Parent and child perspectives on the nature of anxiety in children and young people with autism spectrum disorders: a focus group study

Phenomenology of anxiety disorders in childhood

Epidemiological studies show that typically developing children and adolescents commonly experience anxiety disorders (Muris and Broeren, 2009), and these may be the most common psychological disorder of childhood (Cartwright-Hatton et al., 2006). Anxiety disorders in childhood frequently persist into adolescence and adulthood, often occur with other disorders and can have a negative impact on functioning in areas such as academic and social performance (Rapee, et al, 2009).

Anxiety itself may be best conceptualized on a continuum (Schniering, et al 2000), with many children experiencing transient worry, stress or anxiety. At times anxiety can enhance performance or offer protection from threat, and furthermore, at some developmental stages, certain types of anxiety are so common that they may be seen as normative; for instance, separation anxiety around the ages of 12–18 months. However, anxiety disorders can be differentiated from ‘normal’ fears and worry on the basis of severity of symptoms, persistence and impact on functioning (Beesdo et al., 2009). Current estimates suggest that prevalence of all anxiety disorders is between 2.5% and 5%, affecting twice as many girls as boys (Rapee et al., 2009) with cumulative prevalence rates indicating that 9.9% of children and adolescents under 16 are likely to experience an anxiety disorder (Costello, et al 2003).

Using a cognitive behavioural approach to anxiety in childhood, the clinical features of anxiety may be conceptualized as physiological, cognitive, emotional and behavioural. Physiological changes include somatic complaints, such as nausea, headaches, sweating and muscle tension. The emotional component encompasses feelings of fear, worry, dread or irritability. Cognitive symptoms often reflect catastrophic predictions and expectations of failure to cope. Finally, the predominant behavioural symptom of anxiety is avoidance, although hypervigilance and checking are also common.

Anxiety in ASD

There are now several well-conducted prevalence studies showing that anxiety occurs in children with ASD at a higher rate than in the typically developing population. Although prevalence rates vary from 11% to 84% (White et al., 2009), most studies indicate that around half of children with ASD meet criteria for at least one anxiety disorder (e.g. Simonoff et al., 2008). Knowledge is also emerging of the relationship between anxiety and ASD-specific factors, including impoverished empathy and social skills (Bellini, 2004), theory of mind and central coherence difficulties (Burnette et al., 2005) and sensory over-sensitivity and responsivity (Green and Ben-Sasson, 2010). However, studies examining the phenomenology of anxiety in ASD have generally used measures that were standardized with typically developing children. They may not, therefore, adequately reflect the specific ways in which anxiety may manifest itself in children with ASD. Furthermore, the way that anxious symptoms in ASD are measured and described varies significantly between studies (White et al., 2009), further limiting our understanding of anxiety in this group. Anecdotally, many clinicians recognize that the triggers and presentation of anxiety in children with ASD can be markedly different from those of the typically developing population and may not be captured by existing standardized measures. For example, anxieties associated with changes in routine do not fit easily into any particular category using current diagnostic criteria. There can also be difficulties differentiating between anxiety symptoms and other problems relating to regulation of emotion, particularly anger or low mood; or, indeed, autism symptoms themselves (Muris et al., 1998).

Despite emerging knowledge, little is known about the atypical way that anxiety may present in children and young people with ASD. The difference in presentation of depression between ASD and non-ASD populations has been discussed (e.g. infrequent reporting of typical depression symptoms, such as feelings of worthlessness and guilt, increased aggressive behaviour or self-injurious behaviour; Stewart et al., 2006). To date, a similar investigation on the nature and presentation of anxiety in ASD has not been published, to our knowledge.

The above limitations notwithstanding, there is an emerging literature about the use of cognitive behaviour therapy (CBT), the treatment of choice for typically developing anxious young people, with individuals with ASD (Lang et al., 2010; Moree and Davis, 2010). Recent randomized controlled trials suggest that enhanced and adapted CBT can successfully treat anxiety disorders in high functioning children with ASD (e.g. Wood et al., 2009), although it is not currently clear which adaptations are crucial or which modifications are required for which anxiety disorder (Moree and Davis, 2010). Several issues limit the development of effective interventions. One problem relates to the lack of appropriate measures developed and standardized for children with ASD, which has an impact on the assessment and interpretation of change and outcome in efficacy studies. In addition, the lack of available information about the atypical presentation of anxiety in ASD renders it difficult to implement the AACAP Practice Parameter recommendation (Connolly and Bernstein, 2007) that anxiety treatments are matched to child and family characteristics. Understanding more about the factors triggering and maintaining anxiety in ASD, and how children with ASD and their families experience, think and talk about anxiety, could inform the development and implementation of autism-specific interventions for anxiety for this population.

Therefore, despite many promising advances in the literature, our understanding of anxiety in children with ASD remains relatively limited. Given the limitations of quantitative methodologies using standardized measures developed for typically developing children largely used so far in anxiety research in ASD, the present study uses a qualitative, ‘bottom up’ approach to furthering our understanding of anxiety in this population. We report on data obtained from focus groups with parents and children with ASD as a first step in identifying the triggers, behavioural signs and cognitive processes associated with anxiety in this group through parental perspectives, with the aim of increasing our understanding of the presentation of anxiety in children with ASD.

Participants

Participants were recruited via the National Autistic Society and parent support groups in Berkshire, London and Surrey, UK, via email newsletters sent to all members of those groups. An invitation letter accompanied the information sheets, inviting parents to attend a focus group specifically discussing anxiety in ASD to find out more about the causes and processes of anxiety in this group. Parents were invited to participate if their children had ASD, had any experience of worry or anxiety and were aged 7–18 years. No incentives were provided for participating. No exclusion criteria were set with regard to children’s functioning, comorbid diagnoses or whether they had received treatment for anxiety in the past, as this was an exploratory study and obtaining rich and varied perspectives on the experience of anxiety was prioritized.

Five parent groups in four localities were conducted in total. Altogether, 17 mothers attended, who were parents of 19 children with a diagnosis of ASD; 7 girls (age range 7–18, mean age: 11) and 12 boys (age range 7–16, mean age: 12). Children had received ASD diagnoses from local clinical services: five had a diagnosis of Asperger Syndrome, six had a diagnosis of high functioning autism, one had a diagnosis of atypical autism and seven had a diagnosis of ASD or autism. All parents reported that their children experienced significant levels of anxiety that had an impact on their functioning. Six had received cognitive behavioural or behavioural treatment for anxiety at some stage.

All the children of participating parents were invited to participate in children’s focus groups; however, few accepted. Four boys aged 10–12 years with Asperger’s syndrome and high functioning autism accepted, and one focus group was held to explore the children’s own views. Findings from the children’s focus group are reported in less detail owing to the small numbers recruited.

Topic guide

The topic guide took into account the literature relating to the prevalence, phenomenology and treatment of anxiety in ASD, cognitive behavioural models and treatment of anxiety in typically developing children and the clinical experience of the professionals leading the groups (Appendix 1). Before the research groups, two pilot groups were held with parents from a local parent support group to ensure topics were comprehensive and relevant and to develop appropriate questions for use in the main groups. Once finalized, the same topic guide, adapted to be age-appropriate, was used for the children’s group.

Procedure

Mothers met in groups of 2–4 for about 2 hours. Groups were facilitated by two of the authors (AO and FK), chartered clinical psychologists experienced in working with children and adolescents with ASD. The facilitators outlined the purpose of the group and asked mothers to introduce themselves and talk a little about their children, before using the topic guide to generate discussion. Care was taken not to ask leading questions, and to let discussions develop freely with as little interference from the facilitators as possible, beyond ensuring that all areas were covered. However, at times it was necessary to clarify issues (e.g. when parents were describing angry/oppositional behaviour rather than anxious behaviour). It was acknowledged that it was not always easy to distinguish between the two; the facilitators regularly clarified this by defining anxiety in lay terms as particularly pertaining to worry or stress, rather than aversion or oppositionality. For example, when parents mentioned repetitive behaviours, facilitators clarified whether this worsened at times of anxiety. Facilitators also ensured that the discussions were not monopolized by one member of the group only and made efforts to give all members opportunities to contribute and express their views and experiences. The children’s group met at the same time as the parents’ and followed a similar topic guide and procedure, albeit with more breaks for snacks and games. The sessions were audio recorded using digital voice recorders and later transcribed verbatim.

Ethical considerations

Ethical approval was obtained from King’s College London and the University of Reading Research Ethics Committees. The groups had the potential to raise upsetting memories for participants. The facilitators ensured that boundaries for ‘safe’ discussion and confidentiality were set and maintained throughout the sessions. Information was given to parents about where to seek further support if required.

Data analysis

Following focus group methodology in analyses (Kitzinger, 1995; Krueger and Casey, 2000), transcripts were analysed using thematic analysis (Aronson, 1994). They were initially scrutinized to identify data that related to predefined categories (triggers, behavioural, somatic and cognitive signs). Relevant data from all transcripts were then read and re-read to identify patterns and sub-themes within each predefined category, using an iterative process resulting in the most parsimonious number of sub-themes, yet ensuring that all material could be sub-classified. Once the coding scheme had been established, one complete transcript was double-coded by both researchers to ensure consistency of approach. In accordance with the focus group methodological principle of ‘words not numbers’ (Grudens-Schuck et al., 2004; Fern, 2001; Kitzinger, 1995), transformations of qualitative information into quantitative data were not carried out, as in a group discussion there is difficulty recording every manner of interaction, such as nodding or other communications of agreement or disagreement.

Parents’ groups

Presentation of anxiety

Perceived differences between anxiety in children with ASD and typically developing children

As many parents also had children without ASD, we also enquired whether they perceived any differences between anxiety in their children with and without ASD. Despite some potential limitations of asking parents this question (i.e. they may have limited experiences of typically developing children and/or their other children may be at higher risk for a diagnosis of ASD or an ASD broader phenotype), we were nevertheless interested in finding out what parents perceived as differences, and indeed clear themes emerged. With regards to presentation of anxiety, the behavioural rather than verbal expression of anxiety, and the time delay between an event and expression of anxiety were consistently perceived by the parents in all groups as key differences. Parents reported that anxiety in their children with ASD was typically prolonged and more intense and that their children were more difficult to reassure or soothe. Some parents described their children with ASD as being constantly on the edge of anxiety, or permanently in ‘fight or flight’ mode: ‘He’s never actually happy – it’s [anxiety] his default setting.’ Another theme was that many things usually considered pleasurable (e.g. Christmas) are actually anxiety-provoking for children with ASD owing to their difficulties in coping with change or dislike of surprises.

In terms of triggers, social worries and anxiety in relation to change were identified as different in their children with ASD compared with their siblings or other children without ASD: ‘Social anxiety is inherent, which leads to awkward social situations’; ‘The difficulties with change are different’.

Impact of anxiety

The impact of anxiety on families and the children themselves became particularly apparent when behavioural indicators of anxiety were discussed. Indeed, some parents expressed a view that anxiety often had a more of a negative impact than the autism itself.

Children’s group

Much of what the four boys said in their focus group mirrored the mothers’ comments. When asked ‘what makes you anxious?’ many comments pertained to change in routine, sensory triggers (such as teachers shouting), and social worries (such as being at a friend’s house without their parents and not knowing what to say). Worries around being bullied also featured frequently.

The children were able to identify behavioural signs of anxiety, reporting that they experienced somatic symptoms such as a racing heart, out of control breathing, butterflies, wobbly legs, feelings of nausea and sweating. They also identified some thought processes such as ‘thoughts racing out of control’ and ‘my mind blacking out’; one child referred to ‘the voices in my head’. Identified cognitions included: ‘I’m going to get hurt’; ‘They might insult me’; and ‘Will they like me?’ However, few clearly articulated cognitions were identified and, as a rule, the boys had considerable difficulty explaining why certain situations triggered anxiety: ‘Just felt weird and strange, it’s unlogical’; ‘It was change, it just won’t be the same’.

Limitations

The findings presented here need to be interpreted in light of the study’s limitations. One important limitation is the small sample size and representativeness of the participants, who were a self-selected group from a relatively small geographical area. The lack of an incentive may have resulted in a particularly motivated group. Another limitation was the lack of detailed information about demographics, language skills, cognitive functioning or adaptive skills of the children involved, which was beyond the scope of this exploratory study. However, it is likely that the findings of this study can be tentatively generalized to children in the higher functioning end of the ASD spectrum. A general limitation of focus group methodology is that discussion can be dominated or sidetracked by individuals, or may not be representative of other groups (Kitzinger, 1995; Krueger and Casey, 2000), and therefore we have made an effort to present and focus on themes identified across all the focus groups. At the same time, the interaction of group members is likely to result in richer conversations and wider range of themes discussed (Grudens-Schuck et al., 2004). Another limitation was the challenge of ensuring that parents were at all times referring to anxiety and not non-specific negative affect, such as aversion or discomfort. The facilitators made efforts to ensure that all comments included in this paper specifically pertain to anxiety, but recognize that, owing to the limited ability of their children to express their feelings, parents may at times have been making assumptions about anxiety triggers or anxious behaviours. Thus, although we made every effort to clarify parents’ comments by reiterating definitions of anxiety when necessary, we are aware that parent examples may at times have represented non-anxiety-specific symptoms and behaviours.

Finally, the number of young people with ASD who participated was too small to further explore similarities or differences between parent and child reports of anxiety or differences in anxiety as a function of child characteristics with any confidence.

Theoretical and clinical implications of the focus group findings

The present findings provide further insight into the appropriateness of using standardized measures consisting of predefined symptoms or manifestations of anxiety, developed for the general population, with children with ASD. Although existing measures are useful for comparison purposes, our study shows that when parents and children with ASD are invited to describe their experiences in a semi-structured, non-directed manner, they largely focus on features more specific to the ASD population that are not typically measured on most anxiety measures. Although there are measures that do evaluate broad emotional difficulties in children with developmental disabilities (e.g. the Developmental Behaviour Checklist, DBC; Einfeld and Tonge, 1992), a significant gap remains in the resources available to clinicians and researchers for measuring ASD-specific aspects of anxiety, now well demonstrated to be an extremely common presenting clinical problem in ASD. There are also implications for the inclusion of observational methods in the assessment of anxiety: using multiple informants, multimodal assessment techniques and methods has been recommended (MacNeil et al., 2009). These findings may also be useful for clinicians working with children and young people with ASD, particularly those attempting to engage children with ASD in CBT. Although there is now good evidence that, with adaptations, CBT can be an effective means of treating anxiety in young people with ASD (e.g. Chalfant et al., 2006), clinicians may need to consider whether CBT is the most appropriate intervention, or whether other approaches such as behavioural work through parents or environmental modifications may be more appropriate, as even for verbally able children, accessing or modifying cognitions may prove too challenging. However, clearly CBT is an effective intervention for many children with ASD, and the involvement of parents can further improve outcomes (Moree and Davis, 2010). Our findings support the view that adaptations to CBT, such as adding ASD-specific components to the treatment protocol, are important, relevant and acceptable to parents of children with ASD (Sze and Wood, 2008).

Attempts at developing a theoretical model of anxiety in individuals with ASD are still very much at the stage of theorizing, but two have been proposed to date. Gaus (2007) proposed a model developed for the adult population that includes vulnerability to depression as well as anxiety and highlights social cognition deficits, information processing difficulties, sensory dysfunction and increased risk for negative life events (rejection and bullying) leading to anxiety and depression. Wood and Gadow (2010) more recently proposed a theoretical model of clinical anxiety in ASD that included ASD-related stressors (social confusion, peer rejection, prevention of repetitive behaviours and aversive sensory experiences) leading to two potential mood dysregulation pathways (social anxiety and other anxiety or depression types), which in turn result in increased social avoidance, increased severity of ASD behaviours, increased behavioural difficulties and reduced quality of life. The themes identified by the parents in our study (Figure 1) very closely map onto the factors put forward by Wood and Gadow (2010), and provide some initial support for the face and external validity of this proposed model and its applicability to the real-life experiences of anxious young people with ASD and their families. Wood and Gadow also highlight the importance of integrating these factors in any attempt to systematically understand how anxiety is developed and maintained in this population.

OPEN IN VIEWER

Figure 1.

A summary of key themes identified by parents in relation to triggers, symptoms and impact of anxiety in their children with ASD mapped in relation to the Wood and Gadow (2010) hypothetical model of anxiety in ASD (potentially ASD-specific themes are identified in oval shapes).