Background: Improving maternal and child health (MCH) remains a serious challenge for many developing countries. Geographical accessibility from a residence to the nearest health facility is suspected to be an important obstacle hampering the use of appropriate services for MCH especially in Sub-Sharan African countries. In Burkina Faso, a landlocked country in the Sahel region of West Africa, women’s use of proper healthcare services during pregnancy and childbirth is still low. This study therefore assessed the impact of geographical access to health facilities on maternal healthcare utilization in Burkina Faso. Methods: We used the Burkina Faso demographic and health survey (DHS) 2010 dataset, with its sample of 10,364 mothers aged 15-49 years. Distance from residential areas to the closest health facility was measured by merging the DHS dataset with Geographic Information System data on the location of health centers in Burkina Faso. Multivariate logistic regressions were conducted to estimate the effects of distance on maternal healthcare utilization. Results: Regression results revealed that the longer the distance to the closest health center, the less likely it is that a woman will receive appropriate maternal healthcare services. The estimates show that one kilometer increase in distance to the closest health center reduces the odds that a woman will receive four or more antenatal care by 0.05 and reduces by 0.267 the odds that she will deliver her baby with the assistance of a skilled birth attendant. Conclusions: Improving geographical access to health facilities increases the use of appropriate healthcare services during pregnancy and childbirth. Investment in transport infrastructure should be a prioritized target for further improvement in MCH in Burkina Faso.
The demographic and health survey (DHS) of Burkina Faso 2010 was used for the empirical analysis in our study [27]. The DHS is a nationally representative survey that applies a stratified two-stage cluster sampling design. The sample is stratified into urban and rural areas to represent both areas. In the sampling, the primary survey units (“clusters”) are first selected from larger 13 regional units based on 2006 General Population and Housing Census and then individual households are randomly selected within each cluster. The data were collected between May 2010 and January 2011. In all, 14,947 households from 574 clusters (398 from rural areas and 176 from urban areas) were sampled, and then 14,424 households were actually surveyed. (The response rate was 99.2%). The study population for our analysis was 10,364 women (15–49 years old) who lived in these households and had a live birth in the 5 years preceding the survey [27]. Our main explanatory variable was distance from a residential area to the closest health center. Because the DHS household questionnaires contained no information regarding distance or travel time to health facilities, we used the Geographical Information System (GIS) module of the DHS to calculate these distances. We also obtained geographic information on roads and health centers through the Ministry of Infrastructure of Burkina Faso (for road information) and the Centre National de Recherche Scientifique et Technologique of Burkina Faso (for health center information). For measuring the distance between a residence and the closest health facility, a previous study conducted in Ghana employed the following methods [28]: 1) Euclidean distance (km), the straight-line distance from a residence to the closest health facility; 2) network distance (km), the distance along the road network from a residence to the closest health facility plus the Euclidean distance to the road network from a residence and from the road network to the health facility; 3) network travel time (hour), the distance along the road network from a residence to the closet health facility plus the Euclidean distance multiplied by off-road walking speed to the road network from the residence and from the road network to the health facility; and 4) raster-based travel time (hour), the travel time from a residence to the closest health facility, assuming mechanized or non-mechanized travel on roads and non-mechanized travel on- or off-road depending on the land cover speed [28]. Results obtained by these methods were similar in Ghana [28]. We chose the Euclidean distance to measure the distance from a cluster centroid to the closest health center and then used it as a proxy for the geographical access to a health center from a residential area. The advantage of this method is that it can be generalized for other similar topography and cultural contexts in West Africa [28]. Figure Figure11 shows a map of Burkina Faso, divided into 352 communes. (Burkina Faso is divided into 13 administrative regions, which are subdivided into 45 provinces; the provinces are subdivided into 352 communes) Fig. Fig.22 shows the country’s network of roads. In Fig. Fig.3,3, the red circles indicate the 574 clusters from which households were randomly selected for the survey. Figure Figure44 shows the location of 1520 health centers (purple circles). We did not distinguish the level of health services provided by each center. Finally, we used Quantum-GIS software to calculate the Euclidean distance (km) from each cluster to the closest health center. Burkina Faso divided by communes Road network GIS points of clusters GIS points of health centers In addition to the distance to health facilities, we analyzed if the availability of means of transport at the community-level was associated with the use of maternal healthcare. DHS did not include information on public transport, but asked questions about whether or not the household owned a bicycle and motorbike, which are the popular means of transport in Burkina Faso even among women in seeking healthcare during pregnancy and childbirth [23]. We thus calculated the ownership rates of bicycles and motorbikes per cluster and used them as proxy variables for the community-level availability of means of transport. Multivariate logistic regressions were conducted to analyze the effect of distance to the closest health center on maternal healthcare utilization. Data analysis was carried out using Stata 14.0. Because DHS applied a two-stage cluster sampling design, we used the svy (survey) commands of Stata to correct for unequal sampling probability, clustering and stratification in calculating descriptive statistics and performing regression analysis. We used the following outcome variables: 1) whether the woman made at least one ANC visit during her latest pregnancy (“Received any ANC”); 2) whether the woman made four or more ANC visits during her latest pregnancy (“Received ≥ 4 ANC”); 3) whether the woman used a health facility at the birth-delivery (“Facility delivery”); and 4) whether the woman was attended by a professional health worker, i.e., doctor, nurse, auxiliary nurse or midwife at the birth-delivery (“Delivery by SBA” (skilled birth attendant)). Because all the outcome variables were binary, they were coded 1 if the mother had received appropriate healthcare (ANC, Facility delivery, or SBA) during her pregnancy or at childbirth, or 0 otherwise. Outcome variables for birth-delivery, i.e. 3) “Facility delivery” and 4) “Delivery by SBA, included all the births (14,996) that had taken place during the five years preceding the survey. Therefore, we included mother-level random intercepts into multivariate logistic regressions for 3) “Facility delivery” and 4) “Delivery by SBA to adjust for the correlation of births to the same mother. We used demographic and socioeconomic characteristics at the mother, household, and community levels as control variables. The mother-level variables consisted of age and educational achievement (no education, primary, secondary, and higher). The household-level variables included the religion of the household head (no religion, Muslim, Catholic, Protestant, and traditional religion/animist), asset quintiles. The community-level variables included area dummies (rural or urban). In addition, region dummies (Boucle du Mouhoun, Cascades, Centre, Centre-Est, Centre-Ouest, Centre-Nord, Centre-Sud, Est, Hauts Basins, Nord, Plateau Central, Sahel, and Sud-Ouest) were included to consider regional differences within the country.
N/A