Abstract
BACKGROUND: The sleep diary and wrist-worn actigraphy are widely used to assess sleep disturbances in patients with primary brain tumors (PwPBT) in both clinical and research settings. However, their comparability has not been systematically examined. This study aimed to compare the sleep-wake patterns measured using the subjectively measured Consensus Sleep Diary (CSD) and the objectively measured ActiGraph (AG) actigraphy among PwPBT. METHODS: Sleep-wake patterns were assessed through CSD and AG over 14 consecutive nights across 2 occasions among 30 PwPBT. AG data were processed with AG proprietary and open-source GGIR (GGIR-based approach without the aid of sleep log algorithms), both with and without the assistance of CSD. Thirteen sleep parameters covering sleep-wake times, sleep disruptions, sleep durations, and sleep efficiency were compared using equivalency testing, mean absolute percent error (MAPE), and intra-class correlation. The estimated sleep parameters were correlated with perceived sleep quality and compared across the different sleep measures. RESULTS: Significant between-measure equivalency was claimed for sleep-wake time parameters (P ≤ .05), with acceptable MAPEs (<10%). Sleep disruption parameters such as wake-after-sleep-onset were not statistically equivalent, with a large MAPE (≥10%) between the measures. Sleep efficiency was equivalent, though varied depending on how sleep efficiency was calculated. For most sleep parameters, ICCs were low and unacceptable (<0.50) suggesting incomparability between the measures. Lastly, CSD-derived sleep parameters exhibited a stronger correlation with perceived sleep quality compared to actigraphy measures. CONCLUSIONS: The findings suggest the incomparability of sleep parameters estimated from different measures. Both subjective and objective measures are recommended to better describe sleep health among PwPBT.