Global variation in travel time to the nearest hospital

Abstract

The time it takes to travel to the nearest hospital is an important measure of access to healthcare services, particularly for acute illnesses that require immediate medical attention. Despite the public health significance of this measure, it has never been quantified in a comparable way at a high resolution for the globe. We used a combination of Ministry of Health and open source data on the locations of hospitals to estimate the travel time to the nearest hospital at the 5x5 kilometer level for inhabited areas of the world. We used machine learning algorithms and spatial simulation models to estimate, with uncertainty, the locations of hospitals in each country, and applied a recently published friction map to calculate travel time to the nearest hospital from each 5x5 kilometer raster. We found that 85% (84% to 86%) of all persons were living within 90 minutes of a hospital. We also found large variation in the proportion of individuals living within 90 minutes of a hospital across countries, with most high-income countries having more than 95% of individuals living within 90 minutes of a hospital. We also found a large difference in the mean travel time to the nearest hospital by urbanicity, being 13 (11 to 18) minutes among urban rasters and 82 (78 to 90) minutes among rural rasters.