Abstract
Background: Dissatisfaction with the usefulness of categorical diagnosis as a way of describing the phenomena of psychotic disorders has spawned alternative characterisations. Dimensional representations of “domains of psychopathology” have been the most popular. Positive and negative symptom domains in schizophrenia came first but more comprehensive dimensional measures in more domains have been developed for widely defined psychotic illness.
Aim: To compare the utility of categorical and dimensional representations of psychosis in terms of their ability to predict disability and service needs.
Method: The symptom data from the National Low Prevalence Disorders survey yielded algorithm-derived categorical diagnoses and factor analysis-derived dimensional scores in five domains of psychopathology (described previously). Univariate Analysis of Variance measured the variance in a number of need and disability indicators which was explained by categorical diagnostic systems (ICD, DSM etc) and by the domain scores.
Results: Dimensional scores of dysphoria, positive symptoms, elevation, negative symptoms and substance abuse were superior to categorical diagnosis in predicting service demand, difficult and desperate behaviour, antecedent adaption, personal care, and social and occupational functioning. Only use of support services and illness course were better predicated by any of the categorical diagnoses.
Conclusion: On simple and prosaic measures of utility, dimensional scores of domains of psychopathology perform better than categorical diagnoses. These results add weight to the arguments for the use of dimensional over categorical representations of psychotic disorders.