Concordance Between DSM-5 ASD Criteria and Community ASD Identification Under DSM-IV-TR in a Population-Based Study

Background: The fifth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-5), published in 2013, included revised diagnostic criteria for autism spectrum disorder (ASD). Studies using clinical research samples have shown that some children who met ASD criteria based on DSM-IV-TR will not meet DSM-5ASD criteria, raising concerns about individuals “losing” their ASD diagnosis or no longer qualifying for certain services.

Objectives: To examine the relationship between ASD identification by community professionals and having documented symptoms consistent with DSM-IV-TR or DSM-5ASD criteria.  

Methods: The Autism and Developmental Disabilities Monitoring (ADDM) Network performed population-based ASD surveillance for 8-year-old children living in 14 US communities in 2006 and 2008, using information abstracted from health or education records.  Children’s records were abstracted for clinician review if they contained diagnostic statements or behavioral symptoms suggestive of ASD.  Trained clinicians systematically coded the information corresponding to the behavioral diagnostic criteria for Autistic Disorder, Asperger’s Disorder, or Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NOS), based on the DSM-IV-TR. Clinicians also noted when a child already had an ASD diagnosis or autism classification for special education.  We operationalized the DSM-5 ASD criteria using behavioral symptoms collected from developmental evaluations.  Among 12,277 children with abstracted information, 45% had documentation of ASD by a community professional. We calculated the concordance between community ASD identification and having symptoms consistent with DSM-IV-TR PDD-NOS or Autistic Disorder, and between community ASD identification and having symptoms consistent with DSM-5ASD.

Results: The overall percent concordances between community ASD identification and documented ASD behavioral criteria were 53.6% for any DSM-IV-TR ASD, 74.5% for DSM-IV-TR Autistic Disorder, 35.4% for DSM-IV-TR PDD-NOS or Asperger’s Disorder (excluding children meeting criteria for Autistic Disorder), and 74.3% for DSM-5 ASD.  Similarly, the percent of children identified with ASD varied depending on the ASD criteria met: 49.4% of 10,753 children with documented behaviors indicating any DSM-IV-TR ASD; 67.7% of 6,908 children with documented behaviors indicating DSM-IV-TR Autistic Disorder, 16.6% of 3,845 children with documented behaviors indicating DSM-IV-TR PDD-NOS or Asperger’s Disorder (excluding those meeting Autistic Disorder), and 68.7% of 4,453 children with documented behaviors indicating DSM-5ASD.

Conclusions: The DSM-5 ASD criteria showed higher concordance with community ASD identification practices, suggesting that the DSM-5 ASD criteria may, on the whole, better reflect current ASD identification practices in community settings than the DSM-IV-TR criteria. The lower concordance among those meeting DSM-IV-TR PDD-NOS criteria is consistent with previous clinical studies suggesting that more children meet the minimum symptom threshold for PDD-NOS than clinically appropriate.  Despite the higher overall concordance, fewer children met DSM-5 ASD criteria than the DSM-IV-TR criteria. Because the data collected for this study preceded the development of the DSM-5, additional studies are needed to monitor how community providers adopt the new ASD criteria.