Evaluating and re-demarcating the Hospital Service Areas in Florida

Abstract

The Hospital Service Areas (HSAs) better portray underlying local patterns of hospitalization than administrative units, and offer a promising analysis unit for studies of healthcare market. The widely used Dartmouth HSAs in the U.S. were solely based on Medicare inpatient records about two decades ago. Our analysis used all discharge records from Florida hospitals in the 2011 Healthcare Cost and Utilization Project (HCUP) dataset from the Agency for Healthcare Research and Quality (AHRQ). We first matched Medicare-paid hospitalization records in 2011 to the Dartmouth HSAs for demonstrating the temporal variation of the Medicare-derived HSAs. We then compared the HSA configurations based on the overall hospitalization records to Medicare-derived HSAs in the same year (2011) for assessing the representativeness of the Medicare-derived HSAs. Results indicate the boundaries of the Medicare-derived HSAs have significantly shifted over two decades and are inadequate in representing the overall population. The Huff model was used to generate more solid HSAs than traditional approaches.