Context and intervention

Burkina Faso, ranked next-to-last in the 2007 Human Development Index, launched the BI in 1993. The organisation of public health care follows a classical pyramidal model. First-contact service is composed of primary health centres (CSPS) managed by a village health committee (COGES). Medicines are sold in the EGD (essential generic drug) depots of the CSPSs and the profits are managed by the COGESs. The COGESs use the profits for local operating costs, incentive bonuses to governmental health staff, and salaries for community managers of the EGD depots. The Social Action Ministry is responsible for the official exemption mechanism in Burkina Faso and for delivering the indigence cards. However, the bureaucrats do not have the means to do surveys to decide on the selection of recipients, and the services, which are often unknown to the population, are only available in urban settings. This study was carried out in the Ouargaye District (260 000 inhabitants in 2006, 96% rural), a rural district engaged in subsistence agriculture. In the Centre-East region, where this district is located, 55.2% of inhabitants were living below the poverty line in 2003, compared to 46.4% nationwide. The rate of measles vaccination is 65% for the district, as opposed to 71.1% nationwide. This action research was carried out in 10 of the district’s 25 CSPSs (figure 1). The action consisted of asking the communities to select the indigents who could receive total exemption from payment, to be financed from the BI profits. The definition of the content of the action and the choice of the 10 CSPSs were made by the research and district teams based on three criteria: ethics (not to inflict any financial risk during a pilot project); sustainability (using endogenous funds); and appropriateness (functional COGESs, an action that was participative and adapted to the context).

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Map of the Intervention District.

Planning

In 2003, the authors conducted a study aimed at understanding why exemption measures had not been implemented in Burkina Faso. One conclusion of that study was the need for action research to study community exemption mechanisms.5 The action research approach is, in fact, often recommended to introduce such social change, including in Africa.15 16 Thus, in a first step, the authors (VR) took advantage of a workshop led by the Canadian Coalition for Global Health Research in July 2004 in Canada to develop, with a researcher (AN) from Burkina Faso, the broad lines of this action research. In a second step, two workshops at the national and local levels (led by the Public Health Association of Burkina, AB, and the Regional Director of Health, NC) were organised in Burkina Faso in March 2005. These workshops made it possible to develop the protocol for the action research in a participatory manner with various national and local stakeholders. Then a graduate student (a Burkinabé physician) carried out a feasibility study in August 2005, which suggested, in particular, that the communities’ selections be validated by the village chiefs. In a fourth step, the co-investigators met in November 2005 to finalise the protocol and verify again the interest among decision-makers for such a process. The Minister of Health supported this study because he had been asking since 1992 for it to be organised.17

Content and implementation process

Implementation of the action followed several steps:

• i) Information: All the administrative authorities of the district headquarters were informed, then a tour of the 10 CSPSs was organised to inform the local and traditional authorities, as well as the COGES members and the health personnel.

• ii) Training: Because implementation activities are carried out by researchers in conjunction with the members of the District Health Team (DHT; SO), they were trained in order to implement the action. Then, the health personnel and two people from each COGES were also trained. After a presentation on the study, a workshop involving DHT members, researchers, community member (COGES) and health staff allowed for consensus to be reached on a definition of the indigents who were to be selected in the villages: “someone who is extremely disadvantaged socially and economically, unable to look after himself (herself) and devoid of internal or external resources.” The term “indigent” was translated into the local languages on a consensual basis.

• iii) Creation of village selection committees (VSC): VSCs made up of seven members were set up in each of the 124 villages located in the health areas of the 10 CSPSs. The membership, as designated by the COGESs after the training, had to respect a gender balance. To avoid any local elite capture, the members could not be administrative officers, village chiefs or members of the COGESs.

• iv) Selection of indigents: All the VSCs met several times. They made lists of the persons they considered to be indigent based only on the definition and without any pre-determined criteria. The selection unit was individuals. The process of selection was left under the decision of the members. The majority used their own knowledge of the population, combined sometimes with field visits to households. In accordance with the local context, this list was validated by the traditional chief of the village, the mayor, and finally the COGES, which had the final word.

• v) Distribution of cards: Each indigent received individually an exemption card with his or her photo. This card was validated by the Provincial Social Action Department.

Methods

The objective of this study is to describe and analyse the process of implementing the intervention. Through analysis, it was possible to identify the key characteristics of the intervention, ie its duration, the cost, the organisation of committees, the profiles of those selected as indigents, the financial capacity of the health centres, and finally, the perceptions of stakeholders. Reports and documents from the research project were analysed. Financial capacity was calculated using a method previously used by researchers in the same country.7 18 In each CSPS, the accounting records of the EGD depots were analysed from the expense books (purchase of EGDs, operating expenditures, etc), receipt books (sales of EGDs, user fees, grants, etc), and cash records (cash in the bank and in the safe), as well as the inventory records of drug stocks. Sums accumulated since the start of the BI correspond to the treasury status. Annual profits correspond to revenues (Expenditures + [initial EGD stock –final EGD stock]). A qualitative study was undertaken to analyse the perception of stakeholders concerning the process by means of five focus group discussions (FGD) with male members of VCSs from the 10 CSPSs (n = 43), five FGDs with female members of VCSs (n = 31), one FGD with head nurses of nine CSPSs (n = 9), one FGD with managers of eight COGESs (n = 8), and three individual interviews with the coordinator of the action research and two members of the DHT. All the discussions and interviews were translated where necessary, recorded and then transcribed using word processing software. The data were analysed using content analysis.

Results

Duration and cost

The selection of indigents by the VSCs and their validation took 45 days. The whole process took 4 months. Some expenses were assumed by the DHT (photocopies, printing, transport). The amount spent on the action (excluding the salary of the district-based research coordinator) was 1 500 000 F CFA (2300 €) paid by the research project.

Composition of the VCSs

Of the members of the 124 VCSs, less than 2% were members of the COGESs or religious leaders. Women made up 44% of the members, but only 15% of the VSCs had more than three women among the seven members.

Selection of indigents

The 124 VSCs selected 566 persons. The 10 COGESs retained 269 persons (48%), ie 2.81 per 1000 inhabitants (table 1). Less than 5% of the VSCs selected more than 10 persons, and 15% of them did not select any. Half the indigents were women (134/269) and 63% of the indigents were more than 50 years old.

Financial capacity of the EGD depots

Savings achieved since the launching of the BI (table 2, column b) make it possible theoretically to support between 290 and 4400 indigents per year in the coming years, depending on the CSPS (hypothesis based on one primary health care consultation per year per person). Except for one CSPS, the annual profits (column d) could support 24–560 indigents per year. The mean number of indigents selected by the VSCs represent on average 17% of the financial capacity. Health centres are thus able to support more than six times the number of indigents selected in the intervention (table 2, column f).

Stakeholders perceptions

The action was well received by the stakeholders. They considered that it was beneficial for the worst-off, according to their own (“emic”) criteria. It was a useful incentive for making the COGESs’ legal responsibility towards the indigent a concrete reality. One criterion for the effectiveness of this action was the fact that it was community-based and participative, because, as one VSC member remarked, “there have been other interventions that helped the indigent, but their approach was not like this one, which is much more community-centred.” It was also accepted because it respected the social context by requiring validation of the lists and by choosing as VCS members people who were very involved in the community. In the VCSs, the selection process was consensual. However, the fact that the COGESs only retained half the indigents can be explained according to members of COGES essentially by their fear of using up their accumulated funds and the fact that they do not really know the status of their financial situation. Some COGESs expressed a lack of confidence in the selection carried out by the VCSs and felt they were not rigorous enough in their selection. Some challenges still need to be addressed: ensuring the continuity of the exemption in the health pyramid; having more time and resources for creating the VSCs, providing better information to the COGESs about their financial margins of manoeuvre so that an informed decision can be taken on the number of indigents, and avoiding the risk of social stigmatisation.

Discussion

The first limitation of this study was that it was still not possible to assess the effectiveness of the targeting. The criterion of effectiveness presented at the beginning of this study was focused on the community’s capacity to actually implement the process. The present study thus demonstrated that this was possible. Further research will be required to verify whether the targeting was effective in selecting the worst-off (inclusion or exclusion error) and enabling them to use health care services. Nevertheless, it can be said that it was effective from the point of view of the VSC members who selected the persons they considered to be indigent, but not as much from the perspective of the researchers and health professionals, who expected a larger number of people to be selected. To use the terminology of anthropology,19 this study revealed the “emic” effectiveness more than the “etic” effectiveness of the intervention. The second limitation is that the estimate of the financial capacity to support the indigent is based on a hypothesis of only one visit per year and does not take into account possible fees for more costly treatments, particularly in hospital. That being said, support to the indigent is considered, in the real intervention, as covering all fees, including cases transferred to hospitals. The third limitation, associated with the preceding one, is that this intervention only partly improves the affordability of health services, without acting upon availability and acceptability.20 The exemption reduces only direct costs for the indigent, but they are sometimes also the furthest away (geographically and socially) from health centres and often disempowered. The intervention showed, for example, that the selected indigents also asked for food to take their medications, which were now free. Such an intervention, just in the health care sector, cannot pretend to reduce poverty or the social inequalities of health. Nevertheless, the authors believe this experience of targeting could now be useful to other public services (food security, education, etc) in distributing aid targeted to the worst-off that could complement this access to health care services. However, there remains a serious need for producing data on this subject in Africa, as such studies are still too rare. In fact, an analysis of publications on targeting programs between 1985 and 2003 shows that less than 10% of those programmes were implemented in sub-Saharan African countries.13

To the authors’ knowledge, this is the first study on community-based exemption schemes in the context of the BI in Africa. The study offers good news and bad news. The good news is that it is feasible to implement principle #7 of the BI (as a process) and to adapt it to the context. The selection of individuals perceived as indigent by the community using a community-based process was not very costly and was appreciated by the stakeholders. With respect to gender, the importance assigned to including women in the preparation phase and the directive to ensure an equitable distribution of VSC members contributed most definitely to a gender-balanced selection of indigents, which testifies to the feasibility of gender-balanced interventions called for in this context.21

The bad news is the stringency of the final verdicts of the communities in selecting indigents—only 3/1000, or 6/1000 if retaining the numbers selected by the VSCs. In Cambodia, in a very different context, a community-based targeting approach retained 16% of the population to receive exemptions, knowing that the funding was external.22 In the same region, it appears that pre-identification by the community would be more effective for selecting the worst-off (very poor) than the poor.23 In Burkina, this restrictive selection can be attributed to social and technical reasons. Socially, because their vision of social justice (equity) is more egalitarian than Rawlsian (and its “maximin” principle: the maximum for those who have the minimum), people are more apt to seek social peace and consensus.5 24 25 The selection of indigents was made on the basis of a double burden, economic (lack of financial means) and social (lack of support) because the situation is that of a country ranked next-to-last in the Human Development Index. In a context of widespread poverty, it is extreme poverty that makes the difference for the population. The region’s social organisation is very hierarchical and to be labelled “indigent” can sometimes “dishonour the family”, the authors were told. This social hypothesis had been advanced in a study of three pilot projects in Africa.9 Technically, the (equitable) utilisation of endogenous resources generated by BI user fees is not promoted,7 and therefore that idea definitely hindered the selection of indigents. The COGESs were in conflict of interest, as no third-party purchaser was involved. The effectiveness of the Cambodian experiences can be explained, in part, by significant funding that went beyond the local contributive capacity (table 2) and was mostly exogenous in origin.9 22

Conclusion

This experience teaches us that in the African context of the BI, decision-makers who want to increase service utilisation among the worst-off should definitely consider two possible solutions, because there is still insufficient data9 26 and experience27. First, although endogenous resources may be limited, they do exist (thanks to the BI) and communities can select individuals whom they consider indigent (“emic” effectiveness). Also, the State could legally require (and even enforce, by accompanying measures) that a minimum number of indigents be selected on the basis of a percentage of annual profits. The present research showed, in fact, that the health centres had the capacity to take on six times more indigent persons than the committees selected. Exclusion error could be reduced by increasing the number of beneficiaries. Second, a community-based targeting approach should be tested where support for the worst-off could be funded by a third-party purchaser like the Health Equity Funds.22 28 Further studies need to be undertaken, particularly in Africa, to evaluate the effectiveness of such mechanisms in selecting the worst-off and improving their access to care.