Integrating “Lumpers” Versus “Splitters” Perspectives: Toward a Hierarchical Dimensional Taxonomy of Eating Disorders From Clinician Ratings

In this study, we describe a hierarchical dimensional model of eating-disorder (ED) classification based on the Hierarchical Taxonomy of Psychopathology. Participants were community-recruited adults with an ED (N = 252; 81.9% female). We used a modified version of Goldberg’s method, which involved sequentially extracting latent factors using exploratory structural equation modeling, resulting in a 10-factor hierarchical-dimensional model. Dimensions predicted 92.4% and 58.7% of the variance in recovery outcomes at 6 months and 1 year, respectively. Compared with other illness indicators (e.g., Diagnostic and Statistical Manual of Mental Disorders [DSM] diagnoses, dimensional ED impairment scores, weight/shape overvaluation, and DSM ED-severity specifiers), hierarchical dimensions predicted 0.88 to 334 times more variance in ED behaviors at baseline and 1.95 to 80.8 times more variance in psychiatric impairment at 1-year follow-up. Results suggest that reducing within-disorder heterogeneity for EDs within the broader context of internalizing symptoms provides a powerful framework from which to predict outcomes and understand symptoms experienced by people with EDs.