J.E. Slaven, A. Mnatsakanova, S. Li, J.M. Violanti, C.M. Burchfiel, B.J. Vila, M.E. Andrew
Biostatistics and Epidemiology
Branch, Health Effects Laboratory Division, National Institute for Occupational
Safety and Health, Centers for Disease Control and Prevention,
1095 Willowdale Rd, MS 4050, Morgantown, WV 26505, USA.
Sleep quality has a large impact on daily performance and general health. Among the different methods of objectively
measuring sleep quality, actigraphy continues to be very popular. It can take continuous activity measurements
over several days in order to determine sleep-wake cycles and calculate sleep variables, including the three standard sleep
variables used in determining sleep quality: total sleep time, sleep efficiency, and wake-after-sleep onset, in which study
analyses use the mean of these variables. In the case of wake-after-sleep onset, which calculates the amount of time between
falling asleep and waking up, the average does not characterize wake-after-sleep times as it does not account for the
total number of awakenings or the frequencies of wake-after-sleep times. Instead, we recommend using the entire distribution
of wake-after-sleep onset times, which we will call waiting time distribution, which better characterizes wake-aftersleep
onset than the average value. Sleep quality for each participant was determined by their total sleep time and sleep efficiency.
Non-parametric statistics were utilized to determine differences in waiting time distributions between participants
with good and poor quality of sleep. Discriminant analysis was performed to show that a distribution of waiting
times discriminates better between qualities of sleep than the average wake-after-sleep onset time does. Waiting time distributions
were also fit to standard probability distributions for utility and ease of understanding. Analyses show that the
waiting time distribution categorizes sleep qualities better than the average wake-after-sleep onset variable, as well as giving
more information and better characterizations.