There are a number of widely used assessment measures for child and adolescent eating disorders, most of which have been adapted from instruments designed for use with adult populations. These measures include self-report questionnaires, such as the Children’s Eating Attitudes Test (ChEAT; Maloney, McGuire, & Daniels, 1988), the Children’s Eating Disorders Inventory (EDI-C; Garner, 1991a) and the Eating Disorder Examination Questionnaire (EDE-Q; Fairburn & Beglin, 1994); semi-structured interviews, such as the Children’s Eating Disorder Examination (ChEDE; Bryant-Waugh et al., 1996); and online measures, such as the Development and Well-Being Assessment (DAWBA; Goodman et al., 2000; Moya et al., 2005). These measures are described in more detail below, and see Table 1.
The Children’s Eating Attitudes Test
The Eating Attitudes Test (EAT; Garner & Garfinkel, 1979; Garner et al., 1982) is a self-report screening questionnaire, designed to detect eating disorder psychopathology in adults. The most commonly used version of the measure (the EAT-26; Garner et al., 1982) has 26 individual items that can be used to calculate a total score and three subscale scores (Dieting, Bulimia, and Oral Control). The EAT was developed as a screening tool and has been used in research studies; it has been found to reliably identify potential cases of eating disorders in non-clinical populations (Wood et al., 1992). Normal EAT scores for adolescent females in the UK have been determined, and levels of reliability of the subscales have been shown to be similar to those found in adult populations (Wood et al., 1992). The adult version of the measure has been used with older adolescents but is not suitable for use with younger children (Maloney et al., 1988).
A simplified version of the questionnaire has been developed specifically for use with young people aged 8-13 years (the ChEAT; Maloney et al., 1988). The basic psychometric properties of the ChEAT are similar to those of the adult EAT, making it a potentially valuable measure for capturing children’s eating behaviours and attitudes (Nathan & Allison, 1998). The usefulness of the EAT, however, is limited by high rates of false positives and false negatives (Wood et al., 1992), and its inability to distinguish between cases of anorexia nervosa and bulimia nervosa (Nathan & Allison, 1998), making it unsuitable for use as a diagnostic tool in clinical settings. In fact the ChEAT is mainly used in general population settings.
The Children’s Eating Disorder Inventory
The Eating Disorder Inventory (EDI; Garner, Olmstead, & Polivy, 1983) is a self-report questionnaire that measures eating disorder psychopathology and has psychometric properties that make it suitable for use as a screening instrument for eating disorders in non-clinical populations (Pike, 2005). There are three versions of the measure, the most commonly used to date being the EDI-2 (Garner, 1991b). This version consists of 91 items, mapping onto 11 scales: Body Dissatisfaction, Bulimia, Asceticism, Drive for Thinness, Ineffectiveness, Social Insecurity, Interpersonal Distrust, Perfectionism, Interoceptive Awareness, Impulse Regulation, and Maturity Fears.
The children’s version of the EDI (the EDI-C; Garner, 1991a) was developed by modifying the wording of some of the items of the EDI-2, making it more accessible for younger respondents. While some authors have found the EDI-C to have acceptable psychometric properties, similar to those of the EDI-2 (Thurfjell et al., 2003), research with non-clinical samples has cast doubt over the reliability of some of the subscales when the measure is used with young people and some authors have suggested that a different structure to the original EDI-2 might be more appropriate for children and adolescents (Eklund, Paavonen, & Almqvist, 2005). Like the EAT, the EDI can be useful as a screening instrument and for gathering information about eating behaviours and attitudes, but it is not a diagnostic tool (Pike, 2005).
The Children’s Eating Disorder Examination
The Eating Disorder Examination (EDE; Fairburn & Cooper, 1993) is a standardised, semi-structured, investigator-led interview that is widely used for the assessment of eating disorder psychopathology in adult populations. Frequently referred to as the ‘gold standard’ eating disorder assessment measure (e.g. Wilson, 1993), the EDE has been shown to have good psychometric properties when used with adults (Fairburn & Cooper, 1993; Rizvi et al., 2000; Cooper, Cooper, & Fairburn, 1989; Rosen et al., 1990).
The individual items of the EDE interview schedule assess the frequency and severity of eating disordered attitudes and behaviours over the 4 weeks prior to the interview taking place. Scores on these items can be used to generate subscale scores (the four subscales are Restraint, Eating Concern, Weight Concern, and Shape Concern) and a global score. The interview is also used to gather information about episodes of binge eating, self-induced vomiting, laxative misuse, diuretic misuse, and physical exercise used as a means of controlling weight or body shape. Diagnostic items in the schedule refer to thoughts, behaviours and physical symptoms during the 12 weeks prior to the interview. Responses to these items can be used to generate eating disorder diagnoses based on DSM-IV criteria.
The EDE has been adapted specifically for use with children and adolescents. This version of the interview (the ChEDE; Bryant-Waugh et al., 1996) has the same subscales, global score and diagnostic system as the adult version, but differs slightly in the way the questions are posed and the way in which some items are rated. For example, the rating takes into account the intention to behave in a certain way rather than the actual carrying out of the behaviour (Bryant-Waugh et al., 1996).
Findings from studies that have assessed the performance of the ChEDE in clinical samples have been mixed. The basic psychometric properties of the ChEDE were found to be good in a preliminary investigation by Watkins et al. (2005). However, Bryant-Waugh et al. (1996) found that 4 out of 11 participants (36.4%) with a clinical diagnosis of anorexia nervosa scored below the cut-off for clinical significance on key diagnostic items of the ChEDE. Other studies that assessed clinical samples using the ChEDE found that adolescents with anorexia nervosa received lower scores during the interview than adults with anorexia nervosa and adults and adolescents with bulimia nervosa (Binford, le Grange, & Jellar, 2005; Couturier & Lock, 2006). House et al. (2008) found that the ChEDE failed to identify an eating disorder in 35% of adolescents who were given a clinical diagnosis of either anorexia nervosa or EDNOS.
Couturier and Lock (2006b) modified the EDE in an attempt to make it better suited to the assessment of young people with anorexia nervosa. Use of this modified version did result in higher scores, but some participants denied behaviours that parents or clinicians reported were definitely occurring, and scores were still not as high as those seen in adult populations. Taken together, these findings imply that the ChEDE may have limitations when assessing young people with anorexia nervosa or related EDNOS (House et al., 2008).
In light of this, and in light of the length of time needed to complete the interview (approximately 60 minutes), alternatives to the standard ChEDE have been investigated for the assessment of children and adolescents. Couturier et al. (2007) developed parental and clinician interviews, to be completed in addition to the standard schedule. When they trialled these interviews they found that young people with bulimia nervosa reported more eating disorder symptoms than their parents, while young people with anorexia nervosa and related disorders received significantly lower subscale scores than their parents and clinicians (Couturier et al., 2007). While conducting multiple interviews is a potentially useful way of gathering as much information as possible about eating disorder symptoms, it is time consuming, requiring an hour of interviewing time per informant. In busy clinical settings this may not be a viable option.
A self-report questionnaire version of the EDE (the EDE-Q; Fairburn & Beglin, 1994) has been developed and a number of studies have aimed to determine the psychometric properties of this measure and to assess how useful it might be as an alternative to conducting the EDE interview (Olmstead et al., 2007). The EDE-Q is similar in structure to the EDE, allowing comparisons to be made between the two measures (Fairburn & Beglin, 1994), but the EDE-Q has the advantage of taking less than 15 minutes to complete (Fairburn & Beglin, 1994). A number of studies have compared the two measures with clinical and community samples of adults and have found discrepancies between subscale scores and reports of binge eating (e.g. Fairburn & Beglin, 1994; Black & Wilson, 1996).
Passi, Bryson and Lock (2003) conducted the first study to directly compare the EDE and the EDE-Q in a group of adolescent females with a clinical diagnosis of anorexia nervosa. They found that scores on the four subscales were consistently higher on the EDE-Q than on the EDE interview and that while all subscales were highly correlated across the two measures, scores on three of the four subscales differed significantly from each other across the measures. Binford et al. (2005) compared the EDE and the EDE-Q in a sample of adolescents with bulimia nervosa, partial-syndrome bulimia nervosa and anorexia nervosa. They found the greatest discrepancy between scores on the measures in the bulimia nervosa group.
The authors of these studies suggest that discrepancies between the measures may in part be due to informants being asked to complete the EDE-Q without the detailed instructions that are available to an EDE interviewer. When informants are given additional information prior to completing the EDE-Q, differences between EDE-Q and EDE scores are reduced (Passi et al., 2003; Carter, Aime, & Mills, 2001). Following direct comparisons of the EDE and the EDE-Q, most authors have concluded the EDE interview is superior for the assessment of clinical populations of adolescents with eating disorders. While the EDE-Q can be a useful screening tool and measure of eating disorder psychopathology, unlike the EDE it is not a diagnostic measure.