Regional variation in patient risk factors and mortality after coronary artery bypass grafting

Background: Geographic variations in patient risk factors and operative mortality after coronary artery bypass graft surgery have not been well studied. Methods: Using The Society of Thoracic Surgeons National Cardiac Database, a retrospective cohort study was performed of patients undergoing isolated coronary artery bypass graft surgery from 2004 to 2007 (n = 504,608). Records were sorted into four major geographic regions (Northeast, Midwest, South, and West) and compared with respect to patient risk profiles and outcomes. Using marginal and hierarchical logistic regression, risk-adjusted operative mortality rates were compared across regions and variation assessed within regions, states and hospital referral regions. Results: Patient risk profiles in the Northeast and West appeared similar, as did profiles in the Midwest and South. Risk-adjusted mortality rates were as follows: Northeast 1.63%, Midwest 2.01%, South 2.25%, and West 1.82%. Compared with the Northeast, mortality rates in the Midwest and South were higher, with the following odds ratios (95% confidence intervals): Midwest 1.26 (1.12 to 1.42), South 1.44 (1.27 to 1.62), and West 1.12 (0.98 to 1.28). Major geographic regions accounted for 16.5% of the variation observed in mortality rates; states and hospital referral regions accounted for 17.8% and 65.7%, respectively. Conclusions: Variations in absolute coronary artery bypass graft surgery mortality rates across large regions were subtle, although rates within the Northeast were comparatively lower. Most of the variation was seen at the hospital referral region level. Given that geographic location has not been routinely incorporated into statistical risk model predictions, additional research appears warranted to identify regional "best care" practices and to advance nationwide improvements in cardiac surgical patient outcomes.