What Works and Who Benefits? Identifying Common Components of Parenting Programs for Children Involved with Child Protection Services: A Model-Based Recursive Partitioning Analysis

Background and Purpose.
Studies have shown that evidence-based parenting models such as Parent-Child Interaction Therapy can be effective for preventing child abuse recurrence among families involved with child protection services (CPS). Yet, the uptake of such parent training models among child welfare agencies is slow because of the start-up costs and logistics. Moreover, each model offers essential components unique to it that must be implemented for model fidelity, thus increasing the challenges in the training of staff. To address this problem, Barth et al. (2014) proposed identifying common components from those parenting models empirically shown to be effective in preventing maltreatment which agencies can use to create a composite intervention. Such components include positive parent-child interaction, time-out, coaching, and non-violent disciplinary methods. Based on this idea, this study aims to advance social work science by identifying the components of parenting services and other support services provided to CPS-involved families that were effective in reducing child maltreatment recurrence.

Methods:
We used data from NSCAW-II to examine how combinations of parent training components and support services provided post-investigation to families differentially affect the association between parenting risk (assessed by caseworkers at Wave 1) and any maltreatment recurrence within 18 months. The sample consists of 192 children who remained in-home and whose parents received parent training services from CPS post-investigation. Parent training provided to families included positive parenting, non-violent disciplinary methods, maintaining child-safe home environment, and communication. Support services include financial and food assistance, counseling, and child-care. Our outcome is any maltreatment recurrence within an 18-month period measured by any new reports of maltreatment.

We used Model-based Recursive Partitioning (MOB) to examine how combinations of parent training components and services differentially affect associations between parenting risks and the outcome. MOB uses a data-driven method to find subgroups for which a specified parametric model (for this study, a logistic regression model) has different parameters. Parent training components and services were used as moderating variables.

Results.
For the subgroup of parents who neither receive food assistance nor had parent training on maintaining child-safe home environment, if the parents were at higher parenting risks, their children were more likely to experience maltreatment recurrence (OR = 9.00, p < .05). Among the subgroup of parents who received food assistance, if they were at higher parenting risk, their children were less likely to have recurrence (OR = .78, p = .58). Among the subgroup of parents who received parent training on maintaining child-safe home environment and their children were older than 2 years, if they were at higher parenting risk, their children were less likely to have recurrence (OR = .71, p < .72).

Conclusions and Implications.
This study highlights that providing parent training on maintaining safe-home environment and food assistance is an important combination of parent training component with non-parenting support service that is effective in reducing the risk of maltreatment recurrence. Families with parenting risk may have basic needs that need to be supported by ancillary services to maximize the effectiveness of parent training services.