As the popularity of Medicare Advantage (MA) plans increased, so did the disenrollment from Traditional Medicare (TM). In previous work, we found that increasing rates of disenrollment from TM accounted for two-thirds (67%) of the growth in the share of MA enrollment between 2019 and 2020. In this blog, we update previously reported disenrollment trends through 2020 with more recent data (2021-2024) and assess rates by different population groups. We separately report findings for Medicare-only beneficiaries and Medicare-Medicaid beneficiaries (dual eligibles), as well as for new Medicare beneficiaries and those who are in the last year of life.
Enrollment in MA continues to increase but at a slower rate due to less disenrollment from TM and more switching from MA to TM. Between 2023 and 2024, disenrollment from TM accounted for 53% of growth in new enrollment in MA, down from 67% between 2019 and 2020. Between 2016 and 2024, the TM disenrollment rate reached its highest point in 2021 at 8.1% and stayed at around the 8% mark until declining to 7% in 2024. The MA disenrollment rate, on the other hand, has increased from 1.9% to 2.7% between 2021 and 2024. In 2024, TM-to-MA switching rates were almost 3 times higher than switching rates from MA to TM (similar to our findings in 2020). However, this reflects a decrease from 2020 for Medicare-only enrollees (2020: 3.7 times higher disenrollment rates from TM-MA vs. MA-TM; 2024: 2.7 times higher) and a small increase for Medicare-Medicaid enrollees (2020: 2.5; 2024: 2.9 times).
Dual eligibles, new enrollees to Medicare, and some minority groups have substantially higher disenrollment rates, especially from TM.
The differential between TM and MA disenrollment rates is smaller for beneficiaries in their last year of life. Since 2020, disenrollment from TM exceeded that from MA for end-of-life beneficiaries but started to converge in 2022 for Medicare-only beneficiaries as rates of switching fell for beneficiaries in their last year of life (from 6.5% in 2021 to 5.5% in 2023) (Exhibit 2).
Conclusion