Pulmonary resections performed at hospitals with thoracic surgery residency programs have superior outcomes

Objective: Pulmonary resections are performed at thoracic residency (TR), general surgery residency (GSR), no surgery residency, and no residency hospitals. We hypothesize that morbidity and mortality for these procedures are different between hospitals and that operations performed at TR teaching hospitals have superior results. Methods: Records of adults who underwent pneumonectomy, lobar, segmentectomy, and nonanatomic wedge resections (N = 498,099) were evaluated in an all-payer inpatient database between 2003 and 2009. Hospital teaching status was determined by linkage to Association of American Medical College's Graduate Medical Education Tracking System. Multiple hierarchical regression models examined the in-hospital mortality, occurrence of any complication, and failure to rescue. Results: The mean annual pulmonary resection volume among hospitals was TR (16%), GSR (17%), no surgery residency (28%), and no residency (39%). Unadjusted mortality for all procedures was lowest at TR hospitals (P < .001). Likewise, any complication was least likely to occur at TR hospitals (P < .001). After case-mix adjustment, the risk of any complication after segmentectomy or nonanatomic wedge resection was lower at TR hospitals than in GSR hospitals (P < .001). Among pneumonectomy recipients, TR hospitals reduced the adjusted odds ratio of failure to rescue by more than 25% compared with no surgery residency (P < .001). Likewise, in patients who underwent pneumonectomy, TR centers were associated with reducing the odds ratio of death by more than 30% compared with GSR hospitals (P < .001). Conclusions: In comparison with other hospitals, including GSR hospitals, TR hospitals have lower morbidity and mortality. These results support using hospitals with a TR as an independent prognostic indicator of outcomes in pulmonary resections.