In the present study, all patients who met the diagnostic criteria for a long-term functional psychosis (LFP) were identified within a defined uptake area in the northern part of the county of Uppsala, Sweden. LFP includes patients 1) with productive psychotic symptomatology, not caused by organic disease, for 1 week or more, at least once during the course of the illness; 2) having been affected by a psychosis for a continuous period of at least 6 months on the same occasion; 3) having shown psychotic features or residual symptoms during the index year; and 4) older than 18 years of age. Primarily, all diagnoses were made according to the Diagnostic and Statistical Manual of Mental Disorders (DSM)-III-R. The prevalence of schizophrenia was 4.2 per 1,000 inhabitants. The prevalence of schizoaffective disorder was 0.7 per 1,000 inhabitants and for delusional disorder, 0.1 per 1,000 inhabitants. When the patients were rediagnosed according to DSM-III, DSM-IV and International Statistical Classification of Disease (ICD)-10, it was found that the prevalence of schizophrenia, schizoaffective disorder and delusional disorder was somewhat lower according to the DSM-III criteria, while the same number of patients fulfilled the criteria according to DSM-IV. If ICD-10 was used, it resulted in a broader concept of schizophrenia and a somewhat more narrow concept of schizoaffective disorder. Thus, the introduction of the new parallel diagnostic systems, ICD-10 and DSM-IV, will result in different, but comparable, prevalence estimates concerning schizophrenia, schizoaffective disorder and delusional disorder.