Geographic Barriers to Achieving Universal Health Coverage in a rural district of Madagascar

Abstract

Poor geographic access can persist even when affordable and well-functioning health systems are in place, limiting efforts for universal health coverage (UHC). It is unclear how health facilities and community health workers contribute to achieving UHC. Using geographic information from thousands of patients in a rural district of Madagascar we evaluate how a health system strengthening (HSS) intervention aimed towards UHC affects the geography of primary care access. We find that facility-based interventions (user-fee exemptions, improved readiness) achieved high utilization rates only among populations who lived in close proximity to supported facilities. Scaling only facility-based HSS programs across the district would result in large gaps in health care access for the majority of the population. Community health provided major improvements in service utilization for children regardless of their distance from facilities. Our results have implications for UHC policies and suggest that greater emphasis on professionalized community health programs is essential.

Estimating the local spatio‐temporal distribution of malaria from routine health information systems in areas of low health care access and reporting

“…Our setting is representative of many rural areas in the developing world, where rough landscapes, poor road infrastructures and sparsely distributed populations make it difficult for patients to access health centers. More than 95% of paths were only accessible by foot, and three fourths of the population live more than an hour's walk of a public health center [ 34 , 35 ], a commonly accepted threshold of low geographic access [ 54 – 57 ]. All these factors can lead to significant underreporting of malaria, at levels compatible with estimates presented here.…”

“…Low access to health care was strongly associated with substantial financial and geographic barriers [ 33 ]. For instance, only one fourth of the population lives within an hour's travel of a health center [ 34 , 35 ].…”

“…To create U ij and account for the effect of low health care access on malaria incidence, we used results from a spatio-temporal model of health care utilization in Ifanadiana during the same study period. Details on this model are published elsewhere [ 34 ], and coefficients for each variable included in the final model (adjusted odds ratios) are available in Additional file 1 : Table S1. Briefly, per capita health center utilization rates for each Fokontany were modeled using Binomial regressions in generalized linear mixed models, with a random intercept introduced for the closest health center.…”

“…The model accounts for the exponential decrease in utilization as a function of travel distance from each Fokontany to the nearest health center (Additional file 1 : Figure S1); the positive impact on consultations of programs implemented to reduce financial and geographic barriers (e.g. user fee exemptions, community health strengthening); the positive linear and seasonal trends in utilization rates in the absence of those programs; baseline differences in the type of health center (lower utilization for CSB1 than CSB2); and the positive association with the number of health staff over time in the closest health center [ 34 ]. Based on model predictions of per capita utilization for non-malaria patients, a health center utilization index was produced for each Fokontany-month in Ifanadiana, scaled between zero (no access; set at zero consultations per person-month) and one (full access; set at 0.166 consultations per person-month, equivalent to 2 consultations per person-year, excluding malaria).…”

Estimating the local spatio-temporal distribution of disease from routine health information systems: the case of malaria in rural Madagascar

“…In Ifanadiana, a mountainous landscape, poor road infrastructure and a sparsely distributed population make it difficult for patients to access health centers. More than 95% of paths are not accessible by vehicle, and three fourths of the population live more than an hour's walk of a public health center 36,37 , a commonly accepted threshold of low geographic access [53][54][55][56] . All these factors could have led to significant underreporting of malaria, at levels compatible with estimates presented here.…”