Jama Network
April 22, 2026
Abstract
IMPORTANCE When becoming eligible for Medicare, adults previously covered by Medicaid expansion are typically required to meet an asset test for Medicaid coverage of Medicare cost-sharing and/or premiums through the Medicare Savings Program (MSP). It is unclear whether MSP or overall Medicaid coverage is higher in states that waive the MSP asset test than in states that do not.
OBJECTIVE To examine patterns of Medicaid coverage in the first year of Medicare enrollment among beneficiaries who previously enrolled in Medicaid through expansion.
DESIGN, SETTING, AND PARTICIPANTS This cohort study used Medicare and Medicaid enrollment data from January 1, 2017, to December 31, 2019, to identify beneficiaries who entered Medicare in 2018 and had Medicaid expansion coverage in the month prior to entering Medicare. The analysis included 31 states plus Washington, DC, that adopted Medicaid expansion by January 2018, of which 6 states (Arizona, Connecticut, Delaware, New York, Oregon, and Vermont) plus Washington, DC, waived the MSP asset test. Analyses were conducted between May 2024 and February 2026.
EXPOSURE States without vs states with the MSP asset test.
MAIN OUTCOMES AND MEASURES MSP or any Medicaid coverage 1 year after Medicare entry and at least 6 months or 12 months of MSP or any Medicaid coverage during the first year of Medicare enrollment. Multivariable linear probability models were used to examine the probability of these outcomes among beneficiaries living in states without vs states with the MSP asset test.
RESULTS A total of 266 009 adults with Medicaid expansion coverage entered Medicare in 2018. These beneficiaries had a mean (SD) age of 59 (9) years and included 141 821 females (53%). Overall, 62% of beneficiaries entered Medicare due to reaching 65 years or older; 79% of beneficiaries lived in states with the MSP asset test and 21% lived in states without the MSP asset test. At the 12th month of their first year of Medicare enrollment, 48% had MSP coverage (48% vs 47% in states with vs without the MSP asset test, respectively), 61% had any Medicaid coverage (59% vs 67%, respectively). In adjusted analyses, beneficiaries in states without the MSP asset test were more likely than those living in states with the test to have any Medicaid coverage at the end of their first year of Medicare enrollment (difference, 10 [95% CI, 3-17] percentage points) and to have at least 6 months (difference, 14 [95% CI, 5-22] percentage points) or 12 months (difference, 16 [95% CI, 6-26] percentage points) of Medicaid coverage in their first year, but they were not more likely to have MSP coverage on these measures
Citation
Newhouse, Joseph P. "Medicaid Retention After Transition to Medicare Among Adults With Expansion Coverage." Jama Network (April 22, 2026).