Identifying Homogeneous Subgroups of Children with ASD, ADHD, and OCD

Background:  There is significant overlap in the presence of repetitive behaviors, social skills deficits, obsessive compulsive behaviors, inattention and hyperactivity across autism spectrum disorders (ASD), attention deficit-hyperactivity disorders (ADHD), and obsessive compulsive disorders (OCD).  While the DSM-5 allows for comorbid diagnosis of ASD, ADHD, and OCD, some degree of overlap is expected due to non mutually exclusive diagnostic criteria. The identification of homogeneous subgroups of children who fail to meet the diagnostic threshold for comorbidity may help predict long term functional impairment.  

Objectives:  To identify subgroups with similar phenotypic profiles within a sample of children with a primary diagnosis of ADHD, ASD, or OCD, and to test for differences in functional impairment across the subgroups.

Methods:  The sample consisted of 563 children and youth with  ASD (n=231), ADHD (n=231), and OCD (n=101) participating in the Province of Ontario Neurodevelopmental Disorders (POND) Network. Individual-item data collected using four measures,  the Social Communication Questionnaire (SCQ),  Repetitive Behavior Scale-Revised (RBS-R), Toronto Obsessive Compulsive Rating Scale Parent-Report Version (TOCS), and the Strengths and Weaknesses of Attention-Deficit/Hyperactivity Disorder Symptoms and Normal Behavior Scale (SWAN) were included in a latent class analysis.  ANOVA was used to examine the association between primary diagnosis and latent class membership on Adaptive Behavior Assessment System II (ABAS-II) composite scores.  

Results:  Using a priori criteria, a five latent class model provided the best fit to the data. The largest class (n=183), 85% of whom had a primary diagnosis of ADHD, included individuals with low scores on all but the SWAN scale.  A second class (n=93), 80% of whom had a primary diagnosis of OCD, included individuals with high scores on the TOCS only.  The majority of participants with a primary diagnosis of ASD were classified almost evenly (n=77, 58, and 77) into the remaining three classes (n=116, 75, and 96), with the balance of class members made up predominantly of individuals with ADHD.  Both diagnosis and latent class were significant independent predictors of ABAS composite scores.  Among individuals with a primary diagnosis of ASD, significant differences in the ABAS general composite score were seen across all latent classes.

Conclusions:  The empirical identification of cross-disorder homogeneous subgroups of children and youth with ASD, ADHD, and OCD seems to be informative in predicting functional impairment, over and above the expected prediction ability of index diagnostic categories. This could have implications for the design of cross-diagnostic intervention trials and aid in genetic studies searching for shared and unique variants associated with these disorders.