Predictors and Moderators of Parent Training for Children with Autism Spectrum Disorders and Serious Behavioral Problems

Background:  The Research Units on Pediatric Psychopharmacology--Autism Unit (RUPP) reported the results of a 24-week trial of combined treatment with risperidone/aripiprazole and parent training versus antipsychotic alone in children with autism spectrum disorders (ASD) and serious behavioral disturbances (Aman et al., 2009). The impact of medication only (MED) on the primary outcome measure was d=2.3, with an additional effect of 0.34 from parent training (COMB). Additionally, parent training was associated with lower doses of antipsychotic.

Objectives:  The goal of this analysis was to evaluate potential predictors and moderators of the effect of parent training in this study. Predictors and moderators are important tools for uncovering masked treatment effects and maximizing power in future trials, and may be used clinically to help match patients to treatments. Twenty-two baseline variables, including demographics, were evaluated.

Methods:  Participants were 124 children aged 4-14 years (M = 7.4 years) with ASDs and parent-rated Aberrant Behavior Checklist (ABC) Irritability scores of at least 18. Seventy-five children were assigned to the COMB condition and 49 to MED alone. In addition to the effect found on the on the Home Situations Questionnaire (HSQ), RUPP found an effect of COMB on ABC Hyperactivity/Noncompliance, so it was adopted as an additional outcome measure in the current study. The original ITT model was re-run with the putative predictors and moderators inserted. A random-effects regression model with fixed effects for treatment, time, site and their interactions was employed. Predictor and moderator variables were defined as follows: (a) the variable must temporally proceed the treatment, (b) the variable must be reasonably independent from the treatment, and (c) a significant main effect (predictor) or interaction with treatment (moderator) must be observed.

Results:  One variable was found to predict outcome on the HSQ: higher baseline values on the HSQ predicted greater improvement, regardless of treatment assignment [F(1,548) = 7.23), p = 0.007]. No significant predictors of ABC Hyperactivity/Noncompliance outcome were observed. No variables significantly moderated the impact of COMB on either the HSQ or ABC Hyperactivity/Noncompliance.

Conclusions:  The finding that baseline levels on the HSQ predicts outcome mimics a result from a trial of risperidone versus placebo in children with ASDs. Together, these findings suggest that treatment with antipsychotic brings patients to approximately the same target symptom level, regardless of initial severity. The lack of significant moderators may be encouraging; this suggests that COMB is appropriate for a wide range of children with ASDs. However, it is possible that this study was not adequately powered to explore moderators. The search for predictors and moderators of treatment remains an important part of clinical treatment trials.

Aman, M., McDougle, C., Scahill, L., Handen, B., Arnold, L.E. et al. (2009). Medication and parent training in children with pervasive developmental disorders and serious behavior problems: Results from a randomized clinical trial. Journal of the American Academy of Child and Adolescent Psychiatry, 48, 1143-1154.