Continuity, Change and Predictors of Anxiety Symptoms in a Community Sample of Children and Youth with ASD: A Prospective Follow-up Study

Background:  Many individuals with ASD experience high rates of anxiety, but mixed findings have so far been reported in the literature regarding anxiety’s relationship with other individual characteristics (i.e. age, gender, cognitive and adaptive functioning, ASD symptomatology). Few studies have examined whether the relationship between ASD and anxiety symptoms may be “fractionable”, that is whether ASD social/communication or stereotyped behavioural symptoms may be differentially associated with anxiety. Furthermore, most existing studies have been cross-sectional and continuity and change in anxiety symptoms in prospective study designs has not yet been sufficiently explored. 

Objectives:  (i) to examine continuity and change in caregiver reported anxiety symptoms of young people with ASD; (ii) to examine the predictive value of earlier child characteristics in explaining and predicting anxiety 12-18 months later, considering the two domains of ASD symptomatology independently; and (iii) to explore whether the relationship between anxiety and ASD symptomatology is uni- or bi-directional. 

Methods: Participants were caregivers of 54 children and youth aged 6 to 17 years (88.9% males) with a clinical diagnosis of ASD, Asperger’s syndrome or PDD-NOS recruited from special schools in Singapore. Data was collected on adaptive functioning, anxiety symptoms (using the Spence Children’s Anxiety Scale- Parent Version; SCAS-P) and other emotional or behavioural problems at Time 1 and then 12-18 months later. 241 children participated in the study at Time 1 (Magiati et al., 2015), of whom 164 agreed to be contacted at follow-up and fifty-four (33% response rate) completed the measures at Time 2. Hierarchical regression analyses examined the incremental predictive value of earlier child variables (age, gender, adaptive functioning, autism symptoms and Time 1 anxiety) on Time 2 anxiety symptoms. 

Results:  Mean total and subscale anxiety rates were generally stable from Time 1 to Time 2, except for social anxiety symptoms, which significantly increased over time. Age, adaptive functioning and earlier repetitive/stereotyped behaviors, but not earlier social/communication ASD symptoms, were significant predictors of later total anxiety scores. However, when Time 1 anxiety was included as a covariate, the variance in Time 2 anxiety scores was fully explained by earlier total anxiety scores. The same pattern of findings was found for generalized anxiety, panic/agoraphobia and obsessive-compulsive SCAS-P subscale scores. Only for separation anxiety symptoms did earlier stereotyped behavioral ASD symptoms significantly predict later separation anxiety over and above baseline separation anxiety scores and all other child factors examined. Time 1 anxiety scores, however, did not predict Time 2 stereotyped ASD symptoms, suggesting the relationship may be unidirectional. 

Conclusions:  Anxiety symptoms in young people with ASD remain generally stable over time without intervention, while social anxiety increases. Only Time 1 ASD repetitive behavior, but not social/communication, symptoms predicted anxiety one year later, after age and adaptive functioning were controlled for. However, this relationship was fully explained by earlier anxiety scores predicting later anxiety scores, with the exception of separation anxiety. Given our small sample size and exclusive use of caregiver-report measures, we recommend replication with larger samples and use of multiple-informant measures to validate and extend our preliminary findings.