Characterization of Aggressive Behavior in Children with Autism Spectrum Disorders

Aggression is not well characterized in children with autism spectrum disorders (ASD), although it is commonly claimed that these children engage in more aggressive behaviors than typically developing children. This is in stark contrast to the myriad reports available on the normative and developmental patterns of aggression in typically developing children. Thus, the prevalence and severity of aggressive behavior in a moderately-sized clinical sample of children with ASDs was assessed using a common measure of childhood psychopathology, the Child Behavior Checklist (CBCL) and a relatively new scale of aggression, the Children’s Scale for Hostility and Aggression: Reactive/Proactive (C-SHARP). Participants were n = 223 children with ASDs and n = 115 children with no ASD diagnosis from four sites: Nationwide Children’s Hospital (Columbus, Ohio), Thompson Center for Autism (Columbia, Missouri), University of Illinois at Chicago, and Dayton Children’s Hospital (Dayton, Ohio). Participants were between the ages of 17 and 224 months (mean = 84.74, standard deviation = 45.82). Item- and subscale- level data were reported for the CBCL Aggressive Behavior subscale as well as for the five C-SHARP subscales: Verbal Aggression, Bullying, Covert Aggression, Hostility, and Physical Aggression. Summary data were also reported for the Provocation scale of the C-SHARP, on which rater answer the question, “Who starts it?” Diagnostic, demographic, and other variables were examined as correlates of aggression, including age, sex, comorbid diagnoses, psychotropic medication, IQ, adaptive behavior scores, and language ability. C-SHARP data from the ASD group were also subjected to cluster analysis, in an effort to discover clinically-meaningful subgroups of children based on profiles of aggressive behavior. Children with ASDs did engage in aggressive behaviors; the highest average scores were observed on the Hostility subscale. A larger-than-expected proportion (17%) of the ASD sample received a CBCL Aggressive Behavior subscale t-score above 70. The control group had higher scores than the ASD group on two C-SHARP subscales, Covert Aggression and Physical Aggression. Several correlates of aggressive behavior were observed in the ASD group. Asperger’s disorder was associated with higher scores than the other diagnostic groups on several subscales. Older children had higher scores on Verbal Aggression, Covert Aggression, and Hostility. Members of the ASD group with ADHD diagnoses and/or taking ADHD medication scored higher on most subscales of the C-SHARP. The selected cluster solution comprised five clusters: Low Aggression, Primarily Hostile, Physical, Verbal, and High Aggression. Single-sample and split-sample reliability kappa estimates were above 0.90, indicating excellent agreement. Differences were observed between clusters based on external variables, supporting the validity of the cluster structure. Children with autistic disorder were likeliest to be in the Low Aggression group, but those who were aggressive were over-represented in the Physical group. The Verbal group had a larger-than-expected proportion of children with PDD-NOS, while the High cluster had more children with Asperger’s disorder than expected. Results of this study converge with the limited existing data on aggression in children with ASDs and suggest future directions for areas of focus.