Background. The usefulness of any diagnostic scheme is directly related to its ability to provide clinically useful information on need for care. In this study, the clinical usefulness of dimensional and categorical representations of psychotic psychopathology were compared.

Method. A total of 706 patients aged 16–65 years with chronic psychosis were recruited. Psychopathology was measured with the Comprehensive Psychopathological Rating Scale (CPRS). Lifetime RDC, DSM-III-R, and ICD-10 diagnoses and ratings of lifetime psychopathology were made using OPCRIT. Other clinical measures included: (i) need for care; (ii) quality of life; (iii) social disability; (iv) satisfaction with services; (v) abnormal movements; (vi) brief neuropsychological screen; and (vii) over the last 2 years – illness course, symptom severity, employment, medication use, self-harm, time in hospital and living independently.

Results. Principal component factor analysis of the 65 CPRS items on cross-sectional psychopathology yielded four dimensions of positive, negative, depressive and manic symptoms. Regression models comparing the relative contributions of dimensional and categorical representations of psychopathology with clinical measures consistently indicated strong and significant effects of psychopathological dimensions over and above any effect of their categorical counterparts, whereas the reverse did not hold. The effect of psychopathological dimensions was mostly cumulative: high ratings on more than one dimension increased the contribution to the clinical measures in a dose-response fashion. Similar results were obtained with psychopathological dimensions derived from lifetime psychopathology ratings using the OCCPI.

Conclusions. A dimensional approach towards classification of psychotic illness offers important clinical advantages.