Building Capacity: A Systematic Review of Training in the Diagnosis of Autism for Community-Based Clinicians.

Only ten studies and seven training programmes were found, and most were small pilots. One training report had no stated methods and no before-and-after comparison (high risk of bias) and two were pilots with no before-and-after data (moderate risk). Nearly every result is the clinician's own account of what they do or how confident they feel, collected straight after the training. There is almost nothing here about children: only two trainings reported any family outcome, and those were satisfaction and access, not whether children were diagnosed earlier or did better. The diagnostic agreement figures cannot be compared with each other. Some studies forced clinicians to say autism or not, others allowed them to record uncertainty, and the highest figure of 95% counted only children the clinician had already diagnosed, so it says nothing about the children a trained clinician might miss. Nobody followed clinicians up, so the review cannot say whether new skills, confidence or diagnostic accuracy survive contact with a busy clinic. Titles and abstracts were screened by a single reviewer for most records, with 10% double-checked; the authors state that full double screening would have been better. Only full texts in English were included, the search stopped in August 2023, reference lists of included articles were not searched, and the search missed two studies that an earlier review had included. Cost was reported for only two trainings and did not cover every part of the training, and no training reported anything about whether the changes were sustained.

The authors declare no conflicts of interest, and the review does not say who funded it. In the two included trainings that reported costs, the training was paid for from outside the practice, including a state contract in one, and one programme reported that the ADOS-2 training and materials cost $19,800.

Do not read the 71% to 95% diagnostic agreement as proof that trained community clinicians diagnose autism as reliably as specialists. Each training measured agreement a different way, and the 95% figure counted only children the clinician had already diagnosed, so no child the clinician ruled out was re-tested. The rest of the findings are what clinicians said about their own practice immediately after training, and no study checked whether any of it lasted.