Reunifying Homeless Families with Their Children in Foster Care: Process Findings from the First County Level Pay for Success RCT

Background: The first county level Pay for Success program using a social impact bond (SIB) financing model, the Cuyahoga Partnering for Family Success (PFS) program implements a randomized control trial (RCT) to determine whether homeless families with children in foster care who are prioritized for housing and receive individualized and comprehensive case management services, are reunited with their children more quickly than families in a control group. The program uses evidence-based practices including Critical Time Intervention for intensive case management and Trauma Adapted Family Connections for therapeutic services after reunification. This paper reports on the findings from a two-year process evaluation program, asking three key questions: (1) What are the characteristics of the treatment and control groups? (2) What is the content and dosage of service contacts over time? (3) How has PFS had an impact on service delivery for families?

Methods: A mixed methods process evaluation approach was used to describe the program and impacts of the program participants (who must be homeless and have children in foster care). Quantitative data from the Homeless Management Information System and county child welfare and public assistance services described clients’ demographics, disability status, histories and current use of homeless, public assistance, and child welfare services. Client progress notes provided quantitative information on service dosage including the type, frequency, timing, and length of contacts. Descriptive and inferential statistics were used to compare treatment and control groups where possible. Qualitative data collected included case management progress notes, and semi-structured group and individual interviews with staff. Peer debriefing and member checks were employed to ensure the data reflected staff experiences, and triangulation from multiple data sources supported the rigor of the qualitative data. Qualitative analytic techniques included content analysis of progress notes and thematic analysis of interview transcripts.

Results: Findings suggest that the treatment and control groups are comparable at program entry. Overall, the data support a conclusion that the program helps stabilize families, with more than 84% of treatment group families entering permanent housing options as compared to 11% of the control group. Progress note data indicate that treatment families’ number of service contacts decrease over time, they receive more public assistance, and have less involvement in child welfare than the control group. Interview data indicate that while PFS workers see themselves as coaches and advocates for clients, child welfare workers’ views were mixed; some thought PFS provided necessary coaching toward self-sufficiency, but others felt the program enabled clients. Clients’ experiences with domestic violence were also identified as an important challenge. Taken together, the data indicate that quick housing is a major accomplishment of the program, but more service coordination is needed.

Conclusions and Implications: The findings indicate that the PFS treatment group is becoming more quickly stabilized with regard to housing and receipt of public assistances than the control group. These data will be used to improve collaborative between child welfare workers and housing assistance providers as the program progresses to ultimately improve services and outcomes for vulnerable families while reducing program costs.