School of Medicine, University of Rochester, 601 Elmwood Avenue, Box 601A, Rochester NY, 14642, USA.
Department of Surgery, University of Rochester Medical Center, 601 Elmwood Ave, Box SURG, Rochester, NY 14642, USA.
J Pediatr Surg. 2024 Jan;59(1):74-79. doi: 10.1016/j.jpedsurg.2023.09.014. Epub 2023 Sep 22.
The assignment of trauma team activation levels can be conceptualized as a classification task. Machine learning models can be used to optimize classification predictions. Our purpose was to demonstrate proof-of-concept for a machine learning tool for predicting trauma team activation levels in pediatric patients with traumatic injuries.
Following IRB approval, we retrospectively collected data from the institutional trauma registry and electronic medical record at our Pediatric Trauma Center for all patients (age <18 y) who triggered a trauma team activation (1/2014-12/2021), including: demographics, mechanisms of injury, comorbidities, pre-hospital interventions, numeric variables, and the six "Need for Trauma Intervention (NFTI)" criteria. Three machine learning models (Logistic Regression, Random Forest, Support Vector Machine) were tested 1000 times in separate trials using the union of the Cribari and NFTI metrics as ground-truth (Injury Severity Score >15 or positive for any of 6 NFTI criteria = full activation). Model performance was quantified and compared to emergency department (ED) staff.
ED staff had 75% accuracy, an area under the curve (AUC) of 0.73 ± 0.04, and an F1 score of 0.49. The best performing of all machine learning models, the support vector machine, had 80% accuracy, AUC 0.81 ± 4.1e, F1 Score 0.80, with less variance compared to other models and ED staff.
All machine learning models outperformed ED staff in all performance metrics. These results suggest that data-driven methods can optimize trauma team activations in the ED, with potential improvements in both patient safety and hospital resource utilization.
Economic/Decision Analysis or Modeling Studies.
II.
创伤团队激活级别的分配可以被概念化为分类任务。机器学习模型可用于优化分类预测。我们的目的是展示一种用于预测创伤患儿创伤团队激活级别的机器学习工具的概念验证。
在获得机构审查委员会批准后，我们回顾性地从我们的儿科创伤中心的机构创伤登记处和电子病历中收集了所有触发创伤团队激活的患者（年龄<18 岁）的数据（2014 年 1 月至 2021 年 12 月），包括：人口统计学数据、损伤机制、合并症、院前干预、数值变量和六个“创伤干预需求（NFTI）”标准。使用 Cribari 和 NFTI 指标的联合作为ground-truth（损伤严重度评分>15 或满足 6 个 NFTI 标准中的任何一个=全面激活），在单独的试验中，使用三种机器学习模型（逻辑回归、随机森林、支持向量机）进行了 1000 次测试。模型性能进行了量化并与急诊（ED）工作人员进行了比较。
ED 工作人员的准确率为 75%，曲线下面积（AUC）为 0.73±0.04，F1 得分为 0.49。所有机器学习模型中表现最好的支持向量机的准确率为 80%，AUC 为 0.81±4.1e，F1 得分为 0.80，与其他模型和 ED 工作人员相比，方差更小。
在所有性能指标中，所有机器学习模型均优于 ED 工作人员。这些结果表明，数据驱动的方法可以优化 ED 中的创伤团队激活，从而提高患者安全性和医院资源利用率。
经济/决策分析或建模研究。
II。
