Final Report: Effects of Rural Hospital Closure on Access to Care: Data Analysis

The purpose of this study was to examine the impact of rural hospital closure on access to and costs of medical care. State all-payers hospital discharge data were utilized to test the hypotheses that hospital market areas that experienced hospital closure would have lower admission rates, longer average length of stay and higher costs per admission than hospital market areas where hospitals remained open. The sample for the study included fifteen market areas in nine states where hospitals closed between 1991-1994, as well as fifteen market areas where similar hospitals remained open. The comparison market areas were included to control for medical services utilization trends unrelated to closure. The sample size was limited to 30 because of the unavailability of all-payers data in many states where closure occurred. Bivariate analysis revealed no effects of hospital closure on admission rates, average length of stay and average charge per admission. Admission rates and average length of stay decreased for both closure and comparison market areas. Average charge per admission increased in comparison areas and decreased in study areas; however, this difference in the direction of charges was not found to be statistically significant in bivariate analysis. The sample size used for the bivariate analysis did raise concerns regarding lack of adequate power to detect differences in the outcomes of interest. However, regression analysis using the individual admission as the unit of analysis also did not detect any effects of closure on lengths of stay or charges. No differences in results were identified for market areas that experienced closure and were at the farthest distance from their nearest neighboring hospital. The results of this study must be taken in the context of its limitations, such as small sample size and uneven quality of the state all-payers data. These limitations may have affected the ability of the study to detect meaningful changes in the outcomes of interest. The study limitations primarily resulted from the unavailability of state all-payers data in many areas of the country and the problems that arose in trying to utilize the data. These problems included missing data for certain variables or hospitals and incompatibility of data across states because of a lack of uniformity in data collection and reporting. In addition, the outcomes used in this study (admission rates, hospital charges or lengths of stay) cannot adequately capture changes in access to care that may have led to decreases in health status and quality of life for populations living in areas where hospitals closed.