Lessons Learned from a Cross-Sector Partnership in Harmonizing Healthcare Utilization and Homelessness Data

Health and homelessness are bidirectionally correlated-- experiencing homelessness increases the chances of developing chronic health conditions and incurring lasting injuries, and disability increases risk factors associated with housing loss, as well as limits the suitability of much of the U.S. housing stock. In a contentious sociopolitical climate, expensive supportive housing interventions are met with public pushback, creating a need for local governments to justify spending on increased accessible, affordable housing with in-home health and mental health services for individuals experiencing homelessness. This presentation will outline the process for harmonizing healthcare utilization and Homeless Management Information Systems (HMIS) datasets to facilitate such cost analyses. The presentation will examine the partnership process between a multidisciplinary academic team at a research university, a regional overseer of healthcare systems data, a municipal government, and the non-profit designated to record all homelessness systems data. The study asks the question: what methodological insights were learned from a collaborative cross-sector partnership to evaluate cost offsets of homeless services? The research team will present how localized differences in data ownership and collection may impact fidelity of best practices, as well as sharing methods for adapting best practices based on community needs, limitations, and assets. The presentation will also discuss the analytic strengths and limitations of measuring healthcare utilization costs across three groups of unaccompanied adults: (1) those receiving scattered-site permanent supportive housing services; (2) those receiving street outreach services who did not receive housing placement; and (3) those categorized as undomiciled by healthcare institutions who were not recorded as receiving any support from county homelessness services. The methods emergent from this project may provide guidance for future social work research conducting evaluation with sensitive data across multiple systems, including rigorous adaptations to recognized best practices for harmonizing and analyzing homelessness data. This may also lead to further consideration of how healthcare data can be utilized for housing intervention evaluation.