BACKGROUND AND AIMS Adolescent polysubstance use has been associated with adverse social and health outcomes. Our aim was to measure rates and transitions to polysubstance use during adolescence and identify factors associated with initiation and discontinuation of polysubstance use.

DESIGN Prospective cohort study. Multistate Markov modelling was used to estimate rates and identify correlates of transitions between substance use states.

SETTING AND PARTICIPANTS Adolescent-parent dyads (n = 1927; adolescents in grade 7, age ≈13 years) were recruited from Australian schools during 2010/11 (Wave 1). Adolescents were surveyed annually until 2016/17 (n = 1503; age ≈19 years; Wave 7) and parents were surveyed annually until 2014/15 (Wave 5).

MEASUREMENTS Alcohol, tobacco, cannabis and 3,4-methylenedioxymethamphetamine (MDMA) use outcomes were collected at Waves 3-7. Potential confounders were collected at Waves 1-6 and consisted of sex, anxiety and depression symptoms and externalizing problems, parental monitoring, family conflict and cohesion, parental substance use and peer substance use. Covariates were age and family socioeconomic status.

FINDINGS Few adolescents engaged in polysubstance use at earlier waves (Wave 3: 5%; Wave 4: 8%), but proportions increased sharply across adolescence (Waves 5-7: 17%, 24%, 36%). Rates of transitioning to polysubstance use increased with age, with few (<9%) adolescents transitioning out. More externalizing problems (odds ratio [OR] = 1.10; 99.6% confidence interval [CI] = 1.07-1.14), parental heavy episodic drinking (OR = 1.22; 99.6% CI = 1.07-1.40), parental illicit substance use (OR = 3.56; 99.6% CI = 1.43-8.86), peer alcohol use (OR = 5.68; 99.6% CI = 1.59-20.50) and peer smoking (OR = 4.18; 99.6% CI = 1.95-8.81) were associated with transitioning to polysubstance use.

CONCLUSIONS Polysubstance use in Australia appears to be rare during early adolescence but more common in later adolescence with low rates of transitioning out. Externalizing problems and greater parental and peer substance use are risk factors for adolescent polysubstance use that may be suitable intervention targets.