[Skip to Content]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address 54.158.119.60. Please contact the publisher to request reinstatement.
[Skip to Content Landing]
Views 105
Citations 0
Comment & Response
September 2016

Remaining Questions Concerning Wearable Devices—Reply

Author Affiliations
  • 1Department of Health Promotion and Exercise, National Institute of Health and Nutrition, NIBIOHN, Shinjuku, Tokyo
JAMA Intern Med. 2016;176(9):1409-1410. doi:10.1001/jamainternmed.2016.4762

In Reply We would like to thank Dr Lee and Drs Pai and Budd for their interest in our study.1 They pointed out several methodological concerns and the need to pursue the accuracy of wearable devices. Their questions are probably common to many readers, and we would like to reply to them as an opportunity to help readers understand our study.

We have gone to a great deal of effort to minimize estimation bias from the previous findings and concluded that all 12 devices underestimated total energy expenditure (TEE) with wide variability under free-living conditions. Longer nonwearing time under real living conditions is the primary factor responsible for this underestimation. As all participants strictly complied with the provisions, mean (SD) nonwear time except during sleep was 42 (14) min/d, as revealed from their diaries about daily activities and the device status. The mean (SD) TEE for nonwear time was 28.7 (27.1) kcal/d. Even with the addition of 28.7 kcal/d, the TEE estimated by all devices was lower than the standard TEE determined by the doubly labeled water method. Placement of devices (eg, dominant or nondominant arm, proximal or distal arm, and right or left hip) is also a concern. To minimize bias, the devices on the arm were worn on the nondominant arm accordance with their product’s manuals, and the wearing position for each device was randomly assigned for each participant. In addition, simultaneous wearing of 6 lots of each device confirmed no trend or variability according to placement (CV = 0.8% ± 1.0%, from 0.0% to 3.2%). We should emphasize that the estimated TEE would be more underestimated in real living, because general users have longer nonwear times than the participants included in our study.

First Page Preview View Large
First page PDF preview
First page PDF preview
×