Changes in the transition from cannabis use to use disorder by decade of first use: Findings from the Australian National Mental Health and Wellbeing surveys.

Wilson J.; O'Dean S.; Marel C.; Dobson O.; Sunderland M.; Kershaw S.; Gückel T.; Devine E.; Slade T.; Teesson M. et al. · Addiction (Abingdon, England) · 2026

**Background & Methods** This cross-sectional analysis examined data from the Australian National Surveys of Mental Health and Wellbeing (2007 and 2020-22) including 4,731 cannabis users across two decades. The study used survival analysis to compare cumulative risk of cannabis use disorder (CUD) by decade of first use (<1980s, 1980s-1990s-2000s-≥2010s), with Cox proportional hazards models adjusting for demographics, comorbid substance use disorders, mental health conditions, and service utilization. **Key Findings** • Cumulative risk of CUD following cannabis initiation was lowest in pre-1980 users (1-year: 7%; 5-year: 16%; 10-year: 18%) and significantly higher among those initiating in the 1990s (12%, 25%, 28% respectively) and 2000s (13%, 24%, 26% respectively). • After adjusting for confounders, cannabis users initiating in the 1990s (HR=1.47), 2000s (HR=1.37), and ≥2010s (HR=1.70) demonstrated statistically significantly elevated hazards of developing CUD compared to pre-1980 initiates. • The increased risk trajectory in recent cohorts suggests that enhanced cannabis potency and improved product accessibility are primary drivers of accelerated progression from use to disorder. **Dosage & Administration** Not reported. **Safety & Adverse Effects** Not reported. **Evidence Quality** This study provides moderate evidence based on large, population-representative Australian samples with validated clinical assessments. Limitations include reliance on retrospective self-reported age of first use and CUD onset, potentially subject to recall bias and affecting accuracy of progression timelines. The cross-sectional design prevents causal inference. Confidence intervals overlapped at earlier time points between some cohorts, suggesting borderline statistical significance for certain comparisons. Unmeasured confounders related to product characteristics (THC concentration, CBD content) and individual susceptibility factors were not directly quantified. The findings are generalizable primarily to Australian populations and may not fully represent global cannabis markets or diverse user demographics.

More recent cannabis research