Department of Information Engineering and Computer Science, Feng Chia University, Taichung, Taiwan.
Department of Applied Mathematics, National Chung Hsing University, Taichung, Taiwan.
BMJ Open. 2020 Feb 25;10(2):e033898. doi: 10.1136/bmjopen-2019-033898.
Current mortality prediction models used in the intensive care unit (ICU) have a limited role for specific diseases such as influenza, and we aimed to establish an explainable machine learning (ML) model for predicting mortality in critically ill influenza patients using a real-world severe influenza data set.
A cross-sectional retrospective multicentre study in Taiwan SETTING: Eight medical centres in Taiwan.
A total of 336 patients requiring ICU-admission for virology-proven influenza at eight hospitals during an influenza epidemic between October 2015 and March 2016.
We employed extreme gradient boosting (XGBoost) to establish the prediction model, compared the performance with logistic regression (LR) and random forest (RF), demonstrated the feature importance categorised by clinical domains, and used SHapley Additive exPlanations (SHAP) for visualised interpretation.
The data set contained 76 features of the 336 patients with severe influenza. The severity was apparently high, as shown by the high Acute Physiology and Chronic Health Evaluation II score (22, 17 to 29) and pneumonia severity index score (118, 88 to 151). XGBoost model (area under the curve (AUC): 0.842; 95% CI 0.749 to 0.928) outperformed RF (AUC: 0.809; 95% CI 0.629 to 0.891) and LR (AUC: 0.701; 95% CI 0.573 to 0.825) for predicting 30-day mortality. To give clinicians an intuitive understanding of feature exploitation, we stratified features by the clinical domain. The cumulative feature importance in the fluid balance domain, ventilation domain, laboratory data domain, demographic and symptom domain, management domain and severity score domain was 0.253, 0.113, 0.177, 0.140, 0.152 and 0.165, respectively. We further used SHAP plots to illustrate associations between features and 30-day mortality in critically ill influenza patients.
We used a real-world data set and applied an ML approach, mainly XGBoost, to establish a practical and explainable mortality prediction model in critically ill influenza patients.
目前，重症监护病房（ICU）使用的死亡率预测模型在流感等特定疾病方面的作用有限，我们旨在使用真实世界的严重流感数据集，建立一个可解释的机器学习（ML）模型，以预测危重症流感患者的死亡率。
台湾的一项回顾性多中心横断面研究。
台湾的 8 家医疗中心。
2015 年 10 月至 2016 年 3 月流感流行期间，在 8 家医院因病毒确诊流感而需要 ICU 入院的 336 名患者。
我们使用极端梯度提升（XGBoost）建立预测模型，将其性能与逻辑回归（LR）和随机森林（RF）进行比较，展示按临床领域分类的特征重要性，并使用 SHapley Additive exPlanations（SHAP）进行可视化解释。
数据集包含 336 例严重流感患者的 76 个特征。患者的严重程度明显较高，急性生理学和慢性健康评估 II 评分（22，17 至 29）和肺炎严重指数评分（118，88 至 151）均较高。XGBoost 模型（曲线下面积（AUC）：0.842；95%CI 0.749 至 0.928）优于 RF（AUC：0.809；95%CI 0.629 至 0.891）和 LR（AUC：0.701；95%CI 0.573 至 0.825），用于预测 30 天死亡率。为了让临床医生直观地了解特征利用情况，我们按临床领域对特征进行了分层。在液体平衡域、通气域、实验室数据域、人口统计学和症状域、管理域和严重程度评分域中的累积特征重要性分别为 0.253、0.113、0.177、0.140、0.152 和 0.165。我们进一步使用 SHAP 图说明了特征与危重症流感患者 30 天死亡率之间的关联。
我们使用真实世界的数据集并应用机器学习方法（主要是 XGBoost），为危重症流感患者建立了一个实用且可解释的死亡率预测模型。
