Objective: Like other DSM-IV axis I syndromes, eating disorders are diagnosed without respect to personality, which is coded on axis II. The authors assessed the utility of segregating eating disorders and personality pathology and examined the extent to which personality patterns account for meaningful variation within axis I eating disorder diagnoses.
Method: One hundred three experienced psychiatrists and psychologists used a Q-sort procedure (the Shedler-Westen Assessment Procedure-200) that assesses personality and personality pathology to describe a patient they were currently treating for bulimia or anorexia. Data were subjected to a cluster-analytic procedure (Q-analysis) to determine whether patients clustered into coherent groupings on the basis of their personality profiles. Categorical and dimensional personality diagnoses were then used to predict measures relevant to adaptation and etiology, controlling for axis I diagnosis.
Results: Three categories of patients emerged: a high-functioning/perfectionistic group, a constricted/overcontrolled group, and an emotionally dysregulated/undercontrolled group. This categorization demonstrated substantial incremental validity beyond axis I diagnosis in predicting eating disorder symptoms, adaptive functioning (Global Assessment of Functioning scores and history of psychiatric hospitalization), and etiological variables (sexual abuse history).
Conclusions: Axis I symptoms are a useful component, but only one component, in the accurate diagnosis of eating disorders. Classifying patients with eating disorders by eating symptoms alone groups together patients with anorexic symptoms who are high functioning and self-critical with those who are highly disturbed, constricted, and avoidant, and groups together patients with bulimic symptoms who are high functioning and self-critical with those who are highly disturbed, impulsive, and emotionally dysregulated. These distinctions may be relevant to etiology, prognosis, and treatment.