University of Twente Student Theses


The prediction of early sepsis with wearable devices

Engelbert, P.F. (2021) The prediction of early sepsis with wearable devices.

[img] PDF
Abstract:Objective. Sepsis is a life-threatening organ dysfunction with a mortality rate of 15-25\%. Early recognition and timely treatment are necessary to improve the outcome of this infectious disease. The aim of this research is to investigate the potential of novel features extracted from continuously measured photoplethysmography (PPG) and diaphragm electromyography (EMG) as early signs in the prediction of clinical deterioration in ward patients with severe infections. Our hypothesis is that these novel features may be earlier signs of deterioration than the traditional vital signs. Methods. This research is divided into two sub-studies, where the first sub-study investigated the feasibility of the two chosen wearable devices in the desired study population and the second sub-study consisted of long-term measurements to find out which features are early signs of deterioration. In the second sub-study, the primary goal is to find out whether the parameters were able to distinguish deteriorating and non-deteriorating patients at the general ward up to six hours before the nurse's check-up. The secondary goal is to investigate whether it is possible to predict the overall clinical outcome based on measurements of the first 36 hours after hospital admission. Results. The feasibility study showed that the chosen wearable devices have met the required yield of more than 60\%. The precision and accuracy of the PPG device still left room for improvements and due to limitations in the reference method, the precision and accuracy of the EMG device could not be fully investigated. However, these limitations did not seem to influence the outcome of the long-term study much, as seventeen of the nineteen features gave significant differences in the primary outcome. Five of these features showed significant differences up to six hours before the nurse's check-up. In the secondary outcome, seven of the nineteen features were significantly different and six features showed promising differences. Conclusion. Based on the results of the long-term study, it can be concluded that several of these features show differences up to six hours before the nurse's check-up and that four features showed significantly differences at the first four hours of recording. This indicates that continuously measured PPG and diaphragm EMG can indeed be of added value in the prediction of clinical deterioration. However, before these wearables can be implemented for clinical usage, the wearables and validation methods need to be improved and more research is needed to make a prediction algorithms based on the features used in this research.
Item Type:Essay (Master)
Faculty:TNW: Science and Technology
Subject:44 medicine
Programme:Technical Medicine MSc (60033)
Link to this item:
Export this item as:BibTeX
HTML Citation
Reference Manager


Repository Staff Only: item control page