Defining melancholia: properties of a refined sign-based measure.

We hypothesised that psychomotor disturbance is specific to the melancholic subtype of depression and capable of defining melancholia more precisely than symptom-based criteria sets. We studied 413 depressed patients, and examined the utility of a refined, operationally driven set of clinician-rated signs, principally against a set of historically accepted symptoms of endogeneity. We specified items defining psychomotor disturbance generally as well as those weighted either to agitation or to retardation. We demonstrated the system's capacity to differentiate 'melancholic' and 'non-melancholic' depression (and the comparable success of DSM-III-R and Newcastle criteria systems) by reference to several patient, illness and treatment response variables, to an independent measure of psychomotor disturbance (reaction time) and to a biological marker (the dexamethasone suppression test).