Translation and Psychometric Evaluation of the Arabic Version of the Eating Disorders After Bariatric Surgery Questionnaire (EDABS‑Q‑Arabic18)

Abstract

Background: There is no validated Arabic questionnaire to assess eating disorders (ED) after metabolic and bariatric surgery (MBS) without the need for an interview. We undertook this task.

Methods: The English-published Eating Disorder Examination-Self-report Questionnaire Bariatric Surgery Version (EDABS-Q) was translated and adapted following international guidelines (forward/backward translation, expert panel review, and pretesting). Randomly selected adult patients (n = 1145) who had MBS since ≥ 1 year completed the Arabic questionnaire (EDABS-Q-Arabic). Psychometric properties of EDABS-Q-Arabic were assessed, including face validity (expert panel), construct validity (exploratory factor analysis, confirmatory factor analysis, structural model), internal consistency (Cronbach's α), and discriminant validity [heterotrait-monotrait (HTMT) ratio criterion]. Multiple logistic regression analyses tested associations between patient/surgical characteristics and various factors of the Arabic questionnaire.

Results: Exploratory factor analysis generated a three-factor solution (18 items): 'concerns' about shape/weight/eating (9 items), 'restraint' behaviors (4 items), and 'purging' behaviors (5 items), explaining 22% of the total variance. Confirmatory factor analysis confirmed this factor structure displayed good model-data fit, with comparative fit index (0.96) and Tucker-Lewis index (0.95) both > 0.95 threshold; χ²/df ratio = 1.52 (recommended value ≤ 2); root mean square error of approximation = 0.031 (90%CI:0.022-0.040, p = 1.000) and standardized root mean square residual = 0.047 (recommended values ≤ 0.05). Cronbach's alpha (internal consistency) was 0.80 for 'concerns' (95%CI:0.78-0.82), 0.62 for 'restraint' (95%CI:0.55-0.68), and 0.61 for 'purging' behavior (95%CI:0.51-0.69). EDABS-Q-Arabic18's discriminant validity was excellent, confirming the distinctiveness of each factor, with 0.208 HTMT ratio between 'concerns' and 'restraint' factors, 0.198 between 'concerns' and 'purging', and 0.257 between 'restraint' and 'purging' factors (recommended thresholds < 0.85-0.90). The prevalence of 'concerns' was 98.4%, with patients experiencing mild (31.1%), moderate (48.9%), severe (18.4%), or no (1.6%) concerns. The prevalence of 'restraint' behaviors was high (79.7%) but mostly mild (53.1%) or moderate (22.5%). 'Purging' behaviors had a lower prevalence (44.2%), with 40.1% mild, 3.8% moderate, and 0.3% severe purging. Logistic regression showed that for severe 'concerns', increasing age and pre-operative BMI displayed lower odds OR = 0.98, 95%CI 0.96-1.00, p = 0.013; OR = 0.96, 95%CI 0.93-0.98, p = 0.001 respectively), while current BMI and time since surgery exhibited significant positive associations (OR = 1.19, 95%CI 1.14-1.24, p < 0.001; OR = 1.10, 95%CI 1.02-1.18, p = 0.011, respectively). For severe 'restraint' behaviors, only pre-operative BMI displayed significant association (OR = 1.05-95%CI 1.00-1.09, p = 0.023). Moderate to severe 'purging' behaviors had no significant associations with any patient/surgical characteristics. The type of MBS procedure was not associated with any of the three factors.

Conclusions: The 18-item Arabic version of the Eating Disorder Examination-Self-report Questionnaire Bariatric Surgery Version (EDABS-Q-Arabic18) is a culturally appropriate, valid, and reliable tool for assessing post-MBS ED. Further validation across Arabic-speaking populations is encouraged to strengthen its broader applicability.

Keywords: Arabic translation; Disordered-eating behaviors; Eating disorders; Eating pathology; Metabolic and bariatric surgery; Questionnaire; Validation.

Fig. 1

Flow chart of patients

Fig. 2

Scree plot

Fig. 3

Structural model from the confirmatory factor analysis of the Eating Disorders After Bariatric Surgery Arabic Questionnaire (EDABS-Q-Arabic18). Note: The structural model derived from confirmatory factor analysis (CFA) for EDABS-Q-Arabic18, presenting standardized path coefficients for the 3-factor structure: ‘concerns’ about shape/weight/eating (α = 0.80), ‘restraint’ behavior (α = 0.62), and ‘purging’ behavior (α = 0.61). Model fit indices indicated acceptable fit (χ²/df = 1.52, CFI = 0.96, TLI = 0.95, RMSEA = 0.031, SRMR = 0.047). Significant standardized coefficients (p < 0.05) demonstrate relationships between latent variables and their respective items, as well as covariances among specific items.

Fig. 4

Prevalence of eating disorders after MBS among the sample