Who Benefits? Differential Effects of a Trauma-Focused Parenting Intervention for Infants and Children in Foster Care Using Model-Based Recursive Partitioning Analysis: Illinois "Birth through Three" Waiver Demonstration Project

Background

The testing of the effectiveness of interventions under real-world conditions is important for social work science. Yet, more can be done to use these trials to further understand how participants benefit differently from the interventions to uncover critical factors that can maximize the likelihood of success of the intervention under real-world conditions. Our study looked at a trauma-focused parenting intervention for infants and young children removed from their parents and placed in custody with child protection services. We identified participants’ pretreatment characteristics and risks that interact with the treatment to explain differences in outcomes. We used a novel technique, Model-based Recursive Partitioning (MOB), to examine for differential effectiveness because MOB does not suffer from the limitations (e.g., modeling complex, nonlinear interactions of factors) of techniques such as OLS regression. While evidence has shown that trauma-focused parenting is an effective intervention, understanding how it is useful or less useful for what participants can help in the optimizing of the intervention and in directing additional services to support families.

Methods

Data for the study (n = 414 children) come from the “Illinois Birth through Three” waiver demonstration project. The demonstration tested if children aged birth through three years old, who are placed in foster care experience increased permanency and improved child well-being if they are provided trauma-focused parenting (Child-Parent Psychotherapy or Nurturing Parent Program) compared to similar children who are provided services as-usual. We analyzed with MOB which uses a data-driven method to find subgroups for which a specified parametric model has different parameters. Three generalized linear models (logistic, poisson, and linear regression) were used in accordance with three outcomes: reunification (categorical), days in care (count), and attachment quality (continuous). Baseline moderating factors used included 21 caregiver and child-related demographics and risk factors (e.g., child health, active domestic violence, and social support).

Results

For the outcome of reunification with parents/caregivers, result indicated that the treatment was more effective than the control in reunifying the subgroup of children with adoptive or birth parents and who did not experience physical abuse maltreatment type (OR = 2.25, p < .05). For the number of days in out-of-home care outcome, the subgroup of children with caregivers who had more emotional support spent lesser days in care than the control (b = .13, p <.001). For improving caregiver-child attachment quality, the treatment was more effective than the control for the subgroup of younger caregivers below age 49 (b = 3.13, p < .05).

Conclusions and Implications.

This study showed that the effectiveness of the intervention is different for subgroups of participants that are differentiated by (a) caregivers’ relationship with child, (b) perceived emotional support, (c) their age, and (d) child’s initial maltreatment type. Thus, trauma-focused parenting intervention can be further optimized for these groups of participants (e.g., caregivers with less emotional support) by more targeted focus or putting in additional services to leverage on its effectiveness. MOB is an analytical method that is useful in social work research to examine the differential effects of interventions.