Les auteurs décrivent le déroulement et les résultats de l’application de l’audit clinique à une question relative au programme national d’actions face au suicide : la prise en charge hospitalière des jeunes suicidants. L’audit a été proposé par l’ANAES à des établissements volontaires pour évaluer leurs pratiques et les améliorer au regard des recommandations. Soixante-seize établissements ont recueilli 1 554 observations, 17 ont répété la mesure des mêmes critères 2 ans après. On observe une amélioration des résultats pour les critères qui étaient peu respectés au premier tour ; cette amélioration est cependant concentrée sur quelques établissements. Globalement les établissements respectent les recommandations concernant l’organisation de la prise en charge pendant le séjour, mais réalisent insuffisamment l’évaluation de l’environnement du suicidant et la préparation de sa sortie. La méthode a été bien accueillie, cependant l’expérience montre qu’elle doit être utilisée de façon rigoureuse avec un accompagnement méthodologique et un calendrier structuré.

We describe the implementation and results of a clinical audit of the management of suicidal adolescents in hospital that was carried out as part of the French national suicide prevention programme. The ANAES definition of a clinical audit is : « an assessment method using defined criteria that compares care practices with accepted standards and measures the quality of these practices and the results of care with a view to improving them ». Standards for the audit were derived from the clinical practice guidelines published by ANAES (French National Agency for Accreditation and Évaluation in Health) in November 1998. They comprised 15 criteria covering ressources available and procedures implemented : patient management in hospital (9 criteria), contacts made with the patient’s environment outside hospital (2 criteria) and preparation for discharge from hospital (4 criteria). Participation in the audit was voluntary. Overall, 76 hospitals from the 10 regions of France with a suicide prevention programme took part in the audit and made 1 554 observations. The number of observations per hospital ranged from 1 to 42. Compliance with the criteria was > 80 % for 5 criteria, 50-80 % for 3 criteria, and ≪ 50 % for 7 criteria. A total of 26 hospitals proposed a structured improvement plan, ie, scheduled, ordered and ranked measures with a definition of responsibilities and follow-up. Of these 26 hospitals, 15 had implemented the three ANAES recommendations (setting up a working group for the project, using a grid to analyse results, drafting a structured report) whereas only 5 of the 50 hospitals that did not come up with an improvement plan had done so. An operational outcome thus seems related to compliance with the audit method. Three years after the audit was set up, 17 hospitals took part in a second round. Improvements were noted for 12 criteria, mainly for those giving poor results in the first round. However, these improvements concerned few hospitals (eg just 4 hospitals for the criterion with the worst result in the first round). A worsening of compliance was noted for 3 criteria. In conclusion, hospitals were found to comply with guidelines relating to the management of suicidal adolescents in hospital. However, compliance with guidelines relating to making contact with the patient’s environment outside hospital and preparing for discharge was less good. Although the hospitals taking part in the audit endorsed our method, our experience showed that, to be effective, clinical audits need methodological support and a well-defined time schedule.