Research Article
A Health Status Evaluation Method for Chronic Disease Patients Based on Multivariate State Estimation Technique Using Wearable Physiological Signals: A Preliminary Study
@INPROCEEDINGS{10.1007/978-3-031-06368-8_1, author={Haoran Xu and Zhicheng Yang and Ke Lan and Wei Yan and Zhao Wang and Jiachen Wang and Yaning Zang and Jianli Pan and Muyang Yan and Zhengbo Zhang}, title={A Health Status Evaluation Method for Chronic Disease Patients Based on Multivariate State Estimation Technique Using Wearable Physiological Signals: A Preliminary Study}, proceedings={Wireless Mobile Communication and Healthcare. 10th EAI International Conference, MobiHealth 2021, Virtual Event, November 13--14, 2021, Proceedings}, proceedings_a={MOBIHEALTH}, year={2022}, month={6}, keywords={Health status evaluation Chronic disease management Multivariate state estimation technique Physiological signals}, doi={10.1007/978-3-031-06368-8_1} }- Haoran Xu
Zhicheng Yang
Ke Lan
Wei Yan
Zhao Wang
Jiachen Wang
Yaning Zang
Jianli Pan
Muyang Yan
Zhengbo Zhang
Year: 2022
A Health Status Evaluation Method for Chronic Disease Patients Based on Multivariate State Estimation Technique Using Wearable Physiological Signals: A Preliminary Study
MOBIHEALTH
Springer
DOI: 10.1007/978-3-031-06368-8_1
Abstract
Since chronic disease has become one of the most profound threats to human health, effective evaluation of human health and disease status is particularly important. In this study, we proposed a method based on Multivariate State Estimation Technique (MSET) by using physiological signals collected by a wearable device. Residual was defined as the difference between the actual value of each observed parameter and the estimated value obtained by MSET. The high-dimensional residual series were fused into a Multivariate Health Index (MHI) using a Gaussian mixture model. To preliminarily validate this method, we designed a retrospective observational study of 17 chronic patients with coronary artery disease combined high risk of heart failure whose Brain Natriuretic Peptide (BNP) had changed significantly during hospitalization. The results show that the distribution of residuals estimated by MSET had some regularity, in which the Pearson correlation coefficients between Cohen Standardized Mean Difference (SMD) and Overlapping Coefficient (OVL) of MHI and the change of BNP examination results reached 0.786 and 0.835, with theirp-values less than 0.001, respectively. We preliminarily demonstrated that the model can reflect the level of change in human health status to some extent. This MSET-based approach shows great potential for applications of treatment effect evaluation, and provides abundant information from physiological signals in chronic disease management.
