Examining the Accuracy of Fitness Trackers with Optical Heart Rate Technology: A Comparison between Young and Senior Adults

  • 楊 晁青

Student thesis: Master's Thesis

Abstract

Introduction: To promote the health condition of general population activity trackers might be ideal tools to record and even boost user’s physical activity In recent years the activity trackers with photoplethysmograpphy or optical heart rate sensor can even acquire the real-time heart rate readings which is an important parameter for monitoring and controlling exercise intensity Although the practical benefits of optical heart rate monitors have been highlighted by more and more studies there are scarce studies explore the accuracy of commercial fitness trackers with second-level data Thus this current study aimed to examine the feasibility of applying both devices to monitor instant heart rate during moderate exercises for young and senior adults Methods: This research recruited 20 young adults and 20 senior adults All participants had undergone series of sedentary and moderate physical activities by using indoor aerobic exercise equipment Polar H7 chest strap heart rate monitor was set as the criterion measurement against Xiaomi Mi Band 2 and Garmin VivosmartHR+ The second-by-second heart rate data from both devices were recorded through broadcast heart rate mode Multi-statistical methods included mean absolute percentage error (MAPE) Lin’s concordance correlation coefficient (CCC) intraclass correlation coefficient (ICC) Pearson’s product correlation coefficient and Bland-Altman analysis were performed to examine the accuracy of both devices Results: In agreement with previous researches both devices performed acceptable overall accuracy based on several statistical tests First the MAPE values were all below 10% (Garmin/Young = 3 77%; Garmin/Senior = 4 73%; Xiaomi/Young = 7 69% ; Xiaomi/Senior = 6 04%) Second the reliability of Garmin from correlation tests were moderate to good (CCCyoung= 92; CCCsenior= 80) As for Xiaomi the correlation coefficients were both moderate (CCCyoung= 76; CCCsenior= 73) The results from Bland-Altman analysis revealed that the optical devices would averagely underestimated the heart rate Nonetheless the most important finding from this current study was that some extreme errors occurred unexpectedly in certain participants Conclusion: This study has proved acceptable overall accuracy of selected fitness trackers and age might not be an influential factor on optical heart rate detection However it’s strongly recommended that future researches can lay emphasis on the occurrence rate of extreme errors which might lead to the misjudgment of real-time exercise intensity
Date of Award2019
Original languageEnglish
SupervisorHsueh-Wen Chow (Supervisor)

Cite this

'