[The importance of classifications in psychiatry].

The classifications currently used in psychiatry have different aims: to facilitate communication between researchers and clinicians at national and international levels through the use of a common language, or at least a clearly and precisely defined nomenclature; to provide a nosographical reference system which can be used in practice (diagnosis, prognosis, treatment); to optimize research by ensuring that sample cases are as homogeneous as possible; to facilitate statistical records for public health institutions. A classification is of practical interest only if it is reliable, valid and acceptable to all potential users. In recent decades, there has been a considerable systematic and coordinated effort to improve the methodological approach to classification and categorization in the field of psychiatry, including attempts to create operational definitions, field trials of inter-assessor reliability, attempts to validate the selected nosological categories by analysis of correlation between progression, treatment response, family history and additional examinations. The introduction of glossaries, and particularly of diagnostic criteria, marked a decisive step in this new approach. The key problem remains that of the validity of diagnostic criteria. Ideally, these should be based on demonstrable etiologic or pathogenic data, but such information is rarely available in psychiatry. Current classifications rely on the use of extremely diverse elements in differing degrees: descriptive criteria, evolutive criteria, etiopathogenic criteria, psychopathogenic criteria, etc. Certain syndrome-based classifications such as DSM III and its successors aim to be atheoretical and pragmatic. Others, such as ICD-10, while more eclectic than the different versions of DSM, follow suit by abandoning the terms "disease" and "illness" in favor of the more consensual "disorder". The legitimacy of classifications in the field of psychiatry has been fiercely contested, being variously dubbed "a reductive academic exercise of no relevance to patients", "a dehumanizing labelling system, and a potential source of social and political violence", "a destructive prognostic guide", and so on. Other critics point to various aspects of certain classifications: the abandonment of theoretical concepts, the arbitrary nature of certain categories, the selection of definitions and criteria, the privileged position systematically accorded to the notion of category over that of general dimension. Multiaxial systems such as those proposed in successive versions of DSM or the classifications used in child psychiatry go some way towards meeting these criticisms. They go beyond simple labelling and place the patient in an overall medicopsycho-social setting. Nosographical indicators do not constitute an obstacle to psychopathological understanding. No classifications are capable of satisfactorily fulfilling all needs, namely those of daily practice, research and health statistics. The has led to the development of specialized diagnostic criteria and instruments, as in research for example. It should also be noted in this context that different versions of ICD-10 exist for psychiatrists, general practitioners, researchers and healthcare managers. The greatest danger posed by classifications is the potential reification of hypothetical approaches, arbitrary categorization and the dulling of reflection, all of which have created a need for regular revisions underpinned by field trials.