Home Health Care Use in Medicare Advantage Compared to Traditional Medicare: The Role of Benefit Design

Objective: To compare post-acute and community-admitted home health care use in Medicare Advantage (MA) to traditional Medicare (TM) nationally and by state, and to explore the potential role of home health benefit design in explaining differences between MA and TM.

Design: We compared home health care use between full-year MA and full-year TM beneficiaries in 2011 and 2016, overall and by state using the OASIS linked to the Master Beneficiary Summary File. We adjusted our estimates of home health use for differences in age, gender, race and ethnicity, dual-eligibility, original reason for Medicare entitlement, and eligibility for the Part D low-income subsidy. We also conducted nine interviews with MA plans and home health agencies to understand the role of MA home health benefit design features like cost-sharing, prior authorization, visit limits, and networks. We will also assess length of stay differences.

Findings: Qualitative interviews suggested that MA home health care use, particularly community-admitted use, would be lower and stays would be shorter than in TM because of per-visit payment approaches and utilization management by MA plans. Preliminary quantitative analyses found that MA enrollees were less likely than TM enrollees to use home health in both 2011 (6.5 percent vs 10.1 percent, unadjusted) and 2016 (6.7 percent vs 9.4 percent, unadjusted), though the gap narrowed over the study period. Home health use in MA was still 3.2 percentage points lower in MA than in TM in 2011 and 2.6 percentage points lower in 2016 after adjustment for differences in age, gender, race and ethnicity, dual-eligibility, reason for Medicare entitlement, or eligibility for Part D low-income subsidies. MA enrollees were less likely than TM enrollees to use both post-acute home health (4.1 percent vs 5.2 percent) and community-admitted home health (1.5 percent vs 2.5 percent) in 2016. Finally, home health care use rates in MA varied less by state than in TM in both 2011 and 2016, with states ranging from 3.2 percent to 11.4 percent in MA in 2016 (Hawaii and Maine, respectively), compared to 2.6 to 13.6 percent in TM (Hawaii and Florida, respectively).

Conclusions: MA enrollees have lower home health care use than TM nationally and in most states after controlling for differences in patient demographic factors. Qualitative interviews with MA plans and home health agencies suggested that MA plans control home health care use more tightly, limiting the number of visits, paying on a per-visit rather than a per-episode basis, and managing networks through certification processes. These approaches appear to be successful at reducing home health care use overall and at reducing community-admitted home health use, though it remains unclear whether MA use is too low, TM use is too high, or some combination of the two. By June 2019, the team will explore differences in length of stay between MA and TM adjusted for case mix, variations in home health care use by plan type and cost-sharing structure, and differences between MA and TM in the types of home health agencies used (e.g. quality ratings, profit status).