Objective-Long term evaluation of a quality assurance programme (after an assessment in 1993). Design-Review of medical records. Setting-Emergency area of an orthopaedic, trauma, and plastic surgery unit in a French teaching hospital (Besançon). Subjects-1187 consecutive ambulatory patients' records, from July 1995. Main measures-Occurrence of near adverse events (at risk events causing situations which could lead to the occurrence of an adverse event). Results-71 near adverse events were identified (5.9% of the ambulatory visits). There was a significant decrease in the rate of near adverse events between 1993 (9.9% (2056 ambulatory visits, 204 near adverse events)), and 1995 (5.9% (1187 ambulatory visits, 71 near adverse events)), and significant change in the proportion of each category of adverse event (decrease in departures from prevention protocols). Conclusions-Despite their limitations, the eVectiveness and eYciency of quality assurance programmes seem to be real and valuable. Maintaining quality improvement requires conditions which include some of the basic principles of total quality management (leadership, participatory management, openness, continuous feed back). The organisation of this unit as a specialised trauma centre was also a determining factor in the feasibility of a quality assurance programme (specialisation and small size, high activity volume, management of the complete care process). Quality assurance is an important initial step towards quality improvement, that should precede consideration of a total quality management programme.