Mishra Lipi, Ramaswamy Sowmya Muchukunte, McCallum-Hee Broderick Ivan, Wright Keaton, Croxford Riley, Nagaraj Sunil Belur, Anstey Matthew
Intensive Care Unit, Sir Charles Gairdner Hospital, Nedlands, Western Australia, Australia.
Heailthigence Pty, Perth, Western Australia, Australia.
BMJ Health Care Inform. 2025 Sep 14;32(1):e101354. doi: 10.1136/bmjhci-2024-101354.
Artificial intelligence (AI) holds promise for predicting sepsis. However, challenges remain in integrating AI, natural language processing (NLP) and free text data to enhance sepsis diagnosis at emergency department (ED) triage. This study aimed to evaluate the effectiveness of AI in improving sepsis diagnosis.
This retrospective cohort study analysed data from 134 266 patients admitted to the ED and subsequently hospitalised between 1 January 2016 and 31 December 2021. The data set comprised 10 variables and free-text triage comments, which underwent tokenisation and processing using a bag-of-words model. We evaluated four traditional NLP classifier models, including logistic regression, LightGBM, random forest and neural network. We also evaluated the performance of the BERT classifier. We used area under precision-recall curve (AUPRC) and area under the curve (AUC) as performance metrics.
Random forest exhibited superior predictive performance with an AUPRC of 0.789 (95% CI: 0.7668 to 0.8018) and an AUC of 0.80 (95% CI: 0.7842 to 0.8173). Using raw text, the BERT model achieved an AUPRC of 0.7542 (95% CI: 0.7418 to 0.7741) and AUC of 0.7735 (95% CI: 0.7628 to 0.8017) for sepsis prediction. Key variables included ED treatment time, patient age, arrival-to-treatment time, Australasian Triage Scale and visit type.
This study demonstrates AI, particularly random forest and BERT classifiers, for early sepsis detection in EDs using free-text patient concerns.
Incorporating free text into machine learning improved diagnosis and identified missed cases, enhancing sepsis prediction in the ED with an AI-powered clinical decision support system. Large, prospective studies are needed to validate these findings.
人工智能（AI）有望用于预测脓毒症。然而，在整合人工智能、自然语言处理（NLP）和自由文本数据以加强急诊科（ED）分诊时的脓毒症诊断方面，仍存在挑战。本研究旨在评估人工智能在改善脓毒症诊断方面的有效性。
这项回顾性队列研究分析了2016年1月1日至2021年12月31日期间入住急诊科并随后住院的134266例患者的数据。数据集包括10个变量和自由文本分诊注释，这些数据使用词袋模型进行了分词和处理。我们评估了四种传统的NLP分类器模型，包括逻辑回归、LightGBM、随机森林和神经网络。我们还评估了BERT分类器的性能。我们使用精确召回率曲线下面积（AUPRC）和曲线下面积（AUC）作为性能指标。
随机森林表现出卓越的预测性能，AUPRC为0.789（95%CI：0.7668至0.8018），AUC为0.80（95%CI：0.7842至0.8173）。使用原始文本时，BERT模型在脓毒症预测方面的AUPRC为0.7542（95%CI：0.7418至0.7741），AUC为0.7735（95%CI：0.7628至0.8017）。关键变量包括急诊科治疗时间、患者年龄、到达治疗时间、澳大利亚分诊量表和就诊类型。
本研究证明了人工智能，特别是随机森林和BERT分类器，可用于利用患者自由文本关注点在急诊科早期检测脓毒症。
将自由文本纳入机器学习可改善诊断并识别漏诊病例，通过人工智能驱动的临床决策支持系统增强急诊科的脓毒症预测。需要进行大型前瞻性研究来验证这些发现。
