Parent–child agreement of anxiety symptoms in youth with autism spectrum disorders

Abstract

Best practice for assessment of anxiety symptoms in children suggests that child self-report is an important element to consider. Yet, it is not known if it is a reliable assessment method for children with high-functioning autism spectrum disorders (ASDs). The present study examines the extent to which verbally fluent children with ASD and their parents agree on the child's symptoms of anxiety. Sixty-three children with ASD and their parents independently completed the Screen for Child Anxiety Related Emotional Disorders (SCARED). Intraclass correlations indicated moderate-to-strong agreement on several domains. Associations among child factors (e.g., verbal IQ, metacognitive ability) and parent–child agreement were explored. More advanced verbal ability was associated with better agreement on Separation, School Avoidance, and Total Anxiety; higher metacognitive skills were associated with better agreement on Social Anxiety. Implications for the use of child self-report are discussed.

Introduction

Anxiety disorders are considered to be among the most common psychiatric disorders in childhood (Compton et al., 2004, Dadds and Barrett, 2001, Donovan and Spence, 2000). Epidemiological studies indicate prevalence rates of 5.7–17.7% (Costello & Angold, 1995). The assessment of anxiety disorders in children with autism spectrum disorders (ASDs) is challenging and controversial, given debate in the field regarding which symptoms are core to ASD and which contribute to a unique symptom presentation associated with a co-occurring anxiety disorder (White, Oswald, Ollendick, & Scahill, 2009). This dilemma has also been addressed within the context of “diagnostic overshadowing,” the process by which symptoms are attributed to the primary diagnosis of a developmental disability rather than to a co-occurring psychiatric condition (Reiss & Szyszko, 1983).

Recent research has indicated that children with ASD frequently do exhibit clinically significant symptoms of anxiety (Ghaziuddin, 2005, Simonoff et al., 2008, Tantam, 2000). The prevalence rates of anxiety disorders in individuals with ASD are significantly higher than in the general pediatric population (Gillott et al., 2001, Kim et al., 2000). In a recent epidemiological study of 112 children with ASD, 41.9% of the sample met diagnostic criteria for an anxiety disorder with the following rates of comorbidity: Specific Phobia, 8.5%; Obsessive Compulsive Disorder (OCD), 8.2%; Generalized Anxiety Disorder (GAD), 13.4%; Panic Disorder, 10.1%; and Social Anxiety, 29.2% (Simonoff et al., 2008). In a community sample, parents of 109 children with ASD reported high, but variable rates of comorbidity (Leyfer et al., 2006): Specific Phobia, 44.0%; OCD, 37.0%; GAD, 2.4%; Panic Disorder, 0%; and Social Anxiety, 7.5%. In a clinical sample of 171 children with ASD, Sukhodolsky et al. (2008) also noted high rates of anxiety with 43% of the sample meeting screening criteria for at least one anxiety disorder (full diagnostic criteria were not reviewed); namely, Specific Phobia, 31.0%; GAD, 9.4%; Panic Disorder, 5.3%; and Social Anxiety, 19.9%. Higher rates of anxiety disorders were noted by Muris, Steerneman, Merckelbach, Holdrinet, and Meesters (1998) in their clinical sample of children with ASD, with 84.1% of their sample meeting full DSM-III diagnostic criteria for at least one anxiety disorder. The Muris group noted that their sample met diagnostic criteria for Simple Phobia, 63.6%; OCD, 11.4%; Overanxious Disorder, 9.4%; Panic Disorder, 9.1%: and Social Phobia, 63.6%. While high but variable rates of anxiety disorders have been observed across these studies, the challenge remains integrating the research literature, given the diversity of anxiety assessment instruments used and populations included. Nonetheless, these results suggest that anxiety disorders are prevalent in children with ASD and warrant further examination.

Given that undiagnosed or misdiagnosed anxiety disorders may contribute to prolonged symptom expression and future impairment, there has been increased interest in the assessment of anxiety symptoms in children with ASD (Merlo, Storch, Murphy, Wayne, & Gary, 2005). There have been efforts directed towards modifying existing semi-structured interviews that assess psychiatric symptoms in the general pediatric population to better reflect the common psychiatric symptoms in children with ASD (Leyfer et al., 2006). To date, rating scales that map onto the diagnostic criteria for anxiety disorders as outlined by the DSM-IV (American Psychiatric Association, 1994) have not been modified for use with children with ASD. Commonly used anxiety rating scales in the general pediatric population include the Multidimensional Anxiety Scale for Children (March, Sullivan, & James, 1999), the Revised Children's Manifest Anxiety Scale (Reynolds & Richmond, 1985), the Screen for Child Anxiety Related Emotional Disorders (SCARED; Birmaher et al., 1999), the Spence Children's Anxiety Scale (SCAS; Spence, 1997), and the State-Trait Anxiety Inventory for Children (Spielberger, 1983). Please see Silverman and Ollendick's review of evidence-based assessment measures for anxiety disorders in children for a comprehensive overview (2005).

While the aforementioned self-report measures have yielded good test–retest reliability when used in the general pediatric population, agreement between parent and child ratings of anxiety symptoms has been reported to be low (Achenbach et al., 1987, Engel et al., 1994, Kenny and Faust, 1997, Klein, 1991) to moderate (Nauta et al., 2004). Typically, parents tend to rate their children as experiencing significantly more anxiety than children rate themselves (Krain & Kendall, 2000). Agreement is not thought to differ between clinical and community samples (Seiffge-Krenke & Kollmar, 1998); however, it is thought to be enhanced when behaviors are observable (March, Parker, Sullivan, Stallings, & Connors, 1997). That is, agreement might be higher on behaviors such as proximity to parent and school refusal and weaker for some physiological symptoms (e.g., racing heart), covert internal states (e.g., worry about being negatively evaluated by peers), and symptoms evidenced in environments outside the home (e.g., school).

A number of explanations for disparities across reporters of symptoms have been raised in the general pediatric literature, including the impact of child's gender (i.e., boys demonstrated better parent–child agreement than girls; Sourander, Helstela, & Helenius, 1999), differences in behavior across settings (Achenbach et al., 1987), child concern about presenting in a socially desirable way (Dadds, Perrin, & Yule, 1998), cognitive or developmental limitations (Schniering et al., 2000, Shirk and Saiz, 1992), and biased ratings by parents due to their own psychiatric history (Chapman, Mannuzza, Klein, & Fyer, 1994). The impact of these variables on parent–child agreement is conflicting, leading De Los Reyes and Kazdin (2005) to conclude that discrepancies may be due in large part to attributional biases (e.g., parents attributing problems to the child's disposition and children attributing causes to the environment).

Clinicians struggle with the challenge of making sense of discrepant information (i.e., the “therapist's dilemma,” Hawley & Weisz, 2003), while attempting to create a treatment plan with targeted, functional goals. In fact, in Hawley and Weisz's study of 315 triads of children, parents, and therapists, more than 75% of the triads were unable to achieve consensus on the specific problem domain to be targeted in treatment. The authors posit that the disagreement in symptoms contributes to difficulties in the therapeutic alliance and may compromise the impact of treatment.

In addition to the factors examined in the general pediatric literature regarding parent–child agreement, there may be a unique set of variables that impact the inter-observer agreement when the child has ASD. Youth with ASD usually have difficulty engaging in introspection, identifying emotions, and expressing emotions (Baron-Cohen, 2002, Baron-Cohen et al., 1985, Losh and Capps, 2006, Quay and LaGreca, 1986). As a result, researchers studying the phenomenology or treatment of anxiety symptoms in ASD are reluctant to rely on child self-report as a meaningful index of symptom presentation (Bellini, 2006, Chalfant et al., 2007, Reaven et al., 2009, Sofronoff et al., 2005, Wood et al., 2009). In fact, in a recent study exploring concordance between child self report on a variety of measures and parent report on the Autism Comorbidity Interview—Present and Lifetime, results indicated little agreement and cautioned the interpretation of child self report (Mazefsky, Kao, & Oswald, 2011).

There have been very few published studies in the ASD literature that have specifically looked at parent–child agreement on anxiety ratings scales. Gillott et al. (2001) reported that parents of children with ASD provided higher ratings of their children's social worries (as measured by the Social Worries Questionnaire; Spence, 1995) than their children did, which was the opposite pattern observed in the reports of parent–child dyads, when the child was either typically developing or had a language disorder. Russell and Sofronoff (2004) noted significant differences in parent–child ratings on the Spence Children's Anxiety Scale (SCAS) in their clinical sample of children with Asperger Syndrome. A significant rater × subscale interaction was obtained [F(3,61) = 31.31, p < .001]. Post hoc comparisons indicated that parents rated their children significantly higher on separation anxiety (p < .001), social phobia (p < .001), and GAD (p < .01) than the children rated themselves, but lower on OCD (p < .001). In contrast, in a slightly older population (i.e., community sample of adolescents with Asperger Syndrome), Farrugia and Hudson (2006) reported substantial parent–teen agreement on Total Anxiety Scores on the SCAS (r = .697). Another study examining parent–teen agreement (i.e., an epidemiological study containing adolescents with ASD) used the Youth Self Report and Child Behavior Checklist (Hurtig et al., 2009). Results indicated good agreement on somatic, social, and externalizing problems and modest agreement on internalizing, anxious/depressed, and attention problems scales.

Despite conflicting data on parent–child agreement, parameters for best practice in assessment of internalizing symptoms in children encourage the use of multi-informant report, typically with mothers and children (Silverman & Ollendick, 2005). Silverman and Ollendick's general consensus is that multiple informants aid in correct diagnosis and increase the probability of children receiving needed services. It is assumed that this may also be the case for the ASD population (White et al., 2009). In addition, having children begin to identify and quantify their symptoms of anxiety may be a useful first step in treatment.

The primary aim of this study is to examine agreement on anxiety symptoms between parents and verbally fluent children with ASD, and compare these findings to published results on parent–child dyads. The measure chosen for this study is the Screen for Child Anxiety Related Emotional Disorders (SCARED; Birmaher et al., 1999), which is considered to be one of the best tools for discriminating anxiety from other disorders (Silverman & Ollendick, 2005). In addition, the SCARED was selected because it was brief (41 items), had parent and child versions containing closed-ended statements with forced choice (i.e., potentially a good fit for the ASD learning style), and had existing psychometric and agreement data. The SCARED was also selected given that the language appeared simpler and more concrete than other measures and split symptoms of panic from other disorders. A secondary aim of the study is to explore the impact of factors, such as verbal ability and metacognitive skills, on parent–child symptom agreement. Knowledge gained regarding parent–child agreement will inform the process of assessing anxiety in individuals with ASD and may guide treatment.