Objective: To describe the catchment areas of medical centers with comprehensive stroke management in Bogotá, considering traffic conditions. Methods: We randomly selected centroids for census tracts in Bogotá for the 2005 DANE census. Based on official data, we selected three traffic conditions for weekdays and weekends. Using Google’s Distance Matrix API, we requested times in traffic from points to each of the 14 medical centers with comprehensive stroke management. We sampled a total of 2850 points throughout the week and 1420 for the weekends, for a total of 179.340 transport times. We constructed raster maps using universal kriging and assigned corresponding values to each census track that was not explicitly sampled. Using these values, we constructed choropleth maps and calculated the area and populations covered. Spatial autoregressive models were explored for the association between socioeconomic status and arrival times. All analyses made use of free software. Results: Arrival times to the closest center showed a high spatial autocorrelation. Traffic conditions significantly affected catchment areas, with reductions of up to 73% from maximal to minimal areas. We did not observe a uniform pattern for changes in areas under the various traffic conditions. When mapping the exclusive catchment areas for each center, we observed that 5 medical centers covered the majority of the city’s area and population, with a population to center ratio exceeding international recommendations. Spatial autoregressive models showed a statistically, but not clinically, significant contribution of socioeconomic status to arrival times to the nearest center. Conclusions: Traffic conditions in Bogotá significantly affect optimal coverage areas for medical centers offering comprehensive stroke management. Catchment areas showed that 5 centers covered most of the city, exceeding international recommendations. Socioeconomic status did not show a clinically significant association with transport times to the nearest centers on spatial autoregressive models.
