Panel Paper: Estimating Eligibility for Medicaid Homeless Support Service Benefits and Potential Healthcare Savings in New Jersey

Friday, November 9, 2018
McKinley - Mezz Level (Marriott Wardman Park)

*Names in bold indicate Presenter

Joel C. Cantor1, Sujoy Chakravarty1, Jose Nova1, Taiisa Kelly2, Derek DeLia3 and Richard Brown2, (1)Rutgers Center for State Health Policy, (2)Monarch Housing Associates, (3)Medstar Health Research Institute


Introduction. Providing housing-related support services (HSS) to chronically homeless individuals with disabilities can lead to improved health and reduced avoidable healthcare spending, according to randomized trials and other studies among selected populations. Many such individuals enrolled in Medicaid in the 33 states that expanded eligibility under the Affordable Care Act (ACA), raising opportunities to improve health service utilization patterns and reduce avoidable Medicaid spending from well-targeted HSS benefits. Federal Medicaid guidance suggests that states may be granted authority to cover HSS such as tenancy and dispute resolution education, care coordination, and employment assistance.

Data and Methods. To support the design of a Medicaid HSS benefit for New Jersey, we linked six years (2011-2016) of statewide Homeless Management Information System (HMIS) data with comprehensive Medicaid eligibility, utilization and spending records. Using the linked data we estimate: (1) the number and characteristics of individuals potentially eligible for Medicaid HSS benefits under a range of possible eligibility rules; and (2) the extent of avoidable Medicaid healthcare use and spending among prospective HSS recipients. We focus on hospital inpatient and treat-and-release emergency department (ED) utilization, potentially preventable hospital admissions, 30-day all-cause readmissions, primary care use, as well as overall Medicaid expenditures.

Findings. Over 50% of persons in the HMIS matched to Medicaid data during each study year, both before and after the ACA expansion, while the number of matched beneficiaries increased by nearly 80% post-expansion. In preliminary analysis, we classified about 30% of persons in the linked data as “homeless” based on the type of residence (e.g., emergency shelter, “safe haven”, or place not meant for habitation) and housing program status documented in their final HMIS-recorded encounter in 2016. We found high rates of potentially avoidable healthcare use among Medicaid-covered homeless individuals. Nearly 60% of had at least one ED visit and 10% had six or more ED visits during the year. Further, one-in-five had at least one hospitalization, with over 20% readmitted within 30 days of initial discharge. This group also experienced 34 potentially preventable hospital admissions per 1,000 beneficiaries, and more than a third had no primary care visit during the year. Annual Medicaid spending in this cohort was over $10,000 per beneficiary (adjusted for months enrolled), roughly at the top quartile of the Medicaid spending distribution for NJ.

Preliminary Conclusions and Further Analysis Plans. NJ Medicaid covers a substantial population experiencing homelessness with high rates of potentially avoidable healthcare use and expenditures. Available evidence suggests that well-targeted HSS benefits may encourage more efficient healthcare use and reduce cost. In ongoing analyses informed by discussions with Medicaid and housing officials and stakeholders, we are refining our classification of individuals potentially eligible for HSS benefits and simulating the potential impact of HSS benefits on Medicaid beneficiaries’ health service utilization and cost. These analyses will be completed ahead of the fall APPAM research conference.