Abstract
OBJECTIVE: The targeted vascular module in the American College of Surgeons National Surgical Quality Improvement Program (NSQIP) consists of self-selected hospitals that choose to collect extra clinical details for better risk adjustment and improved procedure-specific outcomes. The purpose of this study was to compare patient selection and outcomes between targeted and nontargeted hospitals in the NSQIP regarding the operative management of abdominal aortic aneurysm (AAA). METHODS: We identified all patients who underwent endovascular aneurysm repair (EVAR) or open AAA repair from 2011 to 2013 and compared cases by whether the operation took place in a targeted or nontargeted hospital. EVAR and open repair as well as intact and ruptured aneurysms were evaluated separately. Only variables contained in both modules were used to evaluate rupture status and operation type. All thoracoabdominal aneurysms were excluded. Univariate analysis was performed for intact and ruptured EVAR and open repair grouped by complexity, defined as visceral involvement in open repair and a compilation of concomitant procedures for EVAR. Multivariable models were developed to identify effect of hospital type on mortality. RESULTS: There were 17,651 AAA repairs identified. After exclusion of aneurysms involving the thoracic aorta (n = 352), there were 1600 open AAA repairs at targeted hospitals (21% ruptured) and 2725 at nontargeted hospitals (19% ruptured) and 4986 EVARs performed at targeted hospitals (6.7% ruptured) and 7988 at nontargeted hospitals (5.2% ruptured). There was no significant difference in 30-day mortality rates between targeted and nontargeted hospitals for intact aneurysms (EVAR noncomplex, 1.8% vs 1.4% [P = .07]; open repair noncomplex, 4.2% vs 4.5% [P = .7]; EVAR complex, 5.0% vs 3.2% [P = .3]; open repair complex, 8.0% vs 6.0% [P = .2]). For ruptured aneurysms, again there was no difference in mortality between the targeted and nontargeted hospitals (EVAR noncomplex, 23% vs 25% [P = .4]; open repair noncomplex, 38% vs 34% [P = .2]; EVAR complex, 29% vs 33% [P = 1.0]; open repair complex, 27% vs 41% [P = .09]). Multivariable analysis further demonstrated that having an operation at a targeted vs nontargeted hospital had no impact on mortality for both intact and ruptured aneurysms (odds ratio, 1.1 [0.9-1.4] and 1.0 [0.8-1.3], respectively). CONCLUSIONS: This analysis highlights the similarities between targeted and nontargeted hospitals within the NSQIP for AAA operative management and suggests that data from the targeted NSQIP, in terms of AAA management, are generalizable to all NSQIP hospitals.