Supporting Evidence Building in Child Welfare: The Evaluation of Project Connect

Last registered on October 16, 2025

View Trial History

Pre-Trial

Trial Information

General Information

Title

Supporting Evidence Building in Child Welfare: The Evaluation of Project Connect

RCT ID

AEARCTR-0007273

Initial registration date

April 23, 2021

Initial registration date is when the trial was registered.

It corresponds to when the registration was submitted to the Registry to be reviewed for publication.

First published

April 26, 2021, 10:41 AM EDT

First published corresponds to when the trial was first made public on the Registry after being reviewed.

Last updated

October 16, 2025, 12:14 PM EDT

Last updated is the most recent time when changes to the trial's registration were published.

Locations

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Primary Investigator

Name

Catherine Kuhns

Affiliation

The Urban Institute

Contact Primary Investigator

Other Primary Investigator(s)

PI Name

Michael Pergamit

PI Affiliation

Urban Institute

Contact Investigator

PI Name

Marla McDaniel

PI Affiliation

The Urban Institute

Contact Investigator

Additional Trial Information

Status

In development

Start date

2021-04-01

End date

2027-03-31

Keywords

Behavior

Additional Keywords

child welfare, substance use

JEL code(s)

Secondary IDs

HHSP233201500064I_HHSP23337014T

Prior work

This trial does not extend or rely on any prior RCTs.

Abstract

Project Connect is a comprehensive home visitation intervention that targets child welfare-involved, substance-affected families with children and adolescents ages 0 to 17 through home-based services and treatment. The program addresses the complex needs of families affected by substance use by providing intensive, long-term services aimed at strengthening families, addressing parental substance use, and helping parents recover while keeping children safe.

The study used a randomized controlled trial (RCT) to evaluate the causal impact of Project Connect services on families experiencing substance use in the Rhode Island child welfare system and to compare the child welfare service trajectories (e.g., placement and length of time in care) of families affected by substance use who received Project Connect services with those who received services as usual. We also collected qualitative data, which included the interviews and focus groups, as part of an implementation study. Families affected by substance use with an open child protective services case in Rhode Island were randomly referred for substance use services – either to Project Connect or to other usual services. The final full study sample includes 230 families, 192 families whose children were in out-of-home care at the start of the study (96 treatment and 96 control) and 34 families who children were in home at the start of the study (17 treatment and 17 control). The majority of our analyses will focus on the 192 “reunification” families. We will examine 24-months post randomization outcomes. If feasible, we will also examine 30-month post randomization outcomes.

External Link(s)

Registration Citation

Citation

Kuhns, Catherine, Marla McDaniel and Michael Pergamit. 2025. "Supporting Evidence Building in Child Welfare: The Evaluation of Project Connect." AEA RCT Registry. October 16. https://doi.org/10.1257/rct.7273-3.0

Sponsors & Partners

There is information in this trial unavailable to the public. Use the button below to request access.

Request Information

Experimental Details

Interventions

Intervention(s)

Project Connect is a home-based intervention for families involved in the child welfare system affected by substance use. Eligible families have children or adolescents ages 0 to 17. The program aims to maintain children safely in their homes (preventing admission to out-of-home care) or to facilitate reunification when children have been placed out-of-home. The program’s theory of change is that child safety, well-being, and permanency are achieved when parents have resources that support recovery, enhance parenting capacity, strengthen family relationships and functioning, and improve families’ living conditions. Project Connect works with families to build those resources through substance use assessment; individual and family counseling; parent education; pediatric nursing assessment and services; relapse prevention planning; assistance with transportation; court advocacy; supplemental visitations; 24/7 emergency on call support; domestic violence support; and aftercare services following DCYF involvement. Project Connect also links families with community resources, including substance abuse treatment agencies; child care; health care; housing; mental health; legal services; and medical services. The Project Connect model emphasizes deep and trusting relationships with caregivers as a key to connecting families to resources and achieving intended outcomes.

Intervention Start Date

2021-04-01

Intervention End Date

2024-06-28

Primary Outcomes

Primary Outcomes (end points)

1. Child reunifications with family from out-of-home care
2. Child reunifications with family from out-of-home care with no placement disruptions within the prior 6 months

Primary Outcomes (explanation)

Child reunification will be measured as the probability that a child who is out of the home at the time of randomization is reunited with their parents. We will examine 1) the share of children reunified at 24 months post randomization; and 2) the share of children reunified at 24 months post randomization with no placement disruptions within the prior 6 months. The latter measure represents reunification stability. We also note that higher rates of stable reunifications imply lower rates of reported abuse and neglect.

Secondary Outcomes

Secondary Outcomes (end points)

1. Differences in the share of children whose caregivers have a re-report of abuse or neglect (reunification families only)
2. Differences in the speed of reunification (reunification families only)
3. Among combined reunification and preservation families at 24 months post randomization, the share of children in a stable in-home placement for the prior 6 months
4. All children are at home at the end of the observation period

Secondary Outcomes (explanation)

A stable, in-home placement will be measured as 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 other permanent situation. The full study sample is 226 families, which includes 34 families who had an open child welfare case but whose children remained at home at the start of the study (i.e., preservation families). The last two secondary outcomes listed above will be an analysis of this full sample.

Experimental Design

The evaluation employed a family level RCT in one site. Randomization occurred between April 2021 and June 2023.Our randomization protocol had caseworkers in the public child welfare agency refer families requiring substance use services to the agency's Central Referral Unit (CRU). The CRU stratified the referrals separating preservation families (i.e., children at-home) from reunification families (i.e., children in out-of-home care). The CRU manager entered information into Urban’s online randomization tool, which provided her the study group assignment. The CRU manager randomized in pairs so the treatment and control groups were of equal size with roughly comparable referral dates. Under our randomization protocol, all families flagged by their caseworker as having a substance use concern and subsequently referred to the agency's CRU were eligible for the study, and randomized.

Experimental Design Details

Not available

Randomization Method

The specific referral and randomization strategy leveraged an existing referral form that caseworkers completed and sent to a central referral unit (CRU) at the public child welfare agency. At the CRU, a staff member reviewed each referral form and randomized all eligible families into either treatment or control. The staff member used a customized online randomization tool designed by the Urban Institute. The randomization tool collected the family’s identification number, whether they were a preservation or reunification family, and their referral form and determined whether a family would be referred to Project Connect or services as usual. The tool used a one-to-one randomization ratio, with one family randomized to treatment for every control family.

Randomization Unit

Randomization took place at the household (family) level.

Was the treatment clustered?

Experiment Characteristics

Sample size: planned number of clusters

N/A

Sample size: planned number of observations

The full study sample is 226 families. The majority of our planned research questions focus only on reunification families – that is, the 192 families whose children were in out-of-home care at the start of the study (96 treatment and 96 control) The study sample also includes 34 families who had an open child welfare case but whose children remained at home at the start of the study (i.e., preservation families). These families are included in one analysis as part of the list of secondary outcomes.

Sample size (or number of clusters) by treatment arms

-96 "reunification" families in the control group
-96 "reunification" families in the treatment group
-17 “preservation” families in the control group
-17 “preservation" families in the treatment group

Minimum detectable effect size for main outcomes (accounting for sample design and clustering)

We estimated Minimum Detectable Effect (MDE) sizes for Intent-to-Treat (ITT) estimates. We calculated the MDE as the impact divided by the comparison group standard deviation (sqrt(m(1-m)), where m = the proportion of the control group with a “successful” outcome (i.e., reunification). The MDE was calculated assuming a randomization ratio of one treatment (Project Connect) to one control (services as usual) with an alpha level of 0.05, a two-tailed test with 80 percent power, and a sample size of 192. This produced a minimum detectable effect size of 0.40.

Supporting Documents and Materials

IRB

Institutional Review Boards (IRBs)

IRB Name

Urban Institute IRB

IRB Approval Date

2025-04-24

IRB Approval Number

0189

Analysis Plan