A new instrument measuring disturbed eating patterns in community populations: Development and initial validation of a five-item scale (EDS-5)

This paper reports on the development and validation of a brief scale, the Eating Disturbance Scale (EDS-5), as a screening instrument for problematic eating disorders in normal populations. A nationwide sample of 6313 subjec ts completed 22 questions about eating patterns. Using principal component analysis, these questions were cut down to five items, forming a scale. Fro m another epidemiological study (N = 835), 51 female subjects responded to these five questions in addition to a clinical interview. The scale yielded a Cronbach alpha of 0.83 and 0.86, respectively, Also, a sensitivity and s pecificity of 0.90 and 0.88 teas found with respect to DSM-IV eating disord ers. The concurrent and construct validity show significant correlations wi th Eating Disorder Inventory subscales measuring disturbed eating (M=0.55, range 0.44-0.73), and even higher correlations with similar factors from th e self-report version of tile Eating Disorder Examination (M=0.73, range 0. 60-0.89). The results suggest that the EDS-5 is sensitive to disordered eat ing patterns and that this instrument appears promising for screening purpo ses in community samples. Copyright (C) 2001 John Wiley & Sons, Ltd and Eat ing Disorders Association.