Increasing clients' power to scale up health services for the poor: THE BAMAKO INITIATIVE IN WEST AFRICA Background paper to the World Development Report Rudolf Knippenberg, Fatoumata Traore Nafo, Raimi Osseni, Yero Boye Camara, Abdelwahid El Abassi, Agnes Soucat Overview 1. The Bamako Initiative, implemented in varying degrees in half of the countries of Sub-Saharan Africa since the late 1980s, has shown that organized communities can help sustain local public health services, not only by contributing resources but also by exerting power over the management and delivery of services. Experiences in Benin, Guinea and Mali, of communities organizing themselves to help run and finance primary health services, have strengthened the delivery of public services by establishing a sustained partnership between the state and organized civil society groups. The action of communities complements, and sometimes substitutes for, the action of over- extended states with poor institutional capacity, and stagnant and/or unpredictable financial resources. The available public infrastructure has been matched to the availability of resources in the household, with the proviso that household contributors can have a say over the running of these services. 2. Significant results have been achieved in these countries in terms of sustained health care availability for the poor, even with a relatively weak voice from the poor in the community. Client power has been organized from the top down; however, locally generated resources are owned by community associations and have been used to improve the quality of services, particularly ensuring the supply of essential drugs. During the more than 10 years of implementation in these three countries, community- based services restored access to primary but also to secondary professional health services for more than 20 million people, raised and sustained immunization coverage, increased the use of services among children and women in the poorest quintile, and led to a sharper decline of mortality in rural areas compared to urban areas in both Guinea and Mali. There was almost no change in female schooling levels over this period. - 2 - 3. Much of the success can be linked to ensuring the supply of affordable essential drugs that are readily available in the health centers under the scrutiny of the committees, and the subsidizing of most essential services for women and children. Another factor has been the involvement of communities in the planning and management of services, particularly immunization and maternal child health interventions. A flexible implementation process and ongoing learning process have contributed to the evolution and expansion of these experiences, integrating new dimensions and feeding into the overall sector reform process and poverty reduction strategies Context and background 4. At the onset of independence, most governments in Sub-Saharan Africa pledged to provide comprehensive health care to all, as part of the new states' emerging social contract. The first two decades after independence saw West African countries such as Benin, Guinea and Mali--which had socialist regimes at the time--experiment with marked progress in the health services infrastructure, where before there were only a few hospitals and some communicable disease services ("Large Endemics" services) dealing with sleep sickness diseases, or leprosy. However, serious disruptions to the situation of the public system occurred during the 1980s as a result of a severe economic recession and financial indebtedness. The health budget in Benin went from US$ 3.31 per capita in 1983 to US$ 2.69 in 1986, out of which more than US$ 2 was earmarked for paying inadequate salaries. In Mali, rural infrastructure was almost non- existent, and in Guinea, health services had almost totally disappeared--except in the capital city, Conakry--during the last years of the Sekou Toure regime. Donors became increasingly involved in augmenting their funding for health, but usually through a project approach, often imposing multiple priorities, strategies and management plans. 5. Following the 1978 declaration of Alma Ata calling for "health for all", a massive amount of work needed to be done. These countries did not have the human resources to start large-scale implementation; Mali had at most one doctor for 25000 people, more interested in hospital work than in rural care delivery. Health Service Providers were generally highly de-motivated. Being posted to rural areas was perceived as a - 3 - punishment, and at the first opportunity one "escaped" to an urban position. Salary payments were increasingly delayed, sometimes by several months. Health centers and hospitals had to make out without running water, electricity, equipment vaccines and drugs. Fuel shortages made it impossible for supervisors or vaccination teams to do their work. The classic image of health centers with long queues of patients and overworked nurses had become an anachronism. The 1980s were a time of decrepit buildings with midwives knitting, the nurse having left the clinic to cultivate his field. 6. The vast majority of poor families in the three countries did not have access to affordable, quality health services or drugs. Purchasing power had plummeted and utilization of services decreased. Immunization coverage was under 15%, and less than 10% of families made at least one visit per year to the existing public health services. The few patients that still visited knew that they would have to pay for the "free services" and that there would be no drugs in the health unit. They were channeled to parallel markets. Out-of-pocket spending on health became the norm. As health services in public health centers were officially "free", clients­mostly illiterate--had little leverage over public health staff. The only available source of drugs and care was the usually informal private sector--mainly drug peddlers--where rural families typically spent U$ 5 per capita out-of-pocket per year. Similarly, communities had little "voice" in policymaking. Political parties and mass movements were largely controlled by the local political elite and their families. There were almost no organized large civil society movements and little sense of entitlement vis-à-vis public services. There was little if no communication between state organizations and traditional solidarity networks. Not surprisingly, U5MR rates were extremely high particularly in rural areas. 7. As the international community pledged to achieve 80% immunization coverage globally by 1990 (UCI 1990), national policy makers in Benin, Guinea and Mali realized that their countries were very unlikely to achieve this ambitious goal. They therefore turned to the experiences of several operations research projects in the region­ specifically the Pahou project in Benin, the Kasongo project in Zaire, and the Pikine project in Senegal--which had developed and field-tested strategies for the effective delivery of primary care services to poor rural populations since the 1970's. In 1987. with support from UNICEF, WHO and the World Bank, policy makers developed national strategies and implementation plans for the scaling up of these strategies through the - 4 - revitalization of community-owned and -managed health center networks. These strategies provided the basis for the launching of the "Bamako Initiative" by the African Ministers of Health in September 1987, at the WHO Regional Committee Meeting in Bamako. Design Changing the power relationship between the key actors 8. Given the practical collapse of government control over service provision, there was little if any potential for a renewed and strengthened "compact" between policy makers and providers. Public spending was low and institutional capacities were very weak. Most government policy makers and donor representatives involved in the design of the Bamako Initiative strategies in Benin, Guinea and Mali were quite young and new in their national positions. They had some field experience in trying to manage primary care services in a context of declining government budgets and support. They had learned "hands on" that communities hold important keys to improving services, but at the same time needed guidance in channeling their efforts to priority health actions. Based on this experience and through much brainstorming between the teams from the three countries, a consensus emerged on more "indirect routes" for policy makers to leverage providers on a large scale. Management systems were re-designed to simultaneously facilitate "direct control by communities and "indirect control" by the policy makers. 9. In all three countries, priority was therefore given at first to reestablish a direct accountability line between the population and professional providers. Communities were placed in the driver's seat in revitalizing "their" health center. From mere recipients of health care, consumers became active partners whose voice counted. A contractual arrangement or partnership was created between the state and communities to deliver basic professional health services by: · decentralizing decision-making and management from the national to the local level; - 5 - · instituting community cost-sharing and co-management of health services, increasing the accountability of local health staff to communities so as to ensure access and quality of care; · ensuring the availability of affordable medicines in a sustainable way, and decreasing the fluctuations linked to uncertainties in government transfers and donor funding--this implied full retention and ownership of funds generated in the community which managed and controlled resources through locally elected health committees. 10. Most development partners present at the time became actively involved in the design and implementation, with a key role played by UNICEF, the WHO and the World Bank, but also by bilateral donors such as Switzerland, Belgium, The Netherlands, Germany and France. They joined the effort either because they believed from the start in its principles and strategies or because they had been convinced by the initial results in the field. Reorganizing the delivery of primary professional care 11. In the three countries, health service delivery was scaled up by initially focusing on the revitalization of health centers as the primary level of interface between populations of 5,000 to15,000 people, and professional health services. The health center was seen as the "hub" of the health system, being the most peripheral facility staffed by professionals--licensed (auxiliary) nurses or midwives with a formal training of 1-2 years--and sufficiently close to villages to provide regular "outreach" and health promotion activities. 12. The service package included professional care to respond to demand for illness management as well as professional public health interventions and actions to support healthy behaviors in families. Emphasis was placed on delivering health interventions that had: (i) a proven potential impact on under five mortality and morbidity; (ii) responded to community perceived needs; and (iii) a relatively low marginal cost of delivery through three basic service modes (see Table 1): - 6 - · intensive periodic efforts for predictable professional services: there was a strong focus on preventive professional services such as immunization and antenatal care, which were offered on a regular basis in the health centers and through periodic outreach to the villages under the responsibility of the health center; · continuous professional services at facility level: the focus was on management of malaria, ARI and diarrhea particularly among children as well as skilled deliveries at health center level; and · family and community-based care: initially with the promotion of ORT and recently with more emphasis on other aspects such as exclusive breastfeeding and the family-based treatment of malaria and acute respiratory infections. Table 1: Services offered as part of Bamako Initiative clinics in 23 countries, 1999 [Insert Table 1] (no Table 1) 13. An analysis of the main constraints to effective use of professional interventions in these countries led to a particular emphasis on some key strategic components of service delivery: · implementation of drug revolving funds to increase the availability of affordable essential drugs (see below); · revitalization of existing health centers, expansion of the network and implementation of outreach by health center staff, with professional services offered periodically to villages within a 15 km radius, so as to increase geographical access; · social mobilization and community-based communication strategies to stimulate demand, especially for professional services such as immunization, and to support healthy behaviors such as oral rehydration therapy for diarrhea; - 7 - · pricing of most effective interventions--preventive interventions as well as care for children for example--below private sector prices both through subsidies from the government and donors, and internal cross-subsidies in the system; · community participation in the analysis of progress and problems in coverage by the most important services, and involvement of communities in planning and budgeting of services; · establishment of a system of defaulter tracing and tracking using "tickler files" and involving community representatives to increase demand and continuity; and · introduction of standardized diagnostic and treatment algorithms following in- service training of all health staff, and followed up by constructive supervision, to improve quality of care. Procurement and supply of quality pharmaceuticals 14. Essential drug policies were introduced in the three countries. In Benin and Guinea, pre-packaged UNICEF kits containing respectively 30 and 50 of the most essential drugs for the health center level, were distributed at half yearly intervals to each health center. The types and quantities of required drugs were calculated based on the local epidemiology, the treatment algorithms (which included standardized drug regimens) for these priority diseases, and expected utilization rates (based on the experience of pilot projects). The plan was to provide these drug kits for 3 years to each facility, for communities to cumulate enough revenue over this period to upgrade the health center, increase utilization, create a capital for the drug revolving fund and establish a buffer or security fund. To establish a sustained procurement mechanism, central medical stores (CMS) were established in the three countries to provide drugs to communities on a "cash-and-carry" basis. In Guinea and Mali, the CMS were "state owned and managed", but in Benin, an autonomous CMS was developed with support from all partners, which functions successfully to this date and has provided a model for similar independent pharmaceutical procurement agencies in other African countries such as Burkina Faso and Cameroon. Cost Sharing - 8 - 15. The cost-sharing systems were based on a community financing scheme where treatments were sold to the patients and revenues were pooled in a community account to be reinvested later in health activities. The price was based on the cost of pharmaceuticals procured as generics at very competitive prices with a variable mark-up depending on the intervention. An average 200-300 % mark up was applied to finance local operating costs, but some "less essential" treatments could have a higher mark-up. This was the case of some treatments for adults (diabetes, impotence, backache for example) which cross-subsidized other interventions. Even with this high mark-up, prices were maintained very low compared to private professional care through the supply and rationalized prescription of generic essential drugs. Immunizations and ORT were free, and curative child care as well as antenatal care and delivery care in rural areas were heavily subsidized. Governments and donors funded the investment in buildings, vehicles, training and drug stocks as well as vaccines, supervision, monitoring, and support systems. In Benin and Guinea, governments also paid the salaries of all staff, and in Mali, that of health supervisors. The communities financed the incentives for the health staff (performance bonuses) and the salaries of some additional support staff in Benin and Guinea. In Mali, even the salaries of health center nurses were financed by communities who had the power to hire and fire. 16. A central principle of the Bamako Initiative that distinguishes it from the global experience with user fees was local retention and community co-management of the entire community financing revenue, under centrally issued guidelines that ensured that the revenue was immediately channeled back into increasing access and quality of the provided health services. User contributions were not considered part of government revenue. The proceeds of user co-payments were owned by the community associations, kept in a commercial bank account, and used only with approval of the committee. In order to facilitate effective (efficient?) "oversight" by communities, management systems and tools--including guidelines for the use of revenue, financial and drug balance sheets, patient registers or cards, and monitoring indicators--were streamlined and simplified, and community committees trained in their use. 17. Decisions on exemption from payment for the poorest members of the community were left to the discretion of the Community Management Committees. They were assumed to best know who was actually unable to pay, taking into account - 9 - extended family networks that continue to be the backbone of the social security system in these countries. However, as experience increased, the focus gradually shifted from serving the general population to better identifying and reaching out to un-reached communities through monitoring and micro-planning. - 10 - Table 2: Reach out approaches used in the Bamako Initiative programs in Benin, Guinea and Mali to benefit poorest groups Disease targeting Geographical Cross subsidies Exempting the targeting poor >Focus on the >Focus on rural >Higher mark up and >Exemptions left to burden of diseases areas. co-payments on the discretion of of the poor: malaria, diseases with lower communities. diarrhea, respiratory >Larger levels of levels of priority. infections, subsidies to poorer >High subsidies for >Exempted malnutrition, regions. child health services. categories include reproductive health. >Free immunization widows, orphans. and oral rehydration therapy as well as promotion activities. Participation of the community in the local management of the health system 18. Communities were also involved in the co-management of pharmaceuticals and revenue. The community pharmacies had double locks, and required both the presence of the health center's chief nurse and a community representative to be opened, as each held the keys to one lock. Bank accounts had double signatures. Members of committees also participated twice a year in the monitoring of health services--analyzing problems and helping to design new actions--and budgeting the use of these revenues, within clear national standard guidelines. Results 19. Over the 12 years or so of implementation in Benin and Guinea, and more than 7 years in Mali, health indicators have improved quite significantly in terms of health outcomes and associated health services utilization, but also in terms of income protection as well as empowerment of communities over professional service providers. - 11 - Improved Health outcomes for the poor 20. In the three countries overall, under five mortality has declined significantly on average but also among the poorest groups. The poor/rich gap has stayed relatively stable in Benin but has decreased in both Mali and Guinea (see Figure 1). Figure 1: Under Five Mortality has decreased in all three countries both on average and among poorest groups Evolution of under 5 mortality in Benin, Guinea and Mali, 1980-2002 400 350 Mali, poor Mali, average Guinea, average 300 thousand 250 Guinea, poor per Benin, average Deaths 200 Benin, poor 150 100 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 Source: Calculated from Demographic and Health Surveys, Benin 1996 and 2001, Guinea 1992 and 1999, Mali 1987, 1996, and 2002. - 12 - 21. The reduction in U5MR has been greater in the poorest groups in Mali leading to improved equity in outcomes (see Figure 2). In Guinea the decline among the rural population has been much steeper than among urban groups (Figure 3). Figure 2 : Improvements in Under Five Mortality, particularly among the poor in Mali Trends in child mortality between 1987 and 1999 400 354 1987 1996 2000 350 342.6 322.4 315.3 300 285.7 274.4 271.8 258 261.2 251.7 250 244.6 239.5 239.8 245.5 231.9 195.8 200 164.8 163.9 150 100 50 0 Poorest Second Middle Fourth Richest Average Source: Calculated from Demographic and Health Surveys 1992 and 1999, (based on births in the last five years before the survey). - 13 - Figure 3 : Improvements in Under Five Mortality, particularly among the poor in Guinea Trends in child mortality between 1992 and 1999 250 229.2 shtribevil0001repshtaeD 1992 1999 200 176.9 150 136.3 107.5 100 98 87.5 66.6 69.7 50 48.4 49.6 0 Neonatal Post-neonatal Infant 1-4 yr Under5 mortality mortality Trends in infant mortality by wealth quintiles 180 1992 160 1999 140 births 120 live 100 1000 80 per 60 40 Death 20 0 Poorest 2 n d 3rd 4th Richest Wealth Quintile Source: Calculated from Demographic and Health Surveys 1992 and 1999 (based on births in the last five years before the survey). - 14 - 22. Immunization levels have increased in all three countries. They have been maintained at a very high level in Benin over the 10 years, both on average-close to 80%, attaining one of highest level of immunization among the poor in the Africa region. Guinea and Mali have maintained a somewhat lower level of immunization largely due to problems of access which still exist. Mali is a very large desert country and access to health services at less than 5km is still below 50% today. However, immunization coverage has nonetheless increased even among the poorest 20% in Mali. This increase being greater than the increases in the richer groups, equity in immunization coverage has somewhat improved (see Figure 4). Figure 4: Steady improvements in immunization coverage in Benin, Mali and Guinea, Benin being among the best performers in Sub-Saharan Africa Evolution of national immunization coverage (DPT3) 1988-1999 100 90 80 70 60 Benin % 50 Guinea Mali 40 30 20 10 0 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 Year Source: WHO, UNICEF and Demographic Health Surveys. - 15 - 23. Coverage with other health interventions has also increased for all three countries. Utilization of health services by children under 5 years of age in Benin increased from less than 0.1 visit per year to more than one visit per year (Figure 5). In Mali rates of exclusive breastfeeding as well as utilization of professional services for antenatal care, deliveries and treatment of diarrhea and ARI have also improved in all groups including the poorest (see Figure 6). Figure 5: Increased utilization of services for children 4 years' old and under Evolution of number of visits for children Benin 1993-2001 70 60 50 children 40 100 for 30 Visits 20 10 0 1993 1994 1995 1996 1997 1998 1999 2000 2001 Source: National Information system MOH Benin. - 16 - Figure 6: Mali increased utilization of various health interventions in all groups including the poorest quintiles Exclusive breastfeeding up to 4 Medical treatment in case of ARI and months per wealth quintile diarrhea per wealth quintile 1996 Trt. for Diarrhea 35 1996 70.0 2000 Trt. for Diarrhea 30 2000 60.0 1996 Trt. for ARI 25 50.0 2000 Trt. for ARI 20 40.0 15 30.0 10 20.0 5 10.0 0 Poorest Second Middle Fourth Richest Average 0.0 Poorest Second Middle Fourth Richest Average Antenatal care by medically trained Attended delivery by medically trained persons per wealth quintile persons per wealth quintile 1987 100.0 90.0 90.0 1996 80.0 80.0 1987 1996 70.0 2000 70.0 60.0 60.0 2000 50.0 50.0 40.0 40.0 30.0 20.0 30.0 10.0 20.0 0.0 10.0 Poorest Second Middle Poorest Second Average 0.0 Poorest Second Middle Fourth Richest Average Source: DHS 24. In an independent evaluation conducted in 1996 in Benin, 75% of the key informants declared being satisfied with the quality of care, although 48% were not fully satisfied. Healthcare users leaving the facility found drug availability to be high--over 80% said drugs were available--and the overall quality of care to be good--91% said care was of good quality. However, while health workers had a good level of technique, it was found that the welcome given to sick people and the general courtesy offered to the population was not perfect (Gilson 2000). - 17 - Protection of household income 25. There is also broad evidence that relative affordability has improved. According to data from focus groups in Benin, improved access led to decreased indirect costs linked to travel. Improved drug availability also reduced the need to visit one or more distant sources of care. Prices have been continuously under the level of other alternative treatment source. Median household expenditures on curative care in a health center in 1989 was found to be US$ 2, less than half the cost of private providers (US$ 5) or a traditional healer (US$ 6.7). Thanks to continuous subsidization by the central level and cross-subsidization between services, curative care prices were set at about average cost of delivery. Price of antenatal care (US$ 1.7) on the other hand was set at much less than average cost (US$ 7). Payments for immunization (US$ 0.16 for the card) were negligible compared to an estimated cost of US$ 11 per immunized child. As the poor benefited largely from preventive services, the cross-subsidization of preventive and promotive services has overall contributed to increased equity. More economically comfortable groups also use the services to a larger extent but this utilization contributed to increased benefits in terms of preventive care for the poorest groups. 26. In 1996, seventy percent of health workers interviewed in Benin confirmed that in accordance with national guidelines, surplus generated by the co-payment system was used to buy essential drugs, small items of equipment and to make payments to locally recruited temporary staff. A proportion was also channeled to pay performance-based incentives to health staff (on the basis of performance in immunization, antenatal care and deliveries). External evaluation found that 91% of the informants found the standard pricing method appropriate. Nearly all of them preferred to maintain the practice of charging everyone the same price for services. 27. Credit was also offered at times, including deferred payments. Sixty two percent of health workers in Benin indicated that they would defer payments if a patient could not pay. However, even though relative affordability improved, and preventive and curative child care services were heavily cross-subsidized, some of the poorest families were still excluded from curative and delivery care, which constitutes and absolute equity issue (see Box 1). - 18 - Box 1: In Benin, health centers have thought about the poor when setting prices. Nonetheless, wherever there is a fee for service, it is clear that not everyone has access to it. "Price levels seen as generally acceptable" 69% of service users 36% of community-based key informants 16% poorest groups "Price levels seen too high": 10% or service users 48% of community-based key informants 69% poorest groups "Price level allows access by all" 98% felt prices allow access by all 87% poorest group felt the levels did not allow access by all. Source: Gilson and others 1999. Empowerment of poor families and communities 28. Focus group discussions in Benin, Guinea and Mali also confirmed some progress in terms of better empowerment of communities, although there is still a perceived exclusion of the poor from the co-management process. Throughout the years, the funds generated by community cost-sharing mechanisms have been managed by the committees with relatively few examples­estimated at less than 10%-- of serious mismanagement. Decisions on levels of performance bonuses were made locally. This sense of community ownership even led to complaints from health providers who did not feel in control any more of their health center but rather perceived themselves as an employees of the community. As they say in Mali, the community has become "the boss of the truck". - 19 - 29. Top-down organization of health committees in the community tends to perpetuate elitism, and the rest of the population can tend to feel marginalized by this process. Clearly most of the management committees have been initially colonized by the local elite including retirees and civil servants who often co-manage the centers in agreement--and sometimes in collusion--with the health personnel. However, over time representativeness of community management committees has improved, following clear guidelines from the policy makers in all three countries. The Ministries of Health support the inclusion of women and young people, as well as representatives of distant villages in the committees; they regularly encourage renewed election processes. A remaining problem is nonetheless the weak "voice" of the poorest families and individuals who have the least time to participate in meetings or other voluntary community support activities. The process of Implementation Incremental scale-up 30. The scaling up of services required an adaptation of the strategies developed in various pilot projects1 to different country contexts. This involved the strengthening of support systems, a phased implementation steered by networking, peer support and an ongoing learning-by-doing process. Through this process, the approach was progressively scaled up in the three countries, evolving from an initial 44, 18 and 1 to 400, 367 and 559 in 2002 respectively in Benin, Guinea and Mali. This increased access of the population to functioning professional services at less than 5km to 86%, 60% and 40% respectively in these three countries (see Table 3). 1 In the early 1980s, successes in delivering primary care services were first analyzed based on the experiences of the Narangwal, Lampang and Bohol projects in Asia, as well as the Danfa, Kintampo, Kisantu, Kasongo and Institute of Child Health Nigeria projects in West Africa. This approach was particularly inspired by the Zaire experience with over 200 "health zones" where communities and nurses jointly managed health services "for less than one dollar per year". The translation of these best practices into a coherent set of service delivery strategies, management systems and instruments was then done in the Pahou pilot project in Benin (1982-1986). - 20 - Table 3: Evolution of the number of community-managed and/or -owned health centers in Benin, Guinea and Mali Year `88 `89 `90 1991 1992 1993 1997 Access 2002 Access <5km <5km Benin 44 200 200 200 250 250 304 80% 400 86% Guinea 18 98 164 214 256 346 42% 367 60% Mali 1 4 11 363 25% 559 40% 31. Health centers are typically staffed with two male nurses and one midwife. In Guinea, some medical doctors are also responsible for large centers. Support to implementation, supervision and monitoring is conducted by district medical officers. Human resources were not initially seen as a major constraint. Health professionals were initially enthusiastic to be able to have access to resources and benefit from additional financial incentives. However, difficulties in recruitment developed as the number of facilities increased, and training efforts slowed down following budget cuts. In Mali, nursing schools significantly decreased their enrollment, leading to no graduations for a few years. In the three countries, the state had also reduced its recruitment, leading to switching salary costs increasingly to the community level. 32. The concept and strategy also evolved overtime. Initially, the approach focused essentially on professionally delivered interventions through facility-based care and outreach. Delivery evolved progressively towards more focus on family care practices, including the promotion of impregnated bed nets, exclusive and prolonged breastfeeding and hygiene, community-based treatment of malaria, and home-based pneumonia management. Benin and Mali are in the process of training and deploying "village level educators" to promote these family practices, and support caretakers through home visits. Innovative interventions have also been progressively integrated. Intermittent presumptive treatment for Malaria in pregnancy and Vitamin A supplementation have been integrated into the activities to be delivered through outreach. Emergency obstetric care and referral systems have been progressively strengthened. In Mali, a cost-sharing system between communities, the local government and the MOH was specifically developed to ensure prompt referral of obstetric emergencies. In Guinea, community- - 21 - based pregnancy insurances (Muriga) were established to cover this risk. A partnership was established with local transport companies for driving women to the hospital. 33. While efforts initially focused mainly at the meso level of implementation (support to the service providers), the government also sustained the process by revising national policies and redesigning national support systems. A legal framework supporting the contractual relationship with communities, cost-sharing arrangements, essential drugs and community participation policies was developed. Community management committees were legalized, and their attributions formalized. The cold chain was reorganized to ensure an uninterrupted supply of vaccines. Training, supervision and communication materials were developed and integrated into the national curriculum for the training of nurses. Administrative and legal guidelines were also developed to specify the status of community associations and committees, their composition and their roles. Peer providers from "best performing health centers and districts" were actively involved in training and supervising their colleagues from new or weaker performing health centers or districts. These processes constituted important entry points for health sector reform and donor coordination. In Mali, this has culminated in a formal National "Health Sector Reform" since the end of the 1990's, which is now feeding into the process of implementation of the national poverty reduction strategy and its associated budgetary framework2. Delegation and Financing 34. The total quantity of additional funds generated by community financing represented only a small fraction of the cost of running the health centers (less than 20% of total recurrent costs). However, the incremental readily available cash was instrumental in protecting the centers from the unpredictability of central funding. Funds generated and retained locally enabled the financing of critical recurrent costs at the periphery on a continuous basis. The "cost items" to be financed directly by the communities were carefully chosen based on best practices. Communities controlled most of the recurrent costs, allowing uninterrupted and reactive management. Changing light bulbs, buying soaps and hiring support staff could be done anytime and only 2 A Mid-Term Expenditures Framework was developed recently to fully integrate the contractual approach into the overall budgetary vision. This expenditures frameworkwill support the implementation of the recently launched National Poverty Reduction Strategy. - 22 - depended on a local decision. Community control over pharmaceuticals, petrol and mobility allowances for outreach allowed preventive activities such as immunization to be run in a sustainable way. Day-to-day immunization activities became independent of donor financing. The control of users over performance bonuses in all three countries and health workers' salaries in Mali also provided substantial leverage for users over the health workers. 35. Community financing of key operational costs also served to buy the communities a "seat at the table" and an increased "voice" in discussions with the policy- makers. This voice appeared essential in adjusting the system to the successive pressures by donors to implement special initiatives and push for the achievement of international goals. Donors and governments had to systematically negotiate new activities with community organizations. For example, when donors and the Government of Benin attempted to accelerate immunization coverage through more "vertical" campaign approaches, community representatives strongly refused to do so. Instead, a consensus was reached to accelerate the scaling up of the Bamako Initiative Strategies to all existing health centers over a period of two years, and to ensure that all centers guaranteed at least 80% immunization coverage in their catchment area. This "community imposed" strategy not only led Benin to reach UCI(Universal Childhood Immunization) goals in 1990, but has allowed it to become one of the few African countries to continuously sustain a high immunization coverage up to the present day. 36. However, the governments did not just give up their responsibilities in guiding and strategically financing the system. The partnership principle also gave a very strong role to the policymaking and service purchasing function. In all three countries, the government--with the support of donors--has continuously provided subsidies to the health centers, particularly providing a larger level of subsidy to capital formation for the drug revolving funds to the health centers located in the poorest regions. In Benin today the public subsidy to health services is largely equalized on a per capita basis between rich and poor populations (see Figure 7). - 23 - Figure 7: Equalized public subsidy to regions in Benin Public subsidy to regions in Benin, per capita, 2001 2.5 2 1.5 1 0.5 0 CouffoBorgouPlateauAtlantiqueLittoral Alibori Donga Mono Zou Collines Atacora 37. Given that richer regions have more infrastructure, this involved a purposeful effort at reallocating resources to the poor regions. In Benin and Guinea, the central level provided motorbikes to health center nurses to conduct outreach. This motorbike however was not simply given to the health center or the community association. It was in fact given to the nurse himself on a "lease-purchasing basis" as part of an overall incentive scheme. Each outreach session conducted by the nurse provided a cash bonus that was saved against the cost of the motorbike. This provided important leverage over providers for both policy makers (who provided and replaced the bikes) and communities (who verified that outreach sessions were actually carried out as planned). The indirect leverage of policy makers on providers was also strengthened by a strong central guidance as to how community financing funds were to be used. Local budgets included "compulsory" cost items to be funded and controlled by the communities, which included gasoline for outreach and performance bonuses linked to achievements in immunization coverage, ANC, and quality of curative care. These cost items had to be first included in the budget. Only the remaining surplus could be used for additional discretionary activities chosen by the communities. - 24 - Co-production of services. 38. The kit system to ensure drug supply proved useful to "jump start the system". Once certain drugs began running out and others cumulated (close to expiration), exchanges between health centers were encouraged under the control of community management committees which owned and co-managed the drugs. This further strengthened community ownership over the management of pharmaceuticals. Drug consumption however rapidly exceeded initial forecasts and a need for additional drugs arose in many centers, which could no longer be resolved through inter-center exchange. Communities began using their "voice" to put pressure on policy makers to find a solution. This triggered the development of "cash and carry systems" at the central medical stores, initially using UNICEF-provided drugs through UNIPAC, and later through the development of competitive procurement. Community representatives were also involved in the planning of services. In Mali, the Government negotiates with communities their adhesion to the community health associations in charge of the new health centers. The hiring and firing capacity of management committees in Mali also allowed communities to exert tight control over the effective presence of health staff. Absent health workers would see their salaries being cut, their belongings confiscated or their house locked. In the three countries "micro-planning" of outreach sessions to the villages was conducted twice a year with the participation of representatives of each village. 39. The central governments on the other hand fully kept the training and supervision functions under their control. Training and supervision were however conducted jointly for health workers and community members. Training modules for different levels of community involvement (pharmaceuticals and revenue management, planning of outreach sessions, monitoring) helped strengthen the capacity of community members to exert control over local providers as a complementary mechanism to regular supervision from the center. Particular attention was also given to developing peer support and control. Peers from well-functioning committees and health centers were frequently co-opted as facilitators and trainers in training and supervision, with an attempt at rebuilding, through regular exchanges and emulation, mechanisms of self regulation and an ethos of community service. - 25 - Monitoring 40. Local monitoring systems have been developed in all three countries in which health center staff, community representatives and district supervisors meet twice a year and jointly review the performance of the health center. Central and regional managers, with health staff and community representatives, jointly compile data from the health management information system in order to: · assess achievements and problems in coverage using indicators of performance for some key health interventions (treatment of child illnesses, immunization, assisted deliveries, antenatal care, family planning, etc.) as well as some efficiency measures, such as the average cost per drug prescription or per outreach session and the average revenue per curative care user; · analyze the evolution of coverage over time and identify plausible causes of coverage problems, based on simple visualization of the dimensions of performance (physical access, availability of drugs, utilization of services, compliance, quality) as a series of steps; and · decide on corrective action for the problems identified on the basis of knowledge of context specific opportunities and constraints, lessons learned in the past and experience of other health centers. As part of this third step, implementation of the corrective action is jointly micro-planned by the health providers and community representatives with specification as to the responsibilities of health staff, communities and supervisors, costing of the actions and budgeting of the funds required from the community financing revenue. 41. Participation of community representatives in these monitoring sessions has also proved important in order to facilitate the understanding of demand-side problems, select appropriate corrective actions and maximize the likelihood of actual implementation of these actions by and with the community. It also helps leverage the action of health providers for implementation of jointly agreed upon supply side activities. After each health center has carried out its own local monitoring, wrap-up sessions are organized at the district and regional levels where representatives of each health center present their results and micro-plans for corrective action. These wrap-up sessions facilitate the - 26 - exchange of experience and emulation in achieving results. The support required from regional/national levels and donors is planned at that time. Finally, the local participatory process provides a platform to voice concerns and needs for central government support. An example of this has been the complaints of community associations in Mali about the shortage of trained health staff, where it proved impossible to find sufficient candidates for additional health center nurse posts funded through debt relief in 2001. In response to pressure from the communities, the MOH in Mali and key donors agreed to support an increase in the capacities to train nurses at the regional level, and to make it more attractive for trained nurses to reside in rural areas through salary increases and hardship allowances. Enforcement and regulation 42. Very early on, a heavy reliance on "decentralized decision-making" and "community co-management" processes to achieve internationally set goals was a reason of concern for the policy makers. They were well aware of the weak management capacity, disfunctioning legal environment and degraded public ethos, as well as unstable governance environment. Combining the overall incentives framework with adequate check and balance mechanisms was a central preoccupation. Hence, health policy makers developed a set of strategies to balance the enhanced power of community representatives with stronger control by the state. Before opening a new center, central planners visited communities and checked whether the minimum conditions of the partnership had been fulfilled, including the establishment of a representative management committee, the rehabilitation of the building and the creation of an adequate pharmacy (including secure doors and windows, as well as double locks). On several occasions, the central team came to villages to deliver a stock of drugs and refrigerators only to find that the conditions had not been fulfilled. After discussion with the local authorities, they left the village with the equipment that was then diverted to another health center. When the team returned to the first village one month later, the conditions had been fulfilled. 43. The bi-annual monitoring sessions--comparing achievements in coverage, costs and revenue between centers--also triggered more in depth "audits", with a drill-down exercise among the weakest performing centers. In Benin, several health centers had - 27 - an average cost of pharmaceuticals per case that was at odds with their average revenue. An in-depth exploration of the issue led to sanctions and replacement of both health staff and committees when found at fault. In Guinea, the State Secretariat for Decentralization was charged with the monitoring and supervision of financial management, as they--in contrast to the MOH--had the legal authority to take legal action in case of fraud. 44. The six months "micro-plans" which specified the activities of the health center also included the specific accountabilities for the implementation of actions to correct the problems identified by monitoring, with verification of implementation at the next monitoring. Performance bonuses were withheld if coverage was incomplete as observed by the "shrinking stacks of MCH cards from one-time users to fully covered" or if there were discrepancies in the management of drugs and/or finances. The discretion of community management committees to pay or withhold salaries (in Mali) or performance bonuses to health center staff (in Benin and Guinea), based on their performance in service delivery and accountability for the management of the drugs and funds handled by the health staff, proved to be a powerful tool to hold health staff accountable. The control has become so powerful today that the MOH of these countries are now engaged in developing mechanisms to counterbalance the abuse of power of some communities, and support and defend the professionalism of health staff. 45. Enforcement also included mechanisms for positive recognition. Best performing nurses or supervisors were appointed as peer supporters ("renforts") to help their colleagues who encountered more difficulties. The status carried benefits in the form of prestige, visibility, participation in higher level meetings and travel. The peer supporter status was renewed yearly, based on the performance of the staff's community, health center or district over the past year. One outcome of this "meritocracy" is the fact that many of today's national and international policy makers in these countries were themselves "peer supporters" in the late 1980 and 1990's. A summary of the various instruments used as implementation levers in these countries is displayed in Table 4. - 28 - Table 4: Examples of instruments used as levers between key actors in the Bamako Initiative INSTRUMENTS CLIENT/PROVIDER CLIENT/POLICY POLICY MAKER MAKER/PROVIDER Delegation- Community financing Community financing of Leased motorbikes financing of local operating district supervision. linked to costs. performance-based payments. Service Community co- Community co- Micro-planning; provision management of management of CMS. essential drugs drugs. Price control of drugs. policy, algorithms. Joint training. Monitoring and Joint monitoring of Advocacy for improved Comparing Health information access and quality by human resource policies. Centers' performance clients and providers. at wrap up of monitoring. Enforcement Performance bonus Associations of Peer meritocracy. and regulation or salary paid by community health Inspection tours and communities. committees. audits of problem health centers. Some remaining challenges for the key actors in the Bamako Initiative Protecting the interests of the poorest families in the client-provider interaction 46. As recognized by an external evaluation of the impact of the Bamako Initiative on equity in Benin (Gilson, 1999), the objective of the program was not set initially at reaching the poorest groups but at restoring access to affordable quality care to the majority of the rural population. This evaluation clearly showed that the approach has improved the "relative affordability" of health services, by improving access and quality of professional care and reducing its costs. Yet at the same time, "absolute affordability" remains a problem for many of the poorest families despite the various "targeting mechanisms" that attempted to orient the services towards better serving the poor. - 29 - 47. Despite sustained subsidies to supply the poorest regions, a significant proportion of the poor remains excluded from services and must rely on self medication, particularly for children illnesses. Most of children deaths still occur at home, a significant challenge in Benin where the majority of women are still illiterate. Clearly, the approach has essentially focused on restoring a reasonable level of supply of professional services but has not sufficiently emphasized support to the demand side, including healthy practices at home and compensation for low financial resources. Based on experience in other countries, options to improve the absolute affordability would be to subsidize the demand side as a complement to existing supply strategies. Multiple mechanisms could be tested, including the establishment of third party payment "poverty funds" that could help the poor access services. Subsidies to existing and nascent micro-insurances could also be envisioned. However at this point (and excepting insurance mechanisms for emergency obstetrical care), experiences with micro-insurance have not proved promising in these countries because of major difficulties. More needs to be done, tested and experimented to better include the poor. Maintaining the voice of the poor within communities 48. According to the same evaluation, Community Management Committees succeeded in strengthening community control over providers, and in giving "power" to the communities. However, the interests of the poorest families (who are often the most illiterate and marginalized, and have little spare time to participate in meetings) are not always represented by management committees. In fact, too little attention has been paid to the process of inclusion of underserved groups (including the poor) in committees, and the actions needed to support them. In the context of implementing national poverty reduction strategies (PRSPs), policy makers have expressed the willingness to take active steps to help those communities identify their poorest and excluded families, help them to organize themselves, and make their "voice heard" in the allocation of budgets and organization of services. Unfortunately, the organizational implications of their political will are not well known and will certainly require as much investment from policy makers in developing the approach as has already been done to improve the supply side. - 30 - Ensuring continuous quality professional services for the rural poor in the contexts of health staff shortages and privatization 49. In Benin, Guinea and Mali, as in many Sub-Saharan Africa countries, the expansion and improvement of professional health services has recently been severely constrained by the shortage of trained health staff. When debt relief and the HIPC (Highly In-Debt Poorest Countries Initiative) made substantial resources available to Mali to recruit and pay additional health staff for the poorest regions, there were not enough candidates. The governments of these three countries are examining some new policy levers such as the placement of schools in poor areas and recruitment of trainees with roots in those regions; recruitment by local governments; as well as the provision of hardship allowances, and other incentives to work in these areas. However, newly opened nursing schools also encounter difficulties in recruiting good secondary school candidates. Social structures also make it difficult for female workers such as nurses and midwives to work in rural areas, as they often have spouses and families to care for in urban areas. In Mali, more than 60% of midwives work in Conakry (the capital), which represents about 25% of the country's population. Moving from client's power to voice in policy development 50. The historical development of the Bamako Initiative has been one of strengthening the "short route" of client empowerment. As seen above, this control has limits, and the system has always balanced that power with control mechanisms from the central level to compensate for the asymmetry of information when it comes to professional services. Yet the progressive inclusion of clients in management and control of service delivery has also created a dynamic that fed into a more present voice of citizens in the policymaking process. Associations of community health committees emerged in Benin and Mali, a few years after the initiation of the approach. These associations now officially represent the communities in many regional and national level policy meetings. In the mid 1990's, the Benin association mobilized pressure from below against an initiative by the Government of Benin to centralize community financing revenue and channel it to the treasury on the argument that there had been leakages of some funds. Verification of the reported leakage demonstrated that this was limited to a few health centers, and constituted a small percentage of the overall community - 31 - financing revenue. The issue was taken up and resolved at the highest levels involving the President of Benin, and the WHO and UNICEF regional directors. The associations in Mali and Benin also succeeded in convincing their governments and donors to phase- in the adjustment of drug prices at the central medical stores at the time of the 50% devaluation of the currency (CFA Franc). They advocated for and obtained additional subsidies to "economically non-viable health centers" in order to mitigate the social impact of the CFA devaluation on the poor. Conclusion 51. Born out of the 1980s' collapse of the post independence social contract promising public health services for all, the Bamako Initiative national strategies in Benin, Guinea and Mali succeeded in developing an innovative approach that reconciled more traditional approaches to local community solidarity and traditional providers' payments with the public good objectives of the modern state. This led to improvements in the access, availability, affordability and utilization of professional health services as well as improved outcomes both for the average population and for poorest groups. To achieve this, policy makers chose consciously to strengthen the power of communities over service providers while balancing this power with a sustained involvement at the central level in subsidizing and regulating services, as well as guiding the process of community management. Even with a somewhat weak level of inclusion of the poor in the client's power mechanisms, significant outcomes were achieved. It would appear that this voice needs to be associated with the use and retention of local resources. Careful attention was given to health providers' incentives. Strategies and instruments were carefully designed based on best practices from the region and adapted over time through an ongoing learning by doing process. Building on their "Bamako Initiative foundations" which mainly strengthened the supply side of professional services, Benin, Mali and Guinea now face the challenge of similarly supporting the demand side through more emphasis on household behaviors, and the protection of the poorest and most vulnerable. The development of poverty reduction strategies represents an opportunity for governments to expand on these foundations and develop additional mechanisms to include the poorest groups in the dynamic already successfully created. - 32 - Sources: MOH Benin 2003, MOH Guinea 2002, MOH Mali 2002, Gilson 1997, Gilson and others 2000, Gilson and others 2001, Mehrotra S., Jarrett S.W. 2002, Knippenberg and others, 1997, Soucat and others 1997, Zhao and others 2003. - 33 - Gilson L., The lessons of user fee experience in Africa. Health Policy Plan. 1997 Dec;12(4):273-85. Review. Gilson L., Kalyalya D., Kuchler F., Lake S., Oranga H., Ouendo M., The equity impacts of community financing activities in three African countries. Int J Health Plann Manage. 2000 Oct-Dec;15(4):291-317. Gilson L., Kalyalya D., Kuchler F., Lake S., Oranga H., Ouendo M., Strategies for promoting equity: experience with community financing in three African countries. Health Policy. 2001 Oct;58(1):37-67. Knippenberg R., Soucat A., Oyegbite K., Sene M., Bround D., Pangu K., Hopwood I., Grandcourt R., Tinguiri K.L., Fall I., Ammassari S., Alihonou E., Sustainability of primary health care including expanded program of immunizations in Bamako Initiative programs in West Africa: an assessment of 5 years' field experience in Benin and Guinea. Int J Health Plann Manage. 1997 Jun;12 Suppl 1:S9-28. Ministére de la Santé Bénin, Revue des Dépenses Publiques dans le secteur de la Santé. 2003. Ministère de la Santé Mali, Santé et Pauvreté au Mali, cadre préparatoire au développement de la Stratégie de Réduction de la Pauvreté. 2002. Ministry of Health Guinea, Guinea, A Country Status Report on Health and Poverty, Health, Nutrition, and Population inputs for the PRSP and HIPC process. 2002. Mehrotra S., Jarrett S.W., 2002 Improving basic health service delivery in low-income countries: 'voice' to the poor. Soc Sci Med. Jun;54(11):1685-90. Soucat A., Gandaho T., Levy-Bruhl D., de Bethune X., Alihonou E., Ortiz C., Gbedonou P., Adovohekpe P., Camara O., Ndiaye J.M., Dieng B., Knippenberg R., Health seeking behaviour and household health expenditures in Benin and Guinea: the equity implications of the Bamako Initiative. Int J Health Plann Manage. 1997 Jun;12 Suppl 1:S137-63. - 34 - Zhao .F, Soucat A. and Traore N., The poor in Mali: Outcomes and Household Determinants of Child Health. 2003.