Autism and Anxiety: Incorporating Self Reports in Evaluating Whether Anxiety Is Elevated in Children and Adolescents with Autism Spectrum Conditions

Background: Only recently has the co-occurrence of autism spectrum conditions (ASC) with behavioral symptoms become a focus of research (Matson & Nebel-Schwalm, 2005). Individuals with ASC diagnoses are often expected to be at elevated risk for depression and anxiety (Howlin, 2000), although studies often lack adequate controls, use small or non-systematically obtained samples, and ignore ASC individuals' self-reports. Given that typically developing children as young as 3½ years can provide reliable reports of self-perceptions (Measelle et al, 2005), ASC individuals' self-reports deserve evaluation. Objectives: We evaluated the reliability and validity of ASC children's self-reports of anxiety and depression. Then, we investigated the association of ASC with anxiety and depression in children using multiple sources of information and multiple control groups. Methods: The main sample consisted of 112 children (all twins). For these analyses, inclusion in the ASC group required above-threshold scores on the SCQ or the ADOS. We administered the Berkeley Puppet Interview (BPI) to children and a companion instrument, the Health and Behavior Questionnaire (HBQ), to primary care providers. Research staff, blind to the child's ASC status, rated each child on a general anxiety scale. We also administered the BPI and the HBQ to a comparison group of twins, matched on cognitive ability, age, gender, and parental income. Results: ASC children provided reliable BPI self-reports, with Cronbach's alpha for the anxiety and depression scales exceeding .70. Evidence for validity came from significant correlations between corresponding HBQ and BPI subscales (e.g., depression, r=.28, p<.05; separation anxiety, r=.37, p<.01). The observer anxiety scale correlated more highly with the child BPI reports on the overanxious subscale (r=.43, p<.01) than with parent reports on the HBQ overanxious subscale (r=.31, p<.05). The correlations among the parent, child, and observer reports were not significantly different between the ASC and comparison groups. In the ASC versus non-ASC cotwin (i.e., discordant pair) comparisons (N=24 pairs), ASC twins reported slightly more depressive and anxious symptoms than their cotwins; however, these differences (ds<=30) were not significant. The parent reports did not indicate significant differences in depressive symptoms or separation anxiety, but did report higher overanxious symptoms in the ASC cotwins (d = .59, p < .05). The blind observer ratings of anxiety showed the smallest ASC versus non-ASC difference, with a non-significant difference of d=21. The ASC group's self-reports of depressive, separation anxiety, and overanxious symptoms were not significantly different from their matched comparison group (N=48, all non-ASC). In fact, the most deviant group appeared to be the non-ASC cotwins of ASC individuals, who self-reported less separation anxiety and less depression than their matched comparison group. Conclusions: Including the often-ignored perspective of ASC children in research appears justified. ASC children may exhibit slightly more anxious and depressive symptoms than their non-ASC cotwins, but the difference is small compared with matched comparison groups. Parents may over-report anxious and depressive symptoms in their ASC children, and non-ASC cotwins may under-report such symptoms.