Combining information from multiple sources in the diagnosis of autism spectrum disorders

Background: Standard case criteria are proposed for combined use of the Autism Diagnostic Interview-Revised and Autism Diagnostic Observation Schedule to diagnose autism and to define the broader category of autism spectrum disorders. Method: Single and combined Autism Diagnostic Interview-Revised and Autism Diagnostic Observation Schedule algorithms were compared to best estimate diagnoses in four samples: U.S. (n = 960) and Canadian (n = 232) participants 3 years and older, U.S. participants younger than 36 months (n = 270), and U.S. participants older than 36 months with profound mental retardation (n = 67). Results: Sensitivities and specificities of 80% and higher were obtained when strict criteria for an autism diagnosis using both instruments were applied in the U.S. samples, and 75% or greater in the Canadian sample. Single-instrument criteria resulted in significant loss of specificity. Specificity was poor in the sample with profound mental retardation. Lower sensitivity and specificity were also obtained when proposed criteria for broader spectrum disorders were applied. Conclusions: The Autism Diagnostic Interview-Revised and Autism Diagnostic Observation Schedule make independent, additive contributions to the judgment of clinicians that result in a more consistent and rigorous application of diagnostic criteria.