Injury Severity Prediction Research Articles

Comparison of continuous vital signs data analysis versus venous lactate for the prediction of lifesaving interventions in patients with traumatic shock.

The prehospital environment is fraught with operational constraints, making it difficult to assess the need for resources such as lifesaving interventions (LSI) for adults with traumatic injuries. While invasive methods such as lactate have been found to be highly predictive for estimating injury severity and resource requirements, noninvasive methods, to include continuous vital signs (VS), have the potential to provide prognostic information that can be quickly acquired, interpreted, and incorporated into decision making. In this work, we hypothesized that an analysis of continuous VS would have predictive capacity comparable to lactate and other laboratory tests for the prediction of injury severity, need for LSIs and intensive care unit (ICU) admission. In this pre-planned secondary analysis of 300 prospectively enrolled patients, venous blood samples were collected in the prehospital environment aboard a helicopter and analyzed with a portable lab device. Patients were transported to the primary adult resource center for trauma in the state of Maryland. Continuous VS were simultaneously collected. Descriptive statistics were used to describe the cohort and predictive models were constructed using a regularized gradient boosting framework with 10-fold cross-validation with additional analyses using Shapley additive explanations (SHAP). Complete VS and laboratory data from 166 patients were available for analysis. The continuous VS models had better performance for prediction of receiving LSIs and ICU length of stay compared to single (initial) VS measurements. The continuous VS models had comparable performance to models using only laboratory tests in predicting discharge within 24 hours (continuous VS model: AUROC 0.71; 95% CI, 0.68-0.75 vs. lactate model: AUROC 0.65; 95% CI, 0.68; 95% CI, 0.66-0.71). The model using all laboratory data yielded the highest sensitivity and sensitivity (AUROC 0.77; 95% CI, 0.74-0.81). The results from this study suggest that continuous VS obtained from autonomous monitors in an aeromedical environment may be helpful for predicting LSIs and the critical care requirements for traumatically injured adults. The collection and use of noninvasively obtained physiological data during the early stages of prehospital care may be useful for in developing user-friendly early warning systems for identifying potentially unstable trauma patients.

Gunshot Wounds to the Head: 29 Cases of Monocenter Children's Trauma Hospital Experience and Predictor Scales in Pediatric Population

Gunshot wounds to the head (GSWTH) in children remain an underexplored area, and all clinical guidelines extrapolated from adult experiences. A key challenge in treating these patients is age stratification, as pediatric survival rates are notably higher than in adults. The objective of the study is to compare 2 groups of patients based on the severity of their condition and to analyze the impact of various factors on the outcomes of these conditions. A single-center retrospective cohort study was conducted. Patient records of individuals (from birth to 18years old) with craniocerebral injuries caused by GSWTH and treated at Morozov Children's City Hospital between 1992 and 2015 were comprehensively reviewed. Data analysis included clinical presentation, computed tomography scan findings, injury site and trajectory, Glasgow Coma Scale and Glasgow Outcome Scale scores, and the St. Louis Scale. A total of 29 pediatric patients (79% male, 21% female, median age 8years) with GSWTH were analyzed. The mortality rate was 17%. A transventricular wound trajectory was found in 17% of patients and was associated with a poor prognosis (P < 0.001). The Glasgow Coma Scale showed insufficient specificity in injury severity assessment. A multiple logistic regression model predicting injury severity based on bullet type, gender, and time to admission had an accuracy of 80%, while a Decision Tree model improved accuracy to 96.6%. A stacking model combining multiple logistic regression and Decision Tree increased sensitivity to 87.5% and explained 65.5% of the variance. The findings emphasize the importance of a multifactorial approach in children with GSWTH, highlighting its effectiveness for precise outcome prediction.

Research on occupant injury severity prediction of autonomous vehicles based on transfer learning

AbstractThe focus of the future of autonomous vehicles has shifted from feasibility to safety and comfort. The seat of an autonomous vehicle may be equipped with a rotational function, and the occupant's sitting position would be diverse. This poses a higher challenge to occupant injury protection during vehicle collisions. The main objective of the current study is to develop occupant injury prediction models for autonomous vehicles that can be used to predict the injury severity of occupants in different seat orientations and sitting positions. The first step is to establish an occupant crash model database with different seat orientations. It is used to simulate the occupant crash injury database of an autonomous vehicle, considering seat rotation and the back inclination angle. The second step is to establish a pre‐training occupant injury prediction model based on the existing database and then train the autonomous vehicle occupant injury prediction model using an in‐house database based on the transfer learning method. Occupant injury prediction models achieve good accuracy (82.8% on the numerical database and 62.9% on the real verification database) and shorter computational time (4.86 ± 0.33 ms) on the prediction tasks. Finally, the influence of the model input variables is analyzed. This study demonstrates the feasibility of using a small‐sample database based on transfer learning for occupant injury prediction in autonomous vehicles.

Prediction of Vehicle–Live Animal Crashes in Britain: Contact and Noncontact Incidents

&lt;div&gt;Animal–vehicle collisions (AVCs) can result in devastating injuries to both humans and animals. Despite significant advances in crash prediction models, there is still a significant gap when it comes to injury severity prediction models in AVCs, especially concerning small animals. It is no secret that large mammals can pose a significant threat to road safety; however, researchers tend to overlook the impact of domestic and small animals wandering along the roads. In this study, STATS19 road safety data was used containing any type of live animal, and a radial basis function (RBF) model was used to predict different severities of injury regardless of whether the animal was hit, or not. As a means of better understanding the factors contributing to severities, regression trees were used to identify and retain only the most useful predictors, removing the less useful ones. A comparison was made between the performance of the trees across a range of severity classes, and the model-fitting results were discussed. Initially, the study was unable to generate satisfactory predictions, but the optimization of the key predictors and the combination of severity classes significantly improved their accuracy. Research findings revealed factors contributing to the severities, which were discussed accordingly. Particular attention was drawn to the pressing safety issue posed by animals crossing A-class single carriageways in rural site clusters. Animals being present on those carriageways without direct vehicular contact significantly contributed to the severity of the injuries sustained. Although the majority of contributing factors were related to human behavior, no evidence of road safety education, training, or publicity interventions specifically targeting AVCs was found in the literature.&lt;/div&gt;

An improved XGBoost model to predict the injury severity of person in road crash

Road traffic accidents are common cause for the human deaths in recent years. The injury severity of person prediction in the road traffic accident is important to find the relevant factors associated with crash. Prediction of injury severity of person in road traffic accident is considered as an important & crucial analysis for driving decisions under the dangerous situations. This study aims to proposes the framework to classify the different classes like No Apparent Injury(O), Possible Injury(C), Suspected Minor Injury (B), Suspected Serious Injury (A), Fatal Injury (K) for injury severity during road crash. In this research paper an improved XGBoost algorithm is used to predict injury severity of person for improving performance like accuracy and loss. The National Highway Traffic Safety Administration (NHTSA) Fatality Analysis Reporting System (FARS) (2017–2019) dataset is used for evaluation of the performance metrics for proposed framework. Initially the dataset is pre-processed by removing missing values and then correlation features selection method is used to find the important strongly correlated features with target Injury Severity. Finally the 44 important attributes are used for training dataset and Injury Severity as Target. The different models like Multinomial Logistic Regression (LR), Extra Trees, Random forest, Naive Bayes and Neural networks, XGB Classifier, improved XGBoost are compared on accuracy, Loss and F1-score metrics. It is seen that improved XGBoost model performs better with 84.06% accuracy and 0.381 loss metric. The study helps to reduce the road accidents by analysing the risk involved in accidents by extent of prediction of injury of person in road accidents.

Injury severity prediction and exploration of behavior-cause relationships in automotive crashes using natural language processing and extreme gradient boosting

Addressing the global challenge of traffic crashes necessitates transcending traditional statistical models, which often fail to fully capture the interactions between factors causing crashes. This oversight restricts the predictive accuracy and adaptability of current methodologies. Additionally, there is a notable gap in research that examines the links between behavior-cause relationships and crash injury severity. Our study deploys Natural Language Processing (NLP) and Frequent Pattern (FP) growth algorithm to mine crash narratives for behavior-cause connections, combines with the predictive strength of eXtreme Gradient Boosting (XGBoost) and the interpretative clarity offered by SHapley Additive exPlanations (SHAP), our approach not only predicts crash injury severity with satisfactory precision but also explains the influence of specific behavior-cause and environment conditions on crash outcomes. The integration of NLP and XGBoost, complemented by SHAP insights, has shown promising results with an accuracy of 0.79, outperforming traditional discrete choice models and competes closely with other machine learning approaches, including Support Vector Machines, Random Forest, Categorical Boosting (CatBoost), and Light Gradient Boosting Machine (LightGBM). Through detailed textual analysis and the establishment of a behavior-cause matrix, identifying five broad crash causes linked to 141 specific crash cause with behaviors, we uncover critical patterns such as the prominence of distracted driving in severe crashes. This comprehensive approach not only fills a critical research gap by linking behavior-cause relationships with injury severity but also sets the stage for developing targeted interventions to enhance road safety.

Use of Reverse Shock Index Multiplied by Simplified Motor Score in a Five-Level Triage System: Identifying Trauma in Adult Patients at a High Risk of Mortality.

Background and Objectives: The Taiwan Triage and Acuity Scale (TTAS) is reliable for triaging patients in emergency departments in Taiwan; however, most triage decisions are still based on chief complaints. The reverse-shock index (SI) multiplied by the simplified motor score (rSI-sMS) is a more comprehensive approach to triage that combines the SI and a modified consciousness assessment. We investigated the combination of the TTAS and rSI-sMS for triage compared with either parameter alone as well as the SI and modified SI. Materials and Methods: We analyzed 13,144 patients with trauma from the Taipei Tzu Chi Trauma Database. We investigated the prioritization performance of the TTAS, rSI-sMS, and their combination. A subgroup analysis was performed to evaluate the trends in all clinical outcomes for different rSI-sMS values. The sensitivity and specificity of rSI-sMS were investigated at a cutoff value of 4 (based on previous study and the highest score of the Youden Index) in predicting injury severity clinical outcomes under the TTAS system were also investigated. Results: Compared with patients in triage level III, those in triage levels I and II had higher odds ratios for major injury (as indicated by revised trauma score < 7 and injury severity score [ISS] ≥ 16), intensive care unit (ICU) admission, prolonged ICU stay (≥14 days), prolonged hospital stay (≥30 days), and mortality. In all three triage levels, the rSI-sMS < 4 group had severe injury and worse outcomes than the rSI-sMS ≥ 4 group. The TTAS and rSI-sMS had higher area under the receiver operating characteristic curves (AUROCs) for mortality, ICU admission, prolonged ICU stay, and prolonged hospital stay than the SI and modified SI. The combination of the TTAS and rSI-sMS had the highest AUROC for all clinical outcomes. The prediction performance of rSI-sMS < 4 for major injury (ISS ≥ 16) exhibited 81.49% specificity in triage levels I and II and 87.6% specificity in triage level III. The specificity for mortality was 79.2% in triage levels I and II and 87.4% in triage level III. Conclusions: The combination of rSI-sMS and the TTAS yielded superior prioritization performance to TTAS alone. The integration of rSI-sMS and TTAS effectively enhances the efficiency and accuracy of identifying trauma patients at a high risk of mortality.

An investigation of heterogeneous impact, temporal stability, and aggregate shift in factors affecting the driver injury severity in single-vehicle rollover crashes

Single-vehicle rollover crashes have been acknowledged as a predominant highway crash type resulting in serious casualties. To investigate the heterogeneous impact of factors determining different injury severity levels in single-vehicle rollover crashes, the random parameters logit model with unobserved heterogeneity in means and variances was employed in this paper. A five-year dataset on single-vehicle rollover crashes, gathered in California from January 1, 2013, to December 31, 2017, was utilized. Driver injury severities that were determined to be outcome variables include no injury, minor injury, and severe injury. Characteristics pertaining to the crash, driver, temporal, vehicle, roadway, and environment were acknowledged as potential determinants. The results showed that the gender indicator specified to minor injury was consistently identified as a significant random parameter in four years’ models and the joint five-year model, excluding the 2016 crash model where the night indicator associated with no injury was observed to produce the random effect. Additionally, two series of likelihood ratio tests were conducted to assess the year-to-year and aggregate-to-component temporal stability of model estimation results. Marginal effects of explanatory variables were also calculated and compared to analyze the temporal stability and interpret the results. The findings revealed an overall temporal instability of model specifications across individual years, while there is no significant aggregate-to-component variation. Injury severities were observed to be stably affected by several variables, including improper turn indicator, under the influence of alcohol indicator, old driver indicator, seatbelt indicator, insurance indicator, and airbag indicator. Furthermore, the year-to-year and aggregate-to-component shift was quantified and characterized by calculating the differences in probabilities between within-sample observations and out-of-sample predictions. The overall results imply that continuing to expand and refine the model to incorporate more comprehensive datasets can result in more robust and stable injury severity prediction, thus benefiting in mitigating the associated driver injury severity.

Identification of the best machine learning model for the prediction of driver injury severity

Predicting the injury severities sustained by drivers engaged in road traffic accidents is a key topic of research in road traffic safety. The current study analyzed the driver injury severity (DIS) using twelve machine learning (ML) algorithms. These models were implemented using 0.70, 0.80, and 0.90 train ratios and 5-, 10- and 15-fold cross-validation. Ten years of accident data (from 2011 to 2020) was obtained from police department of Shillong, India. A total of 693 accidents were documented, with 68% being nonfatal and 32% being fatal. Precision, recall, accuracy, F1 score and area under the curve measures were used to compare the performance of all twelve ML models. Overall, the light gradient-boosting machine model was shown to be the best ML model for predicting the injury severities of drivers engaged in road traffic incidents. Finally, variable importance analysis results showed that cause of accident, collision type and types of vehicles were the most influencing factors in nonfatal and fatal driver accidents. The results also revealed that age and gender were slightly associated with DIS. The findings of the current research could be helpful to road safety agencies for the implementation of suitable countermeasures to increase driver safety in road accidents.

Establishing pediatric age-adjusted shock index cut points in trauma patients younger than 1 year.

Shock Index is used to predict injury severity and adverse outcomes in trauma patients, but pediatric age-adjusted shock index (SIPA) has superior performance in pediatric patients older than 1 year. Pediatric age-adjusted shock index scores younger than 1 year have not been well studied. This project aimed to establish and evaluate SIPA cut point data points for patients younger than 1 year. Using age-based vital signs, we developed cut point values for patients younger than 1 year using our institutional trauma data. All trauma patients younger than 12 months were included, and clinical outcomes were recorded. Pediatric age-adjusted shock index cut points were defined using age-specific vital sign limits (SIPA-VS) and tested against optimal cut points defined by receiver operating characteristic analysis (SIPA-ROC) and a cut point of 1.2 (SIPA-Nordin), which is used for patients aged 1 to 4 years. Student's t test, χ 2 tests, analysis of variance, and test characteristics were used to analyze groups. A total of 609 pediatric trauma patients younger than 12 months were identified from 2018 to 2022. Pediatric age-adjusted shock index scores were calculated for 483 patients. There were 406 patients with blunt trauma and 17 with penetrating. SIPA-Nordin was elevated in 81.6% (n = 397) of patients, compared with SIPA-VS 21% (n = 101) and SIPA-ROC 31% (n = 150). In comparison with SIPA-Nordin, both SIPA-VS and SIPA-ROC score exhibited superior specificity and negative predictive values for multiple outcomes. Elevated SIPA-ROC scores had statistically significant associations with intensive care unit admission, mechanical ventilation, severe anemia, transfusion during hospital admission, and in-hospital mortality. Pediatric age-adjusted shock index is a useful tool in identifying patients at risk for several complications of severe traumatic injury. Pediatric age-adjusted shock index cut points had high negative predictive value and specificity for many outcomes. This study proposes cut point values that may aid in clinical decision making for trauma patients younger than 1 year. Diagnostic Test/Criteria; Level IV.

Factors affecting injury severity in motorcycle crashes: Different age groups analysis using Catboost and SHAP techniques

Objective Motorcycle crashes often result in severe injuries on roads that affect people’s lives physically, financially, and psychologically. These injuries could be notably harmful to drivers of all age groups. The main objective of this study is to investigate the risk factors contributing to the severity of crash injuries in different age groups. Methods This Objective is achieved by developing accurate machine learning (ML) based prediction models. This research examines the relationship between potential risk factors of motorcycle-associated crashes using (ML) and Shapley Additive explanations (SHAP) technique. The SHAP technique further helped interpreting ML methods for traffic injury severity prediction. It indicates the significant non-linear interactions between dependent and independent variables. The data for this study was collected from the Provincial Emergency Response Service RESCUE 1122 for the Rawalpindi region (Pakistan) over three years (from 2017 to 2020). The Synthetic Minority Oversampling Technique (SMOTE) is employed to balance injury severity classes in the pre-processing phase. Results The results demonstrate that age, gender, posted speed limit, the number of lanes, and month of the year are positively associated with severe and fatal injuries. This research also assesses how the modeling framework varies between the ML and classical statistical methods. The predictive performance of proposed ML models was assessed using several evaluation metrics, and it is found that Catboost outperformed the XGBoost, Random Forest (RF) and Multinomial Logit (MNL) model. Conclusion The findings of this study will assist road users, road safety authorities, stakeholders, policymakers, and decision-makers in obtaining substantial and essential guidance for reducing the severity of crash injuries in Pakistan and other countries with prevailing conditions.

Addressing imbalanced data in predicting injury severity after traffic crashes: A comparative analysis of machine learning models

Road traffic crashes are a significant public health concern, leading to substantial human and financial losses. Accurately predicting injury severity is crucial for optimizing rescue efforts and saving lives. This study utilizes various Machine Learning (ML) algorithms, such as Random Forest, Logistic Regression, XGBoost, and Support Vector Machine (SVM), to predict crash severity. The dataset spans from 2015 to 2023, comprising crash data from the City of Chicago, featuring a highly imbalanced ratio of non-severe to severe incidents (1000 to 1). To address class imbalance challenges, the study evaluates various data sampling methods, including Oversampling, Undersampling, and Hybridsampling. Model performance is assessed using AUC-ROC and recall to account for accuracy limitations in imbalanced datasets. Results reveal the inefficacy of conventional data sampling methods where data is highly imbalanced. Consequently, a novel approach was adopted, involving the random removal of observations before applying data sampling methods, leading to a significant improvement in model performance. SVM-SMOTE and ClusterCentroid emerge as the most effective resampling methods. Notably, among all ML models, SVM demonstrates the best overall performance. The final findings of this study aim to assist emergency responders in quickly evaluating the severity of an incident upon receiving a report.

Road Crash Injury Severity Prediction Using a Graph Neural Network Framework

Road Crash Injury Severity Prediction Using a Graph Neural Network Framework

Modeling injury severity of crashes involving golf carts: A case study of The Villages, Florida

Objective Crashes involving golf carts (GCs) have been on an increasing trend in recent years, particularly in the United States. This study focuses on analyzing GC crashes in the Florida community known as The Villages, one of the largest GC-oriented communities in the nation and worldwide. The objective was to evaluate the injury severity of crashes involving GCs in a retirement community where GCs are a common mode of transportation. Methods The ordinal logistic regression (OLR) and Decision Tree Ensemble (DTE) models were used to analyze the injury severity of 616 GC-related crashes. Models’ accuracy parameters were used to check their reliability. Results The analysis revealed that GC crash severity is influenced by various factors. Factors found to be significant by the OLR model in determining injury severity include ejection of one or more occupants from the GC, the extent of damage to the GC, GC speed prior to the crash, roadway characteristics (including divided roadways, traffic control devices, paved shoulders, and T-intersections), and roll-over incidents. The OLR model demonstrated an overall accuracy of approximately 71% in predicting injury severity. The DTE model performed better, with an overall accuracy of 78%. The OLR model’s findings were supported by the DTE model, which identified estimated GC speed, occupant(s) ejection from the GC, estimated GC vehicle damage, intersection type, and type of shoulder as the most important factors influencing GC crash severity. Conclusions Understanding these factors is vital for transportation agencies to develop effective strategies to reduce the severity of GC crashes, ensuring the safety of GC users. This study provides recommendations to transportation agencies on measures to improve the safety of GCs.

Analysis and prediction of injury severity in single micromobility crashes with Random Forest

Urban micromobility represents a significant shift towards sustainable cities, underscoring the paramount importance of its safety. With the surge in micromobility adoption, collisions involving micromobility devices, such as bicycles and e-scooters, have surged in recent years. The second most common crash type involving these vehicles is one that only involves a micromobility vehicle (single micromobility crashes). This study analyzed 6030 single micromobility crashes that occurred in Spanish urban areas from 2016 to 2020. The Random Forest methodology was applied to create a classification model for the purpose of characterizing these crashes, predicting their injury severity, and identifying the primary influencing factors. To address the issue of imbalanced data, resulting from the relatively smaller dataset of fatal and seriously injured crashes compared to slightly injured ones, the Synthetic Minority Oversampling Technique (SMOTE) was applied.The results indicate that certain behaviors, such as not wearing a helmet, riding for leisure, and instances of speeding violations, have the potential to increase injury severity. Additionally, crashes occurring at intersections or at cycle lanes with bad pavement conditions are likely to result in more severe outcomes. Furthermore, the concurrent presence of various other factors also contributes to an escalation in crash injury severity.These findings have the potential to provide valuable insights to authorities, assisting them in the decision-making process to enhance micromobility safety and thereby promoting the creation of more equitable and sustainable urban environments.

Advances, challenges, and future research needs in machine learning-based crash prediction models: A systematic review

Accurately modelling crashes, and predicting crash occurrence and associated severities are a prerequisite for devising countermeasures and developing effective road safety management strategies. To this end, crash prediction modelling using machine learning has evolved over two decades. With the advent of big data that provides unprecedented opportunities to better understand the crash mechanism and its determinants, such efforts will likely be accelerated. To gear these efforts, understanding state-of-the-art machine learning-based crash prediction models becomes paramount to summarise the lessons learned from past efforts, which can assist in developing robust and accurate models. This review paper aims to address this gap by systematically reviewing the machine learning studies on crash modelling. Models are reviewed from three aspects of the application: (a) crash occurrence (or real-time crash) prediction, (b) crash frequency prediction, and (c) injury severity prediction. Further, model intricacies that impact model performance are identified and thoroughly reviewed. This comprehensive review highlights specific gaps and future research needs in three aforementioned model applications, such as improper selection of non-crash events for crash occurrence models, the inability of future forecasting of crash frequency models, and inconsistency in injury severity classes. Critical research needs relating to model development, evaluation, and application are also discussed. This review envisages methodological advancements in machine learning models for crash prediction modelling and leveraging big data to better link crashes with its determinants.

On Scene Injury Severity Prediction (OSISP) model for trauma developed using the Swedish Trauma Registry

BackgroundProviding optimal care for trauma, the leading cause of death for young adults, remains a challenge e.g., due to field triage limitations in assessing a patient’s condition and deciding on transport destination. Data-driven On Scene Injury Severity Prediction (OSISP) models for motor vehicle crashes have shown potential for providing real-time decision support. The objective of this study is therefore to evaluate if an Artificial Intelligence (AI) based clinical decision support system can identify severely injured trauma patients in the prehospital setting.MethodsThe Swedish Trauma Registry was used to train and validate five models – Logistic Regression, Random Forest, XGBoost, Support Vector Machine and Artificial Neural Network – in a stratified 10-fold cross validation setting and hold-out analysis. The models performed binary classification of the New Injury Severity Score and were evaluated using accuracy metrics, area under the receiver operating characteristic curve (AUC) and Precision-Recall curve (AUCPR), and under- and overtriage rates.ResultsThere were 75,602 registrations between 2013–2020 and 47,357 (62.6%) remained after eligibility criteria were applied. Models were based on 21 predictors, including injury location. From the clinical outcome, about 40% of patients were undertriaged and 46% were overtriaged. Models demonstrated potential for improved triaging and yielded AUC between 0.80–0.89 and AUCPR between 0.43–0.62.ConclusionsAI based OSISP models have potential to provide support during assessment of injury severity. The findings may be used for developing tools to complement field triage protocols, with potential to improve prehospital trauma care and thereby reduce morbidity and mortality for a large patient population.

Comparison of age-adjusted shock indices as predictors of injury severity in paediatric trauma patients immediately after emergency department triage: A report from the Korean multicentre registry

IntroductionShock index paediatric-adjusted (SIPA) was presented for early prediction of mortality and trauma team activation in paediatric trauma patients. However, the derived cut-offs of normal vital signs were based on old references. We established alternative SIPAs based on the other commonly used references and compared their predictive values. MethodsWe performed a retrospective review of all paediatric trauma patients aged 1–15 years in the Emergency Department (ED)-based Injury In-depth Surveillance (EDIIS) database from January 1, 2011 to December 31, 2019. A total of 4 types of SIPA values were obtained based on the references as follows: uSIPA based on the Nelson textbook of paediatrics 21st ed., SIATLS based on the ATLS 10th guideline, SIPALS based on the PALS 2020 guideline, and SIPA. In each SIPA group, the cut-off was established by dividing the group into 4 subgroups: toddler (age 1–3), preschooler (age 4–6), schooler (age 7–12), and teenager (age 13–15). We performed an ROC analysis and calculated the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) to compare the predicted values of each SIPA in mortality, ICU admission, and emergent surgery or intervention. ResultsA total of 332,271 patients were included. The proportion of patients with an elevated shock index was 14.9 % (n = 49,347) in SIPA, 22.8 % (n = 75,850) in uSIPA, 0.3 % (n = 1058) in SIATLS, and 4.3 % (n = 14,168) in SIPALS. For mortality, uSIPA achieved the highest sensitivity (57.0 %; 95 % confidence interval 56.9 %–57.2 %) compared to SIPA (49.4 %, 95 % CI 49.2 %–49.5 %), SIATLS (25.5 %, 95 % CI 25.4 %–25.7 %), and SIPALS (43.8 %, 95 % CI 43.7 %–44.0 %), but there were no significant differences in the negative predictive value (NPV) or area under the curve (AUC). The positive predictive value (PPV) was highest in SIATLS (5.7 %, 95 % CI 5.6 %–5.8 %) compared to SIPA (0.2 %, 95 % CI 0.2 %–0.3 %), uSIPA (0.2 %, 95 % CI 0.2 %–0.2 %), and SIPALS (0.7 %, 95 % CI 0.7 %–0.8 %). The same findings were presented in ICU admission and emergent operation or intervention. ConclusionThe ATLS-based shock index achieved the highest PPV and specificity compared to SIPA, uSIPA, and SIPALS for adverse outcomes in paediatric trauma.

Особенности использования лимфоцитарного теста для биологической дозиметрии в ранние сроки после облучения

When eliminating the consequences of large-scale radiation accidents, primary triage of victims is of key importance during the early phase of medical evacuation. Information about lymphocyte counts (blood test) per unit of peripheral blood volume can be used for this purpose. The study was aimed to validate the method of using a lymphocyte test for prediction of acute radiation injury severity in the first days after the exposure associated with the radiation mass casualty incident, given peripheral blood was tested once. We performed correlation analysis of the data of laboratory studies focused on quantifying lymphocytes in peripheral blood of victims during the first days following the Chernobyl disaster and other radiation accidents on the territory of the countries of the former USSR (115 individuals), including radiation accidents with gamma neutron radiation (20 individuals). It was found that with the lymphocyte concentration of 0.2–1.0 × 109/L on day 2 after exposure, the absolute error of estimated dose was ±1.5 Gy in case of gamma exposure and ±1.3 Gy in case of exposure to gamma neutron radiation. When the lymphocyte concentration exceeds 1.0 × 109/L, mild acute radiation syndrome (ARS) is predicted, given the average dose is below 2.0 Gy; when the lymphocyte concentration is less than 0.2 × 109/L&lt; the estimated average dose exceeds 4.0 Gy, which corresponds to severe or extremely severe ARS. Thanks to the lymphocyte test accessibility and simplicity, this biological dosimetry method can occupy a worthy position in the diagnosis of radiation injury associated with large-scale accidents, since the results of cytogenetic tests are not available within first days after the accident.

Crash injury severity prediction considering data imbalance: A Wasserstein generative adversarial network with gradient penalty approach

For each road crash event, it is necessary to predict its injury severity. However, predicting crash injury severity with the imbalanced data frequently results in ineffective classifier. Due to the rarity of severe injuries in road traffic crashes, the crash data is extremely imbalanced among injury severity classes, making it challenging to the training of prediction models. To achieve interclass balance, it is possible to generate certain minority class samples using data augmentation techniques. Aiming to address the imbalance issue of crash injury severity data, this study applies a novel deep learning method, the Wasserstein generative adversarial network with gradient penalty (WGAN-GP), to investigate a massive amount of crash data, which can generate synthetic injury severity data linked to traffic crashes to rebalance the dataset. To evaluate the effectiveness of the WGAN-GP model, we systematically compare performances of various commonly-used sampling techniques (random under-sampling, random over-sampling, synthetic minority over-sampling technique and adaptive synthetic sampling) with respect to dataset balance and crash injury severity prediction. After rebalancing the dataset, this study categorizes the crash injury severity using logistic regression, multilayer perceptron, random forest, AdaBoost and XGBoost. The AUC, specificity and sensitivity are employed as evaluation indicators to compare the prediction performances. Results demonstrate that sampling techniques can considerably improve the prediction performance of minority classes in an imbalanced dataset, and the combination of XGBoost and WGAN-GP performs best with an AUC of 0.794 and a sensitivity of 0.698. Finally, the interpretability of the model is improved by the explainable machine learning technique SHAP (SHapley Additive exPlanation), allowing for a deeper understanding of the effects of each variable on crash injury severity. Findings of this study shed light on the prediction of crash injury severity with data imbalance using data-driven approaches.