Association for Public Policy Analysis & Management

Financial incentives stand out as a mechanism for improving quality.  However, we lack an understanding of how physicians respond when only a subset of their patients are affected.  In this paper, we use the quasi-random assignment of beneficiaries to Accountable Care Organizations (ACOs) to identify how providers respond when facing a patchwork of different financial incentive schemes.  ACOs alter the financial incentives for providers in at least two ways.  First, providers might share some of the financial costs associated with treatment, so if incentives are powerful enough, they will contain costs.  Second, providers are rewarded more directly for improvements in certain dimensions of quality and patient outcomes.  As a result of these twin forces, providers, physicians in particular, should in theory weigh costs against quality in a more meaningful way, thereby generating health improvements at lower costs. 

Yet, much remains uncertain about how the theory links to real-world practice.   First, it is possible that the health outcomes of all patients—regardless of whether they are aligned with an ACO-provider—improve, suggesting that physicians remake their entire style of practice across all patients.  Another possibility is that only ACO-aligned patients are affected, implying that physicians finely calibrate their practice patterns to the financial incentive associated with each patient.  Finally, if the health outcomes of ACO-aligned patients do not change, it is possible that the incentives ACOs place on providers are too weak or financial incentives do not play a meaningful role in clinical decision-making. 

Using Medicare claims data and the quasi-random assignment of Medicare beneficiaries to Pioneer ACOs, we find that when physicians have a sufficient number of ACO-aligned beneficiaries, they respond by reducing the number of procedures performed across all of their patients.  Health outcomes also improve across both ACO and non-ACO aligned beneficiaries.  This finding suggests that financial incentives have blunt and widespread effects where the dominant payment arrangement has spillover effects on the health outcomes of all of their patients.