19348
Close but No Cigar: Factor Structure of the ADHD Rating Scale in Children with Autism Spectrum Disorder
The ADHD Rating Scale is a commonly used measure in clinics and research as a screen for ADHD, because it maps directly to DSM symptoms. In a normative sample, the original factor analysis of the ADHD-IV rating scale found the best fit to be a two-factor solution (inattention and hyperactivity/impulsivity) consistent with the DSM-IV and DSM-5 ADHD subtypes. There are no published studies that examine the ADHD Rating Scale’s factor structure in school-age youths with ASD.
Objectives: To characterize the factor structure of parent and teacher ratings ADHD rating scale in a large sample of youth with ASD and to probe commonly observed relationships between ADHD symptoms and youth characteristics.
Methods: Parents and teachers of 208 youths with an ASD diagnosis (192 males; Age M=10.45 (Range: 6-18 years)); General Conceptual Ability M=100 (Range: 67-158), completed this study. ASD was diagnosed using DSM-IV criteria and confirmed with ADI-R/ADOS. Confirmatory factor analysis (CFA) was used with polychoric correlations using one-, two- (Inattention, Hyperactivity/Impulsivity), and three-factor (Inattention, Hyperactivity, Impulsivity) models, with a robust weighted least squares estimation. We also conducted Pearson correlations with age, executive function, and adaptive behaviors.
Results: According to parents, 57% did not meet clinical cut-off; 16% were Predominantly Inattentive, 6% Predominantly Hyperactivity/Impulsivity, and 21% Combined type. According to teachers, 73% did not meet clinical cut-off; 14% Predominantly Inattentive, 4% Predominantly Hyperactivity/Impulsivity, and 9% Combined type. Correlations revealed significant, but modest, negative relationships for age with parent and teacher ratings of hyperactivity/impulsivity (rs=-.26 and -.29), and teacher ratings of inattention (r=-.22). Partial correlations between parent ADHD symptoms and parent BRIEF scales while controlling for age revealed moderate-to-large effects (r’s>.46). Correlations between parent ADHD symptoms and parent adaptive behavior ratings revealed significant but modest negative correlations (Socialization r=-.23 and Communication r=-.29).
The CFA did not meet Goodness-of-Fit criteria for the 1-factor (Parent: CFI=.907, TLI=.894, RMSEA=.127; Teacher: CFI=.873, TLI=.857, RMSEA=.144) or 2-factor models (Parent: CFI=.946, TLI=.938, RMSEA=.097; Teacher: CFI=.946, TLI=.938, RMSEA=.095), but the 3-factor met partially (Parent: CFI=.952*, TLI=.944, RMSEA=.093; Teacher: CFI=.958*, TLI=.951*, RMSEA=.084). Modification indices suggest items 5, 10, and 15 reduced the 3-factor model fit.
Conclusions: Prevalence rates and relationships with youth characteristics converge with prior studies, but the factor structure observed in community youth did not fit for youth with ASD. Three items require modification to distinguish ADHD from social interaction deficits or more general ASD deficits. This study has significant clinical and research implications for screening/diagnosing ADHD in ASD.