The Picture Anxiety Test (PAT): A new pictorial assessment of anxiety symptoms in young children

Abstract

This study examines the psychometric properties of a newly developed pictorial anxiety assessment, the Picture Anxiety Test (PAT), designed for young children. A sample of 71 5–7-year-old children (clinically anxious and healthy controls) participated in an assessment comparing the PAT to other, established anxiety measures. Children's ratings were also compared to parent ratings of internalizing and externalizing problems. The PAT demonstrated favorable psychometric properties in this sample, including moderate to high internal consistency, high convergent validity with related constructs, and high discriminant validity between diagnostic groups and across pre- and post-treatment. Initial findings suggest that the PAT is a psychometrically sound instrument for assessing child anxiety and avoidance in young children and a valuable addition to existing diagnostic instruments for clinicians working with anxious children.

Introduction

Anxiety disorders are among the most common emotional disorders of childhood and adolescence (Bernstein et al., 1996, Cartwright-Hatton et al., 2006). Although some anxieties and fears are developmentally appropriate (Muris, 2007), severe anxiety can be debilitating for some children, interfering with normal functioning, persisting over time, and continuing into adolescence and adulthood if left untreated (Costello et al., 2003, Keller et al., 1992, Last et al., 1996, Muris et al., 2003). Moreover, childhood anxiety disorders are associated with psychosocial and school problems, increased risk for other externalizing disorders, adult anxiety disorders, depression, and substance abuse (Brückl et al., 2007, Klein, 1994, Strauss et al., 1988).

Although anxiety disorders have the earliest median age of onset of all mental disorders (Kessler et al., 2005, Kim-Cohen et al., 2003), and in spite of the known impact of childhood anxiety (Field, Cartwright-Hatton, Reynolds, & Creswell, 2008), research on anxiety in younger populations lags behind developments in the adult field. Despite an increasing emphasis on early detection, there are only few valid and reliable measurements to assess anxiety symptoms in preschool children. The most prominent and valid method of assessment in clinical child psychology is the structured diagnostic interview (Ollendick and Hersen, 1993, Silverman, 1994). However, these interviews are typically conducted only with the parents if the child is less than 8 years. The format of “classical” structured interviews with “pure” questions and without any child-friendly material reduces compliance in the child and may adversely influence the validity of the child's answers. Another method for identifying emotional disturbances in childhood is through rating scales such as the commonly used Revised Children's Manifest Anxiety Scale (RCMAS; Reynolds & Richmond, 1978) and the Spence Children's Anxiety Scale (SCAS; Spence, 1998). Additionally, two new parent-report scales for measuring anxiety problems in 2–6 years old children have recently been developed. First results showed that both the Preschool Anxiety Scale (PAS; Spence, Rapee, McDonald, & Ingram, 2001) and the Children's Moods Fears and Worries Questionnaire (CMFWQ; Bayer, Sanson, & Hemphill, 2006) yielded moderate to good psychometric properties (Broeren & Muris, 2008).

Researchers generally agree that assessment of anxiety in children should be multi-informant (child, parent, and teacher) and multi-method (i.e., relying on more than one data gathering procedure) (Cartwright-Hatton et al., 2006, Ollendick and March, 2004, Silverman and Carmichael, 1999). The often modest parent–child agreement in clinical assessment (Achenbach and Edelbrock, 1989, Cartwright-Hatton et al., 2006) especially underscores the importance of gathering information from multiple informants. Specifically, parents generally tend to underreport and underestimate the number and severity of children's internalizing symptoms, especially anxiety (Cartwright-Hatton et al., 2006, Stallings and March, 1995). On the other hand, parents with anxiety disorders themselves can tend to overreport anxiety symptoms in their children (Barbosa et al., 2002, Comer and Kendall, 2004).

For effective assessment of the child perspective, child measurements should be age-appropriate and tailored to match young children's cognitive and verbal skills (King et al., 2004, Ollendick et al., 2004). Structured interviews such as the ADIS-C for DSM-IV (Albano & Silverman, 1996) can be difficult for children below the age of 8 years to follow. A few authors successfully overcame some of the inherent limitations associated with standard anxiety interviews and questionnaires for children by using pictures in their measurements (Ernst et al., 2000, Ernst et al., 1994, Muris et al., 2003, Valla et al., 1994, Valla et al., 2000). According to these authors, pictures are well suited for the assessment of internalizing symptoms of young children, as the pictorial format focuses children's attention, stimulates their interest, and improves their participation (Ernst et al., 1994, Valla et al., 1994, Valla et al., 2000). In addition, as human faces are well suited to express emotions (Siegler, DeLoache, & Eisenberg, 2003), and research has shown that children aged 4 years and above are able to identify and discriminate between different emotions (Eisenberg Murphy, & Shepard, 1997), the use of pictures depicting children for assessing anxiety in boys and girls is effective and developmentally appropriate. An extensive literature review revealed three standardized self-report instruments that use the pictorial format for assessing anxiety in children below the age of 7 years: The Pictorial Instrument for Children and Adolescents (PICA-III-R) is a semi-structured interview with pictures covering five diagnostic categories for children and adolescents aged 6–16 years (Ernst et al., 1994). The Dominic questionnaire is a pictorial interview-based scale to assess mental disorders in children aged 6–11 years (Valla et al., 1994). Finally, the Koala Fear Questionnaire (KFQ) is a standardized self-report scale for assessing fears and fearfulness in children between 4 and 12 years (Muris et al., 2003). However, only the KFQ is designed for children below the age of 6 years, and both the Dominic questionnaire and the PICA-III-R are too lengthy to be conducted with children below the age of 6 years. In addition, pictures of these measurements are in black and white (KFQ), or information about this aspect is missing (PICA-III-R, Dominic). The characteristics of these instruments are shown in Table 1.

Given that fear and anxiety symptoms can develop at a very early age, a reliable and valid standardized instrument for assessing internalizing problems in children below the age of 8 years would be a valuable addition to the existing measurements of researchers and clinicians working in the field of childhood anxiety. Based on the strength of pictorial anxiety instruments and for clinical and epidemiological purposes, we developed an anxiety test for children between 4 and 8 years, using a graphic approach. The Pictorial Anxiety Test (PAT) is based on core criteria of the DSM-IV-TR (American Psychiatric Association, 2000), and assesses the tendency toward anxiety and avoidance in specific situations using hypothetical vignettes. As it is difficult for children below the age of 11 to precisely report on the duration and onset of anxiety symptoms, these aspects are omitted from the PAT. Instead, the PAT assesses a wide range of anxieties and fears on a dimensional level, and is supplemental to the structured interview conducted with the parents of the children (e.g., ADIS-C/P for DSM-IV, Albano & Silverman, 1996).

The present study describes development and content of the PAT and examines its psychometric properties in a clinical sample and in a non-clinical control group. We first assessed internal consistency and test–retest reliability. We then predicted that the PAT would demonstrate construct validity by correlating more strongly with other self- and parent-report anxiety measures than with measures of externalizing difficulties. We further predicted that the PAT would show good discriminant validity in that it would be able to discriminate meaningfully between children with and without anxiety disorders and that there would be no differences between the children with different anxiety disorders. Finally, we predicted that the PAT would show utility in assessing the effects of a treatment for anxiety.