Measuring melancholia: the utility of a prototypic symptom approach

BACKGROUND: Melancholia has long resisted classification, with many of its suggested markers lacking specificity. The imprecision of depressive symptoms, in addition to self-report biases, has limited the capacity of existing measures to delineate melancholic depression as a distinct subtype. Our aim was to develop a self-report measure differentiating melancholic and non-melancholic depression, weighting differentiation by prototypic symptoms and determining its comparative classification success with a severity-based strategy. METHOD: Consecutively recruited depressed out-patients (n=228) rated 32 symptoms by prototypic or 'characteristic' relevance (using the Q-sort strategy) and severity [using the Severity-based Depression Rating System (SDRS) strategy]. Clinician diagnosis of melancholic/non-melancholic depression was the criterion measure, but two other formal measures of melancholia (Newcastle and DSM-IV criteria) were also tested. RESULTS: The prevalence of 'melancholia' ranged from 20.9% to 54.2% across the subtyping measures. The Q-sort measure had the highest overall correct classification rate in differentiating melancholic and non-melancholic depression (81.6%), with such decisions supported by validation analyses. CONCLUSIONS: In differentiating a melancholic subtype or syndrome, prototypic symptoms should be considered as a potential alternative to severity-based ratings.