Résumé
Background: Geographic access to healthcare continues to pose a significant challenge for pregnant women in rural areas of sub-Saharan Africa, resulting in consistently high rates of maternal mortality. Geographic barriers can persist even in settings where financial barriers have been reduced and health system strengthening (HSS) efforts are in place. The aim of this study is to gain a precise understanding of spatiotemporal changes in access to and utilization of maternal care services in a rural district of Madagascar benefiting from HSS support.Methods: We collected geolocated monthly information at the village level on antenatal care visits, deliveries and postnatal visits from the registries of 18 public primary health centers in Ifanadiana District, from 2016 to 2018. Similar data were collected from a district-representative cohort via surveys on over 1500 households done in 2016 and 2018. We estimated precise travel time from each village to the nearest health center to understand spatio-temporal variations in maternal care access, and to assess the impact of geographic barriers via statistical analyses while controlling for health system factors.Results: Women who lived within a one-hour walk from a health facility in the HSS catchment area had rates of per capita utilization of most maternal health services were roughly twice that those who lived 1–2 h away and three times higher than those who lived over 2 h away (e.g. relative change for delivery at a health center was 0.60 [0.53–0.67] and 0.40 [0.36–0.45] for women living 1–2 h and over 2 h from a facility, respectively). The exception was the first antenatal care visit (ANC1), for which travel time had more modest effect (e.g. relative change of 0.72 [0.67–0.77] over 2 h). Improvements to primary care services due to HSS in this setting were only observed among women living within two hours from health centers. Statistical models revealed that women’s travel time from a health facility was the strongest determinant of maternal care service utilization.Conclusion: This study shows how a combination of geo-located health system information and population-representative data can help assess the impact of geographical barriers to maternal care in rural areas of sub-Saharan Africa. It highlights that women who live more than 2 h from a health facility had virtually no access to maternal health services despite efforts in place to reduce financial barriers to care and strengthen the health system