Medicare Advantage and post-acute care utilization

Medicare spending on post-acute care has grown substantially in recent years. Prior research has shown that this growth was partly related to Medicare’s payment of post-acute providers as opposed to clinical necessity. Prior to 1997, Medicare paid post-acute providers on a cost-basis, with little incentive for cost control. After the Balanced Budget Act of 1997, prospective payment was implemented for post-acute care providers (with per-diem payments for skilled nursing facilities and episode based payments for inpatient rehabilitation facilities). However, prospective payment provided little incentive to make more efficient use of post-acute care overall and provided different payment rates across settings for similar patients, potentially leading to financial incentives to send patients to particular post-acute settings. Recent years have witnessed considerable growth in the percentage of Medicare beneficiaries enrolled in Medicare Advantage. Unlike traditional Medicare, Medicare Advantage plans negotiate contracts with post-acute providers and have strong financial incentives to lower overall use of institutional post-acute care and length of stay.

In this paper, we investigate differences in post-acute care utilization between traditional Medicare and Medicare Advantage enrollees. We focus on hospital discharges for conditions that are typically high users of post-acute care: hip fracture, lower extremity joint replacement, and stroke.  We investigate discharges between 2008 and 2013 from hospitals that were required to submit information-only claims for Medicare Advantage patients (specifically, the ~80% of short-term acute hospitals that receive medical education or DSH payments from Medicare). We identify skilled nursing facility and inpatient rehabilitation facility stays for Medicare Advantage enrollees in the Minimum Data Set and Inpatient Rehabilitation Facility Patient Assessment Instrument files.  We investigate differences between Medicare Advantage and traditional Medicare in: (1) use of any inpatient rehabilitation or skilled nursing facility care, (2) the relative mix of inpatient rehabilitation versus skilled nursing facility care, (3) the length of post-acute stay, (4) the intensity of rehabilitation received, and (5) differences in hospital readmission rates and return to the community in the 90-days following the initial hospital discharge.

We start by making cross-sectional comparisons of hospital discharges for selected conditions covered by Medicare Advantage versus traditional Medicare controlling for demographic, socio-economic, and health characteristics from enrollment data and the qualifying hospital stay, as well as discharging hospital fixed-effects. Next, we examine the effects of ACA-related changes in Medicare Advantage payments, specifically the freezing of county payment benchmarks in 2010 and 2011, and starting in 2012, reductions in benchmarks for counties with high traditional Medicare spending and increased benchmarks for counties with low traditional Medicare spending.