ADOS-2 Module 4: Psychometric Properties and Diagnostic Performance at an Autism-specialized Clinic.

Abstract

Psychometric properties and diagnostic performance of the ADOS-2 module 4 were evaluated with participants from an autism-specialized clinic. The sample had 331 participants with 226 males and 70 females receiving an ASD diagnosis. The evaluation consisted of the following: (1) A replication of the Exploratory Factor Analysis (EFA) reported by Hus, V., & Lord, C. (2014). The Autism Diagnostic Observation Schedule, Module 4: Revised Algorithm and Standardized Severity Scores. Journal of Autism and Developmental Disorders, 44(8), 1996-2012. https://doi.org/10.1007/s10803014-2080-3 .). (2) Identification of ADOS-2 items best predicting clinical diagnosis using Recursive Feature Elimination (RFE) and comparison of these items to previous findings. (3) Receiver Operating Characteristic (ROC) curve analyses of the effects of age and IQ on diagnostic performance. (4) Comparisons of ADOS-2 revised algorithm scores between females and males and their association with ADI-R scores. The EFA item-to-factor allocation of the ADOS-2 revised algorithm was largely reproduced. When comparing the present RFE to previous RFE findings, the item Quality of Social Responses stood out. ROC curve analysis showed outstanding diagnostic performance for adults with IQ above 70 but for females, ADOS-2 revised algorithm scores were lowered, and, contrary to males, did not correlate with ADI-R scores. Reproducing the item-to-factor allocation and finding outstanding agreement with the diagnostic decision for adults with IQ > 70 showcase the strength of the ADOS-2 revised algorithm. Furthermore, by incorporating, into future revisions, the finding of inter-clinic importance of the item Quality of Social Responses, performance might be further enhanced. Lastly, though, that female scores were lowered and did not correlate with ADI-R indicate a weakness in the ADOS-2 when applied to the adult female population.