Sensitivity and Expected Change of Commonly Used Social Communication Measures in Longitudinal Research of Young Autistic Children.

Study quality varied widely and could not be fully controlled for; on some measures the studies rated 'Poor' showed the largest effects. Studies that did not report means and standard deviations for the outcomes could not be included, and the picture might shift if they were. Measures that appeared in only one or a few studies could not be analysed, so this says nothing about how well they track change. The list of measures searched for came from earlier reviews, so measures developed after that point were not included. How much community therapy children received outside a study was poorly reported, so 'treatment as usual' here does not mean no treatment. The analyses were exploratory and no correction was made for testing many measures and predictors, so some significant findings are likely to be chance. Age and other child characteristics were averaged for each study, so children of quite different ages can sit inside the same average. The paper does not report a total number of children across the 203 studies.

The article states that all data and code are available on request. The text supplied contains no funding statement and no competing-interests declaration, so it is not possible to say who paid for the work or what the authors declared.

Do not read the differences between behavioural intervention, treatment as usual and medication groups as evidence that one treatment works better. Most measures did not separate the groups at all, the comparisons are between groups of studies rather than a direct randomised comparison, and the one contrast favouring medication rested on three effect sizes.