phat 18(15): e4

Research Article

Not All Errors Are Created Equal: Influence of User Characteristics on Measurement Errors of Consumer Wearable Devices for Sleep Tracking

Download56 downloads
  • @ARTICLE{10.4108/eai.24-7-2018.159404,
        author={Zilu Liang and Mario Alberto Chapa Martell},
        title={Not All Errors Are Created Equal: Influence of User Characteristics on Measurement Errors of Consumer Wearable Devices for Sleep Tracking},
        journal={EAI Endorsed Transactions on Pervasive Health and Technology},
        volume={4},
        number={15},
        publisher={EAI},
        journal_a={PHAT},
        year={2018},
        month={7},
        keywords={wearable, sleep, validation, error analysis, Fitbit, EEG, personal informatics},
        doi={10.4108/eai.24-7-2018.159404}
    }
    
  • Zilu Liang
    Mario Alberto Chapa Martell
    Year: 2018
    Not All Errors Are Created Equal: Influence of User Characteristics on Measurement Errors of Consumer Wearable Devices for Sleep Tracking
    PHAT
    EAI
    DOI: 10.4108/eai.24-7-2018.159404
Zilu Liang1,2,*, Mario Alberto Chapa Martell3
  • 1: Kyoto University of Advanced Science, 18 Yamanouchi Gotanda-cho, Ukyo-ku, Kyoto 615-8577, Japan
  • 2: The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
  • 3: CAC Corporation, 24-1 Hakozakicho, Chuo-ku, Tokyo 103-0015, Japan
*Contact email: z.liang@cnl.t.u-tokyo.ac.jp

Abstract

Consumer sleep tracking devices are known to be inaccurate, but there is a lack of understanding of how user characteristics may affect the accuracy of these devices. This study aims to examine the effect of age, gender, subjective sleep quality, sleep hygiene and sleep structure on the accuracy of two consumer sleep trackers, i.e. Fitbit Charge 2 and Neuroon. Sleep data were collected from 27 healthy participants using consumer devices and a medical device concurrently. Analysis found that age, sleep hygiene and sleep structure were significantly associated to the accuracy of consumer sleep trackers, whereas no association was found on gender and subjective sleep quality. Both consumer devices had improved accuracy on total sleep time and sleep efficiency for participants who had longer, deeper and less interrupted sleep. Our findings suggest that consumer devices may not be suited for young adults and for people with short and fragmented sleep.