La traduction française du DSM-5 ou Lost in translation

RésuméLa traduction française du DSM-5 sera publiée (Elsevier) en juin 2015. Cette communication résume l’historique des traductions françaises des DSM depuis le DSM-III en 1980. Nous sommes restés fidèles aux principes antérieurement retenus, à savoir traduire plutôt qu’adapter et respecter – sauf exceptions – le style de la version originale, avec ses répétitions systématiques. De nombreux exemples sont donnés des choix que nous avons effectués. La seconde partie concerne l’utilisation des quelque 65 outils psychométriques recommandés par les auteurs, en libre accès gratuit sur Internet, pour la plupart non traduits et non validés, pour lesquels les règles méthodologiques d’adaptation et de validation ne pourront pas être respectées.

The French translation of DSM-5 will be published by Elsevier Masson in June 2015. This paper summarizes the historical background of the French translations of DSM since 1980 (DSM-III). The translation team has remained true to the principles of previous translations, namely choosing scientific accuracy rather than literary adaptation, and respecting the unique style of the Handbook, with its systematic repetition of phrases and expressions referring to diagnoses. There were very few real problems of translation. Some terms, however, have been the subject of various translation proposals. Other terms have been translated differently depending on context, reflecting the fact that most English and French words are polysemic rather than univocal. Many examples are given of the choices we have made. The second part of the paper is about the emerging measures recommended in DSM-5 section III. These tools (more than 65 questionnaires, inventories or rating scales) can be accessed freely online. Their use is recommended to “enhance the clinical decision-making process, with level 1 cross-cutting self/informant related measures that serves as a review of systems across mental disorders, and level two disorder-specific severity measures”. Most of these assessment measures are not validated in French. The methodology for validating translated inventories or rating scales has been well known since the early sixties. And International guidelines have been recently published in this field. These validation studies are long, difficult and costly. They implicate a tremendous amount of work that will not be achieved for the tools recommended in the DSM-5. The Manual makes no reference to these problems. The validity in uncontrolled conditions of psychometric tools posted on the Internet has to be studied. The equivalence of administering an online version of rating scales or personality tests was assessed in several papers with contradictory results. These results need to be confirmed in the general population and in representative clinical samples.