Diagnosing autism: Australian paediatric research network surveys.

AIM: Autism spectrum disorder (ASD) is a neurodevelopmental disorder with reported prevalence of more than 1/100. In Australia, paediatricians are often involved in diagnosing ASD and providing long-term management. However, it is not known how paediatricians diagnose ASD. This study aimed to investigate whether the way Australian paediatricians diagnose ASD is in line with current recommendations. METHODS: Members of the Australian Paediatric Research Network were invited to answer questions about their ASD diagnostic practice in a multi-topic survey and also as part of a study about parents needs around the time of a diagnosis of ASD. RESULTS: The majority of the 124 paediatricians who responded to the multi-topic survey and most who responded to the parent needs survey reported taking more than one session to make a diagnosis of ASD. Most paediatricians included information from preschool, child care or school when making a diagnosis, and over half included information from speech pathology or psychology colleagues more than 50% of the time. The main reasons for not including assessment information in the diagnostic process were service barriers such as no regular service available or long waiting lists. More than 70% reported ordering audiology and genetic tests more than half of the time. CONCLUSION: Not all paediatricians are following current recommendations for diagnosing ASD more than 50% of the time. While there are good reasons why current diagnostic approaches may fall short of expected standards, these need to be overcome to ensure diagnostic validity and optimal services for all children and their families.

Can questions about social interaction correctly identify preschool aged children with autism?

AIM: We developed a questionnaire to assess social development (SIQ) in preschool children. Social development is often not included in medical assessment, though it may assist in early identification of autism spectrum disorder (ASD). METHODS: Parents of 108 children with ASD, speech and language disorders, or 'developmental concerns', recruited from a clinical developmental assessment and community child health service, completed the SIQ, and also a Childhood Autism Rating Scale (CARS) assessment. Receiver Operator Characteristic (ROC) curves were generated to assess the performance of different questionnaire score thresholds in correctly identifying children with a CARS score of 30 or more. Logistic regression models were used to identify the questions which had the most predictive value for a CARS score of 30 or more. RESULTS: An SIQ score of 14 or more correctly identified children with a CARS ≥ 30 with a sensitivity of 85%, specificity 85%, positive likelihood ratio (LR) 8.3 and negative LR 0.2. Two questions were identified as most predictive of ASD. CONCLUSIONS: The SIQ may assist clinicians in assessing social development and in making decisions about referral for autism assessment. Evaluation of the SIQ at the point of entry to a clinical service is needed.