BACKGROUND We aimed to validate the Eating Disorders Inventory (EDI) in Arabic.
METHODS Subjects were chosen randomly from female school students. Only healthy Saudi students were included. They were asked to fill the EDI and undergo a semi-structured interview by a psychiatrist who was unaware of the EDI scores. Of 146 students approached 12 did not meet our criteria and were excluded; 16 other students were excluded for incomplete responses.
RESULTS The difference between the two diagnostic methods in the proportion of caseness was statistically significantly (P < 0.0001). The scores on Perfectionism, Maturity Fears and Interoceptive Awareness were significantly higher in the Saudi students compared with Canadians (P < 0.05).
CONCLUSIONS The EDI-DT subscale has a sensitivity of 100% and a specificity of 85%. Despite its low positive predictive value of 5%, it may be useful for screening large non-clinical groups for eating disorders.