A Collaborative Autism Screening Project Utilizing Web-based Technology

For young children with autism spectrum disorders (ASD), early detection and screening have garnered considerable attention, in part, due to consensus that children with ASDs identified and enrolled in early intervention programming evince improved outcomes. As a result, several screening instruments have been developed for early detection of ASD. Most ASD screeners used with young children rely on parent report to provide information about symptoms suggestive of the presence of an ASD. Parents are considered ideal informants for completing screening instruments due to their access to children's behavior observed in naturalistic settings, which cannot be replicated in pediatric offices. The primary aim of the research project is to compare the predictive validity of three parent-report ASD screening instruments for preschoolers and young children. We are interested in comparing and contrasting screeners' predictions of results from a standardized diagnostic assessment, which includes a gold standard diagnostic measure. We are also interested in examining if the predictive validity of screening measures differs across racial/ethnic groups and SES. We will use receiver operating characteristic (ROC) analysis to compare and contrast test accuracy across three measures with the aim of identifying which, if any, of the screeners emerges as superior to others. We will contrast test accuracy by testing for differences between areas under the ROC curve (AUCs) generated by each screening measure. A secondary aim of the study targets application of a web-based video analysis tool to confirm results from diagnostic evaluations. Incorporating web-based video technology to establish reliability of gold standard diagnostic test results will allow our group to investigate the utility of a novel web-based video analysis tool to support ASD diagnostic practice from remote locations.