M. M. Sellers1,2, R. L. Hoffman1, C. Wirtalla1, G. C. Karakousis1, R. R. Kelz1 1Hospital Of The University Of Pennsylvania,Center For Surgery And Health Economics Department Of Surgery,Philadelphia, PA, USA 2Mount Sinai School Of Medicine,Depar,New York, NY, USA
Introduction: The feasibility of ranking residency programs based on the clinical outcomes of their graduates has previously been established. Colorectal resection is amongst the most common operations performed by general surgeons. Our objective was to assess general surgery residency program performance in colorectal surgery education using the objective clinical outcomes of patients operated on by program graduates.
Methods: A retrospective cohort study was conducted of patients that underwent a colorectal resection in New York or Florida (2008-2011). After linking with data from the American Medical Association, the cohort included 47,147 patients operated on at 343 hospitals, by 856 surgeons who trained at 74 distinct general surgery residency programs. A hierarchical generalized linear model, risk adjusted for patient, hospital and surgeon characteristics including surgeon specialty, was used to assess the independent association between residency program and adverse events (AE).
Results: The observed AE rates were 3.4% for death, 44.3% for any AE, 39.6% for colorectal-specific AE (CSAE), and 23.4% for prolonged length of stay (pLOS). Patients operated on by surgeons trained in residency programs ranked in the top tertile were significantly less likely to experience an AE than were patients operated on by surgeons trained in residency programs ranked in the bottom tertile (2.8% vs 5.4% for death, 41.6% vs 51.9% for any AE, 36.5% vs 46.3% for CSAE, and 22.9% vs 27.2% for prolonged length of stay (all P < .001)). Adjusted adverse event rates for patients operated on by surgeons trained in top tertile programs were marginally lower than those who were operated on by surgeons trained in bottom tertile programs (see Table). The model C statistics ranged from .76 to .87. The proportion of variation explained by the model ranged from 15.3% to 23.1%.
Conclusion: Unadjusted outcomes of patients treated by surgeons who trained in programs ranked in the top and bottom tertiles differed significantly across all of the outcomes studied. The magnitude of the differences was small after risk adjustment. General surgery program performance is fairly homogeneous when compared by their graduates’ patients’ outcomes following colorectal surgery.
