Rethinking Provider Quality and Behavior: Evidence from Exemplars, Guidelines, and Teams

Abstract

Improving the quality of U.S. healthcare delivery, particularly care delivered by healthcare professionals, is a key priority in research, policy, and practice. My dissertation investigates whether and under what conditions providers learn about and from each other, by using tools from economics and empirical management, as well as data from claims, patient experiences, electronic health records, and physician training. In Chapter 1 (coauthored with J. Michael McWilliams), I study the case of former chief medical residents — chosen by the profession for their teaching, interpersonal, and clinical skill. I find that patients of former chief primary care physicians (PCPs) reported better care experiences than did patients of other PCPs in the same practice, especially on physician-specific measures. My results suggest that the profession possesses meaningful information about physician quality that should be more regularly harnessed to improve quality. In Chapter 2 (coauthored with Maximilian J. Pany, Bethany Sheridan, and Robert S. Huckman), I use a quasi-experimental difference-in-regression-discontinuity study design to compare provider behavior at the diabetes diagnosis cutoff across various team characteristics. I find that larger and more diffuse primary care teams, which may be comprised of physicians, nurse practitioners, and physician assistants, respond more strictly to diabetes diagnosis guidelines. My results imply that managers should consider team structure to influence adherence to clinical practice guidelines. In Chapter 3, I use multilevel mixed effects models to explore the variation in coronary artery bypass graft surgery outcomes attributable to the surgeon dyad — over and above the primary surgeon and the assistant surgeon, and controlling for hospital fixed effects. Though the surgeon dyad does not appear to explain any variation in mortality, I find meaningful effects for rates of acute cardiovascular complications. My results suggest that hospitals and policymakers may want to better understand surgeon dyad-level outcomes and potentially improve match quality between primary and assistant surgeons. In summary, my dissertation contributes to the literature on the determinants and implications of provider quality and explores novel approaches to improve quality, especially as providers increasingly work in groups, teams, and organizations.