Geospatial accessibility analysis is standard in engineering and health, but rare in public policy. AI-enhanced spatial modelling should be embedded early in infrastructure planning.
Türkiye's public–private partnership (PPP) city hospital program expanded tertiary care capacity through large, integrated hospital campuses. Bursa City Hospital, opened in 2019 with 1,355 beds, is among the largest built under this model, yet its peripheral location has raised debate over accessibility implications. This study compares spatial accessibility to hospital services in Bursa before and after the investment, and examines a counterfactual in which the same capacity is located centrally rather than on the periphery. Using a two-step floating catchment area (2SFCA) method, it evaluates travel-time accessibility and population coverage relative to hospital location, integrating spatial tools into infrastructure impact assessment. Findings reveal a marked mismatch between nominal capacity and effective accessibility: nearly 90% of Bursa's population lives more than 15 minutes' drive from the new facility, with limited public transport compounding barriers for those without private vehicles. While the investment expanded catchment coverage overall, gains concentrated around the city centre; the counterfactual shows siting materially shapes accessibility benefits. The Bursa case shows large hospital investments do not automatically yield accessible care; realized accessibility depends on infrastructure scale, site selection, and transport connectivity. Embedding GIS-based accessibility modelling into infrastructure evaluation offers a replicable approach for assessing real-world impact, pushing planning beyond capacity metrics toward systematic geospatial analysis.
Irem Kizilca is a Senior Research Officer at ANU's Infrastructure for Society Institute (I2S).