Dataset Information

Development and validation of a prehospital-stage prediction tool for traumatic brain injury: a multicentre retrospective cohort study in Korea.

ABSTRACT:

Objectives

Predicting diagnosis and prognosis of traumatic brain injury (TBI) at the prehospital stage is challenging; however, using comprehensive prehospital information and machine learning may improve the performance of the predictive model. We developed and tested predictive models for TBI that use machine learning algorithms using information that can be obtained in the prehospital stage.

Design

This was a multicentre retrospective study.

Setting and participants

This study was conducted at three tertiary academic emergency departments (EDs) located in an urban area of South Korea. The data from adult patients with severe trauma who were assessed by emergency medical service providers and transported to three participating hospitals between 2014 to 2018 were analysed.

Results

We developed and tested five machine learning algorithms-logistic regression analyses, extreme gradient boosting, support vector machine, random forest and elastic net (EN)-to predict TBI, TBI with intracranial haemorrhage or injury (TBI-I), TBI with ED or admission result of admission or transferred (TBI with non-discharge (TBI-ND)) and TBI with ED or admission result of death (TBI-D). A total of 1169 patients were included in the final analysis, and the proportions of TBI, TBI-I, TBI-ND and TBI-D were 24.0%, 21.5%, 21.3% and 3.7%, respectively. The EN model yielded an area under receiver-operator curve of 0.799 for TBI, 0.844 for TBI-I, 0.811 for TBI-ND and 0.871 for TBI-D. The EN model also yielded the highest specificity and significant reclassification improvement. Variables related to loss of consciousness, Glasgow Coma Scale and light reflex were the three most important variables to predict all outcomes.

Conclusion

Our results inform the diagnosis and prognosis of TBI. Machine learning models resulted in significant performance improvement over that with logistic regression analyses, and the best performing model was EN.

SUBMITTER: Choi Y

PROVIDER: S-EPMC8756263 | biostudies-literature |

REPOSITORIES: biostudies-literature

ACCESS DATA

Similar Datasets

Project description:ImportanceTraumatic brain injury (TBI) is a massive public health problem. While evidence-based guidelines directing the prehospital treatment of TBI have been promulgated, to our knowledge, no studies have assessed their association with survival.ObjectiveTo evaluate the association of implementing the nationally vetted, evidence-based, prehospital treatment guidelines with outcomes in moderate, severe, and critical TBI.Design, setting, and participantsThe Excellence in Prehospital Injury Care (EPIC) Study included more than 130 emergency medical services systems/agencies throughout Arizona. This was a statewide, multisystem, intention-to-treat study using a before/after controlled design with patients with moderate to critically severe TBI (US Centers for Disease Control and Prevention Barell Matrix-Type 1 and/or Abbreviated Injury Scale Head region severity ≥3) transported to trauma centers between January 1, 2007, and June 30, 2015. Data were analyzed between October 25, 2017, and February 22, 2019.InterventionsImplementation of the prehospital TBI guidelines emphasizing avoidance/treatment of hypoxia, prevention/correction of hyperventilation, and avoidance/treatment of hypotension.Main outcomes and measuresPrimary: survival to hospital discharge; secondary: survival to hospital admission.ResultsOf the included patients, the median age was 45 years, 14 666 (67.1%) were men, 7181 (32.9%) were women; 16 408 (75.1% ) were white, 1400 (6.4%) were Native American, 743 (3.4% ) were Black, 237 (1.1%) were Asian, and 2791 (12.8%) were other race/ethnicity. Of the included patients, 21 852 met inclusion criteria for analysis (preimplementation phase [P1]: 15 228; postimplementation [P3]: 6624). The primary analysis (P3 vs P1) revealed an adjusted odds ratio (aOR) of 1.06 (95% CI, 0.93-1.21; P = .40) for survival to hospital discharge. The aOR was 1.70 (95% CI, 1.38-2.09; P < .001) for survival to hospital admission. Among the severe injury cohorts (but not moderate or critical), guideline implementation was significantly associated with survival to discharge (Regional Severity Score-Head 3-4: aOR, 2.03; 95% CI, 1.52-2.72; P < .001; Injury Severity Score 16-24: aOR, 1.61; 95% CI, 1.07-2.48; P = .02). This was also true for survival to discharge among the severe, intubated subgroups (Regional Severity Score-Head 3-4: aOR, 3.14; 95% CI, 1.65-5.98; P < .001; Injury Severity Score 16-24: aOR, 3.28; 95% CI, 1.19-11.34; P = .02).Conclusions and relevanceStatewide implementation of the prehospital TBI guidelines was not associated with significant improvement in overall survival to hospital discharge (across the entire, combined moderate to critical injury spectrum). However, adjusted survival doubled among patients with severe TBI and tripled in the severe, intubated cohort. Furthermore, guideline implementation was significantly associated with survival to hospital admission. These findings support the widespread implementation of the prehospital TBI treatment guidelines.Trial registrationClinicalTrials.gov identifier: NCT01339702.

Project description:BackgroundChanges in the epidemiology of traumatic brain injury (TBI) in older patients have received attention, but limited data are available on the outcome of these patients after admission to intensive care units (ICUs). The aim of this study was to evaluate the outcomes of patients over 65 years of age who were admitted to an ICU for TBI.MethodsThis was a multicentre, retrospective, observational study conducted from January 2013 to February 2019 in the surgical ICUs of 5 level 1 trauma centres in France. Patients aged ≥ 65 years who were hospitalized in the ICU for TBI with or without extracranial injuries were included. The main objective was to determine the risk factors for unfavourable neurological outcome at 3 months defined as an Extended Glasgow Outcome Scale (GOSE) score < 5.ResultsAmong the 349 intensive care patients analysed, the GOSE score at 3 months was ≤ 4 and ≥ 5 in 233 (67%) and 116 (33%) patients, respectively. The mortality rate at 3 months was 157/233 (67%), and only 7 patients (2%) fully recovered or had minor symptoms. Withdrawal or withholding of life-sustaining therapies in the ICU was identified in 140 patients (40.1%). Multivariate analysis showed that age (OR 1.09, CI 95% 1.04-1.14), male sex (OR 2.94, CI95% 1.70-5.11), baseline Glasgow Coma Scale score (OR 1.20, CI95% 1.13-1.29), injury severity score (ISS; OR 1.04, CI95% 1.02-1.06) and use of osmotherapy (OR 2.42, CI95% 1.26-4.65) were associated with unfavourable outcomes (AUC = 0.79, CI 95% [0.74-0.84]). According to multivariate analysis, the variables providing the best sensitivity and specificity were age ≥ 77 years, Glasgow Coma Scale score ≤ 9 and ISS ≥ 25 (AUC = 0.79, CI 95% [0.74-0.84]).ConclusionsAmong intensive care patients aged ≥ 65 years suffering from TBI, age (≥ 77 years), male sex, baseline Glasgow coma scale score (≤ 9), ISS (≥ 25) and use of osmotherapy were predictors of unfavourable neurological outcome.Trial registrationClinicalTrials.gov Identifier: NCT04651803. Registered 03/12/2020. Retrospectively registered.

Project description:ImportanceThe development and expansion of intracranial hematoma are associated with adverse outcomes. Use of tranexamic acid might limit intracranial hematoma formation, but its association with outcomes of severe traumatic brain injury (TBI) is unclear.ObjectiveTo assess whether prehospital administration of tranexamic acid is associated with mortality and functional outcomes in a group of patients with severe TBI.Design, setting, and participantsThis multicenter cohort study is an analysis of prospectively collected observational data from the Brain Injury: Prehospital Registry of Outcome, Treatments and Epidemiology of Cerebral Trauma (BRAIN-PROTECT) study in the Netherlands. Patients treated for suspected severe TBI by the Dutch Helicopter Emergency Medical Services between February 2012 and December 2017 were included. Patients were followed up for 1 year after inclusion. Data were analyzed from January 10, 2020, to September 10, 2020.ExposuresAdministration of tranexamic acid during prehospital treatment.Main outcomes and measuresThe primary outcome was 30-day mortality. Secondary outcomes included mortality at 1 year, functional neurological recovery at discharge (measured by Glasgow Outcome Scale), and length of hospital stay. Data were also collected on demographic factors, preinjury medical condition, injury characteristics, operational characteristics, and prehospital vital parameters.ResultsA total of 1827 patients were analyzed, of whom 1283 (70%) were male individuals and the median (interquartile range) age was 45 (23-65) years. In the unadjusted analysis, higher 30-day mortality was observed in patients who received prehospital tranexamic acid (odds ratio [OR], 1.34; 95% CI, 1.16-1.55; P < .001), compared with patients who did not receive prehospital tranexamic acid. After adjustment for confounders, no association between prehospital administration of tranexamic acid and mortality was found across the entire cohort of patients. However, a substantial increase in the odds of 30-day mortality persisted in patients with severe isolated TBI who received prehospital tranexamic acid (OR, 4.49; 95% CI, 1.57-12.87; P = .005) and after multiple imputations (OR, 2.05; 95% CI, 1.22-3.45; P = .007).Conclusions and relevanceThis study found that prehospital tranexamic acid administration was associated with increased mortality in patients with isolated severe TBI, suggesting the judicious use of the drug when no evidence for extracranial hemorrhage is present.

Project description:Artificial intelligence (AI) has transformative potential to support prehospital clinicians, emergency physicians, and trauma surgeons in acute traumatic injury care. This scoping review examines the literature evaluating AI models using prehospital features to support early traumatic injury care. We conducted a systematic search in August 2023 of PubMed, Embase, and Web of Science. Two independent reviewers screened titles/abstracts, with a third reviewer for adjudication, followed by a full-text analysis. We included original research and conference presentations evaluating AI models-machine learning (ML), deep learning (DL), and natural language processing (NLP)-that used prehospital features or features available immediately upon emergency department arrival. Review articles were excluded. The same investigators extracted data and systematically categorized outcomes to ensure consistency and transparency. We calculated kappa for interrater reliability and descriptive statistics. We identified 1050 unique publications, with 49 meeting inclusion criteria after title and abstract review (kappa 0.58) and full-text review. Publications increased annually from 2 in 2007 to 10 in 2022. Geographic analysis revealed a 61% focus on data from the United States. Studies were predominantly retrospective (88%), used local (45%) or national level (41%) data, focused on adults only (59%) or did not specify adults or pediatrics (27%), and 57% encompassed both blunt and penetrating injury mechanisms. The majority used machine learning (88%) alone or in conjunction with DL or NLP, and the top three algorithms used were support vector machine, logistic regression, and random forest. The most common study objectives were to predict the need for critical care and life-saving interventions (29%), assist in triage (22%), and predict survival (20%). A small but growing body of literature described AI models based on prehospital features that may support decisions made by dispatchers, Emergency Medical Services clinicians, and trauma teams in early traumatic injury care.

Project description:In combat operations, patients with traumatic injuries require expeditious evacuation to improve survival. Studies have shown that long transport times are associated with increased morbidity and mortality. Limited data exist on the influence of transport time on patient outcomes with specific injury types. The objective of this study was to determine the impact of the duration of time from the initial request for medical evacuation to arrival at a medical treatment facility on morbidity and mortality in casualties with traumatic extremity amputation and non-compressible torso injury (NCTI).We completed a retrospective review of MEDEVAC patient care records for United States military personnel who sustained traumatic amputations and NCTI during Operation Enduring Freedom between January 2011 and March 2014. We grouped patients as traumatic amputation and NCTI (AMP+NCTI), traumatic amputation only (AMP), and neither AMP nor NCTI (Non-AMP/NCTI). Analysis was performed using chi-squared tests, Fisher's exact tests, Cochran-Armitage Trend tests, Shapiro-Wilks tests, Wilcoxon and Kruskal-Wallis techniques and Cox proportional hazards regression modeling.We reviewed 1267 records, of which 669 had an injury severity score (ISS) of 10 or greater and were included in the analysis. In the study population, 15.5% sustained only amputation injuries (n=104, AMP only), 10.8% sustained amputation and NCTI (n=72, AMP+NCTI), and 73.7% did not sustain either an amputation or an NCTI (n=493, Non-AMP/NCTI). AMP+NCTI had the highest mortality (16.7%) with transport time greater than 60 min. While the AMP+NCTI group had decreasing survival with longer transport times, AMP and Non-AMP/NCTI did not exhibit the same trend.A decreased transport time from the point of injury to a medical treatment facility was associated with decreased mortality in patients who suffered a combination of amputation injury and NCTI. No significant association between transport time and outcomes was found in patients who did not sustain NCTI. Priority for rapid evacuation of combat casualties should be given to those with NCTI.