Favorable Selection of Veterans in Medicare Advantage

Background: Prior literature has shown that favorable selection of enrollees in Medicare Advantage (MA) that is not accounted for under the Hierarchical Condition Category (HCC) risk adjustment model can result in significant overpayments to MA plans. Detailed data from the Veterans Health Administration’s (VHA) Nosos risk score can measure previously undetected favorable selection in a unique cohort of patients. Objective: To analyze characteristics associated with Veteran enrollment in MA versus Traditional Medicare (TM) and quantify previously undetected favorable selection. Research Design: Pooled, cross-sectional study using nationally representative VHA survey data and VHA administrative data, 2016–2019. Subjects: Dual VHA-Medicare enrollees. Measures: Enrollee characteristics, risk scores, and VHA inpatient and outpatient costs. Results: Compared with VHA-TM enrollees, VHA-MA enrollees were older [73.8 (8.5) vs. 72.2 (8.2) y; SMD=0.19], more likely to have Medicaid coverage (13.7% vs. 4.7%; SMD=0.31), less likely to use VHA care (67.4% vs. 74.3%; SMD=0.26), and more likely to be in the lowest priority group (facing VHA copays) (21.4% vs. 17.6%; SMD=0.15). However, in 2019, MA enrollees had Nosos scores that were 25.6 percentage points lower (95% CI: −35.7, −15.5) than TM enrollees, signaling a healthier population. In adjusted comparisons, MA enrollees had $453.79 lower VHA costs (95% CI: $832.30, $75.27) than TM enrollees. Conclusions: MA plans enroll Veterans who are healthier and less costly than TM-enrolled Veterans. Our findings underscore plans’ potential selection of enrollees based on characteristics unobservable in the HCC risk adjustment model. Policymakers may consider opportunities to adopt additional risk-adjustment factors specifically for dual VHA-MA enrollees.