Objectives: The purpose of this study was to evaluate short-term and maintenance effects of two models of intervention: cognitive behavioral therapy (CBT) and child-directed play. The trial evaluated both neural (e.g. fMRI) and behavioral outcomes at endpoint and 3-month follow-up, although only behavioral outcome results from endpoint (i.e. 12 weeks) are presented here.
Specific hypotheses: The primary hypothesis for the study was that CBT approaches take advantage of top-down, explicit processing skills in ASD, and as such, would lead to improved performance on measures of social cognition, specifically emotion recognition, compared to child-directed play.
Methods: Participants (n=69), ages 8-11 (mean=9.88, sd =1.25) were randomized into one of two treatments: CBT or child-directed play. Both treatments involved 12, 90-minutes sessions, with concurrent parent and child intervention groups. Participants were fully characterized utilizing research standard diagnostic evaluation methods. Outcomes were conducted at baseline, week 12, and 3 months post-treatment. Outcomes included blinded assessments of emotion recognition (DANVA2, Eyes Test) and social initiation during an analogue playgroup; caregiver reports of social, adaptive, and problem behavior (e.g. Vineland, BASC2); and fMRI tasks (see related submission, 14296) . Linear mixed models were used to evaluate treatment effects on the primary outcome of emotion recognition. Repeated measures ANOVAs conducted to evaluate treatment effects on secondary outcomes.
Results: Analysis of outcomes immediately following treatment indicated improvements in both groups on several treatment outcome variables. The primary hypothesis proposing an advantage for CBT in improving emotion recognition skills was not supported (DANVA2 total correct score, estimate = 0.0437, p=.9858). However, the CBT group did show greater improvement on accuracy of identifying low-intensity (i.e. more subtle emotional cues) items on the DANVA2 (Child Faces low intensity, estimate = 1.03, p=.029). In addition, evaluation of social behavior, as measured by the Children’s Communication Checklist, Social Relations subscale, suggested the CBT group was rated as exhibiting more socially appropriate behavior compared to the play therapy group (F=4.48, df=52, p=.039). No significant improvement in the CBT group over play group was observed for caregiver reported measures of behavior (e.g. BASC2) or adaptive behavior (e.g. Vineland).
Conclusions: This study provides data on the immediate effects of social skills group treatments in an RCT design minimizing expectancy effects of a single treatment or waitlist control designs. Overall, results from this trial suggest a modest advantage for CBT-based treatment models over play-based models for improving emotion recognition and social abilities in school-aged children with autism. Research on the durability, generalization, and potential augmentation of effects continues to be open area of investigations, particularly in light of the modest treatment effects.