The Impact of Mandatory Reductions in Skilled Nursing Facility Cost-sharing for Medicare Advantage Plans

While the traditional Medicare program requires no copayments for the first 20 days of skilled nursing facility (SNF) care following a hospitalization, Medicare Advantage (MA) plans have historically been able to charge first-dollar cost-sharing for SNF care. Since this practice may contribute to favorable selection into Medicare Advantage, the Centers for Medicare and Medicaid Services have progressively regulated cost-sharing for SNF care in MA. CMS regulations implemented in 2014 and 2015 banned some plans from imposing cost-sharing for the first 20 days of SNF care and prohibited remaining plans with low out-of-pocket limits from charging more than $40 per day.

This difference-in-differences study analyzes the impact of these cost-sharing restrictions on use of SNF and hospital care and plan disenrollment. Data sources include plan benefit information, Medicare enrollment data, Minimum Data Set records, and MedPAR inpatient data from 2012-2016. We conducted stratified analyses for beneficiaries age 80 and above, those hospitalized in the past 6 months, and those hospitalized in the past 12 months for a lower joint replacement, surgical hip fracture repair, or sepsis. Outcomes were modeled at a plan-month level using generalized linear models with plan and month fixed effects and robust standard errors clustered at the plan level.

We compared 392 plans (with 2,033,986 members) that had to eliminate or reduce cost-sharing for the first 20 days of a SNF stay to 211 plans (with 1,179,477 members) that already had lower SNF cost-sharing prior to 2014. Among plans with mandatory cost-sharing reductions, the average cost-sharing reduction between 2013 and 2015 for the first 20 SNF days was $-716 and $-565 for plans with higher and lower out-of-pocket limits, respectively.

Prior to mandatory SNF cost-sharing reductions, plans averaged 124.6 SNF days, 5.0 SNF admissions, and 115.9 hospital days per 1,000 members per month. Preliminary results indicate that mandatory cost-sharing reductions were not associated with changes in the average monthly number of SNF days (difference-in-differences estimate: -2.3 [95% CI: -6.1, 1.4], p=0.22) or SNF admissions (difference-in-differences estimate: 0.0 [95% CI: -0.1, 0.1], p=0.62) per 1,000 plan members. Hospital days declined slightly more in plans with mandatory cost-sharing reductions (difference-in-differences estimate: -3.6 days per 1,000 plan members [95% CI: -6.5, -0.8], p=0.01) than in plans without these reductions. In stratified analyses, only two populations demonstrated any increase in SNF use associated with mandatory cost-sharing reductions: beneficiaries with joint replacement had increases in the average monthly number of SNF days (difference-in-differences estimate: 48.1 (3.2, 93.0), p=0.04) and beneficiaries with hip fracture had increases in the average monthly number of SNF admissions (difference-in-differences estimate: 5.6 [95% CI: 1.2, 9.9], p=0.01). These changes represented about a 5% increase relative to 2013 SNF use for these populations. Data will also be analyzed about the number of plan members and SNF users who disenroll from plans before and after cost-sharing restrictions.

These findings suggest that mandatory reductions in cost-sharing did not produce meaningful changes in SNF or hospital use but may have substantially lowered out-of-pocket costs for SNF users in MA plans.