Volume-outcome relationships for head and neck cancer surgery in a universal health care system.

Abstract

We aimed to assess whether surgeon and/or institution resection volume predicts long-term overall survival in head and neck cancer in a publicly funded healthcare system. Population-based retrospective cohort study. Head and neck cancer patients in Ontario, Canada, who underwent a resection confirmed by both hospital-level and physician-level administrative data between 1993 and 2010, comprised our cohort (N = 5,720). Physician and hospital volumes were calculated based on number of cases performed in the year prior by the physician and at an institution performing each case, respectively. A multilevel hierarchical Cox regression model was used to estimate the effect on overall survival of each 25 increase in procedure volume. A crude model without patient or treatment characteristics demonstrated that both surgeon volume (hazard ratio [HR]: 0.927, 95% confidence interval [CI]: 0.879-0.978, P = .006) and hospital volume (HR: 0.980, 95% CI: 0.970-0.991, P = .0003) were associated with improved overall survival. After controlling for clustering and patient/treatment covariates, hospital volume (HR: 0.976, 95% CI: 0.955-0.997, P = .02), but not physician volume (HR: 1.042, 95% CI: 0.941-1.155, P = .43), remained a statistically significant predictor of overall survival. This translates into a 2.4% decrease in the HR for every 25 additional cases performed at an institution. Both high-volume surgeons and hospitals are predictors of better overall survival in head and neck cancer patients. However, the effect is largely explained by hospital volume. This benefit, at the institution level, could potentially be explained by important processes of care that contribute to overall survival.