Mortality prediction by 'Life Threat Index' compared to widely used trauma scoring systems.

INTRODUCTION AND OBJECTIVE: Injuries are an extremely important problem in public life and account for up to one-third of deaths in the entire human population. Available trauma scoring systems provide a good estimation of mortality; however, some factors affect their utility in daily practice. Therefore, a new easily applicable in any given country trauma scoring system has been developed and proposed in this study.

MATERIAL AND METHODS: A retrospective study of the medical records of 485 patients were evaluated, together with diagnostic performance with regard to mortality, was calculated for the Revised Trauma Score (RTS), Injury Severity Score (ISS), The New Injury Severity Score (NISS), Trauma and Injury Severity Score (TRISS), International Classification based Injury Severity Score (ICISS) and the newly-developed Life Threat Index (LTI). Sensitivity, specificity, accuracy, PPV and NPV were calculated for each scoring system, and overall diagnostic performance was estimated with the use of ROC curves.

RESULTS: Apart from RTS, all scoring systems showed similar performance regarding mortality prediction. TRISS and LTI showed the highest sensitivity reaching 0.998 and AUC of 0.89 and 0.87, respectively, which proved its usefulness in predicting mortality.

CONCLUSIONS: LTI proved to be one of the most sensitive in comparison with widely-used and recognized trauma scoring systems. Based on LTI methodology, it can be applied in any given country or region, even without a previously developed trauma database.