An automated patient monitoring using discrete‐time wireless sensor networks. (20th March 2020)

• Record Type: Journal Article
• Authors: Chatterjee, Aditi; Biswas, Jayabrata; Das, Kiranmoy
• Is Part Of: International journal of communication systems. Volume 33:Number 9(2020)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2020-03-20
• Subjects: Bayesian Learning -- joint model -- latent variable -- patient monitoring -- wireless sensor networks; Telecommunication systems -- Periodicals; 621.382
• DOI: 10.1002/dac.4390
• Languages: English
• ISSNs: 1074-5351

Abstract

Summary: In recent years, Internet of Things (IoT) devices are used for remote health monitoring. For remotely monitoring a patient, only the health information at different time points are not sufficient; predicted values of biomarkers (for some future time points) are also important. In this article, we propose a powerful statistical model for an efficient dynamic patient monitoring using wireless sensor nodes through Bayesian Learning (BL). We consider the setting where a set of correlated biomarkers are measured from a patient through wireless sensors, but the sensors only report the ordinal outcomes (say, good, fair, high, or very high) to the sink based on some prefixed thresholds. The challenge is to use the ordinal outcomes for monitoring and predicting the health status of the patient under consideration. We propose a linear mixed model where interbiomarker correlations and intrabiomarker dependence are modeled simultaneously. The estimated and the predicted values of the biomarkers are transferred over the internet so that health care providers and the family members of the patient can remotely monitor the patient. Extensive simulation studies are performed to assess practical usefulness of our proposed joint model, and the performance of the proposed joint model is compared to that of some other traditional models used in the literature. Abstract : We propose a joint statistical model for remotely monitoring the health conditions of a patient. Our approach is based on Bayesian Learning, where the data are collected automatically at discrete time points by some sensor nodes. Our numerical examples illustrate the effectiveness of the proposed approach (less misclassification rate) compared to the traditional models.