The diagnosis of autistic spectrum disorders (ASD) has increased in uniformity and reliability in the last 10 years due to a number of factors, including the agreement between American Psychiatric Association and the World Health Organization committees to propose nearly identical categorizations and the increasingly widespread use in research of standardized diagnostic instruments. Two instruments, the Autism Diagnostic Interview - Revised (ADI-R) and the Autism Diagnostic Observation Schedule (ADOS) have been adopted by most North American and British researchers, as well as translated into French, Spanish, German, Korean, Portuguese, Danish, Norwegian, Swedish and Icelandic. More recently, a screening instrument, the Social Communication Questionnaire (SCQ; formerly the ASQ), has also received attention. However, much work remains to be done to make the instruments more useful for research purposes and clinical service. The present proposal uses analyses of an existing data set of over 1100 individuals with scores on the ADIR, ADOS, intelligence tests, the Vineland Adaptive Behavior Scales, and clinical best estimate diagnoses, with replication of findings in three other large datasets from the University of Wisconsin, Autism Genetics Research Exchange and the International Molecular Genetics Autism Consortium. In addition, data will be collected from the ongoing clinic at the University of Michigan using various manipulations (e.g., testing the effect of """"""""current"""""""" versus """"""""ever"""""""" probes) and small, targeted samples will be collected from currently underrepresented samples (e.g., severe to profoundly retarded individuals without autism). The purpose of the analyses will be to generate severity scores (within and across domains) that are independent of chronological age and language level for each instrument, generate a """"""""certainty of caseness"""""""" rating that takes into account contributions of different instruments across age and language level and allows combination of information across instruments, reduce the number of items and the complexity of the algorithm for each instrument, and test the usefulness of the SCQ as a screening for possible autism in a yet-to-be diagnosed population of children referred to developmental disabilities and general child psychiatry clinics. In addition, the usefulness of a simpler, two domain plus language level approach to conceptualization of the autistic spectrum will be tested.
Showing the most recent 10 out of 38 publications
