The Relationship of Anxiety and Reciprocal Social Impairment in Autism: A Comparison Study of the Scared and SRS

Background:  There is increased recognition and interest in psychiatric comorbidity in Autism Spectrum Disorders (ASD).  Anxiety is frequently seen in ASD but it is not clear if the two conditions are separate or if symptoms of anxiety are part of the essential features of ASD such as social impairment and restrictive, repetitive behaviors. The complex interplay of anxiety and core social communication deficits in ASD warrants additional attention. 

The Screen for Child Anxiety Related Disorders (SCARED) is a well-established screening instrument for anxiety in youth and shows  promising results in its application to children with ASD. The Social Responsiveness Scale (SRS-2) has been  used to identify and study the severity of social communication deficits in ASD  in a quantitative fashion. Comparing results of  these two instruments in  subjects with ASD provides opportunities to better understand the relationship between ASD and anxiety in this population.

Objectives:  The purpose of the study was  to determine how parental and child reports of anxiety and are related to social impairment in subjects with ASD. By examining how individual items and treatment subscales of the SRS-2 relate to measures of anxiety on the SCARED, we hoped to understand better which core symptoms of ASD were most highly associated with parent and child reports of anxiety. 

Methods:  100 patients age 8 to 18 years with a clinical diagnosis of ASD without intellectual disability were recruited from clinic sites.  Participants and parents completed the SCARED child and parent forms and the SRS-2. SCARED parent and child total scores and diagnostic factor subgroups were compared to SRS-2 categories of severity by using the Mann-Whitney U-test or Analysis of Variance. Random Forest models were used to identify individual SRS items important for classifying children as anxious for parent and child SCARED scales separately.  Poisson regression models measured the predictive ability of each of those variables to detect anxiety. 

Results:  Parents rated children with ASD and more severe levels of social impairment as having more anxiety. This pattern holds across all diagnostic subtypes represented by the SCARED. There were no significant correlations between levels of social impairment and child reports of anxiety.

Random Forest analysis associated parents’ reports of restrictive, repetitive behavior and emotional dysregulation on the SRS-2 with high parental ratings of anxiety. However, regression analysis showed none of these items significantly predicted anxiety based on parent SCARED.

Random Forest analysis associated parent reports of preoccupations and lack of observed social behavior on the SRS-2 with high child reports of anxiety. Parental reports of clinging to adults and having unusual sensory interests or stereotypical play significantly predicted child ratings of the presence of an anxiety disorder.

Conclusions:  The SCARED and SRS-2 can be used to understand the nature of anxiety in ASD. Ratings of anxiety in autism correlate with core autistic symptoms as well as externalized behavior. Parents and children associate anxiety with different types of social impairment. The relationship of anxiety and autism may be best understood by non-linear measures.