Are the bipolar disorders best modelled categorically or dimensionally?

Considerable debate exists as to whether the bipolar disorders are best classified according to a categorical or dimensional model. This study explored whether there is evidence for a single or multiple subpopulations and the degree to which differing diagnostic criteria correspond to bipolar subpopulations.

A mixture analysis was performed on 1081 clinically diagnosed (and a reduced sample of 497 DSM-IV diagnosed) bipolar I and II disorder patients, using scores on hypomanic severity (as measured by the Mood Swings Questionnaire). Mixture analyses were conducted using two differing diagnostic criteria and two DSM markers to ascertain the most differentiating and their associated clinical features.

The two subpopulation solution was most supported although the entropy statistic indicated limited separation and there was no distinctive point of rarity. Quantification by the odds ratio statistic indicated that the clinical diagnosis (respecting DSM-IV criteria, but ignoring 'high' duration) was somewhat superior to DSM-IV diagnosis in allocating patients to the putative mixture analysis groups. The most differentiating correlate was the presence or absence of psychotic features.

Findings favour the categorical distinction of bipolar I and II disorders and argue for the centrality of the presence or absence of psychotic features to subgroup differentiation.