Mediating effects of sleep on mental health in older adolescents : Findings from the Burn 2 Learn randomized controlled trial

Abstract

Purpose Our study explored the mediating effect of sleep-related variables on older adolescents' mental health in the context of a school-based physical activity intervention.

Methods We evaluated the Burn 2 Learn (B2L) intervention using a cluster randomized controlled trial, which included two cohorts. Participants for this sub-study were from the second cohort, which included 292 older adolescents (16.0 ± 0.5 years) from 10 secondary schools in New South Wales, Australia. Teachers at intervention schools delivered two high-intensity activity breaks (approximately 10 mins) per week to students during academic lessons. Participants completed measures of mental health (i.e., perceived stress and internalizing problems) and hypothesized mediators (i.e., sleep duration, sleep latency, awakenings, and daytime sleepiness) at baseline (February–April 2019) and post-intervention (August–September 2019). Single mediation analyses were conducted to explore the potential mediating effects of sleep variables on mental health outcomes using a product-of-coefficient test.

Results We observed a small statistically significant effect for perceived stress (β = −0.11, SE = 0.034, p = 0.002), but not for internalizing problems (β = 0.02, SE = 0.051, p = 0.760). There were no significant intervention effects for sleep-related variables. Several sleep-related variables were associated with mental health outcomes but no mediated effects were found.

Conclusion The B2L intervention had a small beneficial effect on perceived stress, however our mediation analyses suggest this was not explained by changes in sleep-related variables. Markers of sleep were associated with mental health constructs, highlighting the importance of sleep for good psychological health. However, in the context of a physical activity intervention, effects on mental health may be driven by other behavioral, neurobiological, or psychosocial mechanisms.