Uncovering Polysubstance Use Patterns in Canadian Youth with Machine Learning on Longitudinal COMPASS Data

This study aims to investigate PSU patterns in a large sample of Canadian youth and explore associated factors using data from COMPASS, a longitudinal health survey of Canadian secondary school students. The study sample consisted of 8824 students from grades 9 and 10 at baseline in 2016/17, followed over 3 years until 2018/19. Leveraging machine learning methods, especially the least absolute shrinkage and selection operator (LASSO) and multivariate latent Markov models, we conducted a comprehensive examination of PSU patterns. Our analyses revealed distinct PSU patterns among Canadian youth, including no-use (C1), alcohol-only (C2), concurrent use of e-cigarettes and alcohol (C3), and poly-use (C4). C1 showed the highest prevalence (60.5%) in 2016/17, declining by 2.4 times over 3 years, while C3 became the dominant pattern (32.5%) in 2018/19. The prevalence of C3 and C4 increased by 2.3 and 4.4 times, respectively, indicating a growing trend of dual and multiple substance use. Risk factors associated with PSU patterns included truancy (ORC2 = 1.67, 95 % CI [1.55, 1.79]; ORC3 = 1.92, 95 % CI [1.80, 2.04]; ORC4 = 2.79, 95 % CI [2.64, 2.94]), having more smoking friends, more weekly allowance, elevated BMI, being older, and attending schools unsupportive in quitting drugs/alcohol. In contrast, not gambling online (ORC2 = 0.22, 95 % CI [−0.16, 0.58]; ORC3 = 0.14, 95 % CI [-0.24, 0.52]; ORC4 = 0.08, 95 % CI [−0.47, 0.63]), eat...
Source: International Journal of Mental Health and Addiction - Category: Addiction Source Type: research