Differential Perceptions of Clinical Anxiety Among Clinicians and Parents in Children with Autism Spectrum Disorder

Background: Recent data using parent-report measures indicate that anxiety is highly prevalent in children with autism spectrum disorders (ASD) (see review by White et al., 2009). However, no studies have examined levels of agreement between parent report measures and clinician diagnosis regarding the presence of clinical anxiety. This topic has important implications regarding the use of parent-reported anxiety measures in clinical practice, as well as conceptualizing anxiety in the ASD literature.

Objectives: To examine the: a) degree of concordance and discordance between parent and clinician assessments of anxiety in children with ASD; and b) child and parent characteristics associated with clinician-parent discordance.

Methods:  Data from 2,262 children (M = 5.77 years, SD=3.48) were obtained from the Autism Treatment Network (ATN) registry. The ATN is a multi-site collaboration among 17 autism centers in the US and Canada that focuses on addressing the medical needs of children with ASD. The ATN enrolls children, ages 2 to 17.9 years, with a diagnosis of ASD as confirmed by the Autism Diagnostic Observation Schedule (ADOS; Lord et al, 2002) and DSM-IV-TR criteria. Clinicians completed customized forms regarding the child’s current psychiatric diagnoses and treatments. Parent report of anxiety was measured using the Child Behavior Checklist Anxiety Problems Scale (CBCL; Achenbach and Rescorla, 2001). Concordance was present when the child received an anxiety disorder by the clinician and the parent reported clinical anxiety. Since only 5% of children had a clinician diagnosis of anxiety disorder in the absence of parent reported clinical anxiety, discordance in this study focused on the absence of a clinician diagnosis accompanied by parent reported clinical anxiety. Simple and multivariate logistic regression analyses examined the child and parent characteristics associated with this discordance.

Results:  Clinicians diagnosed 224 (9.9%) children with an anxiety disorder, whereas parents reported clinical anxiety in 566 (25%) children. Of these 566 children, clinicians provided a diagnosis for 120 (21%) children, resulting in a discordance of 446 (79%). Results from the bivariate regressions indicated that discordance was more likely if the child was younger, non-Caucasian, had lower cognitive, language, and adaptive functioning, greater ASD severity, increased externalizing behaviors, and lower caregiver education. Multivariate analyses indicated that younger age, greater language impairment, and lower caregiver education were the strongest predictors of discordance (all p < 0.05).

Conclusions:  Concordance between parents and clinicians regarding the presence of anxiety in the child with ASD was low. Discordance was more likely if the child was younger, had significant language impairments, and their caregiver was less educated. These data underscore the urgent need for developing a unified construct of anxiety for certain profiles of children with ASD, such as those who are lower functioning. Validation studies of parent-report measures of anxiety in children with ASD are also needed.

We acknowledge the members of the Autism Treatment Network (ATN) for use of the data and the families who participated in the Registry. The ATN is funded by Autism Speaks and a cooperative agreement (UA3 MC 11054) from HRSA to the Massachusetts General Hospital.