Objective: We produced a Japanese version of the Eating Disorder Inventory-2 (EDI-2) and investigated its reliability and validity.

Methods: EDI-2 Japanese version was administered to 246 patients with eating disorders and 100 control subjects. The internal consistency as well as discriminant validity were examined.

Results: The alpha reliability coefficients were high (0.71 to 0.92) in all subscales except the asceticism subscale which was rated 0.61. The anorexia nervosa binge-purging type and bulimia nervosa showed significantly higher scores in all the EDI-2 subscales when compared to the controls, and also significantly higher scores in seven EDI-2 subscales compared to anorexia nervosa restricting type (AN-R). AN-R showed significantly lower scores in the body dissatisfaction subscale. Lower scores in Japanese than Western patients for drive for thinness as well as marked difference in perfectionism and maturity fears suggested cross-cultural issues between Japan and western countries.

Conclusions: Japanese version of EDI-2 is a measure instrument that can be expected to have a satisfactory level of internal consistency except the asceticism subscale as well as high validity as a tool for the evaluation of the psychopathology of eating disorder.