Efficacy of psychosocial interventions for Autism spectrum disorder: an umbrella review

International audienceINTRODUCTION: The wide range of psychosocial interventions designed to assist people with Autism Spectrum Disorder (ASD)makes it challenging to compile and hierarchize the scientific evidence that supports the efficacy of these interventions. Thus, weperformed an umbrella review of published meta-analyses of controlled clinical trials that investigated the efficacy of psychosocialinterventions on both core and related ASD symptoms.METHODS: Each meta-analysis that was identified was re-estimated using a random-effects model with a restricted maximumlikelihood estimator. The methodological quality of included meta-analyses was critically appraised and the credibility of theevidence was assessed algorithmically according to criteria adapted for the purpose of this study.RESULTS: We identified a total of 128 meta-analyses derived from 44 reports. More than half of the non-overlapping meta-analyseswere nominally statistically significant and/or displayed a moderate-to-large pooled effect size that favored the psychosocialinterventions. The assessment of the credibility of evidence pointed out that the efficacy of early intensive behavioral interventions,developmental interventions, naturalistic developmental behavioral interventions, and parent-mediated interventions wassupported by suggestive evidence on at least one outcome in preschool children. Possible outcomes included socialcommunication deficits, global cognitive abilities, and adaptive behaviors. Results also revealed highly suggestive indications thatparent-mediated interventions improved disruptive behaviors in early school-aged children. The efficacy of social skills groups wassupported by suggestive evidence for improving social communication deficits and overall ASD symptoms in school-aged childrenand adolescents. Only four meta-analyses had a statistically significant pooled effect size in a sensitivity analysis restricted torandomized controlled trials at low risk of detection bias.DISCUSSION: This umbrella review confirmed that several psychosocial interventions show promise for improving symptomsrelated to ASD at different stages of life. However, additional well-designed randomized controlled trials are still required toproduce a clearer picture of the efficacy of these interventions. To facilitate the dissemination of scientific knowledge aboutpsychosocial interventions for individuals with ASD, we built an open-access and interactive website that shares the informationcollected and the results generated during this umbrella review.PRE-REGISTRATION: PROSPERO ID CRD42020212630