Effect of On-Site Cardiac Surgery Program on General Thoracic Surgery Outcomes.

OBJECTIVE Limited data exist exploring the relationship between multispecialty surgical collaboration and outcomes in general thoracic surgery. To address this, the Nationwide Inpatient Sample (NIS) was analyzed to determine whether the presence of an on-site cardiac surgery program is associated with improved general thoracic surgery outcomes. METHODS The NIS (1999-2008) was utilized to identify 389,959 patients who had a lobectomy, pneumonectomy, or esophagectomy. Short-term outcomes of patients undergoing these procedures were compared between hospitals with and without an on-site cardiac surgery program. Univariate and multivariate analyses were performed to determine patient and hospital predictors of mortality and morbidity. RESULTS During the study period, patients undergoing lobectomy (n = 314,130), pneumonectomy (n = 34,860), or esophagectomy (n = 40,969) were identified. Univariate analysis demonstrated lower mortality for lobectomy (P < 0.001) and esophagectomy (P < 0.001) but not pneumonectomy (P = 0.344) in hospitals with a cardiac surgery program. All-cause morbidity was significantly lower for all 3 procedures in hospitals with a cardiac surgery program. However, multivariate analysis demonstrated that a cardiac surgery program was not an independent predictor when adjusted for known confounders, particularly procedure volume and hospital academic teaching status. CONCLUSIONS The presence of an on-site cardiac surgery program is not in and of itself associated with improved general thoracic surgery outcomes. The presence of a cardiac surgery program is likely a surrogate for other known predictors of improved outcomes such as hospital teaching status and procedure volume.