On the way to expert systems

Publisher Website
Google Scholar
Abstract
Operationalized diagnostics deal with the standardized assessment of psychiatric symptoms as well as diagnostic criteria. As a diagnostic system based on criteria, the DSM-III was chosen to identify operationalized diagnoses based on the Present State Examination (PSE-9) and some additional DSM-III specific items. By relating PSE symptoms to the diagnostic criteria of DSM-III, an easily applicable expert system leading to DSM-III diagnoses was developed. In two samples of 30 schizophrenic and 51 depressive patients the DSM-III computer diagnoses are contrasted with the ICD-8 diagnoses of the PSE/CATEGO system. In defining a “case”, only minimal differences between the two computer programs were found. In the sample of schizophrenics, CATEGO led to 114 (81%) diagnoses and the DSM-III program to 112 (79%) diagnoses; for the depressive patients 43% cases were identified by CATEGO and 45% by the DSM-III algorithm. Comparing the diagnosis of “acute schizophrenic disorders”, both programs arrived at similar percentages. (CATEGO: 51%; DSM-III: 57%). However, CATEGO is limited to two different subtypes (295.2 and 295.3), whereas the DSM-III program covers the total range of possible schizophrenic subtypes. Furthermore, the DSM-III program identified residual subtypes of schizophrenia in 23% of the diagnostic decisions. In the short-term course of the schizophrenic patients, CATEGO identified 27%–43% with affective diagnoses with high stability per cross-section. Using the DSM-III algorithm affective diagnoses were rather rare (maximum of 17%), marking unstable changes from acute to residual states of the psychosis. In the cohort of affective patients the correspondence between both programs was quite good, especially for the patients with a diagnosis of depression, but in total the DSM-III program requires stricter criteria for affective — especially manic — disorders, whereas CATEGO needs a somewhat higher symptom level for anxiety syndromes to reach the diagnostic threshold.