Investigation of the Validity and Reliability of Two Smart Bands Selected That Count Steps at Different Walking Speed

Abstract

Whether activity monitor smart wristbands that have become widespread can accurately estimate the step count while walking is a matter of curiosity. The current study aims to analyze the validity and reliability of step count (SC) estimation at normal walking and fast walking pace under controlled conditions of two selected smart wristbands of the leading wearable device vendors, Xiaomi (Mi4) and Huawei (H4). Twenty healthy adult male and twenty healthy adult female were included in the study and analyzed separately. The mean age of male and female participants was 22.25 and 21.62 years, with BMI values of 24.22 kg/m2 and 21.42 kg/m2, respectively. The above-ground walking protocol consisted of four separate five-minute tests: Normal Walking Test, Normal Walking Retest, Fast Walking Test, and Fast Walking Retest. In the study, the analyses were performed by using activity monitor measurements and criterion measurements (the number of steps determined from video recordings), compatibility of test-retest measurement values, error indicators (MPE and MAPE), Intraclass Correlation Coefficients (ICC), and Bland-Altman limits of agreement. According to the current study results, it was revealed that the MAPE values recorded for Mi4 and H4 smart wristbands for both normal and fast walking pace in male and female participants were <5%, which was deemed excellent. According to all analyses, the H4 device was found to be valid and reliable, but according to ICC and Pearson Correlation analyses, the Mi4 device was not found to be valid and reliable at fast walking pace.