Wearable Device versus Polysomnography for the Assessment of Sleep Characteristics in Patients with Sleep Disorders

Abstract

Objective: To compare sleep efficiency (SE), total sleep time (TST), and sleep stages recorded by a wearable device (WD) and polysomnography (PSG) in Thai patients with sleep disorders. Materials and Methods: Patients aged ≥18 years scheduled for PSG were included in this cross-sectional study. All research subjects completed questionnaires and wore a WD (Fitbit Alta HR®) on the same night they underwent PSG study. The data from the WD were transferred to a mobile phone and analyzed independently of PSG results, which were scored by sleep technicians. Bland-Altman plots and intraclass correlation coefficients (ICCs) were used for the analyses. Results: Data from 55 patients (33 males, 22 females) were analyzed, with four patients excluded due to data errors. The mean differences between WD and PSG for SE (%) and light sleep were 8.4±23.8 and 43.6±26.4, respectively, both statistically significantly (p<0.05). The ICCs for SE and light sleep were -0.03 and -0.04, indicating poor reliability. However, the mean differences for TST, deep sleep, and REM sleep between the two methods were not statistically significant (p>0.05), with ICC values of 0.17, 0.16, and 0.25, respectively, all considered poor correlations. Conclusion: In patients with sleep disorders, sleep characteristics measured by the WD and PSG showed some differences and weak correlations. As technology advances, the accuracy of wearable devices may improve. Further studies are needed to evaluate different devices and populations.