Introducing Diagnostic Classification Modeling as an Unsupervised Method for Screening Probable Eating Disorders

Abstract

Screening for eating disorders (EDs) is an essential part of the prevention and intervention of EDs. Traditional screening methods mostly rely on predefined cutoff scores which have limitations of generalizability and may produce biased results when the cutoff scores are used in populations where the instruments or cutoff scores have not been validated. Compared to the traditional cutoff score approach, the diagnostic classification modeling (DCM) approach can provide psychometric and classification information simultaneously and has been used for diagnosing mental disorders. In the present study, we introduce DCM as an innovative and alternative approach to screening individuals at risk of EDs. To illustrate the practical utility of DCM, we provide two examples: one involving the application of DCM to examine probable ED status from the 12-item Short form of the Eating Disorder Examination-Questionnaire (EDE-QS) to screen probable thinness-oriented EDs and the Muscularity-Oriented Eating Test (MOET) to screen probable muscularity-oriented EDs.