Medicaid Enrollment following the ACA Eligibility Expansion among Persons using Homelessness Services: An interrupted time-series analysis

Background Persons experiencing homelessness bear high rates of morbidity, injury, and mortality. Medicaid offers an opportunity to provide support, but barriers persist to enrolling and maintaining enrollment among this vulnerable population. The objective of this study was to determine the initial and longer-term effects of the Affordable Care Act (ACA) Medicaid eligibility expansion on Medicaid enrollment among persons observed to be unhoused or housing insecure through recorded housing services utilization. Methods We applied interrupted time-series analyses of linked administrative data from two expansion states – New Jersey (NJ) and Pennsylvania (PA); Homeless Management Information System data were linked to monthly Medicaid enrollment files of non-elderly adults (aged 18-64) utilizing shelter, street outreach and other housing assistance from January 2011 to December 2016. The study outcome was a binary measure of Medicaid enrollment status in month of Homelessness Management Information System service exit overall and stratified by race/ethnicity. Results ACA Medicaid eligibility expansion was associated with a level change in the likelihood of enrollment of 7.5 percentage points (pp) in NJ and 8.5 pp in PA. The trend in enrollment post-expansion also increased by 0.6 pp/month in NJ. Compared to no homeless-related service use in the year, being recorded with one month with a shelter stay or other homelessness assistance services in the year was associated with a higher likelihood of Medicaid enrollment (14.8 pp higher in NJ and 6.9 pp higher in PA), and likelihood of enrollment was highest when two or more months with homelessness assistance services were used in the year (18.6 pp higher in NJ and 12.8 pp higher in PA). However, the effect of the policy change attenuated back to the pre-ACA trend in both states by the end of 2016. Results were similar across race/ethnicity stratifications. Conclusions We found significant increases in the likelihood of Medicaid enrollment after the ACA Medicaid expansions in the months immediately following the expansion. Additional months with homeless shelter stay or other housing services for unhoused persons were associated with a higher likelihood of Medicaid enrollment; this suggests the need for further investigation into the potential of leveraging staff-client relationships at homelessness assistance programs during future health policy initiatives.