Does Parent-Child Interaction Therapy Reduce Future Physical Abuse?: A Meta-Analysis

Background:  Over 650,000 children were victims of child abuse and neglect in 2012 with 60% of abuse cases involving physical abuse. Victims of childhood physical abuse are at higher risk for experiencing a range of mental and behavioral difficulties across the life span. Parent-Child Interaction Therapy (PCIT) is a parenting intervention which alters the family system bymodifying the behavior of both parent and child. The purpose of this meta-analysis is to examine the effectiveness of PCIT at reducing future physical abuse (physical abuse recurrence, child abuse potential, and parenting stress) for abusive or at-risk families.

Methods:  A systematic search of 11 databases was conducted to identify eligible studies. Eligible studies employed experimental or quasi-experimental design and reported physical abuse recurrence, or a standardized measure of abuse potential or parenting stress. Five studies were retained and coded by two independent coders. Statistical analysis produced descriptive information and effect sizes of each intervention for conceptually distinct outcomes of interest to this review. Hedges’s geffect sizes and 95% confidence intervals (CIs) were calculated to allow interpretation, comparison, and analysis of results across studies. To ensure statistical independence of effect sizes, only one effect size was calculated per study per outcome. When multiple measures were used to assess the same outcome construct in a study, a study-level average was calculated.

Results:  PCIT treatment effect sizes were derived from all five studies and included a total of 495 parent-child dyads. Four studies used randomization to create groups. Only one study looked at physical abuse recurrence within the parent-child dyad. Results showed a large effect size estimate (g= 0.77; 95% CI: 0.22 – 1.32; p< 0.01), indicating a statistically significant treatment effect. Five studies examined child abuse potential using the Child Abuse Potential Inventory. Standardized mean differences effect sizes ranged from 0.07 to 0.96. Results indicated a random-effects weighted mean effect size estimate of 0.31. The magnitude of the overall weighted mean effect size is positive and nearly statistically significant (95% CI: -0.002 – 0.620; p= 0.05). Four studies examined parenting stress outcomes using the Parenting Stress Inventory. Standardized mean differences effect sizes ranged from 0.15 to 0.97. Results indicated a random-effects weighted mean effect size estimate of 0.35. The magnitude of the overall weighted mean effect size is positive and statistically significant (95% CI: 0.04 – 0.65; p= 0.03).

Conclusions: The meta-analytic results tentatively support the hypothesis that PCIT is an effective intervention for reducing future physical abuse among physically abusive parent-child dyads. The current analysis underscores the clinical benefits of using objective measures of future physical abuse as opposed to proxy risk measures of child abuse potential or parenting stress. In the current meta-analysis, rigorous objective outcomes produced the most sizable treatment effects. Our results also highlight the importance of identifying effective social work interventions for child-welfare involved families. PCIT goes beyond the didactic parenting classes typically offered to child-welfare involved families by involving both the parent and the child in therapy sessions, teaching parent and child to play and listen together.