Abstract:
Novel consumer-grade devices provide the opportunity to track sleep, but their use in research is limited by a lack of validation. This study aimed to validate sleep trac...Show MoreMetadata
Abstract:
Novel consumer-grade devices provide the opportunity to track sleep, but their use in research is limited by a lack of validation. This study aimed to validate sleep tracking in free-living conditions by comparing total sleep duration measured by popular wearable devices with sleep diaries. Twenty-seven heathy volunteers of mean age 25 (± 9.49), 71% male, and with wide range of skin tones (Fitzpatrick scale from 1 to 6) wore 5 devices for 2 consecutive nights and provided sleep diaries. Devices included the Garmin Vivoactive 4 (GV4) and 4S (GV4S), Fitbit Sense (FS), Withings ScanWatch (WS), and the Oura Ring (OR), which measured sleep using both a standard and beta software. Agreement was assessed using the Spearman's correlation coefficients and Bland Altman plots. Correlation ranged from 0.41 for GV4 to 0.76 for FS, while limits of agreement ranged from (-125, 71) minutes for OR-ß to (-115, 256) minutes for GV4. Pair-wise comparisons showed that the absolute percentage error was not significantly different in most cases, except for GV4 (larger than FS, WS, OR and OR-ß) and for OR-ß (lower than OR). No association was found between the absolute error and skin tone, body mass index or wrist circumference. This data shows moderate to good agreement between wearable-enabled sleep tracking and sleep diaries in free-living conditions.
Published in: 2023 Computing in Cardiology (CinC)
Date of Conference: 01-04 October 2023
Date Added to IEEE Xplore: 26 December 2023
ISBN Information: