Predicting blood transfusion following traumatic injury using machine learning models: A systematic review and narrative synthesis.

BACKGROUND: Haemorrhage is a leading cause of preventable death in trauma. Accurately predicting a patient's blood transfusion requirement is essential but can be difficult. Machine learning (ML) is a field of artificial intelligence that is emerging within medicine for accurate prediction modelling. This systematic review aimed to identify and evaluate all ML models that predict blood transfusion in trauma. METHODS: This systematic review was registered on The International Prospective register of Systematic Reviews (CRD4202237110). MEDLINE, Embase and the Cochrane Central Register of Controlled Trials were systematically searched. Publications reporting a ML model that predicted blood transfusion in injured adult patients were included. Data extraction and risk of bias assessment was performed using validated frameworks. Data was synthesised narratively due to significant heterogeneity. RESULTS: Twenty-five ML models for blood transfusion prediction in trauma were identified. Models incorporated diverse predictors and varied ML methodologies. Predictive performance was variable but eight models achieved excellent discrimination (AUROC >0.9) and nine models achieved good discrimination (AUROC >0.8) in internal validation. Only two models reported measures of calibration. Four models have been externally validated in prospective cohorts: the Bleeding Risk Index, Compensatory Reserve Index, the Marsden model and the Mina model. All studies were considered at high risk of bias often due to retrospective datasets, small sample size and lack of external validation. DISCUSSION: This review identified twenty-five ML models developed to predict blood transfusion requirement after injury. Seventeen ML models demonstrated good to excellent performance in-silico but only four models were externally validated. To date ML models demonstrate the potential for early and individualised blood transfusion prediction but further research is critically required to narrow the gap between ML model development and clinical application. LEVEL OF EVIDENCE: Systematic Review Without Meta-Analysis, Level IV.

Evaluating patient factors, operative management and postoperative outcomes in trauma laparotomy patients worldwide: a protocol for a global observational multicentre trauma study.

INTRODUCTION: Trauma contributes to the greatest loss of disability-adjusted life-years for adolescents and young adults worldwide. In the context of global abdominal trauma, the trauma laparotomy is the most commonly performed operation. Variation likely exists in how these patients are managed and their subsequent outcomes, yet very little global data on the topic currently exists. The objective of the GOAL-Trauma study is to evaluate both patient and injury factors for those undergoing trauma laparotomy, their clinical management and postoperative outcomes. METHODS: We describe a planned prospective multicentre observational cohort study of patients undergoing trauma laparotomy. We will include patients of all ages who present to hospital with a blunt or penetrating injury and undergo a trauma laparotomy within 5 days of presentation to the treating centre. The study will collect system, patient, process and outcome data, following patients up until 30 days postoperatively (or until discharge or death, whichever is first). Our sample size calculation suggests we will need to recruit 552 patients from approximately 150 recruiting centres. DISCUSSION: The GOAL-Trauma study will provide a global snapshot of the current management and outcomes for patients undergoing a trauma laparotomy. It will also provide insight into the variation seen in the time delays for receiving care, the disease and patient factors present, and patient outcomes. For current standards of trauma care to be improved worldwide, a greater understanding of the current state of trauma laparotomy care is paramount if appropriate interventions and targets are to be identified and implemented.

Defence Medical Services' REBOA training course.

Resuscitative Endovascular Balloon Occlusion of the Aorta (REBOA) enables temporary haemorrhage control and physiological stabilisation. This article describes the bespoke Defence Medical Services (DMS) training package for effectively using REBOA. The article covers how the course was designed, how the key learning objectives are taught, participant feedback and the authors' perceptions of future training challenges and opportunities. Since the inaugural training course in April 2019, the authors have delivered six courses, training over 100 clinicians. For the first time in the UK DMS, we designed and delivered a robust specialist endovascular training programme, with demonstrable, significant increases in confidence and competence. As a result of this course, the first DMS REBOA-equipped forward surgical teams deployed in June 2019. Looking to the future, there is a requirement to develop an assessment of skill retention and the potential need for revalidation.

Use of a modified Delphi process to develop research priorities in major trauma.

PURPOSE: The burden of major trauma within the UK is ever increasing. There is a need to establish research priorities within the field. Delphi methodology can be used to develop consensus opinion amongst a group of stakeholders. This can be used to prioritise clinically relevant, patient-centred research questions to guide future funding allocations. The aim of our study was to identify key future research priorities pertaining to the management of major trauma in the UK. METHODS: A three-phased modified Delphi process was undertaken. Phase 1 involved the submission of research questions by members of the trauma community using an online survey (Phase 1). Phases 2 and 3 involved two consecutive rounds of prioritisation after questions were subdivided into 6 subcategories: Brain Injury, Rehabilitation, Trauma in Older People, Pre-hospital, Interventional, and Miscellaneous (Phases 2 and 3). Cut-off points were agreed by consensus amongst the steering subcommittees. This established a final prioritised list of research questions. RESULTS: In phase 1, 201 questions were submitted by 65 stakeholders. After analysis and with consensus achieved, 186 questions were taken forward for prioritisation in phase 2 with 114 included in phase 3. 56 prioritised major trauma research questions across the 6 categories were identified with a clear focus on long-term patient outcomes. Research priorities across the patient pathway from roadside to rehabilitation were deemed of importance. CONCLUSIONS: Consensus within the major trauma community has identified 56 key research questions across 6 categories. Dissemination of these questions to funding bodies to allow for the development of high-quality research is now required. There is a clear indication for targeted multi-centre multi-disciplinary research in major trauma.

Early Identification of Trauma-induced Coagulopathy: Development and Validation of a Multivariable Risk Prediction Model.

OBJECTIVE: The aim of this study was to develop and validate a risk prediction tool for trauma-induced coagulopathy (TIC), to support early therapeutic decision-making. BACKGROUND: TIC exacerbates hemorrhage and is associated with higher morbidity and mortality. Early and aggressive treatment of TIC improves outcome. However, injured patients that develop TIC can be difficult to identify, which may compromise effective treatment. METHODS: A Bayesian Network (BN) prediction model was developed using domain knowledge of the causal mechanisms of TIC, and trained using data from 600 patients recruited into the Activation of Coagulation and Inflammation in Trauma (ACIT) study. Performance (discrimination, calibration, and accuracy) was tested using 10-fold cross-validation and externally validated on data from new patients recruited at 3 trauma centers. RESULTS: Rates of TIC in the derivation and validation cohorts were 11.8% and 11.0%, respectively. Patients who developed TIC were significantly more likely to die (54.0% vs 5.5%, P < 0.0001), require a massive blood transfusion (43.5% vs 1.1%, P < 0.0001), or require damage control surgery (55.8% vs 3.4%, P < 0.0001), than those with normal coagulation. In the development dataset, the 14-predictor BN accurately predicted this high-risk patient group: area under the receiver operating characteristic curve (AUROC) 0.93, calibration slope (CS) 0.96, brier score (BS) 0.06, and brier skill score (BSS) 0.40. The model maintained excellent performance in the validation population: AUROC 0.95, CS 1.22, BS 0.05, and BSS 0.46. CONCLUSIONS: A BN (http://www.traumamodels.com) can accurately predict the risk of TIC in an individual patient from standard admission clinical variables. This information may support early, accurate, and efficient activation of hemostatic resuscitation protocols.

Outcomes following trauma laparotomy for hypotensive trauma patients: A UK military and civilian perspective.

BACKGROUND: The management of trauma patients has changed radically in the last decade, and studies have shown overall improvements in survival. However, reduction in mortality for the many may obscure a lack of progress in some high-risk patients. We sought to examine the outcomes for hypotensive patients requiring laparotomy in UK military and civilian cohorts. METHODS: We undertook a review of two prospectively maintained trauma databases: the UK Joint Theatre Trauma Registry for the military cohort (February 4, 2003, to September 21, 2014) and the trauma registry of the Royal London Hospital major trauma center (January 1, 2012, to January 1, 2017) for civilian patients. Adults undergoing trauma laparotomy within 90 minutes of arrival at the emergency department (ED) were included. RESULTS: Hypotension was present on arrival at the ED in 155 (20.4%) of 761 military patients. Mortality was higher in hypotensive casualties (25.8% vs. 9.7% in normotensive casualties; p < 0.001). Hypotension was present on arrival at the ED in 63 (35.7%) of 176 civilian patients. Mortality was higher in hypotensive patients (47.6% vs. 12.4% in normotensive patients; p < 0.001). In both cohorts of hypotensive patients, neither the average injury severity, the prehospital time, the ED arrival systolic blood pressure, nor mortality rate changed significantly during the study period. CONCLUSIONS: Despite improvements in survival after trauma for patients overall, the mortality for patients undergoing laparotomy who arrive at the ED with hypotension has not changed and appears stubbornly resistant to all efforts. Specific enquiry and research should continue to be directed at this high-risk group of patients. LEVEL OF EVIDENCE: Prognostic/Epidemiologic, level IV.

British Military surgical key performance indicators: time for an update?

BACKGROUND: Key performance indicators (KPIs) are metrics that compare actual care against an ideal structure, process or outcome standard. KPIs designed to assess performance in deployed military surgical facilities have previously been published. This study aimed to review the overall performance of surgical trauma care for casualties treated at Role 3 Camp Bastion, Medical Treatment Facility, Afghanistan, in light of the existing Defence Medical Services (DMS) KPIs. The secondary aims were to assess the utility of the surgical KPIs and make recommendations for future surgical trauma care review. METHODS: Data on 22 surgical parameters were prospectively collected for 150 injured patients who had primary surgery at Camp Bastion between 1 May 2013 and 20 August 2013. Additional information for these patients was obtained using the Joint Theatre Trauma Register. The authors assessed data recording, applicability and compliance with the KPIs. RESULTS: Median data recording was 100% (IQR 98%-100%), median applicability was 56% (IQR 10%-99%) and median compliance was 78% (IQR 58%-93%). One KPI was not applicable to any patient in our population. Eleven KPIs achieved >80% compliance, five KPIs had 80%-60% compliance and five KPIs had <60% compliance. Recommendations are made for minor modifications to the current KPIs. CONCLUSION: 78% compliance with the DMS KPIs provides a snapshot of the performance of the surgical aspect of military trauma care in 2013. The KPIs highlight areas for improvement in service delivery. Individual KPI development should be driven by evidence and reflect advances in practice and knowledge. A method of stakeholder consultation, and sequential refinement following evidence review, may be the right process to develop the future set of DMS KPIs.

Scanning and War: Utility of FAST and CT in the Assessment of Battlefield Abdominal Trauma.

OBJECTIVE: To determine utilization and accuracy of focused assessment with sonography for trauma (FAST) and computed tomography (CT) in a mature military trauma system to inform service provision for future conflicts. BACKGROUND: FAST and CT scans undertaken by attending radiologists contribute to surgical decision making for battlefield casualties at the Joint Force, Role 3 Medical Treatment Facility at Camp Bastion (R3), Afghanistan. METHODS: Registry data for abdominally injured casualties treated at R3 from July to November 2012 were matched to radiological and surgical records to determine diagnostic accuracy for FAST and CT and their influence on casualty management. RESULTS: A total of 468 casualties met inclusion criteria, of whom 85.0% underwent FAST and 86.1% abdominal CT; 159 (34.0%) had abdominal injuries. For detection of intra-abdominal injury, FAST sensitivity (Sn) was 0.56, specificity (Sp) 0.98, positive predictive value (PPV) 0.87, negative predictive value (NPV) 0.90, and accuracy (Acc) 0.89. For CT, Sn was 0.99, Sp 0.99, PPV 0.96, NPV 1.00, and Acc 0.99. Forty-six solid organ injuries were identified in 38 patients by CT; 17 were managed nonoperatively. A further 61 patients avoided laparotomy after CT confirmed extra-abdominal wounds only. The negative laparotomy rate was 3.9%. CONCLUSIONS: FAST and CT contribute to triage, guide surgical management, and reduce nontherapeutic laparotomy. When imaging is available, these data challenge current doctrine about inadvisability of nonoperative management of abdominal injury after combat trauma.

Intrasaccular caecal adenocarcinoma presenting as enlarging right inguinoscrotal hernia.

INTRODUCTION: Colorectal cancer and inguinal hernias are both common surgical pathologies in the elderly but rarely co-exist. Where the conditions overlap, there can be difficulties in both diagnosis and treatment. PRESENTATION OF CASE: A 78 year old man with unexplained iron deficiency anaemia was investigated for gastrointestinal cancer. He was found to have enlarging bilateral inguinoscrotal hernias. CT colonoscopy revealed a herniated caecal carcinoma contained within the scrotum. An open excision was performed. DISCUSSION: Iron deficiency anaemia without obvious bleeding is associated with colorectal cancer and should be thoroughly investigated. Asymptomatic inguinal hernias are often ignored by patients. However, a change in an existing hernia may be associated with neoplasia. CONCLUSION: Dual pathology raises new difficulties in assessment, investigation and management of colorectal cancer.