Bundled payments for care improvement initiative – insights from the test pilots of payment reform

Abstract

Background: The Medicare Bundled Payments for Care Improvement (BPCI) pilot program aims to reward high-value providers by setting a global payment target for particular episodes of care. The representativeness of BPCI participants will influence the ability of this pilot to inform policy decisions.
Methods: We linked the Medicare lists of participants in the risk-bearing portion of BPCI Model 2, encompassing acute and post-acute care, to the American Hospital Association resource file and the 2013 Hospital Value-Based Purchasing quality performance data. We classified episode-initiating hospitals by the number of bundles in which they were participating into “narrow”, “medium” and “comprehensive”. The analysis described the characteristics of hospitals in each of these categories.
Results: The 105 hospitals with linkable data were predominantly large, urban, non-profit, teaching hospitals. These hospitals were quite similar to the general population in terms of disproportionate share, Medicare, and Medicaid percentages. Most participants selected a narrow number of bundles, with the majority selecting a single bundle around joint replacement. There were only minor differences in quality between Model 2 participants and non-participants.
Conclusions: Informing the decision about whether to scale the BPCI program nationally will require evaluation of the pilot’s performance by participants’ characteristics to understand in what conditions and for which providers the program is most effective.