Predicting Adverse Events During Six-Minute Walk Test Using Continuous Physiological Signals

• DOI: 10.3389/fphys.2022.887954
• Publication Date: 2022-06-06T07:59:02Z
• Authors: Jiachen Wang; Yaning Zang; Qian Wu; Yingjia She; Haoran Xu; Jian Zhang; Shan Cai; Yuzhu Li; Zhengbo Zhang
• Keywords: 03 medical and health sciences; machine learning; wearable devices; 0302 clinical medicine; Physiology; QP1-981; 6-min walk test; adverse events; physiological signals; 3. Good health

Background and Objective: The 6-min walk test (6MWT) is a common functional assessment test, but adverse events during the can be potentially dangerous lead to serious consequences low quality of life. This study aimed predict occurrence 6MWT, using continuous physiological parameters combined with demographic variables. Methods: 578 patients respiratory disease who had performed standardized 6MWT wearable devices from three hospitals were included in this study. Adverse occurred 73 (12.6%). ECG, signal, tri-axial acceleration signals, oxygen saturation, variables scales obtained. Feature extraction selection signals 2-min resting 1-min movement phases. 5-fold cross-validation was used assess machine learning models. predictive ability different models compared. Results: Of 16 features selected by recursive feature elimination method, those related blood most important heart rate numerous. Light Gradient Boosting Machine (LightGBM) highest AUC 0.874 ± 0.063 Logistic Regression 0.869 0.067. mMRC (Modified Medical Research Council) scale Borg lowest performance, an 0.733 0.656 respectively. Conclusion: It feasible event Wearable sensors/systems for monitoring provide additional tools patient safety 6MWT.