• Jacqueline V Mclennan, Kevin C Mackway-Jones, and Jason E Smith.
• Manchester University, Oxford Rd, Manchester, M13 9PL, UK; Academic Department of Military Emergency Medicine, Royal Centre for Defence Medicine (Academia & Research), Medical Directorate, ICT Centre, Birmingham Research Park, Vincent Drive, Edgbaston, Birmingham, B15 2SQ, UK; Royal Stoke University Hospital, University Hospital North Midlands, Stoke on Trent, Staffordshire, ST4 6QG, UK. Electronic address: Jackie.mclennan@uhnm.nhs.uk.
• Injury. 2018 Feb 1; 49 (2): 184-190.

BackgroundThe predominant cause of preventable trauma death is bleeding, and many of these patients need resuscitation with massive blood transfusion. In resource-constrained environments, early recognition of such patients can improve planning and reduce wastage of blood products. No existing decision rule is sufficiently reliable to predict those patients requiring massive blood transfusion. This study aims to produce a decision rule for use on arrival at hospital for patients sustaining battlefield trauma.MethodsA retrospective database analysis was undertaken using the UK Joint Theatre Trauma Registry to provide a derivation and validation dataset. Regression analysis of potential predictive factors was performed. Predictive factors were analysed through multi-logistic regression analysis to build predictive models; sensitivity and specificity of these models was assessed, and the best fit models were analysed in the validation dataset.ResultsA decision rule was produced using a combination of injury pattern, clinical observations and pre-hospital data. The proposed rule, using a score of 3 or greater, demonstrated a sensitivity of 82.7% and a specificity of 88.8% for prediction of massive blood transfusion, with an AUROC of 0.93 (95% CI 0.91-0.95).ConclusionsWe have produced a decision tool with improved accuracy compared to any previously described tools that can be used to predict blood transfusion requirements in the military deployed hospital environment.Copyright © 2017 Elsevier Ltd. All rights reserved.

21 keywords...

hide…