Introduction
Research Questions
- Does Project Connect increase reunification of children placed in out-of-home care?
- Does Project Connect speed up reunification of children placed in out-of-home care?
- Does Project Connect increase permanency outcomes among reunification families—that is, the likelihood that a child is not in foster care and is living in a permanent family setting, which may include living with parents, living with kin, adoption, or another permanent situation?
- Does Project Connect increase the likelihood that children placed in out-of-home care are stably reunified from 6 to 12 months postrandomization?
Child welfare cases involving parental substance use are among the most complex, often resulting in the poorest outcomes. In 2022, about 1 in 4 reported cases of child maltreatment nationally had a caregiver with drug use as a risk factor. Families affected by drug use are less likely to achieve successful reunification, and children who enter foster care because of parental drug use are less likely to achieve permanency after they are reunified.
Effective evidence-supported interventions for families in the child welfare system affected by substance use (and drug use in particular) remain relatively rare. To help strengthen the evidence base, this report presents impact findings from an evaluation of Project Connect—a home-based intervention for families involved in the child welfare system affected by substance use in Rhode Island. The evaluation used a randomized controlled trial (RCT) design to test whether Project Connect produced better outcomes for children and families compared with other available services.
Purpose
The purpose of this report is to share findings from an RCT evaluation that examined whether Project Connect had a significant impact on reunifications and permanency among children in out-of-home-care in Rhode Island. The evaluation is part of a larger project, Supporting Evidence Building in Child Welfare, aimed at increasing the number of evidence-supported interventions for child welfare populations. We identified Project Connect for its focus on substance use, for its wide reach serving families involved in the child welfare system throughout Rhode Island, and because its model is unique in the state.
Two previous quasi-experimental studies of Project Connect demonstrated some effectiveness, such that families participating in the program had higher rates of reunification than families who did not. Project Connect had never undergone an RCT evaluation to rigorously test if its services cause an increase in reunifications compared with services as usual. The current study is the program’s first RCT.
Key Findings and Highlights
The study focused on reunification at 6 months postrandomization and permanency at 12 months postrandomization among children who were in out-of-home care at the time of randomization. This roughly corresponds to end of treatment (6 months) and a longer-term outcome posttreatment (12 months). At 6 months and 12 months postrandomization we found families randomized to Project Connect were statistically significantly less likely to be reunified compared with families randomized to services as usual. The results also showed no statistically significant differences in the share of children discharged to permanency (reunification, adoption, or guardianship) among families randomized to Project Connect compared with families randomized to services as usual. Regression-adjusted findings are below:
At 6 months postrandomization, 2 percent of children whose families were randomized to Project Connect were reunified compared with 8 percent of children whose families were randomized to services as usual.
This difference in the shares reunified grew by 12 months postrandomization, as 15 percent of children whose families were randomized to Project Connect were reunified compared with 28 percent of children whose families were randomized to services as usual.
On average, a lower share of children whose parents engaged in Project Connect achieved permanency at one year postrandomization compared with children in the control group.
We discuss possible explanations for the results. Further analysis is needed, but we speculate that Project Connect’s emphasis on parent or caregiver stability and support might contribute to increased surveillance of parent behavior, with Project Connect staff having more frequent opportunities to observe parents compared to services as usual. Because of this, there might be more reluctance to recommend reunification to the courts compared with families receiving services as usual who may experience less surveillance on average. We also hypothesize that, because Project Connect first focuses on parent or caregiver well-being (e.g., greater stability and improved parental functioning), it might take longer to observe improvements in reunification rates and stability (i.e., reunifications less likely to be disrupted) beyond the 12 months we were able to test in the current evaluation. Within 12 months postrandomization, though, the results show that Project Connect produces fewer and slower reunifications.
We discuss study limitations including implications of the COVID-19 pandemic, the absence of information about services families in the comparison group may have received, and the study’s length that limited our ability to analyze long-term outcomes beyond 12 months postrandomization.
Methods
The study employed an RCT design using child welfare administrative records. From April 2021 to June 2023, families involved in Rhode Island’s Department of Children, Youth, and Families (DCYF) who were referred by their DCYF caseworker for substance use services were randomly assigned to one of two groups: Project Connect (treatment group) or to other services as usual (control group). We compared families assigned to each group based on their demographics, placement histories, case records, and reports of abuse or neglect.
The full study sample included 226 families—113 who were randomly referred to Project Connect and 113 randomly referred to services as usual. For the impact analyses we focused on the 192 families (96 treatment; 96 control) whose children were in out-of-home care at the time they were referred to services. The treatment and control groups were statistically equivalent except for differences in the mother’s age and focal child’s age, which we adjusted for in our impact analyses.
Citation
McDaniel, Marla, Catherine Kuhns, Hannah Daly, Annabel Stattleman-Scanlan, Mattie McKenzie-Liu, and Michael Pergamit. (2025). Evaluation of Project Connect: A Program for Families Involved in the Child Welfare System Affected by Substance Use. OPRE Report #2025-043, Washington, DC: Office of Planning, Research, and Evaluation, Administration for Children and Families, US Department of Health and Human Services.
Glossary
- DCYF:
- The Department of Children, Youth, and Families
- Out-of-home placement:
- For children who are removed from their parent’s care, an out-of-home placement serves as a temporary living arrangement for the child outside of their home. This placement can be with other family (kin care), foster care, or in an institutional setting.
- Randomized controlled trial (RCT):
- A randomized controlled trial is a form of evaluation that measures the effectiveness of an intervention by directly comparing a group that receives the treatment to the group that does not. Groups are determined via a random method.
- Reunification:
- Reunification is a term used to describe a situation in which a child is returned to their family home to live with the family member(s) from whom they were originally removed. In these cases, the family member(s) regain(s) physical custody of the child.
- Substance use:
- Substance use is defined as any person’s continued use of alcohol, prescription drugs, over-the-counter drugs, or recreational drugs. In this report, we use substance use as the blanket statement for both drug and alcohol use.