Abstract
OBJECTIVES: The objectives of this study were to determine (1) if there were significant differences between patients who died at a public hospital and those who died at a university hospital that functions as a private, community hospital, and (2) if those differences were associated with an increased risk of death. METHODS: Chart review collected variables used by the Health Care Financing Administration in mortality analyses to examine how severity of illness data contribute to accurate predictions of death in a public hospital compared with a university hospital. RESULTS: Compared with patients who died at the university hospital, public hospital patients who died had more comorbid disease, were more severely ill, more likely to be emergently admitted, and more likely to be admitted from an extended-care facility. Inclusion of severity of illness with variables previously used to predict mortality significantly improved the accuracy of mortality prediction models for the public hospital but not for the university hospital. CONCLUSIONS: The results suggest that urban public hospitals provide care to more severely ill patients. Administrative data sets may not be adequate to identify these differences between patient populations.