RESUMEN
PURPOSE: Geographic access to inpatient care at the Veterans Affairs (VA) Health Care System is challenging for many veterans with chronic obstructive pulmonary disease (COPD) given relatively few VA hospitals nationwide. Veterans with lengthy travel distances may obtain non-VA care, particularly those dually enrolled in Medicare. Our primary objective was to assess whether distance from VA patients' residence to the nearest VA and non-VA hospitals was associated with 30-day all-cause readmission and the system where patients were readmitted (VA or Medicare). METHODS: Using VA and Medicare administrative data, we identified 21,273 patients hospitalized for COPD between October 2008 and September 2011 and dually enrolled in VA and fee-for-service Medicare. Outcome variables were dichotomous measures denoting readmission for any cause within 30 days following discharge and whether the readmission occurred in a non-VA hospital through Medicare. Distance to the nearest hospital was defined as the number of miles between patients' residence ZIP code and the ZIP code of the nearest VA and non-VA hospital accepting Medicare, respectively. Probit models with sample selection were applied to examine the relationship between hospital distance and outcome measures. FINDINGS: Respective distances to the nearest VA and non-VA hospital were not associated with 30-day all-cause readmission. Greater distance to the nearest VA hospital was associated with a greater conditional probability of choosing non-VA hospitals for readmission. CONCLUSIONS: COPD patients with poor geographic access to VA hospitals did not forgo subsequent inpatient care following their index hospitalization, but they were more likely to seek non-VA substitutes.