DSM-5 Criteria Applied to Toddlers Diagnosed with ASD by DSM-IV-TR

Background: The currently proposed DSM-5 diagnostic criteria for autism spectrum disorders (ASD) include substantial revisions. The most significant changes include combining the subcategories (Autistic Disorder, Asperger’s Disorder, and PDD-NOS) into one dimensional category of ASD, combining the social and communication domains into one, and requiring two rather than one restrictive and repetitive behaviors (RRBs) (at DSM5.org). Because of strong evidence of the relationship between early diagnosis and intervention with more positive outcomes (Myers & Johnson, 2007), it is important to have diagnostic criteria that have adequate sensitivity for children under three years. Concerns have been raised about the proposed DSM-5 criteria’s sensitivity for very young children (Worley & Matson, 2012) since repetitive behaviors may not be manifest in this age group. 

Objectives: In order to address concerns about sensitivity of the proposed criteria in toddlers, this current study examined if toddlers who received an ASD diagnosis under the DSM-IV criteria would maintain their diagnosis with the proposed DSM-5 criteria. 

Methods: Children (n=234) between the ages of 16 and 30 months (M=25.77, SD=4.58) who were part of a multi-site study examining the sensitivity and specificity of the Modified Checklist for Autism in Toddlers (M-CHAT; Robins, Fein, & Barton, 1999) and who received an ASD diagnosis were included in the study. Parent report and direct observation measures included various editions of the ADI, the ADOS Module 1, DSM-IV-TR criteria, and the examiner’s additional behavioral observations. Items in these measures that reflected the proposed DSM-5 criteria were used to create an algorithm to determine if participants who met the DSM-IV-TR diagnosis for an ASD also met the proposed DSM-5 criteria. 

Results: Preliminary results suggested that 29% of toddlers who previously met an ASD diagnosis no longer did so with the new criteria; the majority of toddlers did not meet criteria B2 (routines, rituals, and/or inflexibility) and B3 (restrictive interests) in the proposed DSM-5 criteria. Relaxing criterion B by requiring one instead of two RRBs increased sensitivity; 17% of toddlers who previously met an ASD diagnosis no longer did so with the currently proposed DSM-5 criterion A and a relaxed criterion B (1 out of 4 RRBs). Under both systems, children with diagnoses of PDD-NOS disproportionately lost the diagnosis, compared to children with a diagnosis of Autistic Disorder.    

Conclusions: Because of the significant implications of early detection and intervention of ASD on outcome, it is important that the proposed DSM-5 criteria reflect the presentation of ASD in toddlers. Requiring two RRBs may negatively impact the early detection of ASD because these behaviors are not as apparent in toddlers. Requiring one RRB instead of two would significantly increase sensitivity.