Patterns and Predictors of Marijuana Use from High School to Age 50

Introduction: Considerable heterogeneity has been found in developmental trajectories of marijuana use from adolescence through early young adulthood. Empirical studies of marijuana use trajectories among the US general population that reach beyond young adulthood have not been available, but would provide needed information on the proportion of the general population that engages in moderate versus heavy use, how long such involvement continues, ages of pivotal transition between use levels, and how sociodemographic characteristics are associated with various trajectories. This study models marijuana use trajectories from ages 18 through 50 among individuals in 7 national cohorts of US high school seniors.

Methods: Frequency of past 12-month marijuana use was assessed at up to 11 time points from 16,683 individuals participating in the Monitoring the Future study. Respondents were high school seniors from the classes of 1976-1982, followed through age 50 (calendar years 2008-2014). Repeated measures latent class analysis in SAS was used to identify and estimate the prevalence of distinct patterns of change over time in no (0), moderate (1-39 occasions), or heavy (40+ occasions) past 12-month marijuana use and to explore sociodemographic characteristic associations with latent class membership probability.

Results: Seven latent classes were identified: never users (41.8%), under age 23 (under-23) moderate users (20.0%), under-30 moderate users (12.7%), generally moderate users (7.2%), under-23 heavy users (7.4%), under-30 heavy users (5.7%), and consistent heavy users (5.2%). Evidence of developmental progression by age was seen in each class (e.g., among consistent heavy users, the probability of using on 40+ occasions a year rose from 0.62 at age 18 to 0.86 at age 27/28, and then decreased to 0.67 by age 50). The odds of membership in all heavy use classes vs. never users decreased across cohorts. Yet, the odds of membership in shorter duration heavy use classes vs. consistent heavy use also decreased across cohorts. Membership in any of the three heavy use classes compared with never users was significantly associated with age 18 sociodemographic characteristics. Sociodemographic characteristics measured during young adulthood significantly differentiated between duration of heavy- and moderate use classes.

Conclusions: Results of this study identify distinct trajectories of marijuana use from age 18 to 50. The findings have significant implications for prevention, treatment, and policy efforts by providing an empirical description of marijuana use across the life span and ages of pivotal transition between use levels.