Seeing the Big Picture with Multisector Data: Factors Associated with Exiting from Federal Housing Assistance by Exit Type

Alastair I. Matheson
Danny V. Colombara
Public Health – Seattle & King County
University of Washington

Annie Pennucci
Tyler Shannon
King County Housing Authority

Andy Chan
Seattle Housing Authority

Megan Suter
Amy A. Laurent
Public Health – Seattle & King County

This work was funded by the U.S. Department of Housing and Urban Development under funding opportunity FR-6400-N-58.

The U.S. Department of Housing and Urban Development (HUD) has a goal of increasing the proportion of households that exit HUD-supported housing for positive reasons, but little is known about factors associated with different exit types. Learning which tenants are likely to leave for positive or negative reasons can inform policies and programs that aim to encourage positive exits.

The authors linked data from two large public housing authorities (PHAs) to an existing multisector data system that contains behavioral health, homeless services, and Medicaid claims data. The authors used logistic regression to examine factors associated with exiting from housing assistance and used multinomial regression to explore factors associated with exit type, both at the head-of-household level.

The analysis consisted of 8,266 exits: 2,610 negative, 4,538 neutral, and 1,118 positive. Male gender, homelessness within the previous 3 years, and having a behavioral health crisis event were all associated with increased odds of exit. Being older than age 25, increased time in housing (more than 6 years), larger household size, having a single-caregiver household, and having a disability were all associated with decreased odds of exit. Being of working age (25–61) was associated with positive exits but not negative exits. Heads of household in single-caregiver households, who had disabilities, experienced behavioral health crisis events, or had recent emergency department visits were all more likely to have negative exits and less likely to have positive exits compared with neutral exits.

Linked administrative datasets offer PHAs a means of routinely obtaining information about physical or mental health conditions, prior homelessness, or other factors that could influence outcomes for their tenants. This approach can serve as a model to PHAs for cross-sector partnerships with health departments and other groups interested in the health and well-being of housing assistance recipients.