Improving risk adjustment in critically ill trauma patients:: The TRISS-SAPS score

Objectives: To test (a) the prognostic performance of TRISS and SAPS II scoring systems in a large sample of trauma patients admitted to Austrian ICUs, and (b) the hypothesis that the prognostic performance of TRISS could be improved by adding SAPS II. Methods: Prospective multicenter cohort study comprising 5,538 trauma patients out of 35,637 patients admitted to 31 ICUs in Austria over a 4-year period. Results: Separately, TRISS and SAPS II showed lack of calibration in the cohort of trauma patients. The database was then split into two equal samples, development and validation. Using the development sample, a new scoring system was developed, with vital status at hospital discharge as the dependent variable and TRISS and SAPS II as independent variables. The prognostic performance of the new TRISS-SAPS system was then assessed in the validation cohort: Both, discrimination (as shown by area under the ROC curve), and calibration (using Hosmer-Lemeshow goodness-of-fit statistics) was excellent. Conclusions: We improved risk adjustment in critically ill trauma patients by combining TRISS and SAPS II. This new scoring system might aid in evaluating and comparing specialized trauma ICUs.