Objective: To examine the longitudinal relationship between surgeon volume and in-hospital mortality for coronary artery bypass graft (CABG) surgery in New York State and to explain changes in mortality that occurred over time.
Design: Observation of clinically risk-adjusted operative mortality over time.
Setting: All 30 New York State hospitals in which CABG surgery was performed for 1989 through 1992.
Patients: All 57,187 patients undergoing isolated CABG surgery in New York State in 1989 through 1992 in the 30 hospitals.
Main outcome measures: Actual, expected, and risk-adjusted mortality.
Results: Risk-adjusted in-hospital mortality decreased for all categories of surgeons. Low-volume surgeons (< or = 50 operations per year) experienced a 60% reduction in risk-adjusted mortality in the 4-year period, whereas the highest-volume surgeons (> 150 operations per year) experienced a 34% reduction. The percentage of patients undergoing CABG surgery by low-volume surgeons decreased from 7.6% in 1989 to 5.7% in 1992, a 25% decrease.
Conclusions: The overall decline in risk-adjusted mortality could not be explained by shifts in patients away from low-volume surgeons to high-volume surgeons. The proportionately larger decrease in risk-adjusted mortality for low-volume surgeons could not be explained by changes in patient case mix or by improvements in the performance of surgeons with persistently low volumes. Part of the decrease was a result of the exodus of low-volume surgeons with high risk-adjusted mortality (in all years studied), the markedly better performance of surgeons who were new to the system (especially in 1991 and 1992), and the performance of surgeons who were not consistently low-volume surgeons (especially in 1992).