The psychiatric classification systems ICD-10 and DSM-III-R were compared by applying both sets of diagnostic criteria to the same sample of patients suffering from affective and psychotic disorders. Four independent raters assessed diagnoses according to both systems to 100 written case records which had been prepared in a traditional, standard format. The International Diagnostic Checklists (IDCL) were employed to rate relevant psychopathological signs and symptoms and to apply diagnostic decision rules. The results showed that ICD-10 yielded a generally higher reliability for all main disorders except for bipolar disorder. Overall reliability was kappa = 0.53 for diagnoses according to DSM-III-R and 0.59 for diagnoses according to ICD-10. Agreement was best for affective disorders, moderate for schizophrenia and inacceptable for schizoaffective disorder. Insufficient boundaries were found in both systems between schizoaffective disorder on one side and schizophrenia and bipolar disorder on the other side. The different duration criteria for schizophrenia of six months in DSM-III-R and one month in ICD-10 tend to have considerable consequences for frequency rates of schizophrenia in a typical clinical setting.

No abstract

The compatibility of the classification systems ICD-10 and DSM-III-R was investigated in a sample of 100 psychiatric inpatients with severe affective and psychotic disorders. Four independent raters assessed diagnoses by means of checklists from psychopathological descriptions of case records. Congruence between ICD-10 and DSM-III-R was good for depressive disorders with agreement rates above 80% and fair for bipolar disorders with rates above 60%. Less consistent findings were obtained for schizophrenia with rates of 57.5% for ICD-10 diagnoses and 82.6% for DSM-III-R diagnoses. Agreement for schizoaffective disorder was below 40% and thus insufficient. The results indicate that even small differences in the definitions of mental disorders may result in considerable inconsistencies.

No abstract

No abstract