[Organizational and operational work of death analysis committees: review of experiences]

Among deaths occurring in hospital, some are unexpected at the time of admission and unexplained by the course of the illness at the time of death. Health care givers agree that analysis of the circumstances of death is a powerful method for improving the quality of medical care but are nevertheless reluctant to set up mortality conferences. We present institution-wide experience with mortality conferences, detailing methods used and reviewing adverse care events and avoidable organizational and medical errors associated with unexpected deaths.

We reviewed literature cited in the Medline and Pascal Biomed databases and completed our findings with consultations with key-informants of hospital evaluation units. Institution-wide experiences alone were analyzed.

The death analysis committees were composed of health care givers who review the deaths occurring in their units (internal committees). In rare cases, hospitals set up in addition an external committee for methodological assistance. Implicit criteria were almost always used for death case analysis.

Death analysis by internal committees appears to contribute to physician's knowledge and provokes a global improvement in patient care. This analysis relies however on implicit criteria which leads to minimal reliability. The lack of reliability has two consequences: the impossibility of identifying causes of preventable deaths and related factors, and the impossibility of evaluating the impact of the internal committee. As there is no available epidemiological data, a sentinel system cannot be proposed: a systematic analysis of all deaths is therefore advisable. Other limitations on internal committees are the difficulty of obtaining autopsies and the problem of the confidentiality of death analysis, a serious handicap recognized by all physicians. A national guide on methodology should be developed.