ObjectiveTo assess the ability of the Acute Physiology and Chronic Health Evaluation (APACHE II) system and Trauma-Injury Severity Scoring (TRISS) system in predicting group mortality in intensive care unit (ICU) trauma patients.DesignProspective study.SettingA Canadian adult trauma tertiary referral hospital.PatientsConsecutive trauma patients admitted to the medical-surgical ICU or the neurosurgical ICU.InterventionNone.Measurements and Main ResultsFor each patient, demographic data, mechanism of injury, and surgical status were collected. Revised Trauma Scores and Injury Severity Scores were calculated from emergency room and operative data. The APACHE II score was calculated based on the data from the first 24 hrs of ICU admission. The probability of death was calculated for each patient based on the APACHE II and TRISS equations. The ability to predict group mortality for APACHE II and TRISS was assessed by receiver operating characteristic curve analysis, two by two decision matrices, and calibration curve analysis.Four hundred seventy trauma patients were admitted to the ICU.Sixty-three (13%) patients died and 407 (87%) survived. There were significant differences between survivors and nonsurvivors in age, Glasgow Coma Scale, Revised Trauma Score, Injury Severity Score, and APACHE II score. By receiver operating characteristic curve analysis, the areas under the curves (+/- SEM) of APACHE II and TRISS were 0.92 +/- 0.02 and 0.89 +/- 0.02, respectively. Using two by two decision matrices with a decision criterion of 0.5, the sensitivities, specificities, and percentages correctly classified were 50.8%, 97.3%, and 91.1%, respectively, for APACHE II, and 50.8%, 97.1%, and 90.9%, respectively, for TRISS. From the calibration curves, the r2value was .93 (p = .0001) for APACHE II and .67 (p = .004) for TRISS.ConclusionsBoth APACHE II and TRISS scores were shown to accurately predict group mortality in ICU trauma patients. APACHE II and TRISS may be utilized for quality assurance in ICU trauma patients. However, neither APACHE II nor TRISS provides sufficient confidence for prediction of outcome of individual patients.(Crit Care Med 1996; 24:1642-1648)