Quality, not volume, determines outcome of coronary artery bypass surgery in a university-based community hospital network

Objectives: The present study examined the relationship between hospital and surgeon coronary artery bypass grafting procedural volume, mortality, morbidity, and National Quality Forum care processes in a university-based community hospital quality improvement program.

Methods: The study population consisted of 2218 consecutive patients undergoing isolated coronary artery bypass grafting from 2007 to 2009 in a university-based quality improvement program that emphasizes involvement of all surgeons in the academic quality endeavor. The endpoints included operative mortality, major morbidity, and National Quality Forum-endorsed process measures as defined by the Society of Thoracic Surgeons. The procedural volume was analyzed as a categorical and continuous variable using general estimating equations, which accounted for clustering effects and which were adjusted for Society of Thoracic Surgeons risk scores and the propensity for operation in a low- versus high-volume program.

Results: The annual program volume ranged from 67 to 292 (median, 136; interquartile range, 88-224) and surgeon volume from 1 to 124 (median, 58; interquartile range, 30-89). The mortality rate among the hospitals was 0.47% to 2.23% (0.8% overall), and the observed/expected mortality ranged from 0 to 1.20 (0.41 overall). When comparing low-volume (<200 cases/year) and high-volume centers, no difference was found in the mortality (odds ratio [OR], 1.08; 95% confidence interval [CI], 0.46-2.54, P = .85), morbidity (OR, 1.34; 95% CI, 0.73-2.43), or any of the medication process measures. No difference was found in mortality (OR, 1.59; 95% CI, 0.81-3.13; P = .18), morbidity (OR, 1.20; 95% CI, 0.86-1.66; P = .28), or medication failure (OR, 0.57, 95% CI, 0.3-1.10; P = .10) between the high- and low-volume surgeons (<87). After adjustment for both the Society of Thoracic Surgeons risk score and the propensity score, no association was found for either hospital or surgeon volume with mortality or morbidity. However, a lack of compliance with National Quality Forum measures was highly predictive of morbidity (OR, 1.51; 95% CI, 1.18-1.93; P = .001), regardless of volume, even after adjustment for predicted risk.

Conclusions: In the setting of a university-based community hospital quality improvement program, excellent surgical results can consistently be obtained even in relatively low-volume programs. The surgical outcomes were not associated with program or surgeon volume, but were directly correlated with the focus on quality as manifested by compliance with evidence-based quality standards. Meaningful university affiliation might represent a new quality paradigm for cardiac surgery in the community hospital setting.