Changes in Cardiovascular Spending, Care Utilization, and Clinical Outcomes Associated With Participation in Bundled Payments for Care Improvement – Advanced

BACKGROUND: Bundled Payments for Care Improvement – Advanced (BPCI-A) is a Medicare initiative that aims to incentivize reductions in spending for episodes of care that start with a hospitalization and end 90 days after discharge. Cardiovascular disease, an important driver of Medicare spending, is one of the areas of focus BPCI-A. It is unknown whether BPCI-A is associated with spending reductions or quality improvements for the 3 cardiovascular medical events or 5 cardiovascular procedures in the model. METHODS: In this retrospective cohort study, we conducted difference-in-differences analyses using Medicare claims for patients discharged between January 1, 2017, and September 30, 2019, to assess differences between BPCI-A hospitals and matched nonparticipating control hospitals. Our primary outcomes were the differential changes in spending, before versus after implementation of BPCI-A, for cardiac medical and procedural conditions at BPCI-A hospitals compared with controls. Secondary outcomes included changes in patient complexity, care utilization, healthy days at home, readmissions, and mortality. RESULTS: Baseline spending for cardiac medical episodes at BPCI-A hospitals was $25 606. The differential change in spending for cardiac medical episodes at BPCI-A versus control hospitals was $16 (95% CI, −$228 to $261; P =0.90). Baseline spending for cardiac procedural episodes at BPCI-A hospitals was $37 961. The differential change in spending for cardiac procedural episodes was $171 (95% CI, −$429 to $772; P =0.58). There were minimal differential changes in physicians’ care patterns such as the complexity of treated patients or in their care utilization. At BPCI-A versus control hospitals, there were no significant differential changes in rates of 90-day readmissions (differential change, 0.27% [95% CI, −0.25% to 0.80%] for medical episodes; differential change, 0.31% [95% CI, −0.98% to 1.60%] for procedural episodes) or mortality (differential change, −0.14% [95% CI, −0.50% to 0.23%] for medical episodes; differential change, −0.36% [95% CI, −1.25% to 0.54%] for procedural episodes). CONCLUSIONS: Participation in BPCI-A was not associated with spending reductions, changes in care utilization, or quality improvements for the cardiovascular medical events or procedures offered in the model.