Abstract
IMPORTANCE: Compared with traditional Medicare (TM), Medicare Advantage (MA) insurers have greater financial incentives to reduce the delivery of low-value services (LVS); however, there is limited evidence at a national level on the prevalence of LVS utilization among MA vs TM beneficiaries and whether LVS utilization rates vary among the largest MA insurers. OBJECTIVE: To determine whether there are differences in the rates of LVS delivered to Medicare beneficiaries enrolled in MA vs TM, overall and by the 7 largest MA insurers. DESIGN, SETTING, AND PARTICIPANTS: This cross-sectional study included Medicare beneficiaries aged 65 years and older residing in the US in 2018 with complete demographic information. Eligible TM beneficiaries were enrolled in Parts A, B, and D, and eligible MA beneficiaries were enrolled in Part C with Part D coverage. Data analysis was conducted between February 2022 and August 2024. EXPOSURES: Medicare plan type. MAIN OUTCOMES AND MEASURES: The primary outcome was utlization of 35 LVS defined by the Milliman Health Waste Calculator. An overdispersed Poisson regression model was used to calculate estimated margins comparing risk-adjusted rates of LVS in TM vs MA, overall and across the 7 largest MA insurers. RESULTS: The study sample included 3 671 364 unique TM beneficiaries (mean [SD] age, 75.7 [7.7] years; 1 502 631 female [40.9%]) and 2 299 618 unique MA beneficiaries (mean [SD] age, 75.3 [7.3] years; 983 592 female [42.8%]). LVS utilization was lower among those enrolled in MA compared with TM (50.02 vs 52.48 services per 100 beneficiary-years; adjusted absolute difference, -2.46 services per 100 beneficiary-years; 95% CI, -3.16 to -1.75 services per 100 beneficiary-years; P < .001). Within MA, LVS utilization was lower among beneficiaries enrolled in HMOs vs PPOs (48.03 vs 52.66 services per 100 beneficiary-years; adjusted absolute difference, -4.63 services per 100 beneficiary-years; 95% CI, -5.53 to -3.74 services per 100 beneficiary-years; P < .001). While MA beneficiaries enrolled in UnitedHealth, Humana, Centene, and smaller MA insurers had lower rates of LVS compared with those in TM, beneficiaries enrolled in CVS, Cigna, and Anthem showed no differences. Blue Cross Blue Shield Association plans had higher rates of LVS compared with TM. CONCLUSIONS AND RELEVANCE: In this cross-sectional study of nearly 6 million Medicare beneficiaries, utilization of LVS was on average lower among MA beneficiaries compared with TM beneficiaries, possibly owing to stronger financial incentives in MA to reduce LVS; however, meaningful differences existed across some of the largest MA insurers, suggesting that MA insurers may have variable ability to influence LVS reduction.