OPENING THE BLACK BOX: FEASIBILITY OF COLLECTING RAW ACCELEROMETRY DATA FROM CONSUMER WEARABLE DEVICES FOR RESEARCH WITH CHILDREN.

Authors

R Glenn Weaver, University of South CarolinaFollow
James White III, University of South CarolinaFollow
Olivia Finnegan, University of South CarolinaFollow
Aliye Cepni, University of South CarolinaFollow
Xuanxuan Zhu, University of South CarolinaFollow
Lauren Von Klinggraeff, University of South CarolinaFollow
Hannah Parker, University of South CarolinaFollow
Meghan Bastyr, University of South CarolinaFollow
Nicholas Niako, University of South CarolinaFollow
Elizabeth Adams, University of South CarolinaFollow
Bridget Armstrong, University of South CarolinaFollow
Sarah Burkart, University of South CarolinaFollow
Michael Beets, University of South CarolinaFollow

Abstract

BACKGROUND: Accessing consumer wearable accelerometry data may provide an accurate, reproducible, open-access, and device agnostic method for classifying movement. Recent studies show that raw accelerometry data (x-, y-, and z-axis accelerometry data in ɡ’s) can be collected from consumer wearable devices via application programming interfaces (API) and physical activity estimates produced from applying open-source algorithms are comparable to research grade accelerometers. A key next step is to establish the feasibility of collecting free-living raw accelerometry data from consumer wearables. This study examined the data coverage of consumer wearable raw accelerometry data in comparison to a research grade device. METHODS: Fifty-five 5-12-year-old children (63% male, 72% White) wore two of three wrist-placed consumer wearables (Apple Watch Series 7, Garmin Vivoactive 4, Fitbit Sense) and a wrist-placed research grade accelerometer (Actigraph GT9X) while attending a summer camp and over a one night at home sleep. Raw accelerometry data from the consumer wearables were extracted using a custom application that leveraged the devices’ API. Data were aggregated into 10 second epochs and explored separately for camp and sleep. Data coverage was calculated as the % of epochs that contained at least one raw accelerometry data point. A protocol was considered to be 'complete' if there was at least one reading every 10 seconds for the entirety of the protocol. RESULTS: At camp data coverage was the highest for Actigraph (93%,SD=26%) followed by Apple (86%,SD=33%), Garmin (79%,SD=29%), Fitbit (65%,SD=32%). At camp 92.7% of children had complete Actigraph data, followed by Apple (61.3%), Garmin (35.1%), and Fitbit (19.1%). For sleep accelerometer data coverage was the highest for Actigraph (96%,SD=20%) followed by Apple (80%,SD=16%), Garmin (73%,SD=41%), and Fitbit (40%,SD=36%). For sleep 95.8% of children had complete Actigraph data, followed by Garmin (50%), Apple (15.4%), and Fitbit (0%). Factors that lead to reduced coverage could include inadvertent button push that turns off the API, battery life, and unstable application design. CONCLUSIONS: Raw accelerometry data can be collected from consumer wearable devices but data coverage is poor, currently limiting the application of a device agnostic approach in consumer wearables for the collection of free-living movement data in research studies.

Recommended Citation

Weaver, R Glenn; White III, James; Finnegan, Olivia; Cepni, Aliye; Zhu, Xuanxuan; Von Klinggraeff, Lauren; Parker, Hannah; Bastyr, Meghan; Niako, Nicholas; Adams, Elizabeth; Armstrong, Bridget; Burkart, Sarah; and Beets, Michael (2024) "OPENING THE BLACK BOX: FEASIBILITY OF COLLECTING RAW ACCELEROMETRY DATA FROM CONSUMER WEARABLE DEVICES FOR RESEARCH WITH CHILDREN.," International Journal of Exercise Science: Conference Proceedings: Vol. 16: Iss. 3, Article 4.
Available at: https://digitalcommons.wku.edu/ijesab/vol16/iss3/4