The Methodological Quality and Intervention Fidelity of Randomised Controlled Trials Evaluating Social Skills Group Programs in Autistic Adolescents: A Systematic Review and Meta-analysis.

Only 13 of the 18 trials could be pooled, and nearly half of the studies reported only subscale scores, so the meta-analysis rests on a subset of the evidence. The trials differed substantially in their programs, outcome measures and participants, with very high heterogeneity (up to 99%), which makes a single pooled number hard to interpret. Most trials were small and did not blind the assessment of outcomes or conceal which group participants were allocated to, so the expectations of parents, adolescents and researchers could have inflated the results. Several trials pooled autistic children and adolescents together, so the results do not describe adolescents alone. Participants were mostly male, had an IQ above 70, and most trials were run in the United States, so the findings may not transfer to other groups or settings. All the autism-characteristics data came from parent-completed questionnaires, and teacher reports did not detect any change. Five of the seven authors contacted for missing data did not respond, so fidelity was scored only from what the published papers happened to report.

The review was funded by the Stan Perron Charitable Foundation. No other funding source or competing interest is reported in the supplied text.

Do not read the large overall effect as proof these groups work. After the authors adjusted for signs of publication bias the effect fell from 0.96 to 0.60, the social-skills result rested mostly on the teenagers' own reports, and nothing held at follow-up. The paper also describes behavioural and emotional challenges as significantly reduced when the numbers it reports alongside that claim show p = 0.25.