Quality Wrapped in Volume Inside a Hospital

Studies that documented surgical outcome differences proliferated in the late 1990s. High-volume hospitals (that is, those that performed larger numbers of a particular procedure) tended to have better outcomes than their lower-volume peers (1). Nested within these findings was a similar volume–outcome relationship for surgeons (2). When low- and high-volume hospitals were compared, some outcome gaps were very large. Begg and colleagues (3) reported a 14% absolute difference in 30-day mortality rates among patients who had esophagectomy for esophageal cancer. Other gaps were modest. Schrag and coworkers (4) reported a 2% absolute difference in mortality rates 30 days after colorectal cancer surgery. Investigators have also identified a volume–outcome relationship for coronary artery bypass grafting, which has been weak in some studies and stronger in others (5).

Policymakers noticed this volume–outcome relationship. A seminal Institute of Medicine report recommended that “patients undergoing procedures that are technically difficult to perform and have been associated with higher mortality rates in lower-volume settings receive care at facilities with extensive experience (i.e., high-volume facilities)” (6). At about the same time, a consortium of large, self-insured employers began the not-for-profit Leapfrog Group (http://www.leapfroggroup.org), which encouraged consumers to seek care in institutions that met certain quality standards, including annual procedure volume for some high-risk surgeries (7).

The prevailing explanation for the superior outcomes at high-volume centers is that volume is a proxy for experience, which in turn enhances outcomes. The alternative explanation—patients flock to centers that have superior outcomes, which increases those centers' volumes—is not plausible because hospitals do not routinely report their outcomes. Because experience seems to matter, investigators have sought the key structural and process components that emerge through experience (8).

In this issue, Auerbach and colleagues (9) purport to explain most of the volume–outcome relationships in coronary artery …