Polysubstance use in Canada: prevalence, patterns, and self-reported harms from the 2023 Canadian Substance Use Survey

BACKGROUND: Polysubstance use contributes substantially to substance-related morbidity and mortality in Canada, yet contemporary population-level estimates of alcohol-involved polysubstance use are limited. We estimated national prevalence patterns and examined associations with self-reported harms using 2023 Canadian Substance Use Survey (CSUS) data. METHODS: We conducted a cross-sectional analysis of the 2023 CSUS public-use microdata file, including Canadians aged ≥ 15 years residing in the provinces. Past-year substance-use patterns were categorized as single-substance use, drug-only polysubstance use (≥ 2 drug classes without alcohol), and drug-plus-alcohol polysubstance use (alcohol plus ≥ 1 drug class). Weighted prevalence estimates were calculated using survey weights and 1,000 bootstrap replicate weights. Survey-weighted logistic regression examined associations with any self-reported harm. A primary model adjusted for sociodemographic confounders; a sensitivity model additionally adjusted for self-rated physical and mental health. RESULTS: Past-year polysubstance use was reported by 27.2% (95% CI 26.0-28.5) of respondents and was predominantly alcohol-involved (27.0%, 95% CI 25.7-28.3). Drug-only polysubstance use was uncommon (0.3%, 95% CI 0.2-0.4). Common patterns included alcohol plus cannabis (14.4%) and alcohol plus stimulants (11.6%). Compared with single-substance use, drug-plus-alcohol polysubstance use was associated with reporting any harm (adjusted odds ratio [aOR] 4.80; 95% CI 1.02-22.69). Results were robust in sensitivity analyses adjusting for self-rated health (aOR 4.41) and using a stricter concurrent-use definition (aOR 5.72). CONCLUSIONS: Past-year polysubstance use is common in Canada and is largely driven by alcohol co-use. Integration of alcohol into polysubstance prevention, screening, and surveillance frameworks may strengthen public health responses to substance-related harms.