Eat Weight Disord DOI 10.1007/s40519-015-0191-2
ORIGINAL ARTICLE
The eating disorder examination: reliability and validity of the Italian version Simona Calugi1 • Valdo Ricca2 • Giovanni Castellini2 • Carolina Lo Sauro2 • Antonella Ruocco1 • Elisa Chignola1 • Marwan El Ghoch1 • Riccardo Dalle Grave1
Received: 18 February 2015 / Accepted: 24 March 2015 Ó Springer International Publishing Switzerland 2015
Abstract Objective To examine the psychometric characteristics of the Italian language version of the latest edition of the eating disorder examination (EDE). Methods An Italian version of the EDE (17th edition) was designed and administered to 185 in- and outpatients with eating disorders and 60 age-matched controls. Its internal consistency, inter-rater reliability, short-term (7–23 days) test–retest reliability and criterion validity were evaluated. Results Internal consistency was high for all four original EDE subscales. Inter-rater reliability was excellent for global EDE scores and original subscales (C0.93), and for eating disorder behaviours (C0.89). Test–retest reliability was good for global EDE scores and original subscales (0.57–0.80), objective bulimic episodes and days, vomiting episodes, laxative and diuretic misuse episodes, and excessive exercising (C0.82), but unsatisfactory for subjective bulimic episodes and days. Patients with eating disorders displayed significantly higher EDE scores than age-matched controls, demonstrating the good criterion validity of the instrument. Conclusions The Italian version of the EDE 17.0D has adequate psychometric properties and can therefore be recommended for examining Italian patients with eating disorders in clinical and research settings.
& Simona Calugi [email protected] 1
Department of Eating and Weight Disorders, Villa Garda Hospital, Via Montebaldo, 89, 37016 Garda, Verona, Italy
2
Department of Neuroscience, Psychology, Drug Research and Child Health, University of Florence, Largo Brambilla 3, 50134 Florence, Italy
Keywords Eating disorder examination Validity Reliability Italian translation Psychometric characteristics Eating disorder behaviours
Introduction The eating disorder examination (EDE) [1] is a semistructured investigator-based interview considered the ‘‘gold-standard’’ measure of eating disorder psychopathology [2]. The EDE comprises four subscales (Restraint, Eating Concern, Shape Concern, and Weight Concern) whose scores can be assessed separately, and as a global score (the mean of the subscale scores). It is designed to provide a measure of the range and severity of eating disorder features (binge eating, self-induced vomiting, laxative misuse, diuretic misuse, and driven exercise for weight control) and to generate operational eating disorder diagnoses. The EDE is widely used in clinical and treatment studies for obtaining descriptive information regarding eating disorder features, making eating disorder diagnoses, and assessing outcomes [3–5]. A recent study systematically reviewed research to date on the psychometric properties of the EDE and examined all published studies set up to investigate its validity and reliability [6]. Findings on patients with eating disorders support the internal consistency of the four EDE subscales, with values ranging between 0.51 for Weight Concern [7] and 0.85 for Shape Concern [8]. Inter-rater reliability coefficients of around 0.90 were found for both the four subscales and the eating disorder behaviours [9, 10]. Moreover, short-term (2–14 days) test– retest reliability was found to be good for the global and subscale EDE scores, ranging from 0.51 for Eating Concern to 0.88 for Restraint, but inadequate in terms of the
123
Eat Weight Disord
items assessing subjective bulimic episodes [9, 11]. Patients with eating disorders show a score higher than controls on all subscales, and an improvement in eating disorder behaviours (e.g. binge eating and purging) during treatment is associated with a reduction in the subscale scores [2]. The EDE has been translated into multiple languages. For instance, the validity of the EDE 12.0 edition was assessed in a large sample of patients with anorexia nervosa, bulimia nervosa, and binge eating disorder in Italy. The data showed that the instrument had good validity in assessing the eating disorder psychopathology [12] and is a reliable interview even when administered retrospectively [13]. Moreover, findings from a Spanish language version of the EDE (12th edition) provided preliminary support for the reliability and validity of the instrument for use with Spanish-speaking women [14, 15]. Likewise, a Norwegian version of the Child EDE showed satisfactory psychometric properties, and its authors therefore recommended the tool for clinical purposes and research on young people with eating disorders in Norway [16]. EDE 12.0 edition was the last version of EDE which was translated into Italian. The need for a validation of the new versions of the instrument arose from the significant changes made by the authors from the 16th up to the 17th edition. As compared with EDE 12.0, EDE 16.0 edition had a different way of rating the Dietary Restraint subscale, such that restraint for the purpose of gaining a sense of control in general may be rated in addition to restraint intended to influence shape or weight. A binge eating disorder module was included, as well as a distinction between compensatory and non-compensatory behaviours. There was a new ‘‘importance’’ item designed to detect the over-evaluation of control over eating per second. The latest version of the EDE, the 17th edition, has now superseded its predecessors. Indeed, the previous edition (EDE 16.0D) [17] was modified to reflect the latest DSM-5 diagnoses of eating disorders [18]. These changes included removing the item on menstruation, which is no longer considered a diagnostic criterion for anorexia nervosa. In addition, the diagnosis of binge eating disorder now relies on a 3-month, rather than 6-month, period of binge eating, and the items assessing this feature have therefore been adjusted accordingly. Finally, references to ‘‘whose weight might make them eligible for the diagnosis of anorexia nervosa’’ have been replaced by ‘‘whose weight might be viewed as significantly low’’. In all other significant respects, the tool is the same as the EDE 16.0D. The aims of this study were to propose an Italian language version of the EDE 17.0D and to assess its psychometric properties in a group of Italian-speaking patients and controls.
123
Methods Participants The study sample comprised 185 patients meeting DSM-5 diagnostic criteria for eating disorder recruited between January 2013 and December 2014. The DSM-5 diagnosis was carried out using clinical interview conducted by experts in the field. One-hundred and sixty-five of the patients were recruited from the inpatient unit of Villa Garda Hospital (northern Italy) and 20 from the University of Florence Department of Neuroscience, Psychology, Drug Research and Child Health Psychiatric Unit. Sixty healthy age-matched controls were recruited from the general population in various community settings. The study design was reviewed and approved by the Institutional Review Board of Villa Garda Hospital, Verona, and all participants gave written informed consent for the use of their anonymous personal data. For those under the age of 18 years, additional informed consent was provided by their parents. The EDE The EDE is a semi-structured interview for assessing eating disorder psychopathology and behaviours over the preceding 28 days. The interview assesses the behavioural symptoms of individuals with eating disorders, including binge eating, self-induced vomiting, laxative misuse, diuretic misuse, excessive exercising, and food restriction. The EDE is divided into four subscales (Restraint, Eating Concern, Weight Concern, and Shape Concern), comprising 22 items, and a global score related to the cognitive features of eating disorders. EDE subscale items are rated on a seven-point forced-choice format (0–6), with higher scores reflecting greater severity or frequency. The Italian version was translated from the original English version by experts in the field. The translation process was conducted as follows [19]: (1) forward translation into Italian by a bilingual person; (2) blind backtranslation into English by a bilingual person; (3) discussion of items by the investigator team to identify any discrepancies and to adjust for any inconsistencies; and (4) final approval by the investigators. The accompanying manual was also translated via the same process and used to train a group of clinicians to administer the EDE to participants. The EDE was administered on the first day of admission at the Villa Garda inpatient unit on the 165 inpatients and during the pre-treatment assessment interview in the 20 outpatients. In 21 inpatients, to assess the test–retest reliability of the interview, the EDE was also administered
Eat Weight Disord Table 1 Baseline characteristics of patients and controls
Patients (N = 185)
Control group (N = 60)
T test or v2 test
Age (years)
27.8 (8.4)
29.9 (9.0)
1.69
Gender [n (%) female]
173 (93.5 %)
58 (96.7 %)
0.84
0.361
Marital status [n (%)]
150 (81.1 %)
38 (63.3 %)
13.40
0.001
Single, never married
29 (15.7 %)
22 (36.7 %)
Married or living as such
6 (3.2 %)
0 17.87
\0.001
16.52
0.002
8.35
\0.001
p 0.092
Separated or divorced 55 (29.7 %)
4 (6.7 %)
Middle school
96 (51.9 %)
33 (55 %)
High school diploma
34 (18.4 %)
23 (38.3 %)
78 (42.2 %) 57 (31 %)
25 (41.7 %) 32 (53.3 %)
Student
11 (6 %)
1 (1.7 %)
Homemaker
23 (12.5 %)
2 (3.3 %)
Unemployed
15 (8.2 %)
0
Body mass index (kg/m2)
17.0 (4.7)
22.2 (2.1)
DSM-5 diagnosis [n (%)]
114 (61.6 %)
Education level [n (%)]
Bachelor’s degree Occupation [n (%)] Employed
Other
Anorexia nervosa
41 (22.2 %)
Bulimia nervosa
4 (2.2 %)
Binge eating disorder
26 (14.1 %)
Other specified feeding or eating disorders Duration of eating disorder (years) [median (range)]
8.5 (0–29)
Data are presented as mean (SD), frequency (%), or median (range)
1–3 weeks before admission during the routine pre-admission interview. Statistical analysis We performed all statistical analysis respecting the original four-factor structure of the EDE (Restraint, Eating Concern, Weight Concern, and Shape Concern subscales). Reliability Internal consistency was calculated using Cronbach’s alpha for the four original EDE subscales on the full data set pertaining to the 185 participants. We also calculated Spearman rank correlations among EDE subscales. To measure the inter-rater reliability, a random sample of 21 of the 165 inpatients was selected, and two interviewers scored these participants independently from audiotape recordings of the interview. Spearman rank correlations and intraclass correlation coefficients (ICC) were calculated on global and subscale EDE scores and eating disorder behaviours. To explore test–retest reliability,
Spearman rank correlations were applied to a random subgroup of 21 inpatients interviewed at time 1, and again at time 2, 1–3 weeks later. The second interview was administered before any treatment was offered by different interviewers blind to the initial EDE scores. Criterion validity Criterion validity was measured using either the twosample t test for independent groups or the Mann–Whitney U test, as appropriate, to compare the global and original subscale EDE scores and eating disorder behaviours in a subgroup of 60 patients with eating disorders and in agematched controls.
Results The characteristics of the 185 patients with eating disorders and 60 controls are shown in Table 1. The majority of patients and controls were single females in their mid-20s. As well as a lower body mass index than the control group,
123
Eat Weight Disord Table 2 Descriptive statistics at time 1, inter-rater reliability, and test–retest reliability EDE global score
Mean (SD)
Min–max
Inter-rater reliability Spearman rho
3.5 (1.5)
0.8–5.2
0.97**
3.9 (1.5)
0.4–6
0.93**
EDE subscales Restraint Eating Concern
3.1 (1.4)
0.4–5.2
0.95**
Weight Concern
3.5 (2.1)
0–5.8
0.98**
Shape Concern
3.7 (2.0)
0–6
0.99**
Mean (SD)
Median (min–max)
Inter-rater Reliability Spearman rho
Objective bulimic days
9.7 (12.5)
0 (0–28)
0.94**
Objective bulimic episodes
20.3 (36.2)
0 (0–112)
1.00**
Eating disorder behaviours
Subjective bulimic days
11.0 (12.5)
6 (0–28)
0.89**
Subjective bulimic episodes
13.7 (20.0)
6 (0–84)
0.75**
Vomiting episodes
35.1 (42.9)
25 (0–140)
0.996**
Laxative misuse episodes
2.6 (6.7)
0 (0–28)
0.90**
Diuretic misuse episodes
1.9 (6.4)
0 (0–28)
1.00**
13.0 (13.5)
8 (0–28)
Excessive exercise days
0.96**
Mean (SD)
Min–max
Test–retest reliability Spearman rho
3.6 (1.2)
1.2–5.5
0.80**
Restraint
3.8 (1.3)
1.4–5.8
0.57*
Eating Concern
3.0 (1.5)
1–6
0.80**
Weight Concern
3.5 (1.6)
0.8–6
0.77**
Shape Concern
4.0 (1.6)
0.5–6
0.84**
EDE global score EDE subscales
Mean (SD)
Median (min–max)
Test–retest reliability Spearman rho
Eating disorder behaviours Objective bulimic days Objective bulimic episodes Subjective bulimic days Subjective bulimic episodes
8.2 (12.2)
0 (0–28)
0.99**
16.2 (32.7) 8.9 (9.4)
0 (0–140) 4 (0–28)
0.99** 0.36
8.9 (9.4)
4 (0–28)
0.36
16.2 (32.3)
3 (0–140)
0.94**
Laxative misuse episodes
7.7 (12.1)
0 (0–28)
0.92**
Diuretic misuse episodes
0.9 (4.3)
0 (0–20)
1.00**
Excessive exercise days
11.9 (12.4)
10 (0–28)
0.82**
Vomiting episodes
EDE eating disorder examination 17.0D * p \ 0.01; ** p \ 0.001
the patient group also featured lower percentages of employed persons, graduates, and singles. One-hundred and fourteen patients met the criteria for anorexia nervosa, 41 for bulimia nervosa, 4 for binge eating disorder, and 26 for other specified feeding or eating disorders (atypical anorexia nervosa N = 14; bulimia nervosa of low frequency and/or limited duration N = 9; binge eating disorder of low frequency and/or limited duration N = 3).
123
Internal consistency Items were grouped into the original four subscales derived from Cooper et al. [20], namely Restraint, Eating Concern, Weight Concern, and Shape Concern. Cronbach’s alpha coefficients for subscales were 0.66, 0.65, 0.77, and 0.84, respectively. Spearman rank correlations among EDE subscales were all statistically significant (p \ 0.05) and
Eat Weight Disord Table 3 Global and subscale EDE scores and eating disorder behaviours in patients and controls
EDE global score
Patients (N = 60)
Controls (N = 60)
T test or Mann–Whitney test
p
3.4 (1.3)
0.7 (0.6)
15.28
\0.001
3.6 (1.5)
0.9 (0.8)
11.60
\0.001
EDE subscales Restraint Eating Concern
2.9 (1.4)
0.3 (0.5)
13.36
\0.001
Weight Concern
3.4 (1.7)
0.7 (0.7)
11.44
\0.001
Shape Concern
3.8 (1.7)
0.9 (1.0)
11.64
\0.001 \0.001
Eating disorder behaviours Objective bulimic days
0 (0–30)
0 (0–10)
5.10
Objective bulimic episodes
0 (0–140)
0 (0)
–
–
Subjective bulimic days
2 (0–30)
0 (0–8)
4.83
\0.001
Subjective bulimic episodes Vomiting episodes
2 (0–56) 0 (0–140)
0 (0–8) 0 (0)
4.65 –
\0.001 –
Laxative misuse episodes
0 (0–28)
0 (0)
–
–
Diuretic misuse episodes
0 (0–28)
0 (0)
–
–
Excessive exercise days
1 (0–28)
0 (0–20)
5.22
\0.001
Data are presented as mean (SD) or as median (range) EDE eating disorder examination 17.0D
ranged from 0.35 (between Restraint and Shape Concern subscales) to 0.81 (between Shape Concern and Weight Concern subscales). Similar results were found after adjusting for treatment condition (0.35–0.83).
patients with eating disorders than in the age-matched controls (Table 3). Similar results were found after adjusting the data for the baseline characteristics which differ between the two groups (education level, occupation, and body mass index).
Inter-rater reliability Spearman rank correlations on global and original subscale EDE scores and on eating disorder behaviours were all high, ranging between 0.75 and 1.00 (Table 2). ICCs ranged from 0.976 to 0.995 for the EDE subscales, and from 0.88 to 1.00 for eating disorder behaviours. Test–retest reliability Twenty-one patients were administered EDE at time 1, and again 1–3 weeks later. The second EDE interview was given an average of 11.3 days (SD = 4.1, range 7–23 days) after the initial interview. Test–retest reliability was C0.80 on global and original subscale EDE scores, with the exception of the Restraint subscale, for which it was 0.57. Moreover, test–retest reliability was good to excellent for objective bulimic days and episodes, vomiting episodes, laxative and diuretic misuse episodes, and excessive exercising, but unsatisfactory for subjective bulimic days and episodes (Table 2). Criterion validity Global and subscale EDE scores and eating disorder behaviours were significantly higher in the subgroup of 60
Discussion This study examined the internal consistency, inter-rater and short-term test–retest reliabilities, and criterion validity of the proposed Italian language version of the EDE 17.0D in a large group of patients referred to inpatient and outpatient treatment clinics. The Italian version of the EDE 17.0D confirmed the strong psychometric properties of the tool previously demonstrated elsewhere [6]. Cronbach’s alpha coefficients of the four original subscales were high and similar to those found in previous studies based on clinical samples [4, 7, 8, 20], indicating the high internal consistency of the tool. Furthermore, its inter-rater reliability, measured using both Spearman’s correlation coefficients and ICCs, was excellent in both the global and original subscale EDE scores and the eating disorder behaviours. All Spearman correlation values were greater than 0.9, displaying optimal reliability in terms of recall of objective bulimic and diuretic misuse episodes. Only the correlation of subjective bulimic episodes was weakly lower (0.75). These findings are consistent with those reported by other studies conducted on both patients with binge eating disorder and patients with all diagnostic categories of eating disorders [9, 11].
123
Eat Weight Disord
Test–retest correlations for the EDE subscales were high, and invariably greater than 0.75, except for Restraint (0.57), indicating that on the whole the subscales remained stable over a period of time ranging from 7 to 23 days. Correlations of reported eating disorder behaviours for time 1 versus time 2 revealed more inconsistent findings. The number of objective bulimic episodes and days had excellent test–retest correlations (0.99), indicating that participants recalled these major episodes consistently over time. Similarly, participants’ recall of the number of episodes of self-induced vomiting, laxative and diuretic misuse, and excessive exercising were very high (from 0.82 to 0.94). In contrast, the test–retest correlations for subjective bulimic episodes and days were relatively low and not significant, suggesting that the recall of these behaviours, in which there is a sense of loss of control over eating but the amount of food consumed is not excessive, is not as reliable as the recall of other eating disorder behaviours. Our data are similar to those reported in a study that administered the second interview at a mean of 4 days after time 1 [11], but slightly higher than those of another study on patients with binge eating disorder in which the interval between the interviews was similar to that used in our study [9]. These findings appear to contradict the conclusions that test–retest reliability correlations weaken as the length of time between testing increases [9]. The ability of EDE to discriminate between patients and age-matched controls is confirmed by our research. The higher global and original subscale EDE and eating disorder behaviour scores of our patient group with respect to the original English sample [20] could be ascribable to the greater proportion of eating disorder inpatients in our sample (89 % vs 40 %). The strengths of this study lay in the fairly large sample size, the inclusion of inpatients and outpatients, the agematched control group, and the data gathered from two different eating disorder units. Moreover, to our knowledge, this is the first study to assess the inter-rater reliability and test–retest reliability of scores furnished by the items assessing laxative and diuretic misuse and excessive exercising episodes. The main limitation of the study concerns the sample composition. Our sample consisted of patients from all diagnostic categories of eating disorders and did not assess the validity and reliability of the EDE within distinct patient subgroups (e.g. anorexia nervosa, bulimia nervosa, and binge eating disorder). Furthermore, although we assessed discriminant validity analysing whether the EDE is able to discriminate eating disorder patients from controls, we did not verify whether it discriminates eating disorder patients from other psychiatric patients or whether it correlated (or not) with other measures of eating disorder psychopathology or general psychopathology.
123
Finally, we did not perform on purpose a factor analysis, as the EDE subscales comprise items collected together on a rational basis to represent the major areas of eating disorder psychopathology [20].
Conclusions In conclusion, the Italian version of the EDE 17.0D demonstrated good psychometric reliability and validity, thereby providing Italian clinicians and researchers with a tool suitable for assessing eating disorder features in everyday practice, clinical trials, cross-cultural research, and multicentre studies. Future research should focus on further validation in larger samples, as well as across different ethnic groups, in which the cultural significance of eating and weight may differ. Conflict of interest
None declared.
References 1. Cooper Z, Fairburn CG (1987) The eating disorder examination: a semistructured interview for the assessment of the specific psychopathology of eating disorders. Int J Eat Disord 6:1–8. doi:10.1002/1098-108X(198701)6:1\1:AID-EAT2260060102[3. 0.CO;2-9 2. Guest T (2000) Using the eating disorder examination in the assessment of bulimia and anorexia: issues of reliability and validity. Soc Work Health Care 31(4):71–83. doi:10.1300/ J010v31n04_05 3. Wilson GT (1993) Assessment of binge eating. In: Fairburn CG, Wilson GT (eds) Binge eating: nature, assessment, and treatment. Guilford Press, New York, pp 227–249 4. Beumont PJ, Kopec-Schrader EM, Talbot P, Touyz SW (1993) Measuring the specific psychopathology of eating disorder patients. Aust N Z J Psychiatry 27(3):506–511. doi:10.3109/ 00048679309075810 5. Fairburn CG, Cooper Z, Doll HA, O’Connor ME, Bohn K, Hawker DM, Palmer RL (2009) Transdiagnostic cognitive-behavioral therapy for patients with eating disorders: a two-site trial with 60-week follow-up. Am J Psychiatr 166:311–319. doi:10. 1176/appi.ajp.2008.08040608 6. Berg KC, Peterson CB, Frazier P, Crow SJ (2012) Psychometric evaluation of the eating disorder examination and eating disorder examination-questionnaire: a systematic review of the literature. Int J Eat Disord 45(3):428–438. doi:10.1002/eat.20931 7. Grilo CM, Crosby RD, Peterson CB, Masheb RM, White MA, Crow SJ, Wonderlich SA, Mitchell JE (2010) Factor structure of the eating disorder examination interview in patients with bingeeating disorder. Obesity (Silver Spring) 18(5):977–981. doi:10. 1038/oby.2009.321 8. Byrne SM, Allen KL, Lampard AM, Dove ER, Fursland A (2010) The factor structure of the eating disorder examination in clinical and community samples. Int J Eat Disord 43(3):260–265. doi:10. 1002/eat.20681 9. Grilo CM, Masheb RM, Lozano-Blanco C, Barry DT (2004) Reliability of the eating disorder examination in patients with binge eating disorder. Int J Eat Disord 35(1):80–85. doi:10.1002/ eat.10238
Eat Weight Disord 10. Rosen JC, Vara L, Wendt S, Leitenberg H (1990) Validity studies of the eating disorder examination. Int J Eat Disord 9:519–528. doi:10.1002/1098-108X(199009)9:5\519:AID-EAT2260090507 [3.0.CO;2-K 11. Rizvi SL, Peterson CB, Crow SJ, Agras WS (2000) Test–retest reliability of the eating disorder examination. Int J Eat Disord 28(3):311–316. doi:10.1002/1098-108X(200011)28:3\311:AIDEAT8[3.0.CO;2-K 12. Ravaldi CVA, Truglia E, Zucchi T, Mannucci E, Rotella CM, Faravelli C, Ricca V (2004) The eating disorder examination as a retrospective interview. Eat Weight Disord 9(3):228–231. doi:10. 1007/BF03325072 13. Mannucci ERV, Di Bernardo M, Rotella CM (1996) Studio del comportamento alimentare con un’intervista strutturata: la Eating Disorder Examination. Il Diabete 8:127–131 14. Grilo CM, Lozano C, Elder KA (2005) Inter-rater and test-retest reliability of the Spanish language version of the eating disorder examination interview: clinical and research implications. J Psychiatr Pract 11(4):231–240
15. Grilo CM, Crosby RD, White MA (2012) Spanish-language eating disorder examination interview: factor structure in Latino/ as. Eat Behav 13(4):410–413. doi:10.1016/j.eatbeh.2012.07.006 16. Frampton I, Wisting L, Overas M, Midtsund M, Lask B (2011) Reliability and validity of the Norwegian translation of the child eating disorder examination (ChEDE). Scand J Psychol 52(2):196–199. doi:10.1111/j.1467-9450.2010.00833.x 17. Fairburn CG, Cooper Z, O’Connor M (2008) eating disorder examination (edition 16.0D). In: Fairburn CG (ed) Cognitive behavior therapy and eating disorders. Guilford Press, New York, pp 265–308 18. American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders, 5th edn (DSM-5TM). American Psychiatric Publishing, Arlington 19. Hilton A, Skrutkowski M (2002) Translating instruments into other languages: development and testing processes. Cancer Nurs 25(1):1–7 20. Cooper Z, Cooper PJ, Fairburn CG (1989) The validity of the eating disorder examination and its subscales. Br J Psychiatry 154:807–812. doi:10.1192/bjp.154.6.807
123
ORIGINAL ARTICLE
The eating disorder examination: reliability and validity of the Italian version Simona Calugi1 • Valdo Ricca2 • Giovanni Castellini2 • Carolina Lo Sauro2 • Antonella Ruocco1 • Elisa Chignola1 • Marwan El Ghoch1 • Riccardo Dalle Grave1
Received: 18 February 2015 / Accepted: 24 March 2015 Ó Springer International Publishing Switzerland 2015
Abstract Objective To examine the psychometric characteristics of the Italian language version of the latest edition of the eating disorder examination (EDE). Methods An Italian version of the EDE (17th edition) was designed and administered to 185 in- and outpatients with eating disorders and 60 age-matched controls. Its internal consistency, inter-rater reliability, short-term (7–23 days) test–retest reliability and criterion validity were evaluated. Results Internal consistency was high for all four original EDE subscales. Inter-rater reliability was excellent for global EDE scores and original subscales (C0.93), and for eating disorder behaviours (C0.89). Test–retest reliability was good for global EDE scores and original subscales (0.57–0.80), objective bulimic episodes and days, vomiting episodes, laxative and diuretic misuse episodes, and excessive exercising (C0.82), but unsatisfactory for subjective bulimic episodes and days. Patients with eating disorders displayed significantly higher EDE scores than age-matched controls, demonstrating the good criterion validity of the instrument. Conclusions The Italian version of the EDE 17.0D has adequate psychometric properties and can therefore be recommended for examining Italian patients with eating disorders in clinical and research settings.
& Simona Calugi [email protected] 1
Department of Eating and Weight Disorders, Villa Garda Hospital, Via Montebaldo, 89, 37016 Garda, Verona, Italy
2
Department of Neuroscience, Psychology, Drug Research and Child Health, University of Florence, Largo Brambilla 3, 50134 Florence, Italy
Keywords Eating disorder examination Validity Reliability Italian translation Psychometric characteristics Eating disorder behaviours
Introduction The eating disorder examination (EDE) [1] is a semistructured investigator-based interview considered the ‘‘gold-standard’’ measure of eating disorder psychopathology [2]. The EDE comprises four subscales (Restraint, Eating Concern, Shape Concern, and Weight Concern) whose scores can be assessed separately, and as a global score (the mean of the subscale scores). It is designed to provide a measure of the range and severity of eating disorder features (binge eating, self-induced vomiting, laxative misuse, diuretic misuse, and driven exercise for weight control) and to generate operational eating disorder diagnoses. The EDE is widely used in clinical and treatment studies for obtaining descriptive information regarding eating disorder features, making eating disorder diagnoses, and assessing outcomes [3–5]. A recent study systematically reviewed research to date on the psychometric properties of the EDE and examined all published studies set up to investigate its validity and reliability [6]. Findings on patients with eating disorders support the internal consistency of the four EDE subscales, with values ranging between 0.51 for Weight Concern [7] and 0.85 for Shape Concern [8]. Inter-rater reliability coefficients of around 0.90 were found for both the four subscales and the eating disorder behaviours [9, 10]. Moreover, short-term (2–14 days) test– retest reliability was found to be good for the global and subscale EDE scores, ranging from 0.51 for Eating Concern to 0.88 for Restraint, but inadequate in terms of the
123
Eat Weight Disord
items assessing subjective bulimic episodes [9, 11]. Patients with eating disorders show a score higher than controls on all subscales, and an improvement in eating disorder behaviours (e.g. binge eating and purging) during treatment is associated with a reduction in the subscale scores [2]. The EDE has been translated into multiple languages. For instance, the validity of the EDE 12.0 edition was assessed in a large sample of patients with anorexia nervosa, bulimia nervosa, and binge eating disorder in Italy. The data showed that the instrument had good validity in assessing the eating disorder psychopathology [12] and is a reliable interview even when administered retrospectively [13]. Moreover, findings from a Spanish language version of the EDE (12th edition) provided preliminary support for the reliability and validity of the instrument for use with Spanish-speaking women [14, 15]. Likewise, a Norwegian version of the Child EDE showed satisfactory psychometric properties, and its authors therefore recommended the tool for clinical purposes and research on young people with eating disorders in Norway [16]. EDE 12.0 edition was the last version of EDE which was translated into Italian. The need for a validation of the new versions of the instrument arose from the significant changes made by the authors from the 16th up to the 17th edition. As compared with EDE 12.0, EDE 16.0 edition had a different way of rating the Dietary Restraint subscale, such that restraint for the purpose of gaining a sense of control in general may be rated in addition to restraint intended to influence shape or weight. A binge eating disorder module was included, as well as a distinction between compensatory and non-compensatory behaviours. There was a new ‘‘importance’’ item designed to detect the over-evaluation of control over eating per second. The latest version of the EDE, the 17th edition, has now superseded its predecessors. Indeed, the previous edition (EDE 16.0D) [17] was modified to reflect the latest DSM-5 diagnoses of eating disorders [18]. These changes included removing the item on menstruation, which is no longer considered a diagnostic criterion for anorexia nervosa. In addition, the diagnosis of binge eating disorder now relies on a 3-month, rather than 6-month, period of binge eating, and the items assessing this feature have therefore been adjusted accordingly. Finally, references to ‘‘whose weight might make them eligible for the diagnosis of anorexia nervosa’’ have been replaced by ‘‘whose weight might be viewed as significantly low’’. In all other significant respects, the tool is the same as the EDE 16.0D. The aims of this study were to propose an Italian language version of the EDE 17.0D and to assess its psychometric properties in a group of Italian-speaking patients and controls.
123
Methods Participants The study sample comprised 185 patients meeting DSM-5 diagnostic criteria for eating disorder recruited between January 2013 and December 2014. The DSM-5 diagnosis was carried out using clinical interview conducted by experts in the field. One-hundred and sixty-five of the patients were recruited from the inpatient unit of Villa Garda Hospital (northern Italy) and 20 from the University of Florence Department of Neuroscience, Psychology, Drug Research and Child Health Psychiatric Unit. Sixty healthy age-matched controls were recruited from the general population in various community settings. The study design was reviewed and approved by the Institutional Review Board of Villa Garda Hospital, Verona, and all participants gave written informed consent for the use of their anonymous personal data. For those under the age of 18 years, additional informed consent was provided by their parents. The EDE The EDE is a semi-structured interview for assessing eating disorder psychopathology and behaviours over the preceding 28 days. The interview assesses the behavioural symptoms of individuals with eating disorders, including binge eating, self-induced vomiting, laxative misuse, diuretic misuse, excessive exercising, and food restriction. The EDE is divided into four subscales (Restraint, Eating Concern, Weight Concern, and Shape Concern), comprising 22 items, and a global score related to the cognitive features of eating disorders. EDE subscale items are rated on a seven-point forced-choice format (0–6), with higher scores reflecting greater severity or frequency. The Italian version was translated from the original English version by experts in the field. The translation process was conducted as follows [19]: (1) forward translation into Italian by a bilingual person; (2) blind backtranslation into English by a bilingual person; (3) discussion of items by the investigator team to identify any discrepancies and to adjust for any inconsistencies; and (4) final approval by the investigators. The accompanying manual was also translated via the same process and used to train a group of clinicians to administer the EDE to participants. The EDE was administered on the first day of admission at the Villa Garda inpatient unit on the 165 inpatients and during the pre-treatment assessment interview in the 20 outpatients. In 21 inpatients, to assess the test–retest reliability of the interview, the EDE was also administered
Eat Weight Disord Table 1 Baseline characteristics of patients and controls
Patients (N = 185)
Control group (N = 60)
T test or v2 test
Age (years)
27.8 (8.4)
29.9 (9.0)
1.69
Gender [n (%) female]
173 (93.5 %)
58 (96.7 %)
0.84
0.361
Marital status [n (%)]
150 (81.1 %)
38 (63.3 %)
13.40
0.001
Single, never married
29 (15.7 %)
22 (36.7 %)
Married or living as such
6 (3.2 %)
0 17.87
\0.001
16.52
0.002
8.35
\0.001
p 0.092
Separated or divorced 55 (29.7 %)
4 (6.7 %)
Middle school
96 (51.9 %)
33 (55 %)
High school diploma
34 (18.4 %)
23 (38.3 %)
78 (42.2 %) 57 (31 %)
25 (41.7 %) 32 (53.3 %)
Student
11 (6 %)
1 (1.7 %)
Homemaker
23 (12.5 %)
2 (3.3 %)
Unemployed
15 (8.2 %)
0
Body mass index (kg/m2)
17.0 (4.7)
22.2 (2.1)
DSM-5 diagnosis [n (%)]
114 (61.6 %)
Education level [n (%)]
Bachelor’s degree Occupation [n (%)] Employed
Other
Anorexia nervosa
41 (22.2 %)
Bulimia nervosa
4 (2.2 %)
Binge eating disorder
26 (14.1 %)
Other specified feeding or eating disorders Duration of eating disorder (years) [median (range)]
8.5 (0–29)
Data are presented as mean (SD), frequency (%), or median (range)
1–3 weeks before admission during the routine pre-admission interview. Statistical analysis We performed all statistical analysis respecting the original four-factor structure of the EDE (Restraint, Eating Concern, Weight Concern, and Shape Concern subscales). Reliability Internal consistency was calculated using Cronbach’s alpha for the four original EDE subscales on the full data set pertaining to the 185 participants. We also calculated Spearman rank correlations among EDE subscales. To measure the inter-rater reliability, a random sample of 21 of the 165 inpatients was selected, and two interviewers scored these participants independently from audiotape recordings of the interview. Spearman rank correlations and intraclass correlation coefficients (ICC) were calculated on global and subscale EDE scores and eating disorder behaviours. To explore test–retest reliability,
Spearman rank correlations were applied to a random subgroup of 21 inpatients interviewed at time 1, and again at time 2, 1–3 weeks later. The second interview was administered before any treatment was offered by different interviewers blind to the initial EDE scores. Criterion validity Criterion validity was measured using either the twosample t test for independent groups or the Mann–Whitney U test, as appropriate, to compare the global and original subscale EDE scores and eating disorder behaviours in a subgroup of 60 patients with eating disorders and in agematched controls.
Results The characteristics of the 185 patients with eating disorders and 60 controls are shown in Table 1. The majority of patients and controls were single females in their mid-20s. As well as a lower body mass index than the control group,
123
Eat Weight Disord Table 2 Descriptive statistics at time 1, inter-rater reliability, and test–retest reliability EDE global score
Mean (SD)
Min–max
Inter-rater reliability Spearman rho
3.5 (1.5)
0.8–5.2
0.97**
3.9 (1.5)
0.4–6
0.93**
EDE subscales Restraint Eating Concern
3.1 (1.4)
0.4–5.2
0.95**
Weight Concern
3.5 (2.1)
0–5.8
0.98**
Shape Concern
3.7 (2.0)
0–6
0.99**
Mean (SD)
Median (min–max)
Inter-rater Reliability Spearman rho
Objective bulimic days
9.7 (12.5)
0 (0–28)
0.94**
Objective bulimic episodes
20.3 (36.2)
0 (0–112)
1.00**
Eating disorder behaviours
Subjective bulimic days
11.0 (12.5)
6 (0–28)
0.89**
Subjective bulimic episodes
13.7 (20.0)
6 (0–84)
0.75**
Vomiting episodes
35.1 (42.9)
25 (0–140)
0.996**
Laxative misuse episodes
2.6 (6.7)
0 (0–28)
0.90**
Diuretic misuse episodes
1.9 (6.4)
0 (0–28)
1.00**
13.0 (13.5)
8 (0–28)
Excessive exercise days
0.96**
Mean (SD)
Min–max
Test–retest reliability Spearman rho
3.6 (1.2)
1.2–5.5
0.80**
Restraint
3.8 (1.3)
1.4–5.8
0.57*
Eating Concern
3.0 (1.5)
1–6
0.80**
Weight Concern
3.5 (1.6)
0.8–6
0.77**
Shape Concern
4.0 (1.6)
0.5–6
0.84**
EDE global score EDE subscales
Mean (SD)
Median (min–max)
Test–retest reliability Spearman rho
Eating disorder behaviours Objective bulimic days Objective bulimic episodes Subjective bulimic days Subjective bulimic episodes
8.2 (12.2)
0 (0–28)
0.99**
16.2 (32.7) 8.9 (9.4)
0 (0–140) 4 (0–28)
0.99** 0.36
8.9 (9.4)
4 (0–28)
0.36
16.2 (32.3)
3 (0–140)
0.94**
Laxative misuse episodes
7.7 (12.1)
0 (0–28)
0.92**
Diuretic misuse episodes
0.9 (4.3)
0 (0–20)
1.00**
Excessive exercise days
11.9 (12.4)
10 (0–28)
0.82**
Vomiting episodes
EDE eating disorder examination 17.0D * p \ 0.01; ** p \ 0.001
the patient group also featured lower percentages of employed persons, graduates, and singles. One-hundred and fourteen patients met the criteria for anorexia nervosa, 41 for bulimia nervosa, 4 for binge eating disorder, and 26 for other specified feeding or eating disorders (atypical anorexia nervosa N = 14; bulimia nervosa of low frequency and/or limited duration N = 9; binge eating disorder of low frequency and/or limited duration N = 3).
123
Internal consistency Items were grouped into the original four subscales derived from Cooper et al. [20], namely Restraint, Eating Concern, Weight Concern, and Shape Concern. Cronbach’s alpha coefficients for subscales were 0.66, 0.65, 0.77, and 0.84, respectively. Spearman rank correlations among EDE subscales were all statistically significant (p \ 0.05) and
Eat Weight Disord Table 3 Global and subscale EDE scores and eating disorder behaviours in patients and controls
EDE global score
Patients (N = 60)
Controls (N = 60)
T test or Mann–Whitney test
p
3.4 (1.3)
0.7 (0.6)
15.28
\0.001
3.6 (1.5)
0.9 (0.8)
11.60
\0.001
EDE subscales Restraint Eating Concern
2.9 (1.4)
0.3 (0.5)
13.36
\0.001
Weight Concern
3.4 (1.7)
0.7 (0.7)
11.44
\0.001
Shape Concern
3.8 (1.7)
0.9 (1.0)
11.64
\0.001 \0.001
Eating disorder behaviours Objective bulimic days
0 (0–30)
0 (0–10)
5.10
Objective bulimic episodes
0 (0–140)
0 (0)
–
–
Subjective bulimic days
2 (0–30)
0 (0–8)
4.83
\0.001
Subjective bulimic episodes Vomiting episodes
2 (0–56) 0 (0–140)
0 (0–8) 0 (0)
4.65 –
\0.001 –
Laxative misuse episodes
0 (0–28)
0 (0)
–
–
Diuretic misuse episodes
0 (0–28)
0 (0)
–
–
Excessive exercise days
1 (0–28)
0 (0–20)
5.22
\0.001
Data are presented as mean (SD) or as median (range) EDE eating disorder examination 17.0D
ranged from 0.35 (between Restraint and Shape Concern subscales) to 0.81 (between Shape Concern and Weight Concern subscales). Similar results were found after adjusting for treatment condition (0.35–0.83).
patients with eating disorders than in the age-matched controls (Table 3). Similar results were found after adjusting the data for the baseline characteristics which differ between the two groups (education level, occupation, and body mass index).
Inter-rater reliability Spearman rank correlations on global and original subscale EDE scores and on eating disorder behaviours were all high, ranging between 0.75 and 1.00 (Table 2). ICCs ranged from 0.976 to 0.995 for the EDE subscales, and from 0.88 to 1.00 for eating disorder behaviours. Test–retest reliability Twenty-one patients were administered EDE at time 1, and again 1–3 weeks later. The second EDE interview was given an average of 11.3 days (SD = 4.1, range 7–23 days) after the initial interview. Test–retest reliability was C0.80 on global and original subscale EDE scores, with the exception of the Restraint subscale, for which it was 0.57. Moreover, test–retest reliability was good to excellent for objective bulimic days and episodes, vomiting episodes, laxative and diuretic misuse episodes, and excessive exercising, but unsatisfactory for subjective bulimic days and episodes (Table 2). Criterion validity Global and subscale EDE scores and eating disorder behaviours were significantly higher in the subgroup of 60
Discussion This study examined the internal consistency, inter-rater and short-term test–retest reliabilities, and criterion validity of the proposed Italian language version of the EDE 17.0D in a large group of patients referred to inpatient and outpatient treatment clinics. The Italian version of the EDE 17.0D confirmed the strong psychometric properties of the tool previously demonstrated elsewhere [6]. Cronbach’s alpha coefficients of the four original subscales were high and similar to those found in previous studies based on clinical samples [4, 7, 8, 20], indicating the high internal consistency of the tool. Furthermore, its inter-rater reliability, measured using both Spearman’s correlation coefficients and ICCs, was excellent in both the global and original subscale EDE scores and the eating disorder behaviours. All Spearman correlation values were greater than 0.9, displaying optimal reliability in terms of recall of objective bulimic and diuretic misuse episodes. Only the correlation of subjective bulimic episodes was weakly lower (0.75). These findings are consistent with those reported by other studies conducted on both patients with binge eating disorder and patients with all diagnostic categories of eating disorders [9, 11].
123
Eat Weight Disord
Test–retest correlations for the EDE subscales were high, and invariably greater than 0.75, except for Restraint (0.57), indicating that on the whole the subscales remained stable over a period of time ranging from 7 to 23 days. Correlations of reported eating disorder behaviours for time 1 versus time 2 revealed more inconsistent findings. The number of objective bulimic episodes and days had excellent test–retest correlations (0.99), indicating that participants recalled these major episodes consistently over time. Similarly, participants’ recall of the number of episodes of self-induced vomiting, laxative and diuretic misuse, and excessive exercising were very high (from 0.82 to 0.94). In contrast, the test–retest correlations for subjective bulimic episodes and days were relatively low and not significant, suggesting that the recall of these behaviours, in which there is a sense of loss of control over eating but the amount of food consumed is not excessive, is not as reliable as the recall of other eating disorder behaviours. Our data are similar to those reported in a study that administered the second interview at a mean of 4 days after time 1 [11], but slightly higher than those of another study on patients with binge eating disorder in which the interval between the interviews was similar to that used in our study [9]. These findings appear to contradict the conclusions that test–retest reliability correlations weaken as the length of time between testing increases [9]. The ability of EDE to discriminate between patients and age-matched controls is confirmed by our research. The higher global and original subscale EDE and eating disorder behaviour scores of our patient group with respect to the original English sample [20] could be ascribable to the greater proportion of eating disorder inpatients in our sample (89 % vs 40 %). The strengths of this study lay in the fairly large sample size, the inclusion of inpatients and outpatients, the agematched control group, and the data gathered from two different eating disorder units. Moreover, to our knowledge, this is the first study to assess the inter-rater reliability and test–retest reliability of scores furnished by the items assessing laxative and diuretic misuse and excessive exercising episodes. The main limitation of the study concerns the sample composition. Our sample consisted of patients from all diagnostic categories of eating disorders and did not assess the validity and reliability of the EDE within distinct patient subgroups (e.g. anorexia nervosa, bulimia nervosa, and binge eating disorder). Furthermore, although we assessed discriminant validity analysing whether the EDE is able to discriminate eating disorder patients from controls, we did not verify whether it discriminates eating disorder patients from other psychiatric patients or whether it correlated (or not) with other measures of eating disorder psychopathology or general psychopathology.
123
Finally, we did not perform on purpose a factor analysis, as the EDE subscales comprise items collected together on a rational basis to represent the major areas of eating disorder psychopathology [20].
Conclusions In conclusion, the Italian version of the EDE 17.0D demonstrated good psychometric reliability and validity, thereby providing Italian clinicians and researchers with a tool suitable for assessing eating disorder features in everyday practice, clinical trials, cross-cultural research, and multicentre studies. Future research should focus on further validation in larger samples, as well as across different ethnic groups, in which the cultural significance of eating and weight may differ. Conflict of interest
None declared.
References 1. Cooper Z, Fairburn CG (1987) The eating disorder examination: a semistructured interview for the assessment of the specific psychopathology of eating disorders. Int J Eat Disord 6:1–8. doi:10.1002/1098-108X(198701)6:1\1:AID-EAT2260060102[3. 0.CO;2-9 2. Guest T (2000) Using the eating disorder examination in the assessment of bulimia and anorexia: issues of reliability and validity. Soc Work Health Care 31(4):71–83. doi:10.1300/ J010v31n04_05 3. Wilson GT (1993) Assessment of binge eating. In: Fairburn CG, Wilson GT (eds) Binge eating: nature, assessment, and treatment. Guilford Press, New York, pp 227–249 4. Beumont PJ, Kopec-Schrader EM, Talbot P, Touyz SW (1993) Measuring the specific psychopathology of eating disorder patients. Aust N Z J Psychiatry 27(3):506–511. doi:10.3109/ 00048679309075810 5. Fairburn CG, Cooper Z, Doll HA, O’Connor ME, Bohn K, Hawker DM, Palmer RL (2009) Transdiagnostic cognitive-behavioral therapy for patients with eating disorders: a two-site trial with 60-week follow-up. Am J Psychiatr 166:311–319. doi:10. 1176/appi.ajp.2008.08040608 6. Berg KC, Peterson CB, Frazier P, Crow SJ (2012) Psychometric evaluation of the eating disorder examination and eating disorder examination-questionnaire: a systematic review of the literature. Int J Eat Disord 45(3):428–438. doi:10.1002/eat.20931 7. Grilo CM, Crosby RD, Peterson CB, Masheb RM, White MA, Crow SJ, Wonderlich SA, Mitchell JE (2010) Factor structure of the eating disorder examination interview in patients with bingeeating disorder. Obesity (Silver Spring) 18(5):977–981. doi:10. 1038/oby.2009.321 8. Byrne SM, Allen KL, Lampard AM, Dove ER, Fursland A (2010) The factor structure of the eating disorder examination in clinical and community samples. Int J Eat Disord 43(3):260–265. doi:10. 1002/eat.20681 9. Grilo CM, Masheb RM, Lozano-Blanco C, Barry DT (2004) Reliability of the eating disorder examination in patients with binge eating disorder. Int J Eat Disord 35(1):80–85. doi:10.1002/ eat.10238
Eat Weight Disord 10. Rosen JC, Vara L, Wendt S, Leitenberg H (1990) Validity studies of the eating disorder examination. Int J Eat Disord 9:519–528. doi:10.1002/1098-108X(199009)9:5\519:AID-EAT2260090507 [3.0.CO;2-K 11. Rizvi SL, Peterson CB, Crow SJ, Agras WS (2000) Test–retest reliability of the eating disorder examination. Int J Eat Disord 28(3):311–316. doi:10.1002/1098-108X(200011)28:3\311:AIDEAT8[3.0.CO;2-K 12. Ravaldi CVA, Truglia E, Zucchi T, Mannucci E, Rotella CM, Faravelli C, Ricca V (2004) The eating disorder examination as a retrospective interview. Eat Weight Disord 9(3):228–231. doi:10. 1007/BF03325072 13. Mannucci ERV, Di Bernardo M, Rotella CM (1996) Studio del comportamento alimentare con un’intervista strutturata: la Eating Disorder Examination. Il Diabete 8:127–131 14. Grilo CM, Lozano C, Elder KA (2005) Inter-rater and test-retest reliability of the Spanish language version of the eating disorder examination interview: clinical and research implications. J Psychiatr Pract 11(4):231–240
15. Grilo CM, Crosby RD, White MA (2012) Spanish-language eating disorder examination interview: factor structure in Latino/ as. Eat Behav 13(4):410–413. doi:10.1016/j.eatbeh.2012.07.006 16. Frampton I, Wisting L, Overas M, Midtsund M, Lask B (2011) Reliability and validity of the Norwegian translation of the child eating disorder examination (ChEDE). Scand J Psychol 52(2):196–199. doi:10.1111/j.1467-9450.2010.00833.x 17. Fairburn CG, Cooper Z, O’Connor M (2008) eating disorder examination (edition 16.0D). In: Fairburn CG (ed) Cognitive behavior therapy and eating disorders. Guilford Press, New York, pp 265–308 18. American Psychiatric Association (2013) Diagnostic and statistical manual of mental disorders, 5th edn (DSM-5TM). American Psychiatric Publishing, Arlington 19. Hilton A, Skrutkowski M (2002) Translating instruments into other languages: development and testing processes. Cancer Nurs 25(1):1–7 20. Cooper Z, Cooper PJ, Fairburn CG (1989) The validity of the eating disorder examination and its subscales. Br J Psychiatry 154:807–812. doi:10.1192/bjp.154.6.807
123
