Psychometric properties of the Czech version of the tool - screening for early eating disorder signs (SEEDS-CZ)

Design: Validation study.

Methods: The group of respondents consisted of young women aged 15–21 (mean = 17.7). From the total of 230 subjects, 106 had type 1 diabetes and 124 were without disease. Psychometric analysis was implemented with regard to factor analysis and reliability. With respect to reliability, the internal consistency was determined by the Cronbach alpha coefficient. In order to verify the domains of the SEEDS-CZ questionnaire, exploration factor analysis with promax rotation and confirmation factor analysis were used. Convergence validity was assessed by correlation with independent testing of the same or similar areas.

Results: On the basis of the exploratory factor analysis, three factors arose in the Czech version. Factor 1 (Body Image) was saturated by seven items, Factor 2 (Feelings) was saturated by seven items, and Factor 3 (Quality of life) was saturated by five items. Contrary to the original version, item 5 was assigned to Factor 1. The correlation between factors was strong. Correlations with the Diabetes Eating Problem Survey – Revised version, the Eating Attitude Test-26, and the Rosenberg scale were significant. The Cronbach alpha coefficient was 0.89 for the tool as a whole.

Conclusion: The SEEDS-CZ evaluation tool is conceptually consistent with the original version of the SEEDS. From the psychometric point of view, it is suitable for use, particularly with diabetics.