Measuring sleep parameters of naval sailors: A comparison between subjective self-report and wrist actigraphy

Wrist actigraph and self-report activity logs were used in a Royal Canadian Navy's at-sea exercise to track sleep patterns of naval personnel. In this study, we compared sleep parameters obtained from two measurement methods and investigated their intrinsic biases. The results revealed a strong agreement between two methods for recording sleep offset times, but a relatively poor agreement for parameters that include substantial periods of transition between sleep and wake states. Overall, self-reported sleep durations were substantially longer than actigraphic estimates (mean bias of -30.6 min; limits of agreement -95.9 to 34.8 min), and the discrepancy was mainly caused by differences in two methods to track sleep onset latency and Wake-up After Sleep Onset (WASO). Based on a customised activity log, a strong positive correlation (rho = 0.75, p < .001) between self-report and actigraphy was observed for sleep duration estimates, which confirmed the effectiveness of the activity log in field studies. Between two participant groups with different work schedules, the agreement between self-report and actigraphy was consistently better for day workers than watch keepers. The findings inform future sleep research planning that involves naval personnel in field settings.