A Comparison of Physical Activity Levels, Sleep Disrupting Behavior, and Stress/Affective Distress as Predictors of Sleep as Indexed by Actigraphy

Abstract

Background: Several behaviors have been reported to interfere with sleep in otherwise healthy adults, including low physical activity (PA) levels. However, few studies have compared low PA with the other behavioral risk factors of objective sleep impairment, despite the behavior tending to cooccur in highly stressed and affectively distressed individuals. Thus, the authors compared objective and subjective measures of PA and other potential sleep disrupting behaviors as predictors of objective sleep (sleep onset latency, actual sleep time, total sleep duration, awake time, and sleep efficacy) at baseline (T1) and 3 months later (T2). Methods: A community-derived sample of 161 people aged 18–65 years were asked about PA, other behavior (ie, night eating, electronic device use, watching television, caffeine and alcohol use), stress, affective distress (ie, anxiety, depression), and demographics including shift work and parenting young children in an online questionnaire at T1 and T2. PA and sleep were also monitored for 24 hours each at T1 and T2 using actigraphy. Results: Multiple regression analyses indicated that sleep at T1 was associated with PA (ie, total number of steps, metabolic equivalents/time, time spent travelling) after controlling mean ambient temperature and relevant demographics. At T2, longer sleep onset latency was predicted by parenting young children and night time television viewing; shorter sleep duration was predicted by female gender; and awake time and sleep efficacy were predicted by alcohol intake after controlling T1 sleep measures, demographics, and mean ambient temperature. Conclusion: The risk factors for objective sleep impairment included parenting young children and watching television at night, whereas better sleep outcomes were associated with greater engagement with PA.