The TRISS method of auditing trauma deaths necessitates audit of patients with minor injuries who die of their underlying medical problems. Using an anatomic definition of injury as a criterion for audit, as suggested by Wesson et al. at the Hospital for Sick Children in Toronto, excludes patients with minor injuries but necessitates audit of patients who expired due to systems problems rather than in-hospital patient care. We propose combining the TRISS and Toronto methods in order to identify the deaths truly appropriate for detailed review of hospital care. Fifty-four trauma deaths over a 22-month period were audited and categorized as frankly preventable, potentially salvageable, or nonpreventable. Considering only in-hospital care, the deaths designated as potentially salvageable by audit were likely to be identified by both TRISS and Toronto, while deaths targeted by only one system were more likely to be nonpreventable by audit. The predictive value of this combination of methods (84.6%) was better than Toronto (52.4%) or TRISS (54.5%) using audit results as the standard for comparison. This simple computerized method may serve as a practical and inexpensive method of targeting deaths for in-depth review.