Key points are not available for this paper at this time.
Purpose: The commercialization of sleep activity tracking devices has made it possible to manage sleep quality at home. However, it is necessary to verify the reliability and accuracy of wearable devices through comparison with polysomnography (PSG), which is the standard for tracking sleep activity. This study aimed to monitor overall sleep activity using Fitbit Inspire 2™ (FBI2) and to evaluate its performance and effectiveness through PSG under the same conditions. Patients and Methods: We compared the FBI2 and PSG data of nine participants (four male and five female participants; average age, 39 years) without severe sleeping problems. The participants wore FBI2 continuously for 14 days, considering the period of adaptation to the device. FBI2 and PSG sleep data were compared using paired t -tests, Bland–Altman plots, and epoch-by-epoch analysis for 18 samples by pooling data from two replicates. Results: The average values for each sleep stage obtained from FBI2 and PSG showed significant differences in the total sleep time (TST), deep sleep, and rapid eye motion (REM). In the Bland–Altman analysis, TST ( P = 0.02), deep sleep ( P = 0.05), and REM ( P = 0.03) were significantly overstated in FBI2 compared to PSG. In addition, time in bed, sleep efficiency, and wake after sleep onset were overestimated, while light sleep was underestimated. However, these differences were not statistically significant. FBI2 showed a high sensitivity (93.9%) and low specificity (13.1%), with an accuracy of 76%. The sensitivity and specificity of each sleep stage was 54.3% and 62.3%, respectively, for light sleep, 84.8% and 50.1%, respectively, for deep sleep, and 86.4% and 59.1%, respectively for REM sleep. Conclusion: The use of FBI2 as an objective tool for measuring sleep in daily life can be considered appropriate. However, further research is warranted on its application in participants with sleep-wake problems. Keywords: wearable, sleep, tracking, polysomnography, validation study
Lim et al. (Wed,) studied this question.