Diagnostic value of the DSM and ICD categories of psychosis: an evidence-based approach

Abstract

Background: It is generally assumed that the distinction between affective and non-affective psychosis occasioned by modern diagnostic criteria provides a useful symptomatic contrast. Method: In a sample of 708 patients with chronic psychosis, the distinction of lifetime DSM-III-R and ICD-10 diagnoses of affective versus non-affective psychosis was used as a diagnostic test to detect lifetime presence of depressive, manic, positive, negative and disorganisation symptoms. Results: A manic or depressive affective diagnosis was a perfect test to diagnose the presence of manic and depressive symptoms, as evidenced by very high diagnostic likelihood ratios. However, this test result was based solely on the inclusion criterion that patients with affective psychosis must have affective symptoms (guaranteeing high specificity and high likelihood ratios), and ignored the fact that patients with non-affective psychosis also had high affective symptom scores (low sensitivity). Furthermore, a non-affective psychotic diagnosis was a very poor test to diagnose correctly the presence of positive, negative and disorganisation symptoms in comparison with an affective psychotic diagnosis. In general, the DSM-III-R categories performed somewhat better as a diagnostic test than those of ICD-10. Conclusion: The evidence for true diagnostic value of the distinction between affective and non-affective psychotic diagnoses is weak. Rather, the distinction appears to obscure natural overlap between the symptom dimensions of the different diagnostic categories.