Preventing Homelessness in Persons with Severe Mental Illness: A Randomized Trial of Critical Time Intervention Following Hospital Discharge

Purpose: Homelessness among persons with severe mental illness (SMI) is distressingly common. Critical Time Intervention (CTI) is a time-limited psychosocial intervention designed to prevent recurrent homelessness among persons with SMI by enhancing continuity of care during the “critical time” of transition from institutional to community living. Data from an earlier randomized controlled trial (RCT) demonstrated the effectiveness of CTI in reducing the risk of recurrent homelessness among homeless men with SMI following discharge from a shelter to housing in the community. In the current study, we adapted and tested the model with men and women with SMI following psychiatric hospitalization.

Methods: We conducted an RCT to test the effectiveness of CTI in reducing the risk of homelessness among adults with SMI over an 18-month follow-up period after discharge from inpatient psychiatric treatment. CTI is a manualized psychosocial intervention providing targeted emotional and practical support over nine months during a “critical period” of transition from institution to community. Participants were recruited at the point of discharge from two publicly-operated psychiatric hospitals. Inclusion criteria included a diagnosis of psychotic disorder and a history of homelessness within 18 months prior to hospital admission. Diagnoses were assessed by trained interviewers using the Structured Clinical Interview for DSM Disorders (SCID) Participants were randomly assigned (half to each group) to receive either usual discharge planning and other community services or nine months of CTI plus usual services. CTI was delivered over nine months by trained bachelors level workers operating under weekly professional supervision. Homelessness was assessed via subject self-report collected via bi-monthly interviews over the follow-up period. The primary outcome was homelessness at the endpoint of the study (18 months following hospital discharge).

Results: 690 persons were screened between April 2002 and October 2006. 150 persons entered the study and were randomized, following provision of informed consent. 71% of subjects were male, and the mean age was 37. 62% were African-American and 15% were Latino. 96% of the sample were diagnosed with schizophrenia or schizoaffective disorder and the majority had concomitant substance abuse diagnoses. 70% of subjects had five or more lifetime psychiatric hospitalizations. 59% of subjects had been homeless for 30 or more days out of the 90 days preceding their most recent hospital stay. There were no significant differences between the experimental and control groups on these or other descriptive variables at baseline. Complete follow-up date was obtained for 75% of the experimental group and 80% of the controls. In an intent-to-treat analysis, 1.8 % (1/57) of those assigned to the CTI group were homeless at the study endpoint compared with 11.9% (7/59) of the controls (p=.06, two-sided Fisher's exact test).

Implications: CTI is one of the few carefully specified interventions designed to prevent homelessness in high-risk individuals. This is one of a series of studies currently underway to assess its effectiveness with different populations in a variety of service delivery settings. The model may be relevant to other “critical periods” for high-risk populations being served by mental health and social service systems.