Identifying high-quality bariatric surgery centers: hospital volume or risk-adjusted outcomes?

BACKGROUND: Payers and professional organizations are expanding accreditation and "centers of excellence" programs in bariatric surgery. Rather than directly measuring outcomes, most programs rely on procedure volume. We sought to determine whether risk-adjusted outcomes or hospital volume were better at predicting future hospital morbidity with bariatric surgery. STUDY
DESIGN: We identified all patients who underwent gastric bypass in the New York State Inpatient database (n = 32,381 patients, n = 105 hospitals). Morbidity was ascertained using a previously validated combination of diagnostic and procedure codes. We first calculated the risk-adjusted morbidity and volume at each hospital during a 2-year period (2003 to 2004). We then ascertained the proportion of hospital-level variation explained by each measure using hierarchical modeling techniques. Finally, we compared the ability of each measure to predict future performance, as assessed with risk-adjusted morbidity, in the next 2 years (2005 to 2006).
RESULTS: Risk-adjusted morbidity explained 83% of future hospital-level variation in morbidity compared with only 21% for hospital volume. When comparing the "best" with the "worst" hospital quartiles, risk-adjusted morbidity predicted a more than fourfold difference in future performance (1.7% versus 7.2%; odds ratio [OR]: 4.5; 95% CI, 3.5 to 5.9). Hospital volume predicted only a twofold difference (2.5% versus 4.5%; OR: 1.9; 95% CI, 1.5 to 2.4) from the best to worst quartile.
CONCLUSIONS: Risk-adjusted morbidity is much better than hospital volume at predicting future performance with bariatric surgery. Rather than focusing on volume, accreditation and centers of excellence programs should focus more on directly measuring outcomes.