doi:10.3850/978-981-08-7615-9_RE07


Development of a Home-based Frailty Detection Device using Wireless Sensor Networks and Artificial Neural Networks


Yu-Chuan Chang1, Chung-Chih Lin1, Jing-Siang Huang1, Chun-Chang Chen2 and Ren-Guey Lee2

1Department of Computer Science and Information Engineering, Center for Healthy Aging Research, Chang Gung University, Taoyuan 333, Taiwan, ROC

2Department of Electronic Engineering and Communication Engineering, National Taipei University of Technology, Taipei 106, Taiwan, ROC

ABSTRACT

Frailty is one of the greatest gerontological challenges faced by modern societies with ageing populations. Compared with their age-matched non-frail counterparts, frail elderly people have a much higher risk of falling related injuries, which could result in disability, hospitalization, institutionalization or even death. For this reason, early detection of frailty could help delay the onset of frailty and reduce its adverse outcomes. There are several existing frailty measurement methods, however, results are often confounded by experimenter bias and the inconveniences imposed on subjects.

The purpose of this study is to integrate wireless sensor technologies and artificial neural networks to develop a system for frailty measurement and detection that allows users to collect and manage their personal frailty information automatically. The system consists of five parts: (1) an electronic reaction meter (using an LED screen) to measure the subject’s reaction time; (2) an electronic pressure chair (using a pressure sensor) to detect slowness in movement, weakness and weight loss; (3) an electronic pad to measure the subject’s balancing ability; (4) a functional reach measuring instrument to measure body extension; and (5) a Home-based Information Gateway, which collects all the data and predicts the trend of frailty.

We designed two experiments to test the overall system performance: (1) to test the validity and reliability of the tools, we compared them with traditional frailty measurement methods. The results show that there is a correlation between frailty results from our test and results from the drop rules test, close-eyes-one-leg stand, functional reach, weight, time up and go and the 30 seconds sit-to-stand test. (2) We collected 320 cases to evaluate the sensitivity and specificity of our frailty prediction algorithm. The sensitivity and specificity of this system are 69.66% and 82.4% respectively. The positive predicted value is 82.11% and the negative predicted value is 70.07%. These results show that our system is a high specificity prediction tool that can be used to assess frailty.

Keywords: Frailty, Wireless Sensor Networks, Artificial neural networks.


     Back to TOC

FULL TEXT(PDF)