[Comparison between DSM III and CIM 9 in the data processing on computerized system. Study of 4000 records]

The authors expose the results of a confrontation between ICD 9 and the D.S.M. III's classification published by APA. This work is the issue of a collected data system'use (S.D.P.A.) which is a basic tool for epidemiological programm, applied in military psychiatrics department (actually 18 services have got computer sets). It is evident that military structure (centralized dans homogeneous), and the size of the samples, are good conditions for psychiatric epidemiology. The authors recall the different stages and the difficulties on the field, for such a wide scale application. The use of two principal world-known classifications raise up unanticiped remarks: indeed, we consider the logical structure of D.S.M. III's axis I as better than ICD 9 classification. Meanwhile, the axis III, IV and V caused a disappointment, for both clinicians and programm's responsibles. In fact, there were some statistical biases and a little gain of informations. The former system (applied for 1976) was more efficient. Some hypothesis about that insufficient acceptability and reliability are proposed as also an important project for system's modifications from 1989.