Performance of Trauma and Injury Severity Score (TRISS) adjustments: an integrative review

Objective: Identify studies that made adjustments to the equation of Trauma and Injury Severity Score (TRISS) and compared the discriminatory ability of both modified and original equations. Method: An integrative review of studies published between 1990 and 2014 using the following databases: LILACS, MEDLINE, PubMed and SciELO, based on searches using the term "TRISS". Results: 32 studies were included in this review. Of 67 adjustments to TRISS equations identified, 35 (52.2%) resulted in improved accuracy of this index in the prediction of survival probability for trauma patients. Adjustments of TRISS coefficients to study population were frequent, but did not always improve the predictive ability of the analyzed models. Replacement of physiological variables of the Revised Trauma Score (RTS) and changes in the Injury Severity Score (ISS) in the original equation presented varied performance. An alteration to the method of age inclusion in the equation, and the insertion of gender, comorbidities and trauma mechanism, presented a tendency towards improved performance of the TRISS. Conclusion: Different proposals of adjustments to the TRISS were identified in this review and indicated, in particular, RTS fragilities in the original model and the need to change the method of age inclusion in the equation to improve the predictive ability of this index.