An International Multi-Site Investigation of the Measurement Properties of the Spence Children's Anxiety Scale – Parent Version (SCAS-P) to Report Anxiety Symptoms in Children and Adolescents with ASD

Background: Elevated anxiety symptoms are present in many young people with ASD and 30-50% are diagnosed with at least one anxiety disorder, but reporting and identifying anxiety symptoms and difficulties in young people with ASD presents many challenges. Research in this field has been hampered by small samples of largely help-seeking participants with existing anxiety conditions, as well as limited research examining the measurement properties of existing anxiety screening tools when used to measure anxiety in individuals with ASD. Currently, no validated informant-based screening measure exists specifically developed for use with this population.

Objectives: This study aimed to (a) examine caregiver-reported rates and symptoms of anxiety in a large international multi-site sample of young people with ASD; (b) investigate the SCAS-P’s reliability and validity in measuring anxiety symptoms in this population; and (c) examine the relationship between SCAS-P reported anxiety symptoms and age, gender and measures of autism symptom severity, adaptive functioning and/ or intellectual functioning. 

Methods: Participants were 679 6-18 year old children and young people with ASD (591 males) whose data were pooled together from nine different research studies in Singapore (N=241), the UK (N=373) and the USA (N=65). All participants had a clinical/ professional diagnosis of autism, ASD, Asperger’s Syndrome or PDD-NOS and most had at least one measure of autism symptom severity available (i.e. Autism Diagnostic Observation Schedule, Developmental Behavior Checklist-Autism Algorithm, Social Communication Questionnaire or Social Responsiveness Scale). IQ data were available for 211 participants and another 238 had available adaptive functioning scores. All but 21 participants were non-help seeking and were recruited from special schools, other community settings or from ASD diagnostic settings at the time of referral for a diagnosis of ASD, but not for anxiety difficulties.

Results: Preliminary analyses show that the internal consistency for the SCAS-P was excellent for the Total score (α =.94) and acceptable to good for all subscales (all α >.70). All item-total correlations were significant at p<.001 and the mean item-total correlation was of a medium effect size, while item-subscale correlations were also highly significant and of medium to large effect sizes. There was no significant association between chronological age and total anxiety symptoms and no gender differences were found. The positive association between SCAS-P Total score and autism symptom severity scores was statistically significant and of small to medium effect size. No significant relationship between SCAS-P Total scores and IQ or adaptive functioning scores was found. SCAS-P Total and subscale scores were significantly higher for the 21 participants with ASD and known anxiety difficulties when compared to the other participants recruited from community settings. There were some site/ country differences in caregiver-reported mean scores.

Conclusions: Preliminary analyses show that the SCAS-P is promising with regards to its measurement properties when used to measure caregiver-reported anxiety symptoms in young people with ASD. Our findings are discussed with a focus on advancing assessment of anxiety in ASD.