Abstract 15072: Thirty Day Episode of Care Spending Following Heart Failure Hospitalization Among Medicare Beneficiaries With Heart Failure

Background: Heart failure (HF) is one of the costliest conditions in the US with immense public health interest to examine and improve the value of care. We sought to characterize current payments spanning the index hospitalization through 30-days post-discharge for Medicare beneficiaries with HF. Methods: Using Medicare fee-for-service administrative claims data, we identified patients hospitalized with HF from 2016-2018 with the following primary discharge diagnoses (ICD-10 codes): systolic HF (50.2 and 50.4), diastolic HF (50.3), hypertensive heart disease (HHD) with HF (I11), and HHD with HF and chronic kidney disease (CKD) (I13). Coding patterns over time across these four groups, mean 30-day episode of care spending overall, and proportion of total costs allocated to the index hospitalization and post-acute care were analyzed. Results: The study sample included 174,539 patients hospitalized with systolic HF; 163,071 diastolic HF; 221,820 HHD with HF; and 359,950 HHD with HF and CKD. Over time, there was a substantial increase in the use of the HHD with HF +/- CKD codes, accounting for 52% of HF hospitalizations in 2018 (Figure). We found substantial spending on 30-day episode for HF with nearly 2-fold variation across mean spending. Overall, 30-day episode spending was significantly higher for HHD with HF and CKD as compared to other diagnoses. Across all codes, the index hospitalization accounts for 70% of total episode spending while 30% is accounted for by post-acute care spending (Table). Conclusions: This patient episode-level analysis of contemporary Medicare beneficiaries suggests several areas to improve the value of HF care. The 30-day episode of care spending for heart failure hospitalizations is substantial with more than 2-fold variation. The impact of alternative payment models on HF outcomes and spending remains an important area for ongoing investigation.