Agreement among diagnostic instruments for autism spectrum disorders and clinical judgment in preschool aged children

Background: Past research indicates that there is strong agreement between clinical judgment based on DSM-IV criteria, CARS, and ADOS-G when classifying toddlers with ASD; however, the agreement between the ADI-R, and the CARS, ADOS-G, and clinical judgment is poor (Ventola et. al., 2006). The diagnostic agreement rates when classifying older preschool children with ASD have not been sufficiently studied.

Objectives: To examine the diagnostic agreement rates among the ADOS, ADI-R, CARS and clinical judgment in older preschool children with ASD.

Methods: Participants were 107 children who initially screened positive on the M-CHAT (Robins, et al., 2001), and were evaluated, at age 2. At a second time point about two years later (mean age 53 months), all children were reevaluated, and classified as ASD or non-spectrum by the ADOS, ADI-R, CARS, and clinical judgment based on DSM-IV symptomatology. Current results are based on this second evaluation.

Results: Kappa analyses revealed very good agreement between the ADOS and clinical judgment (k = .808, p = <.001) as well as moderate agreement between the ADOS and CARS (k = .553, p = <.001) and the CARS and clinical judgment (k = .442, p = <.001). The ADI-R displayed fair agreement with the CARS (k = .382, p = <.001), the ADOS (k = .236, p = .002), and clinical judgment (k = .262, p = <.001).

Conclusions: This replication found that the ADI-R continues to have relatively weak agreement with other diagnostic measures and clinical judgment in a sample of preschool aged children, suggesting limitations with its use in younger children. Among the factors affecting this lower agreement may be the measure's reliance on only parental report, the inclusion of past behaviors in the diagnostic algorithm, and its lack of PDD or ASD cut off scores.