The agreement between the referrer, practitioner and research diagnosis of autistic spectrum conditions among children attending child and adolescent mental health services.

We aimed to explore the levels of agreement about the diagnoses of Autistic Spectrum Conditions between the referrer, CAMHS practitioner and a research diagnosis, as well as the stability of the practitioner's diagnosis over time in a secondary analysis of data from 302 children attending two Child and Adolescent Mental Health Services over two years. Kappa coefficient was used to assess the agreement between the referrer and research diagnosis. Kendall's tau b coefficient was used to assess the agreement between the practitioner and the research diagnosis assigned using the Development and Well-Being Assessment, as well as the agreement between the referrer's indication of presenting problems and the practitioner diagnosis. Diagnostic stability was explored in children with and without a research diagnosis of Autistic Spectrum Condition. There was a moderate level of agreement between the referrer and research diagnosis (Kappa = 0.51) and between practitioner's and research diagnosis (Kendall's tau = 0.60) at baseline, which reduced over the subsequent two years. Agreement between the referrer and practitioner's diagnosis at baseline was fair (Kendall's tau = 0.36).The greatest diagnostic instability occurred among children who practitioners considered to have possible Autistic Spectrum Conditions but who did not meet research diagnostic criteria. Further studies could explore the approaches used by practitioners to reach diagnoses and the impact these may have on diagnostic stability in Autistic Spectrum Conditions. Standardised assessment using a clinically rated diagnostic framework has a potential role as an adjunct to standard clinical care and might be particularly useful where practitioners are uncertain.