Preventable trauma deaths: evaluation by peer review and a guide for quality improvement. Emergency Medical Study Group for Quality.

Abstract

Background: The trauma-injury severity-score (TRISS) methodology was developed in the United States to calculate the probability of survival for trauma patients presenting for emergency care.

Objective: We assessed the utility of using the TRISS methodology to identify preventable trauma deaths and compared the results to peer review using explicit standards.

Methods: Explicit peer review standards were developed by a focus group. The validity of these standards was evaluated by comparing the results of peer review performed by two independent expert panels. All trauma cases admitted to 10 centers in Japan between April 1, 1994, and March 31, 1996, were evaluated using the TRISS methodology. Cases with an expected probability of survival of more than 0.5 were considered preventable. These cases were subjected to peer review. Patients who were dead on arrival were excluded from analysis.

Results: Of 3,125 patients who were not dead on arrival, the TRISS methodology identified 2,525 as having a probability of survival greater than 0.5. In this group, 189 patients died; thus, 25.3% of all deaths were considered preventable by the TRISS method. Peer review found that only 11.2% of the deaths were preventable; thus, only 46.6% of preventable deaths identified by TRISS were confirmed by peer review. Agreement between the two expert peer review panels was very good (kappa = 0.62).

Conclusion: TRISS can be used as a screening tool to identify potentially preventable trauma deaths. Peer review is appropriate to confirm preventability and to identify potential medical errors.