DSM-5 under-identifies PDDNOS: Diagnostic agreement between the DSM-5, DSM-IV, and Checklist for Autism Spectrum Disorder

Agreement between the DSM-5, DSM-IV, and Checklist for Autism Spectrum Disorder was assessed in 125 children with autism spectrum disorder (ASD), which included high and low functioning autism (HFA and LFA) and pervasive developmental disorder not otherwise specified (PDDNOS), and children with other clinical disorders (e.g., ADHD, mental retardation, and oppositional-defiant disorder). Diagnostic specificity (true negatives) was 100% for the DSM-5 and 97% for the DSM-IV. Sensitivity (true positives) was high for LFA and HFA (DSM-5 98%, DSM-IV 100%), but only 27% of children with PDDNOS were identified by the DSM-5 as having ASD. The unidentified children had significant autism symptoms on an autism severity measure compared to controls. Overall, children with ASD identified by the DSM-5 increased from 75% to 91% when one less symptom was required for an ASD diagnosis, with only a slight change in specificity (97%), similar to previously published research. Diagnostic agreement between the DSM-5 and DSM-IV for the entire sample increased from 86% to 94% when one less DSM-5 symptom was required. Therefore, under-identification of ASD and low agreement between the DSM-5 and other measures may be resolved if one less symptom is required for an ASD diagnosis.