BACKGROUND: To investigate whether diagnostic agreement across different diagnostic systems improves in a sample of chronic patients suffering from functional psychosis compared to first-admitted patients.

SAMPLING AND METHODS: Among 353 patients with a history of functional psychosis, a subset of 100 individuals (35 women and 65 men) were randomly sampled and assessed using the Operational Criteria Checklist for Psychotic Illness and Affective Illness (OPCRIT). Based on the OPCRIT diagnoses the subjects suffering from schizophrenia and schizophrenia spectrum disorders according to seven diagnostic systems were identified. Diagnostic agreement was assessed using unweighted kappa-statistics and pairwise concordance rates (CR).

RESULTS: High diagnostic agreement of schizophrenia was observed across the ICD-10 and DSM systems (CR >0.70, kappa >0.70), which all had a significantly lower concordance to the St. Louis Criteria (SLC), research diagnostic criteria and Schneider's first rank symptoms (FRS) (0.32< CR <0.66; -0.10< kappa <0.51). Agreement on schizophrenia across all systems was observed for one fourth of the subjects. Elimination of the diagnostic impact of 'co-occurrence of psychotic and affective symptoms' excluded FRS standalone individuals from the sample, increased overall homogeneity and resulted in a dichotomized sample according to SLC (46 positive vs. 47 negative). SLC status could be predicted in 78% of cases by four items relating to family history and psychosocial function previous to the onset of illness. Similarly high pairwise CR were observed for schizophrenia spectrum disorders across all diagnostic systems.

CONCLUSIONS: This study demonstrates that diagnostic agreement is higher among chronic patients than that observed in subjects with a recent onset of psychosis, although considerable discordance is also observed in this chronic sample. However, the discordance among chronic patients with functional psychosis largely derives from the different emphasis that diagnostic systems place on co-occurrence of psychotic and affective symptoms. This may have serious epistemological consequences, thus underlining the conventional nature of the present schizophrenia diagnoses and the need for biologically founded diagnostic criteria.