Hospital-Physician Integration and Cardiac Surgery Outcomes: A U-Shaped Relationship?

Abstract

Hospital-physician integration has been increasingly considered a potential solution for the underlying challenges hospitals face as they are adapting to value-based healthcare services. We adopt an activity-based measure of integration (ABI) to investigate the relationship between integration and care outcomes, namely in-hospital mortality risk, length of stay (LOS), and 30-day readmission risk. ABI is measured based on a group of physicians who handle a specific procedure, coronary artery bypass graft (CABG), that is, CABG physicians, and is operationalized as the proportion of cases handled by the CABG physicians who single-site (or concentrate all their activity) at a focal hospital. To test hypotheses that posit a U-shaped relationship between ABI and patient care outcomes, we utilize patient-visit level information for Florida patients who underwent cardiac surgery performed by CABG physicians during 2011–2014. We find that ABI has a U-shaped relationship with both mortality risk and LOS, such that patient mortality risk and LOS are minimized at ABI tipping points of about 55%. In contrast, 30-day readmission risk continues to decrease as ABI increases. We also find that hospital teaching status and bed utilization moderate the relationship between ABI and LOS, such that the U-shaped relationship is flatter, basically linear in teaching and/or high-utilization hospitals. Our results suggest that a medium level of integration could be desirable, since a strategy of high integration trades off potentially higher patient volumes and revenues for suboptimal care outcomes. Overall, this study offers new insights for theory and practice, as the non-linear association between integration and care outcomes has not been investigated previously.