LAUSR

Medicare's Bundled Payments for Care Improvement Advanced Model (BPCI-A) is a voluntary Alternative Payment Model in which participating hospitals are held accountable for ninety-day episodes of care. To meet spending targets, hospitals must either decrease utilization or attract a less sick patient population; this could lead to the elimination of necessary care or avoidance of patients with medical or social vulnerability. We used publicly available data on BPCI-A participation, along with Medicare claims from the period 2017-19, to examine patient selection, changes in Medicare payment, and key clinical outcomes among three groups: patients with frailty, patients with multimorbidity, and patients with dual enrollment (both Medicare and Medicaid). We found no consistent change in patient selection associated with BPCI-A participation. Patients with frailty, multimorbidity, or dual enrollment were more expensive at baseline, but Medicare payments decreased similarly in these groups compared with lower-risk patients. There were no differential negative changes in clinical outcomes between BPCI-A participants and nonparticipants among patients with medical or social vulnerability.