Revision of TRISS for intubated patients

The TRISS system is an important, widely used method for predicting survival in trauma patients. One significant shortcoming of TRISS is its inability to include intubated patients in survival analysis because a respiratory rate and a verbal response are not obtainable. This report describes one approach to this problem. Data from 994 patients with blunt trauma were examined. Like TRISS, survival probability was calculated using a logistic regression model that included age and Injury Severity Score (ISS); however, the best motor response and systolic blood pressure were used in place of the Revised Trauma Score (RTS). With this model, the sensitivity, specificity, and misclassification rate were 57%, 98.9%, and 3.6%, respectively. For TRISS, the sensitivity, specificity, and misclassification rate are 58.8%, 99.3%, and 3.0%, respectively. Thus, our model has predictive performance comparable with TRISS. More importantly, it is applicable to intubated patients who are not pharmacologically paralyzed. Further investigation with larger data bases is necessary.