Exchanging Housing Dollars for Health Care Savings: The Impact of Housing First on Health Care Costs

Abstract

This paper assesses the effects of the Housing First model on the utilization of public health services and claims among chronically homeless individuals and in the process addresses the problem of attribution to natural stabilization on account of regression to the mean. The study uses Medicaid claims data from participants in a Massachusetts “Housing First” program (n = 1,342) and non-participants (n = 44,022) who included chronically homeless individuals. A control group design, propensity score matching techniques, and difference in differences were employed to estimate effects on utilization of and claims for emergency room visits, inpatient hospitalization, mental health encounters, and pharmacy and other claims. We find that both groups registered statistically significant pre–post improvements in utilization of health care services and claims, but these are greater among the Housing First cohort. The findings suggest that Housing First significantly reduces utilization of health care services beyond regression to the mean and that the resultant health care savings offset a significant portion of the cost of housing. Policy proponents and opponents of Housing First must recognize that the model has a significant portion of its costs offset by savings from lower Medicaid claims.