Comparison of risk-adjusted 30-day postoperative mortality and morbidity in Department of Veterans Affairs hospitals and selected university medical centers: vascular surgical operations in women.

Abstract

Women with peripheral vascular disease requiring vascular operations are less well studied than their male counterparts. The surgical outcomes of female vascular patients in the Department of Veterans Affairs (VA) and private sector hospitals have not previously been compared, and their preoperative risk profile, postoperative morbidity, and mortality need to be better elucidated. Patients undergoing vascular operations at 14 private sector and 128 VA hospitals, from October 2001 through September 2004, had their preoperative characteristics, operative data, and 30-day postoperative morbidity and mortality compared, as part of the Patient Safety in Surgery (PSS) Study. Logistic regression analysis was performed to develop predictive models for morbidity and mortality, which allowed for a comparison of risk-adjusted outcomes between the two hospital groups. There were 458 vascular surgical operations performed in women in the VA, and 3,535 vascular operations were performed in women in the private sector. Eighteen of 45 preoperative comorbidities and laboratory variables differed considerably between the institutions, and 16 of 18 were adverse among the private sector patients. The unadjusted 30-day mortality rate was higher in the private sector compared with the VA (5.2% versus 2.4%, p=0.008); the unadjusted morbidity rate was higher in the private sector compared with the VA sector (23.4% versus 13.3%, p < 0.0001). After risk adjustment, there was no marked difference between the VA and the private sector in mortality (p=0.12), but the difference in morbidity rates remained pronounced, with an odds ratio of 0.60 for VA versus private sector (95% CI=0.44, 0.81). Compared with their VA counterparts, women undergoing vascular operations at private sector hospitals had a higher incidence of preoperative comorbidities; after risk adjustment, mortality did not differ substantially. Despite risk adjustment, the incidence of postoperative morbidity in the VA patients was considerably lower, suggesting unidentified differences in the hospital populations, their processes of care, or both.