Modifiable predictors of insufficient sleep durations: A longitudinal analysis of youth in the COMPASS study

Abstract

The purpose of the current study was to simultaneously examine commonly proposed risk and protective factors for sleep deprivation over time among a large cohort of Ontario and Alberta secondary school students. Using 4-year linked longitudinal data from youth in years 1 through 4 (Y1[2012/2013], Y2[2013/2014], Y3[2014/2015], Y4[2015/2016]) of the COMPASS study (n=26,205), the likelihood of students meeting contemporary sleep recommendations was tested based on their self-reported substance use, bullying victimization, physical activity, and homework and screen time. Models controlled for the effect of student-reported gender, race/ethnicity, grade, school clustering, and all other predictor variables. Relative to baseline, students became less likely to meet the sleep recommendations if at follow-up they had initiated binge drinking, experienced cyber bullying victimization, or were spending more time doing homework, with other factors held constant. The likelihood of reporting sufficient sleep increased if students had begun engaging in resistance training at least three times a week. No longitudinal effect was observed when students increased their caffeine consumption (energy drinks, coffee/tea), initiated cannabis or tobacco use, experienced other forms of bullying victimization (physical, verbal, or belongings), engaged in more moderate-vigorous physical activity, or increased their screen use of any type. Few of the commonly purported modifiable risk and protective factors for youth sleep deprivation held in multinomial longitudinal analyses. Causal conclusions appear premature, with further research required to confirm the targets likely to be most effective in assisting more youth in meeting the sleep recommendations.