Rio de Janeiro Stroke Care Access Dataset: SUS Hospitalizations, Healthcare Infrastructure, Population and Road Accessibility

This hierarchical HDF5 dataset integrates 117,706 curated hospitalization records for ICD-10 I60–I64 from the Brazilian SUS hospital information system (2017–2023), 2022-census-calibrated population weights over 110,664 geospatial demand cells, and December 2023 CNES healthcare establishment registrations in Rio de Janeiro, Brazil. It contains estimated grid-to-hospital travel-time and geographic-distance matrices for 195 routed hospitals, source road-network arrays derived from OpenStreetMap, alternative registration-based hospital-network scenarios, eligibility masks, and embedded provenance. The dataset supports scientific investigation of geographic access, spatial inequality, and multiobjective facility-location problems. Hospitalization records are administrative events rather than unique patients; service-registration fields do not certify operational 24-hour stroke treatment; modeled road travel times are not clinical treatment times.