From population outcomes to demand curves: A systematic review of cross-substance effects of tobacco, cannabis, and alcohol policies.

Freitas-Lemos R.; de Lima Bovo A.; Bruckner L.; Mason M.; Baia F.; Henriques M.; Nascimento V.; Fontes R. · Preventive medicine · 2026

**Background & Methods** This systematic review synthesized evidence from policy evaluation and behavioral economics studies examining how tobacco, cannabis, and alcohol policies affect use of these and other substances. Searches across four major databases (PubMed, PsycINFO, CINAHL Complete, Web of Science) through February 2026 identified 95 studies contributing 250 evaluations of cross-substance policy effects, including 245 real-world policy evaluations and 5 behavioral economic studies. **Key Findings** - Across all evaluated policy interventions, substance use responses were most commonly independent (n=105 evaluations, 42%), followed by complementary effects where substances are used together (n=71, 28%), and substitution effects where reducing one substance increases another (n=39, 16%). Four evaluations reported mixed relationships, and 31 reported no measurable own-substance effect. - The direction and magnitude of cross-substance effects varied significantly by policy domain (price vs. regulatory approaches) and specific substance combinations, with heterogeneity analyses frequently revealing subgroup-specific patterns not apparent in pooled estimates, suggesting that policy impacts are population-dependent. - Behavioral economic findings from experimental studies were broadly consistent with real-world policy evaluations, though experimental evidence remains limited, indicating that laboratory-derived demand curves may reasonably predict actual population responses to substance policies. **Dosage & Administration** Not reported. **Safety & Adverse Effects** Not reported. **Evidence Quality** This review presents moderate quality evidence with significant limitations. The heterogeneity in study designs, outcome measures, and substance combinations limits definitive conclusions, and the preponderance of real-world observational evaluations (98%) over experimental evidence restricts causal inference. Many included studies likely had different methodological rigor and potential confounding variables. The finding that 31 evaluations reported no own-substance effect raises questions about publication bias and study quality. The authors appropriately acknowledge that comprehensive assessments of cross-substance outcomes remain limited and call for more rigorous, routine evaluation of unintended policy consequences across substance combinations.

More recent cannabis research