Assessments of Sleep in Youth with Insomnia: A Comparison Between Polysomnography, Sleep Diary, and Actigraphy

Phoebe Hoi Ching Chan, Forrest Tin Wai Cheung, Xiao Li, Ngan Yin Chan, Joey Wing Yan Chan, Yun-Kwok Wing, Shirley Xin Li

Abstract

Introduction: Previous studies suggested poor correlations between subjective and objective sleep assessments in adults with insomnia. However, there were yet studies to systematically compare sleep diary (SD) and actigraphy with polysomnography (PSG) in youths with insomnia. Although PSG is not routinely used as an assessment tool for insomnia, it could provide valuable data that supplement the assessment of insomnia. This current study aimed to compare SD, actigraphy, and PSG as the assessments of sleep in youth with insomnia and healthy controls.

Methods: Ninety participants (age = 20.1±2.2 66% female) with symptoms of insomnia (i.e., complaint(s) of difficulty initiating or maintaining sleep, or early morning awakening at least three times a week in the past three months, with Insomnia Severity Index score > 8), and thirty-six healthy sleepers (age = 19.6±2.21 69% female) were recruited. All participants completed SD and actigraphy monitoring (Actiwatch Spectrum Plus, Philips Respironics) for eight consecutive days, during which one adaptation night followed by one night of PSG were concurrently conducted. Actigraphy data were exported at a low threshold. Total sleep time (TST), sleep efficiency (SE), sleep onset latency (SOL), and Wake after sleep onset (WASO) were extracted from SD, actigraphy, and PSG. Data was analysed using Spearman correlations and mixed repeated ANCOVA followed by Bonferroni post-hoc comparisons.

Results: Weak to moderate correlations between the three measures were found for SOL in the healthy group (r=.29 to .50, all p<.05). In the insomnia group, good concordance was found for SOL between SD and actigraphy (r=.56, p<.001). In contrast, concordance between PSG and SD (r=.30, p=.002), and between PSG and actigraphy (r=.35, p<.001) were lower, where actigraphy (24.4±19.6) and SD (35.4±27.6) significantly overestimated SOL compared to PSG (16.9±17.6). For TST, SD and actigraphy showed acceptable concordance in both healthy (r=.64, p<.001) and insomnia groups (r=.49, p<.001), but both measures showed poor concordance with PSG. In particular, TST was underestimated by an average of 58 minutes by actigraphy compared to PSG in the insomnia group. Lastly, the correlations between the three measures for WASO and SE were insignificant in both groups. For the insomnia group, actigraphy (78.4±26.6) significantly overestimated WASO when compared with PSG (25.7±44.5) and SD (9.2±13.8), whilst SE was underestimated by actigraphy compared to PSG and SD.

Conclusion: Our findings showed that PSG tends to have suboptimal concordance with SD and actigraphy for assessing sleep in individuals with insomnia. Whilst one-night PSG might have the limitation of not capturing night-to-night variability commonly observed in insomnia, the overestimation of WASO and underestimation of SE by actigraphy as compared to PSG raised the concern about the reliability of wake detection by actigraphy. Future research should explore the appropriate sensitivity settings of actigraphy and there is a need to consider concurrent use of multiple measures when assessing sleep in youth with insomnia.

Recommended citation: Chan, P. H. C., Cheung, F. T. W., Li, X., Chan, N. Y., Chan, J. W., Wing, Y. K., & Li, S. X. (2022). Assessments of Sleep in Youth with Insomnia: A Comparison Between Polysomnography, Sleep Diary, and Actigraphy. Sleep Medicine, 100, S109-S110.