Premium
A new approach to eating‐disorder classification: Using empirical methods to delineate diagnostic dimensions and inform care
Author(s) -
Forbush Kelsie T.,
Chen PoYi,
Hagan Kelsey E.,
Chapa Danielle A. N.,
Gould Sara R.,
Eaton Nicholas R.,
Krueger Robert F.
Publication year - 2018
Publication title -
international journal of eating disorders
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.785
H-Index - 138
eISSN - 1098-108X
pISSN - 0276-3478
DOI - 10.1002/eat.22891
Subject(s) - psychology , eating disorders , psychopathology , anxiety , mood , clinical psychology , multilevel model , distress , categorical variable , psychiatry , machine learning , computer science
Objective Despite changes to the diagnostic criteria for eating disorders (EDs) in the DSM ‐5, the current diagnostic system for EDs has limited ability to inform treatment planning and predict outcomes. Our objective was to test the clinical utility of a novel dimensional approach to understanding the structure of ED psychopathology. Method Participants ( N = 243; 82.2% women) were community‐recruited adults with a DSM‐5 ED assessed at baseline, 6‐month, and 1‐year follow‐up. Hierarchical factor analysis was used to identify a joint hierarchical‐dimensional structure of eating, mood, and anxiety symptoms. Exploratory structural equation modeling was used to test the ability of the dimensional model to predict outcomes. Results At the top of the hierarchy, we identified a broad Internalizing factor that reflected diffuse symptoms of eating, mood, and anxiety disorders. Internalizing branched into three subfactors: distress, fear–avoidance (fears of certain stimuli and behaviors to neutralize fears, including ED behaviors designed to reduce fear of weight gain), and body dissatisfaction, which was nested within distress. The lowest level of the hierarchy was characterized by 15 factors. The hierarchical model predicted 60.1% of the variance in outcomes at 6‐month follow‐up, whereas all DSM eating, mood, and anxiety disorders combined predicted 35.8% of the variance in outcomes. Discussion A dimensional approach to understanding and diagnosing EDs improved the ability to prospectively predict clinical course above‐and‐beyond the traditional categorical ( DSM ‐based) approach. Our findings have implications for endeavors to improve the prediction of ED prognosis and course, and to develop more effective trans‐diagnostic treatments.