Adaptation and validation of the Eating Disorder Examination‐Questionnaire in English among urban Indian adolescents

Abstract

Objective Eating pathology is a salient issue in India, with clinical features reported among adults and adolescents. However, there are currently no validated measures of disordered eating in the Indian context. The present study therefore aimed to validate a culturally appropriate English language version of the Eating Disorder Examination Questionnaire (EDE‐Q), a widely used measure, among adolescents in India. Method We adapted and examined the factor structure, reliability, and construct validity of the EDE‐Q among an urban sample of 1,413 Indian adolescents (mean age = 13 years; 45% girls). Results In contrast to the original four‐factor model, exploratory factor analysis identified a two‐factor solution for girls and boys, which was verified by confirmatory factor analysis; a “Preoccupation and Control” subscale and a “Weight and Shape Concerns” subscale. The total scale comprised 15 items for girls and 18 items for boys. Test–retest reliability and internal consistency were satisfactory for the girls' and boys' scales (Cronbach's α = .91 for both) and their comprising subscales (αs ≥ .80). Concurrent validity was established through medium‐high significant correlations with measures of body image and internalization of appearance ideals. Combining items common to both versions of the scale led to a psychometrically acceptable model which was largely invariant across both genders, thus facilitating gendered comparison. Discussion Findings indicate the reliability and validity of the EDE‐Q among English‐speaking urban Indian adolescents. This will facilitate further research examining the prevalence and nature of eating pathology among adolescents in India.

Introduction to a special issue on eating disorders in Asia

“…Similar to nearly all prior investigations in other parts of the world (e.g., Becker et al, 2010), the putative four-factor structure of the EDE-Q was unable to be replicated, but other aspects of the measure demonstrated strong reliability and validity. Specifically, Lewis-Smith et al (2021) reported that the culturally adapted Indian version of EDE-Q had a two-factor structure and was a satisfactory and appropriate tool for the measurement of the clinical features of EDs. Otani et al (2021) reported that the Japanese version of EDE-Q had a three-factor structure and demonstrated sufficient reliability and validity.…”

Assessing eating disorder symptoms in low and middle-income countries: a systematic review of psychometric studies of commonly used instruments

“…The control group was undergraduate and graduate students from the China University of Geosciences in Wuhan without eating disorders 84 19.9 ± 3.2 Female/Male Eating Disorder Examination Interview 62 Restraint Eating Concern Shape Concern Weight Concern Discriminant validity: AN BN Individuals without EDs Statistically significant difference Diagnostic performance: AUC not assessed Reliability Internal consistency: Cronbach's alpha coefficient: good Test–retest: Spearman coefficient: strong correlation Penelo [ 31 ] Mexico Spanish School children of both sexes (under 18 years old) of the Mexican population in Nayarit, in urban and rural areas. Participants were excluded if they did not answer the complete questionnaires 2928 15.1 ± 1.7 Female/Male Eating Disorder Examination Questionnaire Self-report 38 Restraint Eating Concern Shape Concern Weight Concern Convergent validity: Questionnaire on Influences of Aesthetic Body Ideal: moderate correlation EDI-2: weak to strong correlation Children’s Eating Attitudes Test: moderate correlation Construct validity: Confirmatory factor analysis: χ2/df: adequate fit RMSEA: good fit CFI: mediocre adjustment Factor loadings: minimum level Measurement invariance: Model equal factor loading: χ2/df: adequate fit RMSEA: acceptable fit CFI: mediocre adjustment Model equal intercepts χ2/df: inadequate fit RMSEA: acceptable fit CFI: mediocre adjustment Reliability Internal consistency: Omega coefficient: good Test–retest: ICC: good Cohen’s Kappa coefficient: discrete Becker [ 28 ] Fiji Fijian The population of all available Fijian ethnic origins, aged 15–20, enrolled on Forms 3 to 6 in the 12 secondary schools registered in an administrative sector of the Fijian Ministry of Education in October 2006 were included 523 16.6 ± 1.0 Female Eating Disorder Examination Questionnaire Self-report 28 Restraint Eating Concern Shape Concern Weigh...…”

“…The included studies used the three versions of the EDE-Q: the EDE-Q 4.0 [ 30 , 31 ], the EDE-Q 6.0 [ 22 , 23 , 25 – 29 ], and the EDE-QS [ 21 ]. The EDE-Q was translated by eight studies [ 26 – 28 , 30 – 32 , 34 , 36 ]. Two did not carry out the translation process and used a version of the questionnaire that was already translated into Spanish [ 33 , 35 , 57 ].…”

“…The number of participants in the studies varied from 298 [ 36 ] to 2,928 [ 35 ]. Six studies were conducted in female and male participants [ 26 – 28 , 31 , 35 , 36 ], three studies in female participants [ 30 , 32 , 34 ], and only one study in male participants [ 33 ]. The mean age of the participants varied from 10 [ 36 ] to 26 [ 33 ] years old.…”

“…The participant inclusion criteria varied according to the target population chosen in the studies. In one study, all participants were recruited through convenience sampling at a hospital among patients with an eating disorder diagnosis [ 30 ], in six studies the participants were chosen through convenience sampling at secondary schools [ 26 , 28 , 30 – 32 , 35 ], in three studies they were recruited at universities [ 27 , 33 , 34 ], and in one study they were recruited at sports centers [ 33 ]. In three studies, the exclusion criterion was participants who did not answer three or more questionnaires [ 24 , 30 , 31 ].…”