Objectives: The aim of the study was to investigate who would meet criteria for a draft DSM-5 diagnosis in data that was not collected specifically for the purpose of diagnosis according to DSM-IV-TR or ICD-10.
Methods: To investigate this, a new DSM-5 ASD algorithm was designed and tested using the Diagnostic Interview for Social and Communication Disorders (DISCO). First, sensitivity and specificity was tested using a clinically-matched sample [N=82] and further tested using an independent dataset [N=115]. Second, the output of the DISCO DSM-5 algorithm was compared with that of the DISCO ICD-10 algorithm using a dataset of individuals ranging in ability and age (N=200). This included a subset with Gillberg’s Asperger Syndrome (Gillberg et al., 2001) and a subset with ICD-10 Atypical Autism. Finally, modifications were made to the DSM-5 algorithm to improve sensitivity.
Results: Sensitivity and specificity was good for identification of cases with clinical autism (AUC <.87-.89 for the validation sample). Convergence was high between DISCO outputs for DSM-5 ASD and ICD-10 Childhood Autism (96%) and between DSM-5 ASD and Gillberg’s Asperger Syndrome (90%). No age or gender effects were found. Across all samples, 13%, mostly high functioning, did not meet DSM-5 criteria. Removing the age of onset criterion increased sensitivity without reducing specificity. Adjusting the social communication criterion increased sensitivity but with loss of specificity.
Conclusions: Use of the DISCO DSM-5 algorithm indicates that the proposed DSM-5 changes should identify the majority of individuals who currently receive a diagnosis of ICD-10 Childhood Autism or Gillberg’s Asperger Syndrome.
See more of: Clinical Phenotype
See more of: Symptoms, Diagnosis & Phenotype