Menu
Menu
Menu
Menu
Background: In 2006, the Parliament of Burkina Faso passed a policy to reduce the direct costs of obstetric services and neonatal care in the countrys health centres, aiming to lower the countrys high national maternal mortality and morbidity rates. Implementation was via a partial exemption covering 80% of the costs. In 2008 the German NGO HELP launched a pilot project in two health districts to eliminate the remaining 20% of user fees. Regardless of any exemptions, women giving birth in Burkina Fasos health centres face additional expenses that often represent an additional barrier to accessing health services. We compared the total cost of giving birth in health centres offering partial exemption versus those with full exemption to assess the impact on additional out-of-pocket fees. Methods. A case-control study was performed to compare medical expenses. Case subjects were women who gave birth in 12 health centres located in the Dori and Sebba districts, where HELP provided full fee exemption for obstetric services and neonatal care. Controls were from six health centres in the neighbouring Djibo district where a partial fee exemption was in place. A random sample of approximately 50 women per health centre was selected for a total of 870 women. Results: There was an implementation gap regarding the full exemption for obstetric services and neonatal care. Only 1.1% of the sample from Sebba but 17.5% of the group from Dori had excessive spending on birth related costs, indicating that women who delivered in Sebba were much less exposed to excessive medical expenses than women from Dori. Additional out-of-pocket fees in the full exemption health districts took into account household ability to pay, with poorer women generally paying less. Conclusions: We found that the elimination of fees for facility-based births benefits especially the poorest households. The existence of excessive spending related to direct costs of giving birth is of concern, making it urgent for the government to remove all direct fees for obstetric and neonatal care. However, the policy of completely abolishing user fees is insufficient; the implementation process must have a thorough monitoring system to reduce implementation gaps. © 2012 Ben Ameur et al.; licensee BioMed Central Ltd.
In this study, researchers are analysing the results of a natural experiment. A natural experiment is an empirical study in which the experimental conditions are determined by nature or by other factors out of the control of the evaluators [11]. The study took place in the Sahel Region of Burkina Faso about 300 km from the capital Ouagadougou, which has specific geographical and cultural characteristics. This region has four districts: Dori, Sebba, Djibo and Gorom-Gorom. The first two represent the only districts where the NGO intervened during the study; the evaluators had no control in selecting these districts for the pilot project. In order to select a third district for comparison with Dori and Sebba we examined the level of facility-based delivery (expressed as a percentage of all institutional deliveries) in Gorom-Gorom and Djibo. According to Annuaire statistique 2009 (Ministère de la santé – Secrétariat général, Direction générale de l’information et des statistiques sanitaires), 22% of institutional deliveries in Gorom-Gorom had skilled care, while this proportion was 52.4% in Djibo. Djibo was chosen as the comparison district because its percentage of skilled care was much closer to the proportions observed in the Dori and Sebba districts. Dori and Sebba served as the case districts with a full user fee exemption provided by a pilot project of the NGO HELP, and Djibo as the control district with the government-mandated partial user fee exemption. The table below describes the main characteristics of the three districts (Table (Table1).1). Districts characteristics (source: Plans d’Action 2010 des Districts Sanitaires de la Région du Sahel) (*) The official cost of delivery corresponds to « acts, medicines and consumables, and observation » (Ministère de la santé, 2006a). (**) Annuaire statistique 2008, 2009, 2010, 2011 Ministère de la santé – Secrétariat général, Direction générale de l’information et des statistiques sanitaires. NGO= Non government organization. CSPS= Centre de santé et de protection sociale, i.e. “first line health centre”. F CFA= Franc de la Communauté Financière Africaine. « facility-based delivery » refers to facility-based childbirths assisted by professionals. According to the Annuaire Statistique where the indicator comes from, it is called in French: Accouchements assistés par du personnel de santé. In every district, six health centres were selected for a total of 18 health centres (30% of the total centres). This selection sample was based on two previous studies in the region that reflected situational diversity [12][13]. Only two exceptions were made because of access constraints due to the rainy season at the time of data collection. For these two cases, two other CSPS with similar characteristics were chosen. The selection sample includes only women who gave birth in health centres and had no complications, as the study focused on expenses related to an institutional birth. In the full exemption districts (Dori and Sebba), the study population was women who gave birth in health centres two months prior to the start of the study. In the partial exemption district (Djibo), the study population was women who gave birth in health facilities in the six months prior to the start of the study. This longer recall period was needed to gather a large enough sample size because the number of women using facility-based services in that district was very low. At the CSPS level, the sampling was done using maternal registers available in each health centre. A random sample of approximately 50 women per health centre was selected from the registers. The survey size was limited to 50 per health centre due to the limited financial resources available for the study. In Dori and Sebba it was possible to achieve this target. However, in Djibo, the low rate of facility-based births (2010) together with an exceptional number of emigrants made it more difficult to reach our targeted sample size. Thus, the final sample in this district was 270 women rather than 300. The total sample was 870 women who gave birth in 18 health centres. Twenty-one women were excluded from the original sample as they were found to have given birth at home or before reaching the health centre (see Table Table22). Sampling (1) Source: maternal registers available in health centres Period of delivery: *March 2010- May 2010; **December 2009- May 2010. Survey instruments were developed based on a similar survey conducted in another district of Burkina Faso in 2006 by IMMPACT researchers and in 2010 by VR and AB in Ouargaye [14]. The instruments measure medical household expenditures related to institutional birth at the point of use. The medical expenditures include: user fees, drugs and consumables, laboratory fees, and cleaning products. Traditional poverty measurements based on consumption or incomes were particularly difficult in the sample district settings. Instead, household characteristics and asset ownership are widely used as indicators of wealth [15]. Consequently, a series of questions related to the specific characteristics of rural households in Burkina Faso was used to gauge household economic status. Questionnaires and consent forms were developed in French and then translated into the local language, Peulh, following the double-translation method. The instruments were pretested in the Dori district in a CSPS not included in the sample. The pre-test assessed the validity of the data-collection instruments and procedures, as well as the sampling procedures. This process identified content and logistical issues that led to revising some of the questions and the data collection process. Sahel Regional Authorities approved the survey. Fieldwork was conducted over a five-week period starting in May 2010 by six trained local interviewers fluent in Peulh and French. Each interviewer interviewed +/− 50 women over a ten-day period. The household survey questionnaires were administered to all study participants. Respondents gave verbal consent to the interview and were assured of data confidentiality. Two research coordinators carefully supervised data collection during the entire fieldwork. Data was input and validation was performed with the double entry method using Epi Data software, and the data set was then converted to SPSS 17 and STATA 11 for analysis. Analysis of variance (ANOVA) was performed to compute means of medical expenses and compare them between districts and other explanatory variables (education, distance, etc.). A poverty proxy was developed using household data indicators. Through Principal Components Analysis (PCA), households/women were assigned to poverty quintiles [16]. This allows a classification of households from the poorest (rated as 1) to the least poor (rated as 5) in order to show socio-economic differences. As we did not have information about household consumption and/or income, it is not possible to measure catastrophic expenditures related to births health centres. Instead, we relied upon the concept of “high delivery health care spending per household”, equivalent to “excessive spending per household”. It was estimated by analyzing the expenditures dispersion within the two samples (full exemption vs partial exemption group). By using the statistical outlier method (or Tukey method) it was possible to identify the value threshold. Supposing Q1 and Q3 are respectively the first and third quartiles of the distribution of delivery health expenditure of the sample, then a high-spending or outlier household within the group is one for which the value of delivery health expenditure is greater than Q3+k*(Q3-Q1), where k is a constant (varied from 0.5, 1 or 1.5). Having three values for k, rather than one, allows more flexibility in defining the outlier households at different expenditure cut-off points and permits testing the sensitivity of the results. In fact, the smaller k is, the stricter is the approach [17]. The Ministry of Health of Burkina Faso examined and approved the ethics component of this research project and authorized the study. Ethical approval was given prior to data collection.
N/A
