Smart wearable devices that capture physical activity data are increasingly used for health research and show potential for augmented cognition. These devices must be tested to understand their function before use in research and everyday life. However, there are few standards for the evaluation of step count comparisons between devices. We completed a technical function evaluation of two consumer-grade devices – Fitbit Versa 3 and generation 2 Oura Ring – against research-grade gold standard ActiGraph devices – wGT3X-BT and GT9X-Link. We compared data analysis methods to evaluate smart wearable physical activity data to inform development of standards and guidance for data analysis. Based on this effort, we suggest the use of Median Absolute Percent Difference along with Spearman’s Rho as a correlation measure and Bland-Altman plots to visualize the agreement. This combination of measures provides a multi-perspective view of step counts and can assist researchers in determining limitations and best uses for smart wearable devices.
CITATION STYLE
Boles, K. K., Gallimore, M. R., Howland, C., Emezue, C., & Reeder, B. (2023). Technical Function Evaluation of Two Smart Wearables and Data Analysis Methods for Step Counts. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 14019 LNAI, pp. 71–88). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-35017-7_6
Mendeley helps you to discover research relevant for your work.