文献速递重症综合管理与镇静镇痛

急诊科资源分配改进的机器学习框架开发与验证

Development and validation of a machine learning framework for improved resource allocation in the emergency department.
PubMed 文献 · Am J Emerg Med · 2024-07-27
一句话结论基于分诊时可得数据构建的144个投票集成模型,平均AUC 0.82、准确率0.76,可预测急诊患者资源需求以改善分诊资源分配。

本研究为回顾性研究,分析Medical Information Mart for Intensive Care IV中的391,472次急诊就诊,按80%训练、20%测试划分队列。利用分诊时可得数据(生命体征、年龄、性别、到达方式、用药史、主诉),通过Azure AutoML为144个目标(实验室检查、影像学检查及急诊期间所需药物)各构建投票集成模型。144个框架平均AUC 0.82、准确率0.76;特征重要性分析显示主诉等特征在不同目标间具有较高聚合重要性。作者认为该机器学习模型预测急诊患者资源需求准确性高,可改善资源有限急诊科的患者流动与资源分配。

要点

  • 回顾性分析391,472次急诊就诊,80%训练、20%测试,为144个资源目标各建投票集成模型。
  • 144个模型平均AUC 0.82、准确率0.76,评估指标还包括F1分数、精确率和召回率。
  • 预测目标涵盖实验室检查组合、影像学检查及急诊期间所需药物。
  • 输入特征为分诊时可得数据:生命体征、年龄、性别、到达方式、用药史、主诉。
  • 主诉等特征在不同目标间具有较高聚合特征重要性。

适用人群与场景

  • Medical Information Mart for Intensive Care IV中记录的急诊就诊患者。

研究设计与关键结果

  • 回顾性研究,分析391,472次急诊就诊,数据分为训练队列(80%)和测试队列(20%)。
  • 使用Azure AutoML为144个目标列(实验室检查组合、影像学检查及急诊期间所需药物)各创建投票集成模型。
  • 144个模型平均AUC 0.82、准确率0.76,评估指标包括AUROC、F1分数、准确率、精确率和召回率。
  • 特征重要性分析显示主诉等特征在不同目标间具有较高聚合特征重要性。

证据类型、等级与局限

  • 证据类型:回顾性研究。
  • 局限:原文未报告局限性。

对床旁实践的改变

  • 作者认为,该机器学习模型预测急诊患者资源需求准确性高,可极大改善资源有限急诊科的患者流动与资源分配。

提炼说明

  • 本条依据论文摘要整理;摘要未涉及「诊断标准与关键阈值」「处置流程与药物剂量」「红旗征与不能漏的情况」「指南或共识出处」,故未列出这些栏目;全文请点「阅读原文」。
英文原文(摘要)· 点击展开

OBJECTIVE: The Emergency Severity Index (ESI) is the most commonly used system in over 70% of all U.S. emergency departments (ED) that uses predicted resource utilization as a means to triage [1], Mistriage, which includes both undertriage and overtriage has been a persistent issue, affecting 32.2% of total ED visits [2]. Our goal is to develop a machine learning framework that predicts patients' resource needs, thereby improving resource allocation during triage.

METHODS: This retrospective study analyzed ED visits from the Medical Information Mart for Intensive Care IV, dividing the data into training (80%) and testing (20%) cohorts. We utilized data available during triage, including patient vital signs, age, gender, mode of arrival, medication history, and chief complaint. Azure AutoML was used to create different machine learning models trained to predict the 144 target columns including laboratory panels and imaging modalities as well as medications required during patients' ED visits. The 144 models' performance was evaluated using the area under the receiver operating characteristic curve (AUROC), F1 score, accuracy, precision and recall.

RESULTS: A total of 391,472 ED visits were analyzed. 144 Voting ensemble models were created for each target. All frameworks achieved on average an AUC score of 0.82 and accuracy of 0.76. We gathered the feature importance for each target and observed that 'chief complaint', among others, had a high aggregate feature importance across different targets.

CONCLUSION: This study shows the high accuracy in predicting resource needs for patients in the ED using a machine learning model. This can greatly improve patient flow and resource allocation in already resource limited emergency departments.

来源:Am J Emerg Med,查看原文。版权归原作者所有。

参考来源

  1. El Ariss AB, Kijpaisalratana N, Ahmed S, et al. Development and validation of a machine learning framework for improved resource allocation in the emergency department. Am J Emerg Med. 2024. PMID: 39127019. doi:10.1016/j.ajem.2024.07.040

供执业医师抢救与诊疗参考,不替代临床判断;剂量请结合体重、肝肾功能和本院药品规格复核,以指南原文为准。