A Test of Competing Mediators Linking Trouble Sleeping to Cannabis Use in Adolescents and Emerging Adults

This study examined day-level associations between trouble sleeping and three cannabis-use indices (likelihood/quantity of use and impaired control). We evaluated behavioral and cognitive mediators of the association between trouble sleeping and cannabis outcomes. Youth (N = 86, ages 15-24, 48.8% female, 58.8% White, 18.6% Latine) who regularly used cannabis were recruited for an intervention study. This preregistered secondary data analysis leveraged data from a 1-week ecological momentary assessment (EMA) study completed prior to intervention. Trouble sleeping, cannabis use, and impaired control over use were assessed each morning; negative affect, risk-taking propensity, and cannabis craving were assessed multiple times and aggregated to create a daily average. Multilevel structural equation modeling evaluated hypothesized temporally sequenced associations and putative mechanisms at the day (i.e., within) and person (i.e., between) level. In bivariate analyses at the person level, there were large-effect associations between trouble sleeping and craving and negative affect, and between craving and cannabis-use likelihood and quantity (rs from .34 to .48). In multilevel analyses at the day level, participants were less likely to use cannabis the next day after reporting more trouble sleeping (β = −.65, p < .001). Trouble sleeping was not directly associated with subsequent cannabis-use quantity or impaired control, or indirectly via negative affect, risk-taking propensity, or craving. Trouble sleeping had differential relations with cannabis-use indices at the day and person levels. To promote youth health and reduce cannabis use, future research may consider the unique, person- and situation-driven mechanistic processes at play. Public Health Significance This analysis used temporally sequenced data to examine if variations in youths' trouble sleeping related to next-day cannabis-use behaviors, and if deviations in negative affect, risk-taking propensity, and craving explained associations between trouble sleeping and cannabis behaviors. On days when youth reported more trouble sleeping than their own average, they were less likely to use cannabis the following day. This study is the first to investigate daily mechanisms relating sleep to cannabis use among youth, and provides insight for future research through investigating real-world mediational pathways linking sleep to cannabis use.