members, or those in the informal sector, continue to pay the same fixed sum of roughly US$ 1.14 per month as was stipulated in the original 1967 law. Contrary to expectations, this approach to financing does not seem to provide subsidies from wealthy to poor, from healthy to sick, or from smaller to larger families. Table 2 below shows findings that cast considerable doubt as to the equity enhancing effect of the NHIF. In 1990/1991 low-income individuals contributed 18 percent of resources to the fund but the facilities they frequented only received 3 percent of total paid claims. The reason for this is that individuals with low-income are unable to pay the large additional fees they would need to pay out-of-pocket to seek care from more expen- sive private facilities (balance billing). In effect, low-income individuals are subsidizing the middle and high-income members of the NHIF. Table 2: NHIF Contributions Received and Claims Paid by Member Income Category FY 1989/1990 FY 1990/1991 Contributions Claims paid Contributions Claims Member income rec'd (%) (%) rec'd. (%) paid (%) Low 23 1.4 18 3 Middle and High 27 98.6 32 97 Source: Akumu, 1992. (continued on next page) 14 Resource Allocation and Purchasing in Africa Box 1 (continued) Assessment of both the strengths and considerable weaknesses of the NHIF system gener- ates lessons for RAP arrangements for the rest of the region. By providing insurance coverage for hospitalization, NHIF provides a degree of financial protection for households. Increasing voluntary enrollment of the informal sector indicates that households are willing to pay for this type of risk coverage. The NHIF, as a predominant purchaser of health services in Kenya, exerts significant influ- ence on both government and private provision of hospital services. Allowing multiple com- peting providers to receive reimbursement has stimulated an expansion in the number of providers who can service beneficiaries leading to increased access in Kenya's urban areas. The purchasing power of the NHIF has the potential to be used strategically to influence appro- priate growth of the private sector. A major weakness of the NHIF is that poorly developed compliance and monitoring systems have enabled fraud and abuse by both providers and the covered population. In some cases, providers have been known to overcharge clients and members are known to share NHIF mem- bership cards with non-members. Weak capacity to manage these insurance systems is a con- siderable impediment to effective implementation and ongoing management of the NHIF. The per diem reimbursement system provides incentives to providers to increase lengths of stay while basing reimbursement rates on facility characteristics such as the number of beds does not encourage quality care. The NHIF should take advantage of its considerable purchasing power to drive improvements in efficiency and quality by implementing stronger systems to monitor fraud and abuse, to control utilization, and to reward quality of care. Box 2: The link between health insurance and the existence of private for-profit providers Evidence seems to indicate that health insurance encourages the development of private providers. For example, in Kenya the private sector has grown from a few private providers at independence in 1963 to about 1,500 providers in 1993. Currently, the private sector accounts for 50 percent of all hospitals and provides 36 percent of all hospital beds. Approximately 21 percent of all health centres and 51 percent of all other outpatient treatment facilities are pri- vately owned. New facilities offering routine hospital services are mostly privately owned and for-profit. This rapid development of private, for-profit hospitals is unique in SSA (Shaw & Griffin, 1995). The most plausible explanation for these developments is the existence of national health insurance, supplemented by private insurance schemes. Kenya established an NHIF in 1967 which covers more than 25 percent of the population. Contrary to volatile public subsidies, payments by public and private insurance schemes provide a stable source of income for pri- vate providers. Although this evidence is not conclusive it provides support for the argument that health insurance encourages the development of private providers. Current Practices and Trends in Resource Pooling 15 health insurance provides financial incentives stimulate the development of medical aid soci- to providers to serve insured individuals, who eties through tax incentives. This is the case in are typically wealthier than uninsured individ- Zimbabwe (see Box 3). They function as a uals. If a government wants to mobilize more form of social insurance in that premium con- resources through a social insurance scheme tributions are income based or flat fees and not and the sector has an upward sloping supply based on the risk characteristics of individuals curve for medical services you would expect an or households. Premium contributions from increase in the size of the health industry and members of a given plan are pooled to provide improved health access as a consequence of financial coverage for the benefits included in such improved financial incentives for the package. Medical aid societies also have providers. This was the case in Kenya and is a the potential to stimulate innovation in the possible scenario for other African countries. insurance market by encouraging competition However, most African countries do not among competing funds. have an elastic supply for medical services, at As implemented in Zimbabwe, medical aid least not in the short-run. A number of SSA societies have not explicitly targeted the poor. countries continuously suffer from chronic The concept of medical aid societies, however, shortages of medical personnel. In 2000, a is similar to some of the CBHI schemes that are World Bank study reported that the shortage of being introduced for the informal sector in human resources is the single most important SSA. Lessons from the development of risk problem in African health care (McLaughlin, pooling schemes in the region include that it is 2000). Under these circumstances, the intro- important to implement sound systems to duction of public insurance schemes would monitor fraud and abuse and that well shift resources towards the insured rather than designed provider payment mechanisms can increasing overall access to the health system. provide powerful incentives to both improve In this case, the introduction or expansion of quality and overall costs. social insurance would have a negative impact on equity, at least in the short-run. Promoting risk pooling schemes for the informal sector Setting up Medical Aid Societies Poor people not only face financial barriers to As previously mentioned, medical aid societies access proper health care but they are also very are generally private non-profit entities organ- vulnerable to high cost health events which ized to provide health insurance, or prepay- may force them into further poverty. This ment schemes, to a particular industry or pop- requires solutions to protect such households ulation (Quigley 1997). In Africa, medical aid against financial risk. In SSA the majority of societies primarily collect and pool resources public systems have failed to provide financial and provide financial protection to formal sec- protection for the poor. The previous sections tor working populations. Usually they are gov- described schemes that pool resources and pro- erned by a Board of Directors comprised of tect primarily formal sector households against member firms and sometimes with representa- financial risk. This section will present three tives of the covered population. When mem- prepayment schemes that pool resources and bership is employment based, members con- provide some degree of financial protection for tribute through wage deductions that are the informal sector. sometimes supplemented by employer contri- It should be noted that traditional solidarity butions. Some governments have chosen to schemes have existed for a long time in most 16 Resource Allocation and Purchasing in Africa Box 3: Medical aid societies in Zimbabwe In Zimbabwe, health insurance is provided to 7 percent of the population, or approximately 800,000 people, by 25 medical aid societies (Quigley 1997). The market is dominated by Com- mercial/Industrial Medical Aid Societies (CIMAS), organized to serve the commercial and indus- trial sector and by the public service sector, organized to serve government workers. Other med- ical aid societies focus on small groups, railway workers, or specific geographic areas. In the past, there was little competition among medical aid societies but this is changing as employers search for lower cost solutions to providing financial coverage for their employees. While the majority of enrollees are from the formally employed sector, voluntary enrollment is also possible. Generally, participating employers choose one medical aid society to be its health insurance carrier and offer different insurance packages to employees based on their income. Employers, as such, pay a significant share of the contribution. Packages vary from the most basic services that entitle beneficiaries to coverage for services in the public sector to the most extensive services that provide access to the full range of care including tertiary care through the private sector. Additionally, members have free choice of providers covered within their package and providers are paid fee-for-service by CIMAS and bill patients for the uncov- ered balance. Medical aid societies have the potential to use their purchasing power to imple- ment cost-saving strategies such as negotiated low rates and payment mechanisms that transfer financial risk to providers but CIMAS and the other medical aid societies in Zimbabwe have not done so. They are essentially passive payers. Medical aid societies are facing many challenges. Currency devaluations have increased the cost of imported goods, most importantly from pharmaceuticals. The HIV/AIDS epidemic has caused an increase in demand for services for opportunistic infections, though antiretroviral ther- apy is not covered. Providers are mounting pressure for increased fees and there is a significant level of fraud and abuse by both providers (false billing and over-utilization) and patients (shar- ing membership cards with non-members). Recent estimates are that fraud and abuse amount to up to 33 percent of all claims expenses. The result of these factors altogether is that insurance premiums have been rising faster than inflation and employers are increasingly concerned that they will no longer be able to afford to provide coverage for their employees. To control costs associated with fraud and excessive use, CIMAS management considered a range of strategies employed by managed care organizations in the U.S. (Campbell et al., 2000) that focused on: the role of the primary care provider; incentives for providers and patients; composition of the enrolled population; selective contracting with a network of providers; active care management strategies; communications and education; and continuous measure- ment and improvement. CIMAS management decided to introduce a pilot project named Health Guard that focused on the relationship between primary care providers (PCP) and patients. In this pilot, the PCP is responsible for coordinating medical care (including specialty and inpatient care) for CIMAS members in his/her panel. CIMAS was able to select the PCPs that are part of the network. All care reimbursed by CIMAS had to be based on a referral from the PCP. To participate in the pilot study, patients receive extended drug benefits and PCPs receive additional payment to compensate for additional responsibilities. CIMAS management did not choose to change payment to PCPs from fee-for-service to capitation (as is the practice in many models of managed care). Instead, they chose to introduce gradual reforms that includ- ed extensive monitoring of utilization and referral patterns. Results are not yet available. Current Practices and Trends in Resource Pooling 17 African countries. They are prepayment case of illness or disability. These results seem schemes which are often used to help during a to indicate that CBHIs increase health access catastrophic event, such as burial. (higher utilization) and welfare (risk-sharing) In a government-stimulated community for its members. However, the study also con- health fund (CHF) model, the government, cludes that CBHIs fail to protect the poorest of often with donor resources, provides matching the poor because they cannot afford the premi- funds to community-managed prepayment ums. Furthermore, CBHIs typically suffer from schemes that offer access to a limited package a relatively small risk pool (500 to 2,000 mem- of ambulatory care services. This was the case bers) and lack funds to cover catastrophic risk in Tanzania. In the provider-initiated model, (Jutting, 2001). health providers offer prepayment schemes to The experience of CHFs in Tanzania (see households in the communities they serve that Box 4) contain lessons for the design and provide access to a wide range of services, as implementation of similar schemes for the occurred in Uganda. In the CBHI model, informal sector in SSA. From household sur- households prepay into a community-managed veys in Kilosa, it appears that the population is fund for coverage for a defined package of willing to prepay for participation in a pool for services delivered by a range of local providers. coverage of a range of services if the services This model was utilized in Rwanda. they will have access to are perceived to be of For most participants in such schemes it is adequate quality (Kihombo, 2002). Suggestions the first time that they had access to health to improve the functioning of the CHF which insurance. Such risk-sharing agreements repre- have the potential of making similar approach- sent a significant welfare gain for the insured es more attractive to households include: population, since individuals who take out vol- expanding coverage to include higher cost serv- untary insurance prefer to pay monthly premi- ices such as hospitalizations; allowing choice of ums rather than being exposed to the risk of a provider, and introducing incentives to improve major health expenditure. It is, however, service quality as perceived by patients. unclear whether these schemes improve overall It is obvious that CHFs include only those health access for poor people. A survey of 258 people who can pay the premium, not the CBHIs in developing countries, including 131 poorest people. To reach the poorest, govern- CBHIs in Africa, shows that little is known ment subsidies would be necessary. In Tanza- about the health impact of CBHIs. The study nia, the Government's matching grant program concludes that there is no evidence that CBHIs helps reach those that can still pay, not the positively impact health status and financial poorest. protection, particularly of poor people (ILO, CHFs are also interesting in that they create 2002). a local collective power, and it has purchasing However, another study, undertaken in capacity. rural Senegal arrives at a more favourable con- The design and implementation of provider clusion. The study analyzes four CBHIs which based prepayment schemes in Uganda (see Box have been operating between three to ten 5) provides lessons for the rest of SSA. These years. Using household survey data, the study schemes provide coverage for both outpatient shows that insured members have higher uti- and inpatient services and provide financial lization rates than non-members. Furthermore, protection against catastrophic costs for those members report "less personal worries" as a households that can afford to and choose to result of having health insurance. Without enroll. The restriction that only groups can CBHIs they would have been forced to rely on enroll controls adverse selection and co-pay- their social network or money lenders in the ments mitigate moral hazard. Schemes begin 18 Resource Allocation and Purchasing in Africa Box 4: Community Health Funds in Tanzania: a government stimulated health insurance scheme that covers primary health care for the informal sector As a response to growing concern about the financial burden user fees impose on poor house- holds and the potential barriers to access care they create, the Government of Tanzania began implementing CHF in 1996. By December 2001 CHFs were established in 20 districts (Kihom- bo 2002). In exchange for an annual prepayment fee (equivalent to US$ 5 per year), house- holds who choose to enroll in a CHF are entitled to access ambulatory care services in partic- ipating public facilities without paying customary user fees. The Government provides matching funds equivalent to the amount the CHF collects from households as an incentive to community members to enroll in the program and as a mechanism to increase financial resources for health services in public facilities. CHFs are managed by District Boards that include representatives from communities, district councils, and the MOH. Stakeholder man- agement and oversight of CHFs was supposed to improve responsiveness of health providers to households by increasing accountability. Overall goals of the CHF are to improve health outcomes through better access to quality services, improved financial protection through pooled prepayment of premiums, and improved responsiveness of providers to clients through changed incentives (Shaw, 2002). The initial design phase of the CHF envisioned paying providers based on a capitation pay- ment and enabling both public and private providers to participate. Because of concern that this approach may jeopardize funding for public hospitals making it impossible to cover salaries of civil servants, the CHF chose a different model to pay providers. Instead, CHF funds were used to improve the physical condition of facilities and to improve availability of equip- ment, drugs, and medical supplies. Shaw indicates that these improvements enhanced worker motivation and therefore contributed to improved quality of services. More recently, CHF funds have been used to provide bonuses to staff (Shaw, 2002). When initially designed, CHFs were expected to attract approximately 30 percent of house- holds with hopes that eventually up to 70 percent would enroll (Shaw 2002). Enrollment rates have fallen far short of these expectations and many households that initially enrolled have dropped out (see Table 3 below). Table 3: Enrollment and Dropout Rates in the CHF in Kilosa District by August 2001 Total # households Enrollment rate 1999-2000 Dropout rate 56,519 4.3 77.7 Source: Kihombo, 2002. with little knowledge of the expected utiliza- the prepayment schemes to the general popula- tion of their covered population making it dif- tion. In addition, it clearly shows that poor ficult to price premiums to cover actual costs in people are familiar with the way insurance the first years of implementation. There is a schemes operate, appreciate their value, and need for better marketing of the advantages of are willing to pay for financial protection Current Practices and Trends in Resource Pooling 19 Box 4 (continued) Researchers (Shaw, 2002; Kihombo, 2002) have hypothesized that there are a number of reasons households have been slow to enroll and have decided to drop out after partici- pating. Since CHFs do not cover inpatient services, they are not providing households with financial protection for catastrophic health events when people need to be hospitalized. Since user fees for ambulatory care services in public facilities are relatively low, CHF insurance coverage does not necessarily provide expected benefits that exceed the cost of joining. A recent survey of households in Kilosa district shows insignificant differences in satisfaction with perceived quality of providers between non-joiners, drop-outs, and those currently enrolled. However, households enrolled in the scheme report more satisfaction with availability of drugs than the other two groups who cite unreliable access to drugs as a primary reason for not joining or dropping out. Of households surveyed, those currently enrolled in the CHF report an average of 8.21 outpatient visits per household in compari- son with an average of 2.93 visits among non-joiners and 3.89 visits among drop-outs (Kihombo 2002). If enrolled households had paid user fees, annual expenditures would have exceeded the amount paid to enroll in the CHF. In contrast, both those households that had never joined and those that had dropped out would have paid less than the value of the annual premium. One explanation for these differences in utilization rates is adverse selection, which means that households that know that they will use more health services than the average house- hold, choose to join. A second explanation is moral hazard ­ that the presence of insurance coverage stimulates households to use more services than they would if they were paying full fees. A third and important explanation is that the availability of health insurance coverage through the CHF removes a financial barrier to access and enables households to obtain the care they need. Evidence from Kilosa indicates that household wealth, measured as total household expen- diture, is a determinant of whether households decide to join the CHF scheme. Average income of enrolled households is Tsh 59,023, in comparison with Tsh 46,273 for drop-outs and Tsh 33,102 for non-joiners (Kihombo 2002). These data indicate that the lowest income households may face financial barriers to paying the annual premium and that the decision to drop out may also be determined by resources available to households. These data indicate that the poorest households in Kilosa are not being served by the CHF, though it is also important to note that the majority of households in Kilosa are considered poor. against uncertain events if they can be assured Expanding membership beyond the current of access to good quality services. level remains a challenge. The experience of CBHI schemes in Rwanda The sustainability of community risk-pool- (see Box 6) indicates that low income individ- ing schemes needs to be tackled by govern- uals will prepay for services if they can be ments, especially in countries with high HIV assured of quality services, can trust scheme prevalence. Re-insurance might need to be managers with their contributions, and can introduced to ensure the sustainability of afford the premiums and co-payments. those schemes. Governments also need to 20 Resource Allocation and Purchasing in Africa Box 5: Provider based prepayment schemes in Uganda offer access to ambulatory and inpatient care. The MOH Health in Uganda acknowledged that it could not provide access to an acceptable and affordable level of health care services to the entire population and led a process to search for new sources and mechanisms to finance care. The National Health Policy and Health Ser- vices Strategic Plan for 2000/2001 through 2004/2005 emphasizes broadening the financing base for the sector and providing support for strategies to promote increased efficiency, fairness, risk pooling and protection for poor and vulnerable groups (Matsiko et al., 2001). In 1996, Government and donors helped establish the first provider-based prepayment scheme in Kisiizi hospital, a NGO in the western district of Rukungiri. By 2001, eleven provider-based prepay- ment schemes had been established with varying degrees of success (Matsiko et al., 2001). The Kisiizi Hospital Health Society (KHHS) is considered one of Uganda's success stories. Kisiizi has a 200 bed hospital and a community based health care program funded largely by user fees (70 percent) with some support from donors and from national government pro- grams such as immunizations and TB control (Walford et al., 1997). In 1996, the KHHS was established to offer households the opportunity to prepay for access to health services. To man- age potential problems with adverse selection, the KHHS scheme is only available to Engozi societies (all residents of the region of KHHS belong to Engozi societies which, in exchange for prepayment, finance transportation to hospitals, fund funerals, and offer small loans). In exchange for a premium that varies by size of household and a co-payment at each visit, imple- mented to control moral hazard, members of the KHHS have access to a wide range of out- patient and inpatient services with a few exclusions (eye glasses, normal deliveries, and self inflicted injuries). Of the 87 Engozi societies in the region, 36 were able to convince at least 60 percent of households to join in the first year (Walford et al., 1997). By 2000, 13.3 percent of the total catchment area population of 60,000 was enrolled in the scheme (Matsiko et al 2001). Table 4 presents a progression of enrollment from 1996 through 2000. consider providing targeted subsidies to the The Bamako Initiative's main principles are very poor so they can benefit from the that (i) it is a financial contribution to a self- schemes. managed system which includes the communi- ty; (ii) the money is kept at the local level; and (iii) it is used to replenish the local drugs Building on the Bamako Initiative: revolving fund, and for local maintenance of A pooling mechanism with local the health infrastructure. The Bamako Initia- ownership tive has been implemented differently in vari- ous countries, most countries applied it only as As a result of the Bamako Initiative imple- a cost-recovery scheme with the implementa- mented in 1988, which promoted community tion of user fees for health services and medi- financing and management of recurrent costs, cines, while other countries also stressed com- 27 African countries had introduced a cost munity management of the money. Every recovery scheme by the mid­1990s (Shaw and country kept the money at the local level and Griffin, 1995). Current Practices and Trends in Resource Pooling 21 Box 5 (continued) Table 4: Individuals covered by KHHS Year Number of Individuals Covered 1996 1,536 1997 5,768 2000 ~ 8,000 Sources: Walford et al 1997; Matsiko et al 2001. The chairman of the Engozi group collects and delivers premiums to KHHS each quarter. Premiums were priced at below the cost of services expected to be utilized by members with a commitment by donors to fund the difference during the initiation phase of the scheme. After one year of experience with the scheme, members were found to utilize fewer outpatient visits (almost half) but more than double the number of inpatient admissions than predicted (Wal- ford et al., 1997). KHHS interpreted this increase in admissions over what was expected as evi- dence that the customary user fees were a barrier to access and that membership in the KHHS reduced this barrier. KHHS intends to increase premiums to cover expected costs once mem- bership increases and the population grows to appreciate the value of the scheme. In early 2001, focus groups and interviews were conducted with households who live in regions of Uganda where prepayment schemes are available (Matsiko et al 2001). This inves- tigation included the area served by Kisiizi and can give some insights into whether prepay- ment schemes were being accessed by the poorest households. Of the 201 households inter- viewed, 53 were members of schemes. The most common form of income for households was agriculture (48 percent), the majority owned a radio and or cassette player (87 percent), and over half of households owned a bicycle (52 percent). Of the households not enrolled, 30 per- cent stated that the primary reason was that they could not afford the premiums. Among the 148 households not enrolled in any scheme, 71 said that they did not have the cash needed to pay for services when a family member last visited a health facility. This evidence indicates that it is probable that lowest income families are excluded from participation because of the costs of the premiums. It also indicates that customary user fees needed to pay for services pose a barrier to accessing care. benefited from the pooling of funds to buy 76­77). Some of the problems were that local drugs and equipment. management committees typically valued Implementation of the Bamako Initiative in investment over redistribution, and exemp- Benin, Guinea and Mali for example, helped tions did not reach the poorest. the MOHs focus on defining the primary The community financing of key operational health care package, its costs and on improving costs bought communities a seat at the table. the accessibility of primary care. Since the early Governments had to systematically negotiate 1980s the Bamako Initiative helped cover more new activities with community organizations. than 20 million people in those three countries, However, all three countries need to establish under-five mortality declined significantly, even mechanisms to subsidize and protect the poor among the poorest, and immunization levels better and this is a priority in those countries' increased (World Bank, WDR 2004, pp. reform process (WDR 2004). 22 Resource Allocation and Purchasing in Africa Box 6: Community-Based Health Insurance (CBHI) Funds in Rwanda increase utilization and reduce financial barriers to access health care As part of the national effort to rebuild the country, the Government of Rwanda was interest- ed in promoting innovative strategies to generate additional resources to fund health care serv- ices. In Rwanda, public health centers and hospitals earn the majority of their revenues from user fees which imposes a financial burden on patients at the time of need. As a result, uti- lization of formal health services is low and people delay obtaining care until they are very ill (Schneider et al., 2001). An additional result is that there are insufficient funds to operate health services which results in drug stock outs and unmotivated staff. CBHI was the strategy tried in Rwanda to address the issues of low utilization, lack of financial protection, and insuf- ficient resources to fund health services. The MOH chose to test community based prepayment schemes in three districts using an approach that involved community members in design, management, and oversight. This experience has been well documented and evaluated by Pia Schneider and colleagues using study and control districts and before and after comparisons and contains valuable lessons for design, implementation, and management of community based prepayment schemes in resource poor settings. In the first year of the pilot study, starting July 1999, 54 prepayment schemes were initiat- ed that enrolled 88,303 members representing 8 percent of the population of the three districts (Schneider et al., 2001). In exchange for an annual premium, families could obtain access to all preventive and curative services and drugs offered in their chosen public health center and to a limited package of inpatient services from the district hospital. Hospital services were cov- ered only if the patient had a referral from the health center. By prepaying for services at a time that households had income, and after a one month waiting period, families were able to insure themselves against large out-of-pocket payments for services throughout the year. One result was that new case consultations for members were up to five times higher than for non-mem- bers. Strong improvements in the utilization of preventive services were also a result. Immu- nization rates increased 50 percent, prenatal care 25 percent, and there were 45 percent more assisted deliveries among the covered population. Rather than being interpreted as a moral hazard effect of insurance leading to excessive utilization, these results were interpreted as eliminating the gap between needed and obtained health services that existed before the intro- duction of community based health insurance (Schneider et al., 2001). Current Practices and Trends in Resource Pooling 23 Box 6 (continued) With the introduction of CBHI, each of the 54 health centers in the three pilot districts became a partner that offers one prepayment scheme. Families that enroll choose their preferred participating public health center. Each month, scheme managers retain 4 percent of premiums for administration; send 4 percent to the district hospital fund that pools risks for the district and manages funds to cover hospital services for members; send 49 percent to health centers to cover capitation payments; and retain 43 percent in bank assets to cover future payments. Health centers receive a monthly capitation payment for each member who enrolls with them. Hospitals are paid by the district federation per episode for cesarean sections, malaria treatment and non-surgical pediatric cases and fee-for-services for consultations and overnight stays. By the end of a year, 7 percent of premiums were spent on administration, 7 percent on hospital services, and 86 percent on health center level care (Schneider et al., 2001). As part of the evaluation of the impact of the pilots, the MOH wanted to understand what population groups chose to enroll in the community based insurance schemes and whether membership improved financial access without increasing the overall burden of out-of-pocket spending. To answer these questions Pia Schneider and colleagues estimated three demand models using household data (Schneider and Diop, 2001a). Results indicated that the probability of purchasing insurance was not determined by health need or economic factors but by the level of education of the head of the household, family size, district of residence, distance to the health center, and radio ownership. The income quar- tile of families was not shown to be significant and cattle ownership, an indication of house- hold wealth, was also insignificant. The second model looked at the determinant of use of serv- ices and found that members used up to five times the number of curative and preventive services than non-members. The probability of a visit decreased with distance to the health center and increased with severity of illness, but those with coverage sought care when less sick. Results from the third model found that annual per capita contributions of member are up to five times greater than payments by non-members but payments at the time of service are significantly lower for members. This indicates that the presence of insurance changed care seeking behavior of members causing them to access care more frequently and sooner. CHAPTER IV Current Practices and Trends in Resource Allocation Utilizing budget mechanisms to ensure tion formula to distribute funds. Two studies better resource allocation suggest this change resulted in a more equi- table distribution of the government budget in M ost African countries use historical all but two provinces. It also supported an budgeting to distribute their increase in the average share of primary care resources. Under this process, the expenditure from 39 to 54 percent by 1998. current year's budget is based on (Bossert et al., 2000; Gilson et al., 2000). last year's allocation, typically with minor Efficiency improvements could result from a modifications. Historical budgeting ensures more rational approach to resource allocation. that existing capacity is funded but it does not For example, in Malawi, budgetary allocations necessarily ensure that the optimal mix of serv- to hospitals are based on cost-per-bed. This ices is funded, that priority population groups creates a perverse incentive either to increase receive services, or that funds are used effi- beds or to inflate costs; it does not provide ciently and effectively. Moreover, historical incentive for hospitals to be resource-efficient budgeting provides no incentives for outputs (Picazo, 2002). or outcomes. If health facilities are poorly distributed in the first place, such a resource allocation Decentralization: an ongoing trend arrangement will continue to lead to severe inequity. To improve allocation, some coun- A number of countries have begun to decen- tries have started to allocate funds on a needs tralize health services, although most achieved basis rather than on the basis of what was only some form of deconcentration to date. needed in the past. A needs-based approach This move towards decentralization stems requires a country to develop a new resource from the desire to increase local ownership, allocation formula, one that is typically based and to improve efficiency and equity of health on indicators such as population size, age and services. Reform initiatives usually focus on sex profile, and degrees of absolute and rela- giving more political and administrative auton- tive poverty. For example, in 1994, Zambia omy to decentralized districts. They often also introduced a population-based resource alloca- consist of raising local revenue for health serv- 24 Current Practices and Trends in Resource Allocation 25 ices through cost recovery schemes, integrating study on Zambia's decentralization is slightly other health service providers, and introducing more favorable. It concludes that Zambia's needs-based resource allocation (Bossert, decentralization effort's probably improved 2000). overall efficiency since the activity levels have To date, most governments have not been maintained in face of declining funding achieved significant improvements in efficiency (Bossert et al, 2000). More recent analysis and equity as a result of their move towards (Bossert et al, 2003) however concludes that decentralization. For example, the Govern- decentralization may not have had either a ment of Uganda began in 1993 to undertake positive nor negative impact on services. In decentralization by devolving responsibilities Rwanda less than 1% of public resources went for primary care from the central to the district to finance preventive and primary health care level, including the provision of basic health services in health centers in 1998, but in 2000, services and control over health personnel. for the first time, almost one-third of the pub- District funding was provided by three main lic health budget was decentralized and dis- sources: government grants (81 percent), bursed to the regional/prefecture level (Rwan- donor assistance (12 percent), and local rev- da Health Accounts, 2000). While the enue (7 percent). Between unconditional experience of decentralization varies from grants, flexible donor funding, and local rev- country to country in Africa, it seems that it is enue districts are free to decide the allocation more difficult to increase ownership and of between 25 and 60 percent of their total accountability through decentralization than health revenue, which represents a consider- initially expected. able rise in local autonomy compared to the A public expenditure tracking survey earlier system. However, studies suggest that undertaken in Ghana in 2000 (Xiao Ye & districts have allocated fewer funds to public Canagarajah, 2002) showed that only about goods type activities, in particular to primary 20 percent of non-salary public health expen- care, as they progress further into the decen- diture reached the primary care sub-district tralization process. On average, district spend- facilities. A large proportion of the leakage ing on primary health care fell from 33 to 16 occurred between line ministries and district percent in the period from 1995 to 1998. (Akin offices, where public expenditures are usually et al, 2001) The utilization of maternal and turned into materials from cash flows. The child health services has also fallen significant- study underscored the need to set up a consis- ly since the introduction of decentralization tent and transparent recording system all the (Mwesigye, 1999). A World Bank study warns way down to the service provision facilities. that poorly carried out decentralization runs The following table (Table 5) shows the dis- the risk of recreating inefficient centralized sys- connect between the measurement of the tems within each district (Habte et al., 1999). resource flows at the central government and Other countries also reported difficulties in at the facility level. At the central level, the implementing decentralization. In Senegal, Government is committed to provide more resources were allocated to local authorities, resources to basic health care, however, clin- who were not accountable for health outcomes ics receive a large proportion, if not all, of (Diop et al., 2001). In Chad, the Government public non-salary expenditures in kind, and granted 60 percent of its federal budget to the thus don't have much control over it. More- districts. However, the allocation process was over, salaries are distributed directly to not an a needs-basis and the wealthier districts employees by a central government agency. received significantly higher shares of funding Consequently, at the facility level, there is lit- than the poorer provinces (Soucat, 2004). A tle knowledge about what public resources 26 Resource Allocation and Purchasing in Africa Table 5: Ghana: A comparison of Public Financing at Ministry and Facility level Percent of Financial Resources from: Government and donor pooled fund Paid by patients NGO's assistance Based on MOH estimates for sub-district clinics 87 13 Based on sub-district clinic estimates 39 54 7 exist for them. This leakage means that the ture local capacities (Cote, 2003). A health patient bears a much higher proportion of the expenditures tracking survey undertaken in costs than intended by the central govern- 2002 showed that only 30 percent of the public ment. resources that were allocated to local health Burkina Faso undertook an interesting infrastructures actually reached those structures experiment to make health districts more (Bah, 2003). The problem here again is mainly accountable for their results, with the help of one of complex procedures for funds to be sent the World Bank. It involved a performance- to the local level, lack of flexibility in deciding based contract between the central MOH and how to use those funds at the local level, delays the District Health Office in one-half of the in budget voting and allocation, and lack of nation's districts. Because it was a project transparency as mentioned before. financed by an external agency, an exception A recurrent problem is that most MOHs was made to the regular flow of funds, and have not established a clear decentralization money was transferred directly to the district policy which is in harmony with the rest of the level. Indicators such as immunization seem to government's decentralization policy. Often, have improved (World Bank, 2003b), howev- the health sector decentralizes in a vertical way, er, once the project terminated, the money did with no or little links to municipalities' offices. not reach decentralized units any longer, jeop- Such a policy is needed to identify clearly the ardizing acquired results. The World Bank's lines of authority and the flow of funds. In Team Leader held that this was attributable to Malawi for example, there are three uncoordi- several factors: (i) transfers to district bank nated systems which attempt to decentralize accounts had stopped, which had been the pri- with varying degrees of success: the MOH sys- mary source of funding for the action plans; tem, the Ministry of Local Governments which (ii) the planning and performance monitoring funded health services through local authori- process that had been established and financed ties, and the Christian Hospital Association of by the project also stopped; and (iii) the proj- Malawi. (Picazo, 2002) ect management unit which played a key role Decentralization is a trend to support, but in organizing and implementing the process the ways and means for implementing this was disbanded (BTO, 2003). This problem process must be improved in terms of: 1) bet- shows the difficulty projects have in scaling-up ter financial record keeping and information; and integrating lessons learnt into national 2) better flow of funds to the local level; and in programs. terms of 3) a clear consolidated policy and A diagnosis of the decentralization process in strategy. In addition, alternative options for Guinea asserted that one of the main problems strengthening downward accountability is that ministries were more preoccupied with should be considered (Ndegwa and Levy, issues related to administrative processes than 2003), such as contracting decentralized units with providing an enabling environment to nur- with performance criteria. Current Practices and Trends in Resource Allocation 27 Better allocation through primary primary care by decreasing their own primary health care care expenses. Second, governments planned the expansion of primary care without Although a number of countries in Africa have addressing the politically sensitive question of made efforts to expand primary care to reducing funding for other health care facili- increase coverage of basic services to the rural ties. Reducing funding for secondary and terti- population, since the Alma Ata Conference in ary level care has proven to be difficult. As part the late 1970s countries have encountered sev- of health reforms in Zambia funding for hos- eral problems in this effort, three of the main pitals was reduced. After some time, there was ones being: a public outcry about deteriorating hospital conditions which caused the government of a) Expansion: Although as previously men- Zambia to reduce its reforms in scope (Blas tioned, Zambia increased its expenditures on and Limbambala, 2001). Third, most govern- primary care from 39 percent to 54 percent in ments underestimate the timeframe required to the period from 1994 to 1998 (Bossert, 2000) achieve the desired redistribution. In the short a number of other countries had difficulties run, 80 to 90 percent of the costs are fixed, expanding primary care, especially to under- making it hard to change the allocation of served areas. For example, the Government of resources. South Africa aimed to redistribute resources from the relatively prosperous mostly urban b) Recent benefit-incidence analysis challenges Gauteng Province to poorer Northern the conventional wisdom that public primary provinces. Over a five-year period Gauteng's care expenditures mostly benefit the poor. A share was expected to decline from 25 to 17 study (see Table 6) conducted in seven African percent, and the Northern provinces' share to countries showed that on average, 23 percent increase from 6 to 15 percent. Thirty percent of primary care expenditures benefit the richest of this shift was expected during the first year quintile whereas only 15 percent go to the but the plan proved impossible to implement. poorest ones (Gwatkin, 2002). Similar dispari- Gauteng Province received more funds than ties have been reported with respect to certain budgeted (Pearson, 2002). Studies from Ugan- cost-effective interventions which have been da (Akin et al., 2001) and Mauritania designed with the poor in mind. For example, (Hahmed and Soucat, 2004) also document an intervention like oral rehydration therapy sharp declines in the share devoted to primary (ORT) is considered a cost-effective treatment care in the 1990s, contrary to efforts to achieve for diseases concentrated in poor areas the opposite. (Gwatkin, 2001). However, recent research in There are several reasons why the expansion SSA suggests the richest quintile (77 percent) is of primary care proved difficult. First, some more likely to use ORT than the lowest quin- take advantage of external funding to support tile (58 percent) (World Bank, 2002a). Table 6: Inequities in 7 Government Health Care Expenditures Percentage of total benefit gained by Poorest 20% of the population Richest 20% of the population Total government health care expenditures 12 30 Government primary health care expenditures 15 23 Source: Gwatkin, 2002 28 Resource Allocation and Purchasing in Africa c) Hospitals are still used as primary care Policy measures which have structures because the referral system operates considerable public health impact inefficiently. For example, in Malawi, tertiary hospitals devote most of their resources to All African countries have adopted a basic basic or Level I health care, while only 10 to health care package based on cost-effectiveness 15 percent of patients at these facilities receive analysis (World Bank, 1993a). This was useful Level II or III care (Picazo, 2002). This is very in particular in defining what could be done at costly since basic care can be delivered much each level with limited resources. In Dar Es more cheaply by less fancy structures, such as Salaam, the Tanzanian capital, for example, 10 health centres. activities can be undertaken at the community Those problems highlight once again three level, 35 at the dispensary, and so on. (Dar Es common institutional issues: (i) that of a lack Salaam City Medical Office of Health, 2003) of stewardship from the State, which means However, in spite of some progress, many cost- policies need to be taken and applied; still in effective measures were not implemented to the realm of lack of stewardship, there is the extent that was expected. This is the case, inability to tap existing resources in the coun- in particular, for immunization, nutrition and try, for example where the State is unable to tobacco control. self-finance some essential functions, such as Nutrition benefited in the 1980s and 90s immunizations, it could mobilize resources in from large scale successes in Tanzania, Mada- the country, especially out-of-pocket payments gascar and Senegal (Marek et al., 1999) which and private delivery for such purposes; and (ii) shifted responsibility for service delivery to that of weak governance where the separation communities and the private sector, but most of financing from provision of services is a shift governments did not build on those successes, not many governments are likely to want to as it required a shift in paradigm from govern- make on a large scale. ment service delivery to contracting out of Overall, the move towards primary care and services. Although 53 percent of all child cost-effective procedures is taking place in deaths are attributed to malnutrition (Black et Africa and represents a clear step in the right al., 2003), and that there are successes to build direction. But the extent to which resources upon, there is still nothing being done on a have been shifted towards primary care varies grand scale in the continent. from country to country and some African Immunizations showed some progress but countries still devote far too little towards pri- an analysis of 12 West African countries shows mary care and cost-effective procedures. How- that on average, rates basically remained at the ever, even in countries which have succeeded in same levels they were in 1990 (see Table 7 increasing spending on primary care, it should below). This is in part due to the fact that in be emphasized that a shift of resources is not the 1980s, national immunization programs enough to target poor people. The shift must benefited largely from donor support, and this be accompanied by other measures that will was not the case in the 1990s. ensure that these expenditures actually reach Less industrialized countries already poor people. The focus which, up to now has account for half of all deaths attributable to been on universal coverage, needs to shift tobacco. One of the most cost-effective meas- toward targeted coverage to give priority to the ures to control tobacco smoking is to increase poor. taxes on cigarettes, an action only South Africa has taken with success. Current Practices and Trends in Resource Allocation 29 Table 7: DTP3 coverage in 12 West African countries, 1990 vs. 2001 Coverage Country (WHO/UNICEF Best estimates % point change Performance 1990 2001 Mauritania 33 61 +28 Medium-high Guinea 17 43 +26 Medium-low Ghana 58 80 +22 High Mali 42 51 +9 Medium-low Niger 22 31 +9 Low Gambia 92 96 +4 High Cote d'Ivoire 54 57 +3 Medium-low Benin 74 76 +2 Medium-high Senegal 51 52 +1 Medium-low Togo 77 64 ­13 Medium-low Guinea-Bissau 61 47 ­14 Low Burkina Faso 66 41 ­25 Low Source: Naimoli, 2003 Box 7: Adoption of a Cost-effective measure to improve public health in South Africa In 1999, the Tobacco Control Amendment Act gave the country some of the most progressive tobacco control policies in the world. Today, all tobacco advertisements and sponsorships have been banned; smoking at work and in restaurants is illegal, except in clearly demarcated areas; and explicit health warnings are required on all cigarette packs. Excise taxes represent almost 50 percent of the total retail price of cigarettes. Around 80 percent of potential tax revenues from cigarettes are actually collected (excise and sales taxes combined), representing several billion rands a year. As a result, cigarette consumption is on a downward spiral, it decreased from 1.9 billion packs in 1991 to about 1.3 billion packs in 2002. The rate of decline has accelerated especial- ly since 1997, when large tax increases sharply increased the price of cigarettes. Between 1993 and 2000, total tobacco consumption decreased by about 26 percent, the percent of adults who smoke decreased from 33 percent to 27 percent, with the biggest decreases being seen in low income groups, and the percent of young people aged 16 to 24 who smoke has decreased significantly from 24 percent to 19 percent. These gains are the result of decades of steadfast lobbying by the health community and antismoking groups and of the new South African government's commitment to public health. They have come in the face of vehement opposition from the tobacco industry, advertising agencies, hospitality associations, sporting organizations whose events were sponsored by tobacco companies, and, until the early 1990s, an apartheid government with extraordinary close links to the trade itself. Source: de Beyer and Brigden, 2003. 30 Resource Allocation and Purchasing in Africa Governments are thus missing opportunities achieve their objectives, technicians and ana- to build on existing best practices and scale up, lysts have to pay more attention to "the art of to harness existing resources to expand, and to politics". Analytical techniques like stakehold- implement policies which have great impact on er analysis or policy characteristics analysis public health. To do so, Governments would can be used to gain a better understanding of need to focus on their stewardship role to set the actors' positions. Generally, the study sug- and ensure that policies are implemented, to gests to balance strong political analysis with monitor progress, to set standards and ensure rigorous technical analysis (Gilson et al., they are applied, to partner with service 2000). providers from the private sector, among other tasks. Some Trends to Better use external aid for resource allocation Technicians overlook the importance of local politicians in allocating The usefulness of conditionalities resources versus political commitment Most African governments have expressed In order to ensure that health funding reaches their intention to focus scarce resources on poor people, external donors typically attach providing health services to poor people. In conditions to grants and loans. Recent practice, however, as mentioned before, funds research, however, suggests that conditionality are frequently spent disproportionately on has very little effect if the government is not facilities and services which primarily serve fully committed. For example, the Democratic wealthier populations (Gwatkin, 2000). To Republic of Congo received significant exter- understand this imbalance it is important to nal aid, including six structural adjustment emphasize that decisions pertaining to resource loans by the World Bank between 1965 and allocation are made within a wider political 1996 (Easterly, 2001). However, the lack of a context. Funding for health usually competes coherent macroeconomic framework and with other tax-funded government activities. endemic corruption under the Mobutu Gov- Vocal interest groups, typically urban popula- ernment led to a three decade-long decline. In tions, and other vested interests influence allo- 1996, GDP per capita was estimated at US$ cation decisions (Pearson, 2002). 150, less than 40 percent of its 1958 level. For- The difficulties in achieving commitment to eign aid probably has reduced the urgency for large-scale health reform are analyzed in a reform rather than promoted better economic recent study on the dynamics of policy change policies (Devarajan et al., 2001). Thus, com- in South Africa and Zambia during the 1990s. mitment by the government to reform is an In both countries, political transition brought a essential component of any successful interven- government to power which demanded speedy tion. and visible health policy changes and created windows of opportunity for change. The study Sector Wide Approaches Programs shows that reforms were heavily shaped by a (SWAPs): A Trend to Watch few key actors, in particular the Minister of Health. Technical experts and analysts general- In order to deal with the challenge of owner- ly had less influence than politicians on the ship and coordination, some African countries design of the health reform and its implemen- have agreed on a SWAP to managing health tation. The study recommends that in order to funds. SWAPs focus attention on the perform- Current Practices and Trends in Resource Allocation 31 ance of the entire health sector in contrast to governments towards pro-poor policy (Peters performance of individual projects. The ulti- 1998). If the parties cannot establish such a mate vision of SWAPs is to provide budgetary commitment, any externally proposed pro- support, i.e. all sources of public financing poor initiative has a high likelihood of failure. would be disbursed against a comprehensive budget of the MOH in the respective country. Marginal Budgeting for Bottleneck In practice, many donors still provide ear- (MBB) marked funding to support specific compo- nents of the budget that are articulated as part Despite extensive Sector Reforms, the health of a national health strategy. This approach systems in many SSA countries still fail to has required changes in standards and reach large numbers of women and children-- approaches by both donors and clients. especially the poorest and most vulnerable-- Donors are assured a voice in the process of with these interventions. It is increasingly clear developing a national health strategy and that the strategies adopted previously in the countries are required to develop sound health sector have to be different. national health policies and secure their imple- This situation triggers a growing demand for mentation (McLaughlin, 2000). The shortage tools that help answer three questions: of resources for the health sector has often frustrated the intentions of donor earmarked What are the major health systems bottle- financing as Government resources will some- necks hampering the delivery of health times be reallocated away from the externally services, and what is the potential for financed categories by the Government. This their improvement? partly explains the preference for SWAPs in SSA by donors (McLaughlin, 2004). How much money is needed for the Many countries, however, do not have a expected results? national health policy in place and their budg- ets often do not reflect stated priorities. Thus, How much can be achieved in health out- a one-step transition to the new approach is comes by removing the bottlenecks? not feasible as many donors are reluctant to channel their funds through existing systems. The MBB tool recently developed by SWAPs, however, provide direction for future UNICEF, the World Bank and WHO, and test- development work which both sides can agree ed in several countries is a response to this upon and gradually establish over time. To demand. date, there has not yet been evaluation of a The mainstay of MBB is to identify coun- fully implemented SWAP. Experience in try/province specific "implementation con- Ghana, however, shows increased government straints" of health system and estimate the health spending in line with agreements nego- "marginal costs" to overcome them. MBB uses tiated with external partners, which might like- existing information available for selected trac- ly have a positive impact on health outcomes. er interventions to identify the "bottlenecks," SWAPs provide a framework as to how to -- the weakest links in the chain of conditions work towards a broad-based consensus. Bud- and debate various options to address them. geting via a SWAP helps to make conflicting The tool also allows to assess the likely impact interests transparent and encourages a dia- of alternate options on health outcomes based logue between donors and government. How- on available evidence. The MBB approach thus ever, SWAPs only provide a framework; their helps in improving allocative efficiency of gov- success depends on the genuine commitment of ernment health budgets. This makes MBB dif- 32 Resource Allocation and Purchasing in Africa ferent from traditional approaches of pro- What to expect from MBB? gramming and budgeting of health interven- tions. However, the tool is not meant to be used as This Tool Helps Analyze: a quick fix solution for improving health out- Coverage Frontiers, Costs and comes. Adequate local participation and con- Impact of Policy Options sultation are important pre-requisites to identi- fy the bottlenecks for implementing evidence · What New Interventions? (Home based interventions and discuss alternate Based Neonatal Care/Essential strategies for overcoming the bottlenecks based Obstetric Care) on local experiences. The indicators and unit · By whom? (Public/Private sector) costs need to be adapted to local needs. The · How? (Supply or Demand Focus) tool can help to generate alternate scenarios · To Whom? (Geographic/Social using different strategies to enable policy mak- Targeting) ers to choose the best option. The results are · With What? (Input Mix) country specific and temptation to generalize · At what Cost? (for Drugs, Salaries, the scenarios to other countries should be Construction) restricted. · Who Pays? (Public/OOPs) Source: R. Knippenberg, A. Soucat, W. Vanlerberghe et al. CHAPTER V Who is Covered? O verall, many African countries have Africans rely on private care (Castro-Real et undertaken important efforts to al., 2000). improve health access for poor peo- This inequity is especially striking for hospi- ple. The outcome differs from coun- tal care and much less important for primary try to country but benefit­incidence analysis care. The poor­rich benefit ratio (using suggests that there is significant potential for unweighted averages) for hospital care is 11 to improvements. This section will review empir- 34 while it is only 15 to 23 for primary care ical evidence illustrating how new financing (Castro Real et al., 2000). This confirms that and RAP arrangements can be applied to targeting health spending to the poor in Africa improve health access for poor people. would require spending less on hospitals and more on primary care facilities. However, one of the reasons for which governments subsidize Addressing inequities tertiary health services is that no insurance A study carried out a benefit­incidence analy- Table 8: Benefit incidence of public spending on sis to assess whether existing resource alloca- health in selected countries, by population quintiles tion arrangements in seven African countries reach the poor (Castro-Real et al., 2000). The Quintile shares of all health going to results indicate considerable inequity with the richest income quintile receiving twice as many Country The Poorest The Richest benefits as the lowest quintile from govern- Guinea (1994) 4 48 ment health spending. For example, in Guinea, Ghana (1992) 12 33 48 percent of public health expenditure went Cote d'Ivoire (1995) 11 32 to the highest income quintile and only 4 per- Madagascar (1993) 12 30 Tanzania (92­93) 17 29 cent to the lowest one. Out of the seven coun- Kenya (1992) 14 24 tries, only South Africa managed to give the South Africa (1994) 16 17 lowest and the highest incomes a similar share, Unweighted average 15 23 one of the reasons being that wealthy South Source: Castro-Real et al., 2000 33 34 Resource Allocation and Purchasing in Africa market exists. Households cannot insure them- (Ethiopia Survey, 2002). On the other end, selves against the risk of serious illness or more than 60 percent of the children belonging injury and the consequent need for very expen- to the poorest 20 percent of the population sive treatment (Castro Real et al., 2000). were treated outside the public sector for their On the other hand, some countries have most recent ARI or diarrhoea bout in 16 out of made significant headway in tackling this equi- 19 SSA countries (Bustreo et al., 2003). In the ty problem. Ghana, for example, where benefit same way, in Uganda, of the sick children who incidence has increased slightly for the poorest were seeking treatment for diarrhoea, cough or one-fifth of the population, from 12 to 13 per- fever, only 17 percent went to a public institu- cent, and decreased for the richest two quin- tion, 41 percent were brought to a private clin- tiles of the population, from 54 to 50 percent, ic. This latter percentage was lower among the between 1992 and 1998. The middle income caretakers who were less educated (33 percent) group also seems to benefit more from public and higher among the more educated (47 per- health services than before. Similarly, the rural cent) (Ministry Of Health Uganda, 2001). population benefits significantly more in '98, Data for 1997 in Ghana suggests that in gener- getting 61 percent of resources compared to 51 al the patients prefer to use private for profit percent in 1992 (World Bank, Ghana PAD, services for minor ailments because of slow- 2002). ness, unavailability of drugs and poor staff In conclusion, targeting the poor via: 1) a attitude in the public facilities. But for more better supply and provision of services through serious cases, government or mission facilities outreach to reach them where they live; or via are perceived to be better (Ghana Policy Analy- 2) demand mechanisms through subsidies; or 3) sis and Development Group, 1999 referred to some kind of risk pooling mechanisms, is one in World Bank Ghana PAD, 2002). In South of the most urgent challenges governments are Africa public sector primary care is free, yet facing in efforts to improve health indicators. around 30 percent of people without medical insurance still choose to pay out of their own pocket to attend facilities in the private sector. Where else can the poor go for Even in the lowest income quintile this propor- services? tion is estimated to be 20 percent (Palmer et al., 2001). Although the public sector plays an important Who comprises the private sector? Most for- role in health service delivery in Africa, public profit providers work in urban areas, other pri- providers still fail in general to concentrate on vate providers cater to rural as well as to poor the poor, and seem to rather support the better- urban populations. For example in Benin, 60 off population. There is thus a need for gov- percent of service providers are concentrated in ernments to specifically target the poor. the capital city, while only 15 percent of the There is evidence that the poor buy services population lives there (Decaillet and May, from the private sector, but to a different 2000), but mission and church hospitals sup- extent in each country in Africa. On one end, ply crucial health services to disadvantaged in 1997 in Eritrea, the private sector was used and poor areas. In Zimbabwe, about 35 per- as a source of care by fewer than 20 percent of cent of all hospital beds are privately owned, medical care seekers (World Bank, 2001). The and about 96 percent are located in poor and same happened in Ethiopia where a study disadvantaged areas (Gilson et al., 2000). showed that between 1997 and 2000, nearly Although for-profit providers are less com- 80 percent of the population obtained their mon in Africa than in other non-industrialized care for malaria from the public sector countries, they play an important role, and it is Who is Covered? 35 a very rapidly expanding sector: it was esti- urban population (Marek et al., 2004). In mated that in the 1980's and early 1990's the Guinea's rural areas, independent practitioners number of for-profit physicians in eight who do home visits absorb 91 percent of the African countries ranged from 2 (Burundi) to out-of-pocket expenditures (44 percent in 86 (Zimbabwe), significantly below the aver- urban areas), and private clinics absorb anoth- age of 213 for all developing countries. About er 44 percent of urban out-of-pocket expendi- 46 percent of all physicians were private in tures (4 percent in rural areas) (Schwabe et al., Africa (Hanson and Berman, 1998). In Coto- 2003). Those rural independent practioners are nou, the capital of Benin, alone there are 328 likely to be mainly traditional healers and mid- private providers. One-half of the managers of wives. those private structures admit having a second Recent research also suggests that private job, and half of those said it was a job in the providers, in particular NGOs, often provide public sector. It is interesting to note that more health services more efficiently than do gov- than 75 percent of those private providers are ernments. A study on the performance of open or available 24 hours a day. It was also church and mission hospitals in Ghana, Tan- found that the majority of those providers zania and Zimbabwe showed that they were accepted to treat a poor person freely or on average more efficient than their govern- through credit with less than 10 percent said ment counterparts: In Tanzania, private hospi- they would refuse to treat a person who could tals treated twice as many outpatients and not pay (Decaillet and May, 2000). Tanzania inpatients as the government hospitals. In began to encourage private provision of health Ghana, private and government hospitals had care in 1991, and it is estimated that between similar costs but drug availability was about that date and 1996 there was a 36-fold 15 percent higher in private facilities. In Zim- increase in the number of private for-profit dis- babwe, the average doctor to bed ratio was pensaries and that the number of for-profit significantly higher for private providers hospitals increased five-fold (McLaughlin (Gilson et al., 1997). Another case study from 2004). Development of this sector is also seen Zimbabwe suggests that subsidies to non- as a way to reverse the brain-drain, as is the profit providers leads to better improvements case in Ghana. For-profit hospitals have a in equity than those to for-profit providers strong foothold mainly in Kenya and South (Mudyarabikwa, 2000). In Ethiopia (Ethiopia Africa (Shaw and Griffin, 1995). In Ethiopia, Survey, 2002), recurrent costs per patients the private sector and NGOs owned 9 percent treated in 2 public and 3 NGO health stations of hospitals and 28 percent of clinics in 1995, and clinics in 1999 showed that the cost was the numbers increased respectively to 24 and 12.6 and 9.8 Birrs respectively; however, at 41 percent in 2000 (Abay Asfaw, 2003). the health center level the NGO spends 42 Since urbanization is rapidly increasing in Birrs while the public sector spends 38 Birrs; SSA from 34% in 2000 to 51% in 2030 (L'In- finally at the hospital level the recurrent cost telligent, 2002), private health providers will per patient is lowest in the public sector at 21 likely grow quickly to serve those 621 million Birrs, followed by a sugar factory hospital at inhabitants in 2030. 41 Birrs while a private hospital spent the In rural areas, private providers are also most, at 114 Birrs. So evidence is not quite quite present, but more as NGOs and tradi- conclusive on how cost efficient the private tional healers. In Kenya, in 1998, among the sector is compared to the public. For the rural people who sought care for young chil- moment, one should, therefore, look to the dren with diarrhea or cough, 52% went to a private sector to improve quality and coverage private provider, this figure was 55% for the more than to decrease costs. 36 Resource Allocation and Purchasing in Africa What can be done? Already in 1993, the (iii) fail to address the market-level problems World Development Report called for an of anti-competitive practices and lack of increased use of the private sector to improve patient rights (Lilani et al., 2000). equity and efficiency (World Bank, 1993). In Ethiopia, the private sector is constrained Africa is very rich in local traditional organ- in a variety of ways. One constraint is licen- izations, such as age groups, women groups, sure. Providers obtain their licenses only from youth groups, sports groups, which can be regional health bureaus, and this is perceived mobilized to work with communities, as the to be overly centralized and could be delegated case of Senegal illustrates (see Box 8 below). to the zonal level. License renewal is also time Such local organizations are considered part of consuming and bureaucratic (Ethiopia Survey, the private sector too. Thus, partnerships with 2002). the private sector have the potential to improve Some might argue that many governments equity in African health systems. To date, most have a limited capacity to regulate, and this is, countries are starting to implement formal indeed, why the unregulated private market agreements between private providers and the flourishes. So, in countries with very limited government. Generally, there seems to be an capacity, it is more helpful to think in terms of increased appreciation of the importance of what can be achieved in the short- and long- private providers as an untapped resource to term. In the short-term, efforts can be made to: effectively deliver health services. The private (i) increase health related knowledge and infor- for-profit sector could continue to serve the mation to users, (ii) work with commercial and rich urban population making this segment of informal providers (traditional healers, street the population pay for it, while the public sec- drug vendors, etc.) to improve dispensing and tor and not-for-profit private sector could con- treatment practices; (iii) expand good practices centrate on the poor. where they exist; and (iv) encourage local Increased partnership with the private sec- organizations to manage or monitor health tor, however, will require that governments provisions. In the long-term, policy needs to shift their effort to functions such as regulation concentrate on building capacity to (i) develop and leave more service delivery to the private regulatory frameworks and formal accounta- sector. Current regulations of the private sector bility mechanisms to improve health sector are inadequate. In Tanzania and Zimbabwe performance, (ii) contract with providers current regulations: (i) focus on individual against specific outputs, and (iii) establish inputs rather than health system organizations; information systems to monitor performance (ii) aim to control entry and quality rather than (Bloom and Standing, 2001). explicitly quantity, price or distribution; and Who is Covered? 37 Box 8: The government as an active purchaser through strategic contracting in Senegal An example of successful contracting by the public sector for preventive services can be found in community nutrition projects in Senegal. (Marek et al., 1999) This project avoided most of the challenges mentioned by Anne Mills (1997). In Senegal, contract management was dele- gated to a third party, an NGO called Agetip which became the project management unit. This project management entity manages and monitors contracts for the government and is respon- sible for project implementation and results. Transaction costs of introducing and maintaining the contracts amount to 17 percent of total project costs. As the project is focused on improv- ing the nutritional status of malnourished children, it is a clear example of using strategic pur- chasing to benefit the poor. This was further verified by results of a study in Senegal that demonstrated that 79 percent of project expenditures were spent in the poor peri-urban target neighborhoods. High-risk children receive the following services: monthly growth monitoring; weekly nutri- tion and health education services to mothers; referral to health centers and home visits when necessary; food supplementation for the malnourished; improved access to water. Service deliv- ery is contracted out as well as training, supervision and operations research. Services are deliv- ered by a group of previously unemployed young people who create a legal entity, and who come from and live in the target neighborhoods. The youth groups who are the service providers, are supervised by a supervisor (often a pair of unemployed medical doctors) who is, in turn, supervised by Agetip. The project did not use competition to award contracts for overall project management but open tendering was used to choose the supervising NGOs. The youth entities are chosen by their communities and they sign a formal contract with Agetip. Contractual agreements spec- ify the services that are to be provided, the number of beneficiaries served, and the amount of attendance to weekly health and nutrition education sessions required. Performance is assessed by evaluating results as reported in the management information system that is built and main- tained by each service delivery unit. Supervisors and community nutrition workers can be fired if their performance is below expectations. The same four indicators are monitored monthly by the targeted community, the supervisors and Agetip. Local community committees moni- tored in part the service delivery, which increased accountability. Malnutrition rates decreased rapidly and coverage steadily increased. A community based study confirmed that malnutrition rates decreased steadily and that after 17 months of project implementation severe malnutrition disappeared among children of 6-11 months and moder- ate malnutrition fell from 28 to 24 percent among the 6 to 35-month age group. Project man- agers believe that the decrease in malnutrition is caused more by better care through regular growth monitoring and nutritional education for mothers than by the food supplement. CHAPTER VI Current Trends and Practices in Purchasing Health Services N ational Health Accounts revealed influence over critical management decisions that in ESA, 53 percent of health such as hiring and firing or how resources will expenditures go through private pur- be allocated to models with a higher degree of chasers: 33 percent through house- local control. There are some attempts at holds, 11 percent through NGOs, 9 percent autonomization of hospitals, but no data is yet through private insurance and employers (ESA available to assess their impact. Few systems NHA, 2000). have effectively evaluated the impact of these In spite of those figures, governments in reforms. One exception is Burkina Faso where Africa are still in the early stage of the financ- the health system has been decentralized (see ing­providers split, and only recently did timid Box 9). It must be noted that this successful efforts start to purchase services in a strategic experience has since ceased in Burkina itself, as way. mentioned previously. Public­public arrangements: Public­private arrangements Local level purchasing within a decentralized public model International advisors have been actively pro- moting contracting out health services rather As mentioned previously, in recent years, some than direct government provision as a way to countries in SSA have decentralized their pub- improve efficiency and quality of care and to licly funded health systems. This move toward improve access by providing services in regions decentralization has been partly driven by the where public providers are scarce. While gov- belief that local control over resources will ernments in Africa have contracted the private result in better performing health systems sector to provide non-health services such as because local authorities are more accountable laundry, food service, and cleaning there are few to their populations. Design and implementa- examples of governments contracting private tion of "decentralization" in the continent has sector providers to deliver health services and included a variety of models ranging from the fewer cases that have been evaluated. The Gov- establishment of regional bodies that have no ernment of Malawi plans to contract out clean- 38 Current Trends and Practices in Purchasing Health Services 39 Box 9: Innovation in RAP arrangements within a hierarchical public structure: Burkina Faso Case study Presented here is the bottom-up planning, oversight, and resource control approach imple- mented by Burkina Faso (Eichler, 2001). One of the primary strategies to improve perform- ance of primary health care in Burkina Faso is to engage each community to assume leader- ship for their own community's health. Community committees have been established to: formulate priorities for health; identify strategies to solve identified problems; incorporate strategies into annual action plans; oversee operations of community primary health care facil- ities; establish local user fee schedules for cost recovery; manage drug depots as revolving drug funds; and to manage funding from the government, donors and user fees. What is unique is that action plans are developed at the community level and funds flow from the central level to districts and then to communities and are managed at the community level. Communities effectively become purchasers for some inputs important for primary health care service deliv- ery (e.g., motorbikes for health education outreach and essential drugs), control user fees, manage funds for drugs and consultations, and oversee the performance of the primary health care workers whose salaries are covered through the central ministry. Central to this model of decentralization is the bottom-up approach to developing action plans. Each year action plans are developed at the regional, district, and community level (defined as the group of villages served by a primary health care center) according to a clear process established by the central ministry. Action plans define specific activities that will be carried out by each level in the system to address priority problems, the funding required, indi- cators of performance, and procedures for appropriate management of funds. Manuals docu- ment the process to be followed and provide formats for preparing budgets and performance indicators. Staff in regions and districts facilitate the planning process with community health committees to develop community level action plans. Each year a conference is organized with the regions, the central government and donors to discuss and approve plans and to obtain commitments to contribute resources. After this consultative process, plans are revised and funding commitments are made. While defining performance targets that communities perceive to be important is an approach to focus efforts on results, communities are not fully held accountable for achieving the goals established in their action plans. Current management agreements do not contain either rewards for attaining targets or penalties for failing to reach them. Burkina Faso is con- sidering introducing explicit performance based funding to provide more powerful incentives to each level in the public system to attain performance goals. The current information system is adequate for monitoring key primary health care indicators and the management system could be revised to accommodate a performance based payment scheme. (continued on next page) ing, maintenance and other ancillary services Anne Mills (1997) describes examples of con- affecting 5,237 civil servants who will then be tracting with the private for-profit hospital sec- retrenched from the civil service, retrained or be tor in South Africa and Zimbabwe with mixed absorbed by the private sector (Picazo, 2002). results. She cautions that key issues that must be 40 Resource Allocation and Purchasing in Africa Box 9 (continued) Results so far are promising but more experience is needed before conclusions can be drawn about whether this is a model to propose for the rest of SSA. By 2001, all communities in Burk- ina Faso had developed action plans indicating a high degree of inclusion of the poor and rural population. There is wide agreement that one of the contributors to success is the availability of adequate financial resources that are managed at the local level. Data indicate that since implementation of this decentralized model availability of essential drugs has increased slight- ly; immunization coverage has increased markedly (see Chart below); the percentage of women receiving at least 2 prenatal visits has increased from 41 to 52 percent; assisted deliveries have increased from 30 to 33 percent, and use of modern family planning methods remains low with the family planning prevalence rate showing a slight increase from 3.7 to 4.9. The transaction costs of implementing this approach are estimated to range between US$ 0.03 per capita under the lowest cost scenario and US$ 0.12 per beneficiary under the highest cost scenario. If per- formance across key primary health indicators continues to improve it is easy to argue that this investment is justified. Immunization coverage in Burkina Faso (1995­20000 9 0 8 0 7 0 6 0 5 0 4 0 3 0 2 0 1 0 0 9 5 9 6 9 7 9 8 9 9 0 0 BCG CTOQ3 VAR Sources: 1995-1999 Ministry of Health and Eichler 2001. considered include: the capacity of the contract- providers, no formal contracts were drawn as ing agency; the details of the design of contracts; such. Contracting is a new trend. the capacity to implement contracts; and the Good and bad reasons to work with the pri- characteristics of the supply and labor markets. vate sector: Whether the private sector has Although Governments have historically some comparative advantages depends on spe- been providing subsidies to not-for-profit cific situational circumstances and how the Current Trends and Practices in Purchasing Health Services 41 partnership was established. There is, however, that is motivated to achieve social goals, this evidence to show that using the private sector approach to purchasing enables the franchisor can increase equity, especially if NGOs or to take advantage of market power to negoti- Community-based organizations are used to ate low prices and assure quality with suppli- work in areas not covered by the public sector. ers. After products are distributed to fran- Another advantage of using the private sector chisees, final purchases are made either with is the possibility of improving quality of serv- out-of-pocket payments by consumers or by ices, mainly through better accountability to consumers who have some form of third party the client. Finally, the private sector can in payment. Governments can stimulate the cre- many instances increase coverage very fast ation of franchises by making capital available when mobilized. There already exist good and by establishing social priorities. If fran- examples of different types of instruments chises are viable private enterprises, they are upon which to build, such as contracting, leas- sustainable. ing and concessions (Marek and Yamamato, There are a few franchises in Africa, some 2003). are for reproductive health services, such as Private not-for-profit providers have been Mary Stopes in East Africa, others produce proven effective in reaching the poor in areas and distribute condoms or impregnated mos- which have been typically under-serviced by quito nets. Many seem to face the problem of using explicit contractual agreements. For sustainability as they cannot recover all their example, in a community nutrition project in costs they need to have subsidies in order to Senegal, service delivery was contracted out to continue reaching the poor. A promising fran- local youth groups, to cover poor peri-urban chise is that of Cry for the World, in Kenya, areas (see Box 8). which sells drugs (see Box 10). In addition to franchises, there are several large companies which act as purchasers as Private­private arrangements well as financiers for the health of their work- ers. Mining and large agricultural estates have Another approach to increasing the availabili- been known to invest heavily to reach the com- ty of adequate quality health services and munities from where their workers come, espe- essential medicines is to stimulate the develop- cially in regards to AIDS. ment of private providers. One way to do this is through franchises. Franchisors develop a business model that franchisees can buy into. Demand-side strategies to enable Each franchise is a private business. Fran- purchasing among target groups chisees must comply with standards and proce- dures developed by the franchisor, sell prod- A potentially powerful strategy to enable pri- ucts purchased through the franchisor, and ority population groups to purchase health charge prices determined by the franchisor. services is to subsidize their demand. One such Moreover, the franchisees must comply with mechanism is to provide consumers with the franchisor's management and reporting vouchers that enables their access to either dis- systems. This degree of standardization offers counted or completely free services. A project assurances to consumers of quality and consis- in Tanzania provided pregnant women with tency. vouchers to purchase insecticide-treated nets One level of strategic purchasing happens (ITNs) (see Box 11). The mixed results provide when the franchisor purchases products to good lessons for the design and implementa- supply the franchisees. In a social franchise tion of similar schemes in SSA (Marchant et 42 Resource Allocation and Purchasing in Africa Box 10: Expanding access to essential medicines through franchised drug shops in Kenya Cry for the World Foundation (CFW) started a network of franchised drug retail outlets in the districts of Kirinyaga and Mbere in Kenya in early 2000. By 2002, 26 retail shops had been opened. Initially, the shops were operated by community health workers who received train- ing by the franchisor. In order to open an outlet, an outlet manager must provide 20 percent of the financing and take a loan for the additional 80 percent for an interest rate of 18 percent per year. Only drugs on Kenya's essential drug list are sold. Outlet managers are required to distribute only those products that are supplied through CFW and to keep financial and stock management records and records of patient interactions according to the rules of the franchise. Prices are determined by CFW and are considered to be among the lowest in Kenya (Marsden, 2002). At the central level in Nairobi, CFW is responsible for helping to establish new shops, super- vising and developing training materials, purchasing drugs and supervising drug distribution, assuring drug quality control, and managing relationships with the MOH. Supervisors locat- ed at the district level train franchise outlet owners and assure that franchise procedures are being followed. Because franchisees own their shops and earn a return on sales, they have financial incentives to be responsive to population preferences by adapting location, hours of operation, styles of interaction, and physical appearance of their shops. By 2002, a population of roughly 100,000 people were being served by CFW outlets. A study of sales and patient interaction records examined the impact of the franchise on preven- tion and treatment for malaria. Of the 60,000 patient records that were examined, over 20,000 were found to be for malaria. Record reviews found that outlets are strictly adhering to nation- al treatment guidelines which indicates that quality services are being provided. In addition, impregnated bed-nets were sold by outlets. CFW does claim to be reaching the poor, but not the very poor who cannot pay (CFW, 2002). al., 2002), and to help answer the question of ulation groups in SSA but more evidence is whether vouchers can be used to target popu- needed. lation groups to increase utilization of a prod- Overall there are some very good purchasing uct or service. experiences to build upon to reach the poor in Conclusions about the effectiveness of SSA, especially in terms of public­public and vouchers to stimulate demand cannot fully be public­private arrangements. It would be rela- made based on this experience in Tanzania. tively easy to build on the positive experiences What is clear is that in order to be effective, a in order to go to scale. More attention needs to social marketing strategy must target pur- be given to evaluating the schemes and to chasers in households and that incentives to training of both government and private actors use the system in ways other than is intended in contracting. There's also a need to provide must be considered as part of the design and the nurturing policy and administrative envi- implementation process. This example is ronment for those schemes to be implemented included as a strategy that has potential to with the least hurdles possible. stimulate service utilization among target pop- Current Trends and Practices in Purchasing Health Services 43 Box 11: Vouchers to improve access to ITNs among pregnant women in Tanzania: a strategy to stimulate demand side purchasing Vouchers intended to enable pregnant women and children under five to purchase ITNs at a discounted price were tested in the Kilombero Valley region of southern Tanzania in 1999 (Marchant et al., 2002). Sleeping under ITNs has been linked to reduced incidences of anemia and malaria in pregnant women and the young children who tend to sleep with their mothers. As designed, public maternal and child health (MCH) clinics in the region were supposed to give pregnant women and mothers with young children vouchers that provided a 17 percent discount in price of ITNs available at local retail outlets. In effect, this discount reduced the price of a net from US$ 3.8 to US$ 3.1. MCH workers were given training in how to com- municate the importance of ITNs to pregnant clients and in how to explain the way vouchers worked. Outlet retailers submitted the vouchers for reimbursement for the discounted amount plus an additional handling fee. To assess impact, as sample of 505 pregnant women in their third trimester were interviewed at their homes. Among this group, 97 percent attended MCH clinics. In spite of this high atten- dance only 28 percent had heard about the vouchers and only 2 percent reported having been given a voucher. Of the 10 women who reported receiving a voucher, 80 percent had used it to purchase an ITN. Of the remaining 131 women who knew about the vouchers but had not received one, 83 said that they did not want one because the discounted price was still too high for them to afford. Another 29 said that they already had an ITN at home and therefore did not need the voucher. Only a small percentage of women (5 percent) said they did not under- stand how to use vouchers. In contrast to the low uptake among pregnant women, 86 percent of all vouchers issued in the study area were used. What is not clear is how many of these vouchers were actually used by eligible women. Suggested explanations for the low use among the target population group include: MCH staff trying to sell vouchers, retailers refusing to accept vouchers, and non-eli- gible people receiving vouchers (Marchant et al., 2002). It is also possible that eligible women gave false statements about having received the vouchers because they were used to purchase nets for non-eligible family members or they hoped they would receive an additional voucher. In addition, women reported that the majority of purchasing decisions were made by their hus- bands calling into question whether using MCH clinics attended by women is the most effec- tive vehicle to market and distribute vouchers. CHAPTER VII Conclusion: Misconceptions, Constraints, and Future Opportunities Misconceptions This paper highlights a few (7) common misconceptions: 1) The problem is one of money: although funding for health is severely constrained in SSA, there are plenty of resources which are either not used (large amounts of undisbursed donor money, untapped and unregulated private sector providers, low levels of execution of public budget) or used in an inefficient manner; 2) Resources allocated to primary health care are reaching the poor: they usually benefit the wealthy more than the poor, unless proper targeting mechanisms are in place; 3) The public sector is the main provider of health services in Africa. Although donors have up to recently focused mostly on how to improve public sector spending, there is a large share of health spending which is private, which also needs to be utilized more effectively; 4) Public­private partnerships are justified because they draw on the efficiencies of the pri- vate sector. Evidence indicates that public private partnerships are more likely to benefit equi- ty, service quality, and community empowerment than cost control; 5) The private sector can do better than the public sector. This is not always the case. Pub- lic­private partnerships are usually justified when a government wants to improve equity, qual- ity, empower communities, but does not want to deliver that service or perform that function itself, or when a private entity has a comparative advantage in service delivery. Such partner- ships should not be attempted for other reasons; 6) Working with the private sector means neglecting the public sector. On the contrary, for a government to work effectively with the private sector it has to be able to: 1) negotiate con- tracts; 2) set up the right enabling environment for the private sector to operate; and 3) pro- vide norms and ensure that they are implemented -- among other functions. Other functions, particularly those related to service delivery, can be devolved to the private sector; 7) External aid will solve the problem: reliance on external sources of financing is still very high in most countries, but donors seem to provide aid more based on political considerations, and the fragmentation in external sources of financing adds to the countries' problem of man- aging and coordinating. 44 Conclusions: Misconceptions, Constraints, and Future Opportunities 45 Constraints to Effective RAP This paper also highlights some (3) of the constraints faced by governments in Africa: 1) Many governments experience a lack of stewardship. There are many reasons for this, some of which are political and some are technical. Inadequate management of human resource represents a major challenge to African health systems that results in the non-dis- bursement of a considerable amount of available funds. Added to this is the complexity of managing often conflicting donor priorities and the pressure placed by the HIV/AIDS epidem- ic and the result is health systems that are overwhelmed and, thus, perform poorly. Govern- ments, with donor support, have tried to introduce SWAPs to establish priorities for the sector and to manage competing donor agendas. 2) The allocation of resources based on historical budgeting only perpetuates existing inequities. Decentralization is one strategy used to improve resource allocation, but it has not yet generated the expected results. Reallocation of resources toward primary care and away from secondary and tertiary care has also proven to be difficult to implement. Even in coun- tries where some reallocation has occurred, the poorest populations are not necessarily the beneficiaries. There is, therefore, a need to target public services specifically to the poor and to ensure that those that can pay, do pay. 3) Donors need to ensure that their aid helps reinforce health systems rather than create fragmentation and more strain on already stretched human resources. Future Opportunities to improve RAP This paper shows that some African governments have already looked for ways to improve resource generation, provide risk protection, and carry out strategic purchasing. This also paper mentions a number of successful examples and lessons learnt which can be built upon, and are to be encouraged. There are trends which need to be encouraged to improve RAP, and we classified them into three categories: To improve the policy-side of RAP arrangements: (i) There are several ways to improve equity, which include: Risk pooling: A challenge to the increasing utilization of essential services by the poor is providing mechanisms to protect households from the financial burden of high cost health events. Various insurance schemes can provide this needed financial protection and can result in higher utilization of essential services by removing the financial barriers to access that households face. While each risk pooling scheme is different, experiences con- tain the following common lessons to be considered when designing and introducing schemes: (continued on next page) 46 Resource Allocation and Purchasing in Africa · The human resource capacity needed to manage insurance schemes must be considered carefully; · Systems are needed to control fraud and abuse by both providers (excessive billing) and patients (sharing membership cards with non-members); · Payers should use their purchasing power more strategically by using payment mecha- nisms that consider incentives for both provider and patient behavior; · There appear to be advantages to enrolling groups rather than individuals to avoid adverse selection; · Attention needs to be paid to ensuring that services are of adequate quality as perceived by patients; · Schemes should consider expanding coverage to include high cost services such as hos- pitalization; · If catastrophic costs are covered, larger risk pools or social re-insurance schemes will be needed to ensure the financial viability of community schemes; · Choice of provider may make voluntary schemes more attractive to the population; and · More attention needs to be paid to effective marketing of voluntary schemes. Poor people are willing to pay premiums in advance to receive coverage for uncertain future events if they can be assured of access to quality services and that the perceived benefits outweigh the costs. Participation in a community based health insurance scheme tends to increase utilization as compared to the uninsured population. Except in Rwanda where household wealth did not seem to be a determinant of enrollment, studies indicate that having to pay premiums continues to pose a barrier to the participation of poor peo- ple. Evidence from CBHI schemes indicates that overall enrollment rates are low and drop out rates are high. An improved design that increases responsiveness, controls expendi- tures, and effectively markets to the population may increase the success of CBHI schemes in SSA though more time, experience, and evidence is needed. Utilization of targeting instruments: There are a number of tools that can be utilized for pro-poor strategies, such as vouchers and performance-based contracts. Vouchers repre- sent a demand-side strategy that has the potential to increase access among low income people. By providing consumers with vouchers that entitle them to either discounted or free services, consumers are enabled to obtain essential services from a range of outlets. Establishing the rules and system which govern vouchers and reimbursement of providers calls for government leadership and increased capacity along these lines. Performance- based contracts can be used to increase coverage among certain populations. Marginal Budgeting for Bottlenecks: Despite extensive Sector Reforms, the health sys- tems in many SSA countries still fail to reach large numbers of women and children-- especially the poorest and most vulnerable--with these interventions. (continued on next page) Conclusions: Misconceptions, Constraints, and Future Opportunities 47 This situation triggers a growing demand for tools that help answer three questions: · What are the major health systems bottlenecks hampering the delivery of health serv- ices, and what is the potential for their improvement? · How much money is needed for the expected results? · How much can be achieved in health outcomes by removing the bottlenecks? The Marginal Budgeting for Bottlenecks (MBB) tool recently developed by UNICEF, the World Bank and WHO, and tested in several countries is a response to this demand and needs to be monitored and evaluated in the countries where it is being implemented. (ii) Decrease dependency on donors for selected activities. This is especially important for recurrent costs. (iii) Develop policies to work with the private sector in order a) to alleviate the burden of financing health services by the poor who already purchase services from the private sector; and b) to harness those existing private resources to achieve public health goals. (iv) Implement public health policies that have a large impact. Some very cost-effective measures could be implemented by simple policy changes (such as smoking), others need to build on existing successes and go to scale (such as for nutrition and immu- nization) in a sustained way. To improve the organizational-side of RAP arrangements, there are four primary messages: (v) Develop a culture of results rather than process, by using performance-based contracts with lower level authorities as well as with the private sector. By doing so, Governments might integrate this approach into the governments' overall decentralization process. This is a way to do away with a hierarchical bureaucratic relationship and it might yield more results. The idea is to encourage decentralization of fiscal resources based on per- formance targets of lower level authorities. Although many governments in Africa have committed to decentralization, the mechanisms to implement this process need to be improved. The flow of funds should be clearer and traceable, so that all levels know precisely the amount of resources to expect when. Performance-based contracting with districts, as was done in Burkina Faso, shows promise for increasing accountability and results. (vi) less fragmentation in the financing and delivery of health services must be ensured, wherein the Sector Wide Approach Programs (SWAPS) might be a good start; (continued on next page) 48 Resource Allocation and Purchasing in Africa (vii) provide subsidies to the poor so that they are able to pool resources, increased atten- tion to the potential capacity and contribution of health mutuals is one way to achieve this. To improve the institutional-side of RAP arrangements, here are some options for African Ministries of Health: (viii) assume a stronger role as stewards, that is provide regulation, supervision, monitoring and competitive arrangements. Develop mechanisms, guidelines, and ensure adequate training to work with the private sector. The issue is not whether services should be provided by the public sector or by the private sector. What is important is that all the people be covered by quality health services. This paper reveals that both the public and private sectors could greatly improve on ensuring that this important objective is achieved. Governments need to focus on the essential functions that only they can undertake, and harness the potential of private providers for services provision. The pri- vate sector can function effectively on a large scale only if the public sector is able to fill its role as regulator. The literature shows that ministries of health are still trying to do everything -- but that they are not able to do it. There is, therefore, the need to design simple yet precise public­private partnership strategies, a few examples of which are reviewed in this paper. Such strategies must include strengthening the public sector to: 1) set norms, 2) control quality, and to 3) design and implement regulations so as to provide an enabling environment where the private sector can serve social goals. To effectively contract service providers, governments also need the capacity to design, negotiate, manage, and evaluate the performance of contractors. An example of effective contracting that avoided the contracting capacity constraints that usually pres- ent bottlenecks in many MOHs can be seen in a nutrition project in Senegal (Box 8). By contracting a delegated contract management agency, Agetip, it was possible to mini- mize transaction costs and and ensure the work was done without increasing Govern- ment personnel. Franchises are another approach to increasing availability of services and essential medicines. Presented here is the case of Cry for the World, a franchise of drug retail out- let shops in Kenya that provides access to high quality and reasonably priced essential drugs. Because consumers must pay out-of-pocket for drugs, the poorest population does not specifically benefit. This strategy has the potential to improve access to quali- ty products at controlled prices. Third party payment such as subsidized participation in a CBHI scheme could help increase access to franchise services and products by the poorest people. Community organizations constitute an opportunity to increase coverage in Africa which still remains to be acknowledged, harnessed and contracted as service providers by governments. (ix) separate financing from provision of services in order to help with governance problems and to improve accountability. Bibliography Abt Associates (2000a). Health Insurance Atim, Chris. (July, 1998) The Contribution Becoming Popular in Africa. Health- of Mutual Health Organizations to Watch, Fall, Volume 5, No.1. Financing, Delivery, and Access to Abt Associates (2000b). Rwanda National Health Care: Synthesis of Research in Health Accounts 1998, PHR Technical Nine West and Central African Coun- Report No. 53, September 2000. tries, Abt Associates. Akin, John, Hutchinson, Paul and Strumpf, Atim, Chris, Steven Grey, Patrick Apoya, Koleman (2001). Decentralization and Sylvia J. Anie, and Moses Aikins, (2001, Government Provision of Public Goods: September) A Survey of Health Financing The Public Health Sector in Uganda. Schemes in Ghana, Partnerships for MEASURE Evaluation Project, Working Health Reform Plus. Paper 01­35. Audibert Martine and Mathonnat Jacky, Akumu, O. Some Issues of Class Equity at the « Cost recovery in Mauritania: initial les- National Hospital Insurance Fund, July sons », Health Policy and Planning; 15 1992 with data derived from NHIF (1):66­75. Oxford University Press, archives by Kraushaar (1997). 2000. Arhin-Tenkorang, D. (2001, September) Back to Office (BTO) Report of a Health Sec- Health Insurance for the Informal Sector tor Supervision Mission in Burkina Faso, in Africa: Design Features, Risk Protec- February 18, 2003. Schneidman, Miriam. tion, and Resource Mobilization, Com- The World Bank, Washington, D.C. -- mission on Macroeconomics in Health internal document). Working Paper Series, Paper No. WG3:1. Bah, Oumar. "Traçabilité des dépenses Asfaw Abay(2003). Costs of illness, demand publiques appliquées au secteur de la for medical care, and the prospect of santé en Guinée", August 2003, Ministry community health insurance schemes in of Health, Guinea. the rural areas of Ethiopia, Development Bennett, Sara and Gilson, Lucy (2001). Health Economics and Policy Vol.34. Frankfurt, Financing: Designing and Implementing Peter Lang Verlag 2003. Germany. Pro-poor Policies. DFID Health System Resource Centre, London. 49 50 Resource Allocation and Purchasing in Africa Black et al. (2003, June) "Where and why are Chao S. and Kostermans K., (2002) Improving 10 million children dying every year?", health for the poor in Mozambique, The Lancet, vol. 361, N. 9351,. Africa Region, The World Bank. Blas, Erik and Limbambala, Me (2001). The Chawla Mukesh, Ramesh Govindaraj, Peter challenge of hospitals in health sector Berman, Jack Needleman. Improving reform: the case of Zambia. Health Policy Hospital Performance through Policies to and Planning, 16(Suppl 2): 29­43. Increase Hospital autonomy: Method- Bloom, Gerald and Standing, Hilary. (2001) ological Guidelines. August 1996. "Pluralism and marketisation in the Cote, Michel. (2003, July 10) "Diagnostic health sector: meeting health needs in organisationnel PRCI (Projet de renforce- contexts of social change in low and mid- ment des capacités institutionnelles)-- dle income countries", Institute of Devel- Guinée" from SOGEMA/Canada. opment studies Working Paper 136. (SOGEMA = Canadian management con- Bossert Thomas, Mucosa Bona Chitah, Maryse sulting firm working closely on the inter- Simonet, Ladslous Mwanza, Maureen national scene with developing countries Daura, Musa Mabandhala, Diana Bows- in the fields of management, training and er, Joseph Sevilla, Joel Bauvais, Gloria systemization.) Presentation made at the Silondwa, Munalinga Simatele (2000). World Bank. Decentralization of the health system in Cry for the World Foundation ­CFW­ (2002) Zambia, Technical Paper No. 2. Bethesda, A Network of Franchised Drug Shops in MD: Partnership for Health Reform Pro- Kenya, powerpoint presentation to ject, Abt Associates. Forum 5. Bossert Th., M. Bona Chitah, D. Bowser Dar Es Salaam City Medical Office of Health (2003). Decentralization in Zambia: (2003, January). "Minimum package of resource allocation and district perform- health and related management activi- ance, Health Policy and Planning, Vol. 18 ties", January 2003. (4) De Beyer, Joy and Brigden, Linda Waverly. Edi- Bustreo Flavia, Harding April, Axelsson Hen- tors. "Tobacco Control Policy: Strategies, rik. Can developing countries achieve successes and setbacks", World Bank and adequate improvements in child health RITC, 2003, Washington. outcomes without engaging the private Decaillet François, and John May. (2000, May) sector? Bulletin of the World Health Le secteur médical privé à Cotonou, Organization. May 2003. Bénin, en 1999. World Bank, Washing- Castro-Real F, Dayton J, Demery L and Mehra ton, DC. p. 27. K. (2000) Public spending on healthcare Devarajan, Shantyanan; Dollar, David and in Africa: Do the poor benefit? Bulletin of Holmgren, Torgny (2001). Aid and the World Health Organization, 78: Reform in Africa. World Bank, Washing- 66­74. ton, D.C. Campbell, Paul, Karen Quigley, Pano Yer- Deville, Leo, Joseph Decosas, Peter Nhindiri, acaris, and MacDonald Chaora. (2000) Shelagh Leyland, and Bernard McLough- Applying Managed Care Concepts and lin. (1997, November) "Review of the Tools to Middle and Low Income Coun- Health Sector of the Republic of Zimbab- tries: The Case of Medical Aid Societies in we", Health Research for Action Zimbabwe, Harvard School of Public (HERA). Health report. Diop, Francois, Ndiaye, Colonel and Grunde- mann, Christophe (2001). Décentralisa- Bibliography 51 tion et système de santé au Sénégal: Une Financing Reform in South Africa and synthèse. Bethesda, MD: Partnership for Zambia. Major Applied Research, Tech- Health Reform Project, Abt Associates. nical Paper No. 3. Bethesda, MD: Part- Demissie Habte, Mary Mulusa, and Paul nership for Health Reform Project, Abt Hutchinson (1999). Health Care in Ugan- Associates. da: Selected Issues. World Bank Discus- Global Fund to fight AIDS, Tubercolosis, and sion Paper 404. World Bank, Washington, Malaria (2002). Initial Questions on the D.C. Global Fund: FAQ. Geneva. Easterly, William (2001). The Elusive Quest Gilson, Lucy, Joseph Adusei, Dyna Arhin, for Growth: Economist's Adventures Charles Hongoro, Phare Mujinja and Ken and Misadventures in the Tropics. MIT Sagoe, "Should African Governments Press. Contract Out Clinical Health Services to Eichler, Rena (2001, October) Improving Church Providers", in Private Health Immunization Coverage in an Innovative Providers in Developing Countries: Serv- Primary Health Care Delivery Model: ing the Public Interest, ed. Sara Bennett, Lessons from Burkina Faso's Bottom Up Barbara McPake, and Anne Mills, Zed Planning, Oversight, and Resource Con- Books: London and New Jersey, p. trol Approach that holds Providers 276­302, 1997. Accountable for Results. World Bank, Gottret Pablo and Preker Alex (2003, May) Washington, DC. "Accelerating progress towards the MDG Engida, Endale and Mariam, Damen (2000) and health financing issues", DRAFT. Assessment of the Free Health Care Pro- The World Bank, Washington, D.C. vision in Northern Ethiopia. Presented to Gwatkin, Davidson (2000). Health inequalities CERDI, France. CERDI=Centre d'Etudes and the health of the poor: What do we et de Recherches sur le Developpement know? What can we do?. Bulletin of the International. World Health Organization 78 (1): 3­18. Enright, Michael, Daniel Kraushaar, Derek J. Gwatkin, Davidson (2001). Overcoming the Oatway, Ikiara Gerrishon. (1994) Devel- inverse law: Designing Health Programs oping Prepaid Health Programs in Kenya: to Serve Disadvantaged Population A Private Insurance Assessment, PROFIT Groups in Developing Countries. World Project, USAID. Bank, Washington, D.C. Ethiopia Survey: Roll Back Malaria: Study on Gwatkin Davidson (2002). Presentation made Malaria Financing and Expenditures in at the Liverpool School of Tropical Med- Ethiopia, KUAWAB Business Consultants icine, Leverhulme Lecture, September 17, (PLC), Addis Ababa under contract with 2002 the World Bank's Malaria Team, August Habte Demessie, Mulusa Marie, and Hutchin- 2002. son Paul (1999) Health Care in Uganda: Fiedler, Jack, Anne Levin and Dennis Selected Issues. 1999. Mulikelela. (November 1998) A Feasibil- Hahmed, Mohamed and Soucat, Agnes (Févri- ity Analysis of Franchising the PROS- er 2004). Santé et Pauvreté en Mauri- ALUD/Bolivia Primary Health Care Ser- tanie: Analyse et Cadre Stratégique de vice Delivery Strategy in Lusaka, Zambia, Lutte contre la Pauvreté. Ministre de la Technical Report No. 15, Partnerships for Sante et des Affaires Sociales de la Planifi- Health Reform. cation, de la Coopération, et des Statis- Gilson, Lucy et al. (2000). The Dynamics of tiques. Policy Change: Lessons from Health 52 Resource Allocation and Purchasing in Africa Hanson, Kara and Berman, Peter (1998). Pri- goals", Concept Note, Unicef/World vate health care provision in developing Bank/WHO, 2003 countries: a preliminary analysis of levels Konde-Lule, Joseph, and Okello, David. and composition. Health Policy and Plan- (1998). User Fees in Health Units in ning, 13(3): 195­211. Uganda: Implementation, Impact and HSPH, Issue Brief: Policy process of health Scope. Bethesda, MD: Partnership for reform in Latin America, Data for Deci- Health Reform Project, Abt Associates. sion Making, Harvard School of Public Kraushaar, Daniel (1997, February) The Kenya Health. Guidelines for Enhancing the National Hospital Insurance Fund: What Political Feasibility of Health Reform in Can We Learn from Thirty Years' Experi- Latin America. June 2000 ence?, REDSO/BASICS consultant report, Hutchinson, Paul and Ignacio Parker (1997, February 1997. February) Review of the Community Kihombo, Aggrey (2002, October) Communi- Health Fund Pre-Test in Igunga District, ty Health Insurance in Tanzania, Enroll- Government of Tanzania/World Bank ment Determinants: The Case of Kilosa support to Tanzania Health Sector District, Dissertation to be submitted to Reform in Health Care Financing..L'Intel- the Heller School of Social Policy and ligent, No. 2151, April 1­7, 2002,p. 48. Management of Brandeis University, International Labour Office (2002). Extending DRAFT. Social Protection in Health Through Kumaranayake Lilani et al. (2000) « How do Community Based Health Organizations. countries regulate the health sector? Evi- ILO-Universitas, Geneva. dence form Tanzania and Zimbabe", ILO (2002a), International Labor Organiza- Health Policy and Planning; 15 tion, "Key Indicators of Labour Market (4):357­367, 2000. 2001­2002", Geneva. L'Intelligent, No. 2151, April 1­7, 2002, p. 48. International Monetary Fund (2002). Actions Marchant, Tanya, Joanna Armstrong Schellen- to Strengthen the Tracking of Poverty- berg, Tabitha Edgar, Rose Nathan, Salim Reducing Public Spending in Highly Abdulla, Oscar Mukasa, Hadji Mponda, Indebted Poor Countries (HIPCs). Inter- Christian Lengeler. (2002, February) national Monetary Fund, Washington, "Socially Marketed Insecticide-Treated D.C. Nets Improve Malaria and Anaemia in Joint United Nations Program on HIV/AIDS Pregnancy in Southern Tanzania", Tropi- (2002). Fact Sheet: Sub-Saharan Africa. cal Medicine and International Health, Barcelona. Volume 7, Number 2, pp. 149­158. Jutting, Johannes, The Impact of Health Insur- Marek, Tonia, Issakha Diallo, Biram Ndiaye ance on the Access to Health Care and and Jean Rakotosalama. (1999) "Success- Financial Protection in Rural Areas of ful Contracting of Prevention Services: Developing Countries: The Example of Fighting Malnutrition in Senegal and Senegal, Center for Development Madagasgar", Health Policy and Plan- Research report, December 2001. ning; 14(4): 382­389. Knippenberg R., Soucat A., Vanlerberghe W., Marek, Tonia, O'Farrell Catherine, Yamamoto et al., "Marginal budgeting for bottle- Chiaki, Zable Ilyse. "Snapshot of Trends necks: a tool for performance-based plan- and Opportunities in Public­Private Part- ning of health and nutrition services for nerships for Health Service Delivery in achieving the millennium development Africa," to be published in 2004. World Bank, Washington, D.C. Bibliography 53 Marek, Tonia and Yamamoto Chiaki. "Private Ministry of Health and Child Welfare Harare, Health Policy v. Regulatory Options for Zimbabwe (1997, December). National Private Participation," Viewpoint Note Health Strategy for Zimbabwe 1997- No. 264, June 2003. The World Bank, 2007. Washington, D.C. Mudyarabikwa, Oliver (2000). An examina- Marsden, Andrew (2002, January) Cry for the tion of Public Sector subsidies o the Pri- World, Kenya: Logistics and Information vate Health Sector: A Zimbabwe Case Technology Assessment, report for the Study. Equinet Policy Series No. 8. SEAM (Strategies for Enhancing Access Musau, Stephen N., (1999, August) Communi- to Essential Medicines) Project, managed ty-Based Health Insurance: Experiences by Management Sciences for Health. and Lessons Learned from East and Matsiko, Augustine, Elioda Tumwesigye, and Southern Africa, Partnerships for Health Charles Rwabukwari. (2001, April) Reform, Technical Report No. 34. Assessing Health Pre-payment Schemes in Mwesigye, Frederick (1999). Priority Service Uganda, Health Planning Department, Provision Under Decentralization: A Case Ministry of Health, Kampala, Uganda, Study of Maternal and Child Health Care supported by the Delivery of Improves in Uganda. Bethesda, MD: Partnership Services (DISH) project and the World for Health Reform Project, Abt Associ- Health Organization. ates. McLaughlin, Julie. (2000, August) The Evolu- ESA National Health Accounts, 2000. Nation- tion of the Sector-Wide Approach al Health Accounts in Eastern and South- (SWAPs) and Explaining the Correlation ern Africa: a comparative analysis, ESA between SWAps and Reform Initiatives- NHA Country Teams, 2000. Background. Note for the World Bank Naimoli Joseph. (2003) "Benchmarking in the Institute Core Course on Population and health sector: a tool for improving World Reproductive Health and Health Sector Bank operations," May 3, 2003. The Reform, Washington, DC. World Bank McLaughlin, Julie. (2004) Improving Health, Ndegwa Stephen and Levy Brian, June 2003, Nutrition and Population Outcomes for "The politics of decentralization in Africa, the Poor: The Role of the World Bank in a Comparative Analysis", World Bank. Sub-Saharan Africa. Chapter Five: Sus- Nyonator, Frank and Kutzin, Joseph (2000) tainable Financing for Health, Nutrition Financement des systemes de sante dans and Population Interventions. Pending les pays a faible revenue d'Afrique et Publication, World Bank Africa Region, d'Asie. Presented to CERDI (Centre d'E- Washington, DC. tudes et de Recherches sur le Developpe- Mills, Anne (1997). "Contractual Relation- ment international), France. ships Between Government and the Com- Palmer Natasha et al. (2001) A new face for mercial Private Sector in Developing private providers in developing countries: Countries", in Private Health Providers what implications for public health? Bul- in Developing Countries: Serving the leting of the WHO, 2003, 81 (4) quoting Public Interest, ed. Sara Bennett, Barbara Cornell et al. "National Health Accounts McPake, and Anne Mills, Zed Books: -- The Private Sector Report". Technical London and New Jersey, p. 189­213. Report, Health Economics Unit, Universi- Ministry of Health (2001). IMCI Multi-Coun- ty of Cape Town, 2001. try Evaluation Study, Uganda. Pearson, Mark (2002). Allocating public resources for health: developing pro-poor 54 Resource Allocation and Purchasing in Africa approaches. DFID Health Systems (CBHI) on Financial Accessibility to Resource Center, London. Health Care in Rwanda, Partnerships for Peters, David (1998). The sector-wide Health Reform Technical Paper. approach in health: What is it? Where is Schwabe, Christopher et al. (2003, May) it leading? World Bank, Washington, "Revue des dépenses du secteur de la D.C. santé en Guinée", Medical Care Develop- Picazo, Oscar (2002) "Better health outcomes ment International, Washington, D.C. from limited resources: focusing on prior- Shaw, Paul and Griffin, Charles (1995). ity services in Malawi", Africa Region Financing Health Care in Sub-Saharan Human Development Working Paper Africa through User Fees and Insurance. Series, The World Bank, April 2002 World Bank, Washington, D.C. Preker, Alexander; Baeza, Christian; Jakab, Shaw, R. Paul. (2002, August) "Tanzania's Melitta and Langenbrunner, Jack (2002). Community Health fund: Prepayment as Resource Allocation and Purchasing an Alternative to User Fees", World Bank (RAP) Arrangements that Benefit the Institute, Flagship On-Line Journal on Poor and Excluded Groups. World Bank, Health Sector Reform and Sustainable Washington, DC. Financing. Preker, Alexander, Guy Carrin, David Dror, Shirima, R.M. (1997) Community Health Melitta Jakab, William Hsiao, Dyna Fund in Tanzania One Year Pretesting Arhin-Tenkorang (2001, August) Health Experience in Igunga, paper presented to Care Financing for Rural and Low- the Nairobi Regional Seminar for Health Income Populations: The Role of Com- Financing. munities in Resource Mobilization and Soucat, Agnes (mars 2004). Le secteur de la Risk Sharing, Commission on Macroeco- Santé au Tchad: Analyse et Perspective nomics and Health, August 2001. dans le Cadre de la Stratégie de Réduction Quigley, Karen. (1994) Technical Assistance in de la Pauvreté. World Bank, Washington Provider Contracting and Provider Rela- D.C. tions in Managed Care Health Plans, trip Tashobya C.K., Chao S., Nabyonga J., report for the Kenya Health Care Financ- Kadama P.Y., Nguyen S. "The impact of ing Project, USAID. abolition of user fees on the demand for Quigley, Karen. (1997) Zimbabwe: New an supply of health care in Uganda", Health Care Models in a Developing paper presented at the International Country, Health Economics Association 4th World Schneider Pia, Francois Diop, Daniel Maceira, Congress, San Francisco, June 2003. and Damascene Butera (2001, March) Uzochukwu Benjamin, Onwujekwe Obinna, Utilization, Cost, and Financing of Dis- Akpala Cyril, "Effect of the Bamako Ini- trict Health Services in Rwanda, Partner- tiative drug revolving fund on availability ships for Health Reform Technical Report and rational use of essential drugs in pri- No. 61. mary health care facilities in south-east Schneider et al. (2000). Rwanda National Nigeria", Health Policy and Planning, Health Accounts 1998. Bethesda, MD: 17(4), 2002. Partnership for Health Reform Project, Wagstaff Adam. (2002, October) Health Abt Associates. Spending and the MDGs: the case of Schneider Pia and Francois Diop. (2001a, under-five mortality. May) Synopsis of Results in the Impact of Walford, Veronica and Sebastian Olikira Community-Based Health Insurance Baine. (1997, September) Review of Kisi- Bibliography 55 izi Hospital Health Society, Ministry of World Bank (2003b). Partnership within the Health, government of Uganda and public sector to achieve health objectives, Department of International Develop- World Bank, Washington, D.C., March ment, United Kingdom. 2003) Waters, Hugh, Laurel Hatt, and Henrik Axels- World Bank (2004) World Development son (2002, June) Working with the Pri- Report 2004. World Bank, Washington, vate Sector for Child Health, HNP Dis- DC. cussion Paper, The World Bank, June World Health Organization (2000). World 2002. Health Report 2000 (Statistical Annex). World Bank (1993). World development report World Health Organization, Geneva. 1993 -- investing in health. New York: World Health Organization (2002a). Report of Oxford University Press. the Commission of Macroeconomics and World Bank 1993a. Better Health in Africa, Health (page 3, reference 3) (Add) World Bank, 1993 World Health Organization (2002b) The World Bank. (2001, June) Eritrea Health Sec- World Health Report 2002, Annex 5: tor Note, June 2001. Washington, D.C. Selected National Health Accounts indi- World Bank Ghana Project Appraisal Document cators for all Member States, estimates (PAD) for Second Health Program Support for 1995 to 2000. Geneva Project (2002), by Francois Decaillet. World Health Organization (2002c) -- World World Bank (2002). World Development Indi- Health Report Annex 5: Selected National cators 2002. World Bank, Washington Health Accounts indicators for all Mem- D.C. ber States, estimates for 1995 to 2000 World Bank (2002a). Multi-Country Reports Xiao Ye and Sudharshan Canagarajah. (2002, by HNP Indicators on Socio-Economic June) "Efficiency of public expenditure Inequalities. World Bank, Washington, distribution and beyond: a report on D.C. Ghana's 2000 public expenditure tracking World Bank (2002b). World Bank Portfolio survey in the sectors of primary health Indicators. World Bank, Washington, and education", Africa Region Working D.C. Paper Series No. 31. World Bank (2002d). World Bank Operation Database. World Bank, Washington, D.C. World Bank (2003) World Development Report 2003. World Bank, Washington, DC. p. 94.