Abstract

Objectives: To estimate the incidence of adverse events in medical and surgical activity in public and private hospitals, and to assess the clinical situation of patients and the active errors.

Design: Prospective assessment of adverse events by external senior nursing and doctor investigators with ward staff.

Setting: Random three-stage stratified cluster sampling of stays or fractions of stay in a 7-day observation period for each ward.

Participants: 8754 patients observed in 292 wards in 71 hospitals, over 35 234 hospitalisation days.

Main outcome measures: Number of adverse events in relation to number of days of hospitalisation.

Results: The incidence density of adverse events was 6.6 per 1000 days of hospitalisation (95% CI 5.7 to 7.5), of which 35% were preventable. Invasive procedures were the source of half the adverse events, of which 20% were preventable. Adverse events related to the psychological sphere and pain were mostly considered as preventable. Ward staff found it difficult to assess the role of care management in the occurrence of adverse events: 41% of adverse events were expected because of the disease itself, and could have occurred in the absence of the related medical management.

Conclusion: At the national level in France, every year 120 000–190 000 adverse events during hospitalisation can be considered as preventable. Areas such as perioperative period and geriatric units should receive closer attention. As adverse events occurred more commonly in vulnerable patients, who are not specifically targeted by clinical guidance, practising evidence-based medicine is not likely to prevent all cases. Therefore clinical risk management should prioritise empowerment of local staff, provision of favourable conditions within the organisation, and staff training based on simple tools appropriate for ward-level identification and analysis of adverse events.