Skip Navigation
Login or register
About Us  Contact Us
Substance Abuse and Mental Health Services Administration
Add Comment
Share This
No Recommendations Yet Click here to recommend.
The major objective of this study was to identify predictors of the duration of homeless spells among persons with severe mental illness (SMI). This was the first study to incorporate time-varying covariates into the survival analysis model predicting the length of homeless spells. The sample consisted of 215 individuals who had participated in two experiments evaluating the effectiveness of various treatment programs for homeless individuals. Persons who received assertive community treatment exited homelessness sooner than individuals who received brokered case management, outpatient treatment, or services from a drop-in center. More assistance in finding and maintaining housing were especially predictive of shorter homeless spells. Inclusion of variables which varied over time in the model (e.g., income and frequency of service contacts) reduced the impact of many of the demographic variables in the model. This survival analysis demonstrated that the efforts of human service agencies to assist homeless mentally ill individuals do make a difference. In general, persons who received more services exited homelessness sooner. (Authors)