Abstract 19: Episode of Care Payments Across Two Cardiac Conditions: Where Does the Money Go?

Objective: One approach to reduce health care spending and improve coordination of care is to pay for an episode of care rather than individual services. Anchoring these episodes around an index hospitalization is sensible because hospitalizations are a leading contributor to rising healthcare costs and an index admission provides a clear time to begin the episode. Understanding which care settings are responsible for a greater proportion of expenditures can inform efforts to improve efficiencies in the care provided. Our objectives were to: 1) characterize total episode payments for two conditions and 2) examine the care settings accounting for the highest proportions of the 30-day episode of care payment. Study Design: We used Medicare Claims data from 2008 and 2009 to identify hospital discharges for acute myocardial infarction (AMI) and heart failure among fee-for-service beneficiaries ≥ 65 years. We defined an episode as admission plus 30 days. To reflect differences in payments due to clinical care rather than geographic or policy adjustments, we omitted or averaged payment adjustments such as wage index and indirect medical education. For hospitalizations involving transfers, we summed the per diem and diagnosis related group payments and attributed this amount to the transfer-out hospital. Principal Findings: For AMI and heart failure patients respectively, the median unadjusted episode payment was $15,581 (interquartile range (IQR) $11,416, $24,675); and $10,139 (IQR $6,972, $17,827). For AMI, 77% of total payments were for index admission and 23% for post-acute care. Among post-acute care payments, 35% were for readmissions, 30% for skilled nursing facilities (SNF), and 13% for non-acute inpatient stays (i.e., inpatient rehabilitation, inpatient psychiatric facilities, and long-term care facilities). For heart failure, 61% of total payments were for index admission and 37% for post-acute care. Among post-acute care payments for heart failure, 35% were for readmissions, 33% for SNF, and 7% for non-acute inpatient stays. Conclusions: Total episode payments vary by condition. For heart failure, which does not routinely require procedures, 61% of total episode payments are attributable to the index admissions compared to 77% for AMI. Across both conditions, the highest proportion of payments made after discharge were for readmissions, SNF, and non-acute inpatient stays. These results can inform efforts to reduce costs, including payment reforms such as bundled payment initiatives.