Anxiety in Autism and Autism in Anxiety: Symptom Overlap on the SRS-2 Adult Self-Report

Background: There is growing awareness that common autism symptom measures including the ADI-R (van Steensel et al. 2013), ADOS-G (Pine et al., 2008) and SRS (Cholemkery et al., 2014; Puelo & Kendall, 2011) may overlap with other non-diagnostic constructs including anxiety and mood symptoms. On the one hand, knowing about co-occurring social difficulties may be useful for treatment in mood and anxiety disorders. On the other hand, these findings raise questions about the discriminant validity of hallmark symptoms of autism from other associated features. Previous studies in this area have been with children. To our knowledge, there is little or no work in this area for adults, where anxiety and autism symptoms may converge more strongly.

Objectives: We aimed to characterize the overlap of autism and anxiety symptoms using two separate autism symptom measures in ASD, anxious, and non-anxious adults.

Methods: Participants included adults diagnosed with ASD (n=39; ASD group); first-time visitors to a university counseling center who scored high on intake surveys of anxiety (n=56; ANX group); and a sample of university students with no psychiatric history (n=37; NT group). Participants all completed the self-report Social Responsiveness Scale-Second Edition (SRS-2) and the Autism Spectrum Quotient (AQ) in addition to IQ testing and a battery of emotion symptom questionnaires.  

Results: Strong, significant correlations were found between ASD and anxiety measures in all three groups. One-way ANOVA indicated significant main effects for diagnosis on both the AQ and SRS total scores. Post-hoc analyses showed that the ANX group scored significantly higher than the NT group but significantly lower than the ASD group on both measures. The ANX raw scores mean and standard deviation (68 ± 24) indicates that many in this group are scoring in a range thought to indicate autism diagnosis. The SRS-2 Motivationsubscale, which includes items related to social approach and social discomfort (“I avoid starting social interactions with other adults;” “I enjoy making small talk (casual conversation) with others”) did not differentiate between ASD and ANX groups. In both the ASD and ANX groups, the Motivation scale was most strongly correlated with social anxiety measures but these were not significantly correlated in the NT group.

Conclusions: The SRS-2 and AQ have been shown to discriminate between ASD and typically developing (TD) individuals. Our data also indicates higher autism symptom scores in an ASD vis-á-vis anxious sample of adults, but likewise shows that the ANX group had markedly elevated scores on both measures. This seems to be at least partially driven by high scores on the Motivation subscale, which includes a number of items related to social anxiety. Further work to separate autism versus anxiety-related social behaviors can inform clinical intervention—including the need to explicitly treat social anxiety in adults with ASD—as well as research into the overlap of anxiety in autism.