Validity of Wearable Devices for Measuring Steps, Energy Expenditure, and Heart Rate by Age, Sex and Body Mass Index
Sumayyah Musa, Ryan Ellis, Brianne Chafe, Daniel Fuller, Kim Cullen
Background: The results of systematic reviews and meta-analysis examining the concurrent validity of these devices in measuring step counts has shown limited validity for healthy adult populations. Limited research has examined whether concurrent validity of these devices varies by sociodemographic factors including age, gender, and body mass index (BMI).
Objective: The purpose of this paper is to examined the concurrent validity of commercial wearable devices in measuring step count and energy expenditure across age, gender, and BMI groups.
Methods: This study used a sample of 422 studies from publicly available data of a systematic review examining the validity of wearable devices. The proportion of men and women in a study, the mean age, and the BMI category reported for each study were used. The mean percentage error (MPE) of the commercial wearable compared to the criterion was used as the outcome. Separate ANOVA and multiple linear regression models examined the associations between age, gender, and BMI with device concurrent validity.
Results: We included 422 studies carried out between 2013 and 2019 in 13 countries. Our analysis showed a significant difference in the MPE for Step count at p < .05 level within age and BMI groups, and within BMI groups for Energy expenditure. Step count MPE in adults and normal weight individuals was significantly higher than that of older adults and overweight individuals respectively, Energy expenditure MPE in normal weight individuals was lower than that of overweight individuals at the .05 level of significance. All other comparisons were not significant.
Conclusions: Findings from this study may be an issue for wearable devices if users who are trying to lose weight or stay active rely on their data, as they may expend less energy or walk a smaller number of steps due to the general overestimation of these measures in all groups.
Clean code is available here: https://github.com/walkabillylab/wearable-validity-by-age-sex-bmi/blob/main/wearable-validity.md
Data are available here: https://doi.org/10.7910/DVN/O7GQIM or https://github.com/walkabillylab/wearable-validity-by-age-sex-bmi/blob/main/wearable_review_data_validity.csv