A randomized controlled trial to evaluate a supportive housing program for people experiencing chronic homelessness and high rates of justice involvement and emergency department utilization.
Grant Recipient: Urban Institute
Principal Investigators: Sarah Gillespie, Devlin Hanson
Term: 2022 — 2030
Summary: The Denver Housing to Health (H2H) project is a randomized controlled trial (RCT) to evaluate a supportive housing program for people experiencing chronic homelessness and high rates of justice involvement and emergency department utilization. The Denver H2H program consists of a Housing First approach (i.e., quickly getting people into housing without any preconditions or requirements) as well as intensive wraparound services including modified assertive community treatment (ACT), behavioral health services, links to community resources, transportation assistance and referrals, and assistance navigating the criminal justice system.
The Denver H2H program builds on the success of the Denver Social Impact Bond (SIB) Supportive Housing program, which was evaluated in a well-conducted RCT with a population of individuals experiencing chronic homelessness with multiple prior arrests. The Denver SIB study found that, over the three year period after random assignment, the program significantly reduced (i) the use of jails by 28% (from 138 jail days in the control group to 100 jail days in the treatment group); (ii) the use of shelters by 35% (from 269 shelter days in the control group to 174 shelter days in the treatment group); and (iii) the number of emergency department visits (from 15.2 visits in the control group to 9.2 visits in the treatment group). A cost analysis found that the annual reduction in the use of these and other public services amounted to roughly half of the annual cost of the program (estimated at $12,000 — $15,500 per person per year) and that the savings were greater for individuals with more emergency department visits prior to randomization.
Through the Denver H2H evaluation, the researchers seek to evaluate the effectiveness of supportive housing with a more targeted population of individuals experiencing chronic homelessness with multiple prior arrests and multiple prior emergency department visits to determine whether a greater share of the program’s cost can be offset by savings among this population. The researchers will randomly assign 440 such individuals to either (i) receive supportive housing through the program or (ii) a control group that will receive usual care. The primary outcomes of interest are: (i) total healthcare costs, as measured by the total amount of claims billed to Medicaid and (ii) healthcare costs paid by the federal Medicaid program, as measured by the total amount actually paid by Medicaid. Other secondary outcomes of interest include housing stability, shelter use, police contacts and arrests, jail stays, and use of short-term detoxification or sobering services. All outcomes will be measured using administrative data through five years after random assignment.
The study’s pre-specified analysis plan is linked here.