Back to search

University of Peloponesse

Προγνωστικά εργαλεία επιβίωσης σε ασθενείς με τραύμα

Abstract

dc:description

Trauma prediction models are well-established tools of modern comprehensive trauma systems to credibly adjust the risk of mortality and evaluate the quality of trauma care. However, great caution should be exercised when applying models in different trauma population than that being original studied on account of clear evidence of deterioration in their performance.The aim of this study was to externally validate the Trauma and Injury Severity Score (TRISS) and the International Classification of Diseases - Injury Severity Score (ICISS) and compare their prognostic performance with local versions derived from Greek trauma population. In addition, we compare the performance of various anatomic and physiological trauma - scoring systems to determine which of these most accurately predicts in-hospital mortality in our setting.This is a retrospective cohort study of admitted trauma patients conducted in two tertiary Greek hospitals from January 2015 to December 2019. Patient identification was performed by searching through the hospital information database for admittance and discharge diagnoses consistent with injury as given by the 10th version of International Classification of Diseases. Data were retrospectively extracted from medical records of eligible patients and split temporally or geographically into derivation and validation cohort. The probability of survival for each trauma patient was calculated using the original and revised coefficients for TRISS or survival risk ratios for ICISS and their local counterparts derived from derivation cohort. The patients were also assigned a value based on Injury Severity Score (ISS), New Injury Severity Score (NISS), Exponential Injury Severity Score (EISS), Revised Trauma Score (RTS), New Trauma Score (NTS) and Mechanism - Glasgow Coma Scale, Age, and Systolic Blood Pressure (M-GAP). The primary outcome was in-hospital mortality. Models’ prediction was performed using discrimination and calibration statistics.The original TRISS displayed excellent discrimination with an area under the curve (AUC) of 0,912 (95% CI 0,895 - 0,929) and comparable with its revised version (AUC = 0,908, 95% CI 0,889 - 0,926, p = 0,4757) in the validation cohort. Calibration of both versions was poor (Hosmer–Lemeshow test p < .001), tending to underestimate the probability of mortality across almost all risk groups. The AUC for ICISS was 0,817 (95% CI 0,798 - 0,837) in the validation cohort. Calibration was poor using the Hosmer-Lemeshow test (p < .05), but visual analysis showed accurate prediction at patients with low mortality risk, especially below 30%. Recalibration of both models resulted in substantial improvements in discrimination (AUC = 0,929, 95% CI 0,914 - 0,944, p < .001 for TRISS και AUC = 0,834, 95% CI 0,814 - 0,854, p < .001 for ICISS) and excellent calibration. Modification of both models was also conferred additional benefit in their performance. The GAP discriminated and calibrated better than others novel and traditional physiological trauma-scoring systems (AUC = 0,870 95% CI 0,853 - 0,886 vs AUC = 0,862 95% CI 0,845 - 0,880 for MGAP, AUC = 0,851 95% CI 0,830 - 0,872 for NTS και AUC = 0,795 95% CI 0,771 - 0,820 for RTS). Among anatomical models ISS (AUC = 0,899 95% CI 0,886 - 0,913) perform as well as NISS (AUC = 0,895 95% CI 0,882 - 0,909) and EISS (AUC = 0,894 95% CI 0,880 - 0,908).The study performs an external validation of nine injury scoring systems for prediction of mortality in a Greek trauma population. The optimal choice of physiological or anatomical models should be able to strike a balance between accuracy in quantifying injury severity and feasibility in implementation. Therefore, there is a need for national consensus on which model is most appropriate in predicting mortality in our country.

Degree

thesis:*
Grantor dc:publisher
University of Peloponesse
Year dc:date
2022

Author and committee

dc:creator, dc:contributor.*
Authors dc:creator
  • Filippatos, Georgios
  • Φιλιππάτος, Γεώργιος

Subjects

dc:subject × 12

Rights

Language dc:language
gre

Identifiers

dc:identifier.*
Identifier
10.12681/eadd/52570

Chain of custody

source
Harvested from
Greek National Archive of PhD Theses
Base URL
phdtheses.ekt.gr/eadd_oai/request
Last updated
2026-11-18
Source record
OAI-PMH GetRecord
citation

Filippatos, Georgios; Φιλιππάτος, Γεώργιος. Προγνωστικά εργαλεία επιβίωσης σε ασθενείς με τραύμα. University of Peloponesse, 2022. http://hdl.handle.net/10442/hedi/52570