Predictors and moderators of problem-focused social skills training versus resource activation for children with disruptive behavior problems.

Background: Not all children benefit equally from evidence-based treatments for disruptive behavior disorders (DBD). In a previous randomized controlled trial (RCT), computer-facilitated social skills training (ScouT) demonstrated superior efficacy over solution-focused, resource-oriented treatment (STARK) in reducing parent-rated peer-related aggression in children with DBD. This study explores predictors and moderators of these treatments to determine which subgroups respond differently. Methods: The sample included Results: Treatment adherence and most pre-treatment child and family characteristics were not significant predictors or moderators. In moderation analyses, ScouT showed stronger effects than STARK on parent-rated peer-related aggression in children with low levels of parent-rated internalizing symptoms and low levels of clinician-rated DBD symptoms at pre-treatment. No significant group differences were detected in children with high levels of these variables. In hierarchical multiple regression analyses, younger age, higher functional impairment, and higher internalizing symptoms at pre-treatment predicted poorer outcomes, nonspecific to treatment type. Results differed across secondary analyses using teacher- and child-rated outcomes. Conclusion: Moderation analyses support the robustness of RCT findings: ScouT's superior efficacy in reducing parent-rated peer-related aggression seems to hold across various characteristics. However, children with high levels of internalizing symptoms and high DBD symptom severity may need more targeted or longer treatment. The small effect sizes of the moderation effects limit the relevance of the findings for clinical practice. Age, functional impairment, and internalizing symptoms predict outcome irrespective of the treatment modality. These findings should inform future hypothesis-testing studies.