Is reduced social competence a mechanism linking elevated autism spectrum symptoms with increased risk for social anxiety?

Objectives: Approximately 50% of children with autism spectrum disorder (ASD) develop comorbid social anxiety disorder, and this comorbidity predicts poorer treatment outcomes than either syndrome alone. ASD and social anxiety are both associated with reduced social competence as evidenced by difficulties implementing fundamental social skills for successful social interactions, but it remains unclear whether reduced social competence reflects a mechanism that explains the increased risk for social anxiety associated with elevated autism spectrum symptoms.

Design/methods: To address this gap in the literature, the current study combined multi-informant measures (child, parent and teacher report) with a sample of 194 children with and without psychiatric disorders (ages 8-13; 68 girls; 69% White/Non-Hispanic). Autism spectrum traits, social competence and social anxiety symptoms were measured continuously.

Results: Bias-corrected, bootstrapped conditional effects modelling indicated that elevated parent-reported autism spectrum symptoms predicted reduced teacher-perceived social competence (β = -.21) and elevated child self-reported social anxiety (β = .17); reduced social competence accounted for 20% of the autism/social anxiety link (indirect pathway β = .04, ER = .20), and reduced social competence also predicted higher social anxiety independent of autism symptoms (β = -.16; all 95% CIs exclude 0.0, indicating significant effects). Exploratory analyses suggested that these findings were driven primarily by autism spectrum social communication difficulties rather than restricted/repetitive behaviours/interests.

Conclusions: The current findings are consistent with prior work implicating reduced social competence as a risk factor for the development of social anxiety among children with ASD, and extend prior work by demonstrating that this link is robust to control for mono-informant/mono-measure bias, age, sex, SES, majority/minoritized race/ethnicity status, clinical comorbidities, and item overlap across measures.