Bridging Disciplinary Boundaries (January 11 - 14, 2007)

Background: As many as 600,000 families are homeless annually in the United States. Although homeless families are believed to be heterogeneous regarding the intensity of homeless experience and background characteristics, no research had discerned subgroups of homeless families. Additionally, no research has investigated whether the subgroups of homeless families have different utilization of health and social services. This study tested a typology of family homelessness by patterns of shelter utilization and furthermore compared mental health care use and child welfare involvement for the subgroups.

Methods: This study was conducted using administrative data on homeless shelter, public mental health service, and child welfare systems in a large U. S. city. The study group comprised 1,568 families with children who have first entered a homeless shelter. To give an equal opportunity for the development of homelessness patterns, each family's homeless episodes were observed for the same period of three years subsequent to their first homeless episode. Each family's record of shelter use was augmented with relevant data for child welfare and mental health service records before and after their homeless episodes. Cluster analysis was used to examine the existence of different subpopulations of homeless families. This study then examined whether utilization of child welfare and mental health services differ among the subpopulations of homeless families.

Results: Approximately 71% of the study group are transitionally homeless, who enter the shelter system for only one stay and for a short period. Another 8% are episodically homeless, who are frequently move in and out of homelessness. Nearly 21% are long-term homeless, for whom shelters function like long-term housing rather than emergency arrangement. Although the long-term group comprises one fifth of homeless families, they consume more than three fifth of the total shelter days. The prevalence of treated mental disorder for the episodic group is 31% while the prevalence rates for the transitional and long-term groups are 15% and 8%, respectively. More than 20% of the episodic group have a child placed in foster care compared to 16% of the long-term group and 12% of the transitional group.

Conclusions: These findings suggest that long-term homeless families are substantially heavier shelter users and that supported housing and services for these families can be cost-efficient with reduced shelter services offsetting the cost for the programs. Episodic homeless families are more likely to receive mental health care than others and these families may benefit from integration between homeless service and mental health systems and continued follow-up. The transitional homeless families can benefit the most from preventive assistance. The findings also suggest the need for increased planning and coordination between the homeless assistance, child welfare, and mental health systems.