Emergency Care Sensitive Conditions in Brazil: A Geographic Information System Approach to Timely Hospital Access

Abstract

Background: The benefits of treatment for many conditions are time dependent. The burden of these emergency care sensitive conditions (ECSCs) is especially high in low- and middle-income countries. Our objective was to analyze geospatial trends in ECSCs and characterize regional disparities in access to emergency care in Brazil. Methods From publicly available datasets, we extracted data on all patients assigned an ECSC-related ICD-10 code and on the country’s emergency facilities from 2015-2019. Using ArcGIS, OpenStreetMap, and WorldPop, we created catchment areas corresponding to 180 minutes of driving distance from each hospital and estimated their contained population. We then used ArcGIS to characterize space-time trends in ECSC admissions and to complete an Origin-Destination analysis to determine the path from household to closest hospital. Findings: In the North, only 69.1% of the population resided within 180 minutes (171km) of an emergency hospital. In the Northeast and Southeast, there were clusters of increasing trends in ECSCs while the South and Midwest saw the opposite trend. The highest density of emergency centers was in Sao Paulo. Interpretations: Significant geographical barriers to accessing emergency care exist in certain regions of Brazil, especially in the North. One limitation of this approach is that, given the lack of roads and zip codes in this region, geolocation was not possible in many areas and thus we are likely underestimating the burden. Subsequent work should evaluate ECSC mortality data. Funding: This study was funded by the Duke Global Health Institute Artificial Intelligence Pilot Project. Declaration of Interest: None to declare.