Integrating Parent-Child Interaction Therapy into Foster Parent Training

Background/Purpose:  Parent-Child Interaction Therapy (PCIT) is an efficacious intervention for young children with emotional and behavioral problems. PCIT also has been shown to enhance parent-child interactions as well as parenting attitudes and practices, findings that are consistent with its emphasis on dyadic parent-child activities and live parent coaching.  Although they often care for children with significant mental health needs, foster parents rarely receive experiential training with their foster children to address those needs; the trainings they attend are typically didactic, adult-centered and lacking empirical support. 

This study presents results from a randomized trial of a PCIT model that was adapted to increase its accessibility to foster families. Extending prior work indicating that the intervention reduced child externalizing and internalizing symptoms, this study examines parenting attitudes and behaviors based on self-report and observational data. 

Methods:  Eligible children (n=114), ages 3-6, were randomly assigned to a waitlist control condition or to one of two PCIT conditions.  One PCIT group of foster parent-child dyads was asked to participate in two full-day group trainings coupled with 8 weeks of telephone consultation and homework.  The second PCIT group was asked to participate in three full-day trainings plus 14 weeks of consultation and homework. Parents completed the Parenting Stress Index-Short Form (PSI-SF) at baseline as well as 8 weeks and 14 weeks post-baseline.  In addition to a total score, the PSI-SF yields three subscales: parental distress (PD), parent–child dysfunctional interaction (PCDI), and difficult child (DC).  Observational data were gathered during a 20-minute assessment of parent-child interactions at baseline and 14 weeks post-baseline.  Using the Dyadic Parent–Child Interaction Coding System (DPICS), observational data were coded by multiple raters to create two composite parenting measures: (1) labeled praise; (2) negative talk.

There were no significant group differences in demographic characteristics at baseline. Therefore, intervention effects on PSI outcomes were tested using 3x3 (condition x time) mixed-model repeated measures Analysis of Covariance (ANCOVA). Analyses of DPICS outcomes were run as 3x2 ANCOVAs.

Results: The three study conditions did not differ at baseline on any parenting measure. Mixed model ANCOVAs revealed significant condition-by-time interaction effects (p< .01) for the total PSI-SF scale along with the PCDI and DC subscales, denoting an omnibus difference in change over time. Pairwise tests confirmed that the trajectories of the PCIT groups differed significantly from the trajectory of the waitlist group. There was also a significant omnibus difference in post-baseline levels of labeled praise (p= .015). Pairwise contrasts verified that, compared to waitlist parents, the PCIT parents exhibited higher levels of labeled praise at 14 weeks (p< .01). No omnibus differences on the PD subscale or the negative talk measure were detected. Furthermore, there were no differences between the PCIT groups post-baseline on any outcome, suggesting intervention effects did not vary by dosage. 

Conclusions:  Findings suggest that integrating PCIT into foster parent training may improve upon usual care and enhance the parenting attitudes and practices of substitute caregivers. Implications for advancing translational research and evidence-based practice in child welfare will be discussed.