Purpose: Consumer-level activity monitors, such as Fitbit and Misfit devices, are an ever popular and low-cost means of measuring physical activity. However, the accuracy of such devices have not been extensively explored in older adults particularly in free-living conditions, limiting their use within research and clinical settings. This study aims to validate two consumer-level activity monitors in community dwelling older adults. As a secondary aim, this study will explore older adults’ satisfaction of wearing such devices.
Methods: Twenty-five community-dwelling older adults (aged 65+) simultaneously wore 5 devices; 3 consumer-level activity monitors (Fitbit Charge HR – wrist, Misfit Shine – wrist, Misfit Shine – waist) and two reference devices (Actigraph and NL2000i) during a series of walking activities in a lab-setting and then during free-living conditions for 7 days. Accuracy of the devices acheived by comparing the consumer-level devices and reference devices using Bland-Altman Plots and through Interclass Correlation Coefficients. The Assistive Device Subscale of the Quebec User Evaluation of Satisfaction with assistive Technology (QUEST) alongside open ended questions were used to assess participants’ satisfaction with wearing the devices.
Results: There was substantial agreement between the consumer-level activity monitors and reference devices. Compared to the Actigraph, the waist worn Misfit Shine displayed the highest agreement amongst the devices worn (ICC = 0.96, 95% 0.90 to 0.98). The Fitbit Charge HR showed the poorest agreement compared to the Actigraph device (ICC = 0.82, 95% 0.60 to 0.92), and displayed a systematic bias of overestimating steps taken. Participants were highly satisfied with the consumer-level devices, however it was identified that adjustment of the device was a common issue amongst older adults.
Discussion: Despite some variation, consumer-level activity monitors can accurately measure steps of healthy older adults, with excellent agreement between them and reference devices. Importantly, participants were satisfied with the design of the device, though there are some design features that could be improved to make them more acceptable for older adults. As the number of consumer-level activity monitors grow, researchers need to select devices that are not only valid and functional, but also consider whether the devices are acceptable for the target population.