Understanding Repetitive Behaviours: A clinical and cost-effectiveness, multi-site randomised controlled trial of a group for parents and carers of young autistic children.

The trial did not reach its recruitment target (227 families randomised against a planned 250) and only 155 had usable data at the 24-week primary endpoint, so it may have missed a real difference between the programmes. Only 47 families completed the 24-week assessment before COVID-19 restrictions; assessments and both programmes then moved online, so most of the trial ran in a changed form and the authors say they could not determine how much this affected the results. Families attended on average 5 of the 8 sessions, so many did not receive the full programme. Both programmes were group-based, ran for the same number of hours and were led by trained professionals, and parents in both arms reported improvement in stress, wellbeing and family functioning; nothing in the trial separates the content of either programme from the effect of attending a supportive group. The advantage on the targeted behaviour was no longer present at 52 weeks. Each parent chose the behaviour to work on, so what improved differed from family to family. The experts rating the target-behaviour vignettes did not have their agreement checked on the final ratings. Families came from Scotland and north-east England, most were of white background and just under half of parents had a postgraduate degree, so the results may not apply to other populations.

Funded by the National Institute for Health Research Health Technology Assessment programme (HTA 16/111/95). Four authors (Linda Dixon, Ann Le Couteur, Jessica Maxwell and Victoria Grahame) are authors of the Understanding Repetitive Behaviours programme, so they have an interest in its results. Deborah Garland is employed by the National Autistic Society, which trained and approved the leaders of the comparison programme. Another author, Sue Fletcher-Watson, is an editor of the journal. The trial's steering and data monitoring committees were independent of the sponsor and funder.

Do not read this as evidence that the programme changes how autistic children function overall. The primary outcome, a clinician's global rating of improvement, showed no difference between the two programmes (12.35% versus 14.86% of children improved, P = 0.51), and the advantage on the specific targeted behaviour was gone by 52 weeks. Do not read it as evidence the programme is worthless either: children whose parents attended it were more likely to improve on the behaviour their parent had chosen to work on, and both groups of parents reported improvements in their own wellbeing.