[DSM IV and training: the limits]

Since its third edition, DSM has been considered to be an excellent tool for psychiatric research. The primary objective of this classificatory system was to put forward internationally accepted standard definitions. DSM diagnostic criteria are now indispensable for any publication in the scientific literature. It appears however that this work has gradually lost sight of its initial objective and is used as an educational tool for training of clinicians. What are the limits and risks of such a use? Can the DSM IV philosophy be reconciled with the objectives of training? Are the criteria in force for the selection of homogeneous patients groups identical to those which enable knowledge acquisition required for identification of disorders and their treatment? What is the heuristic value of enumerating symptoms and syndromes isolated from any theoretical context? Can symptoms be separated from the patient's history and personality? Is the excessive use of concurrent disorders not likely to be a source of conceptual and therapeutic inflation? Is a purely descriptive approach to psychiatric disorders not likely to run the risk of overestimating them? The points are discussed in succession by the authors.