OBJECTIVE

To analyze the psychometric properties of the Eating Attitudes Test (EAT-26) in a Russian sample of patients with a depressive episode (DE) and recurrent depressive disorder (RDD).

MATERIAL AND METHODS

This multicenter cross-sectional study included 338 people (216, 63.9% women), aged 18—59 years, with diagnoses of DE/ RDD. Concomitant current eating disorder (ED) was diagnosed in 10.1% (n=34) of patients, including 2.4% (n=8) with anorexia nervosa and 7.7% (n=26) with bulimia nervosa. All participants underwent a structured «Mini International Neuropsychiatric Interview» to verify the clinical diagnosis, as well as a psychometric examination to assess EDs using EAT-26. Statistical analysis was performed in RStudio v1.4.1717.

RESULTS

Results of correlation analysis of EAT-26 items revealed that item 26 had a negative adjusted correlation coefficient (–0.052) and was excluded for the analysis of subsequent stages. The six-factor structure of EAT-25 was revealed, including the subscales «Bodily control», «Social pressure», «Dietary restrictions», «Bulimia», «Ritual nutrition» and «Loss of self-control.» Item 19 («I have self-control over food») was inverted, as it had a negative factor load, as well as a negative meaning. The internal consistency of EAT was analyzed on 25 questionnaire items, as well as on individual subscales, and was quite high. When checking the convergent validity of EAT with established diagnoses of ED, it was revealed that patients with ED had a significantly higher score on EAT, as well as on the subscales «Body control», «Dietary restrictions», «Bulimia» and «Loss of self-control» compared with patients without a diagnosis of ED. According to the «Ritual nutrition» subscale, no links with ED were found.

CONCLUSION

The EAT demonstrated good psychometric characteristics in a Russian sample of patients diagnosed with DE/RDD.