Fuzzy logic as a decision support tool for vital signs monitoring

Abstract

The main objective of this research was to investigate whether a fuzzy logic rule-based simple decision support system could be used to detect abnormal health conditions, by processing data received from vital signs monitoring devices. Validation of the approach was undertaken using a synthetic vital signs dataset. An application of the system to predict postural status of a person was demonstrated using real data, to mimic the effects of body position changes while doing normal daily activities. Specifically, this project intended: to design and implement a reliable fuzzy logic rule-based vital signs data processing system, to provide simple decision support functions; to evaluate the performance of the system using synthetic and laboratory gathered data; to determine the sensitivity and specificity of the system in recognizing postural status of an individual, using vital signs. The fuzzy logic system approach was applied to a synthetic dataset to recognise the physiological (normal/abnormal) states of a person based on his/her vital signs. This produced a successful outcome and provides confidence that the same methodology could be used to work on similar clinical problems in reality. Next, the approach was applied to real experimental data obtained from a subject using a tilt table to detect three postural status conditions: Static, Raising, and Lowering. The fuzzy logic prediction system gave results with Sensitivity values of 0.56 to 0.89, Specificity values of 0.75 to 0.97, Accuracy rates of 0.85 to 0.87, and Error rates of 0.13 to 0.15. Applying this type of fuzzy logic approach, a decision system could to some extent be used to inform necessary actions by caregivers or a person themself to take care decisions for their health situation.

Date of Award	2013
Original language	English