Trends in cannabis use disorder among adults in the United States, 2021-2024

AI Summary

CUD prevalence rose from 5.9% (2021) to 7.4% (2024), average 6.8%, with annual adjusted odds ratio 1.09.
Upward CUD trends were consistent across age, sex, race and ethnicity, education, income, employment and metropolitan status, affecting historically disadvantaged groups.
Rising CUD observed in states both with and without medical cannabis laws, with similar annual percentage point increases.

AI Summary

While increases in cannabis use during the early pandemic are documented, less is known about more recent trends in cannabis use disorder (CUD).

We used survey-weighted logistic regression to model past-year cannabis use disorder (CUD; DSM-5 yes/no) as a function of survey year, sociodemographic characteristics, and state medical cannabis law (MCL) status.

Average marginal effects showed higher annual percentage point (pp) changes in CUD overall (pp = 0.54) and among all age groups (pp range = 0.35-0.89); males (pp = 0.69) and females (pp = 0.40); Hispanic (pp = 0.27), non-Hispanic White (pp = 0.68), and non-Hispanic Black (pp = 0.90) adults; across all educational attainment (pp range = 0.35-0.76), annual household income (pp range = 0.49-0.63), employment (pp range = 0.44-0.60) statuses; metropolitan status subgroups (pp range = 0.52-0.63; and among residents of states with and without MCL coverage (p = 0.54 for both).

Basic summary

Addict Behav. 2026 Aug 12;184:108835. doi: 10.1016/j.addbeh.2026.108835. Online ahead of print.

ABSTRACT

BACKGROUND: The COVID-19 pandemic coincided with disruptions to mental health and substance use behaviors in the United States (U.S.), alongside continued changes in cannabis policy. While increases in cannabis use during the early pandemic are documented, less is known about more recent trends in cannabis use disorder (CUD).

METHODS: We used repeated cross-sectional data on adults (aged 18 +) from the 2021-2024 National Survey on Drug Use and Health (n = 186,823). We used survey-weighted logistic regression to model past-year cannabis use disorder (CUD; DSM-5 yes/no) as a function of survey year, sociodemographic characteristics, and state medical cannabis law (MCL) status. Overall and subgroup-specific adjusted marginal linear trends were estimated by including year-by-covariate interaction terms.

RESULTS: The proportion of CUD was 5.9% (2021), 7.0% (2022), 6.9% (2023), and 7.4% (2024), with an average of 6.8%. Increasing survey year was associated with greater odds of CUD (AOR: 1.09, 95% CI: 1.06-1.13). Average marginal effects showed higher annual percentage point (pp) changes in CUD overall (pp = 0.54) and among all age groups (pp range = 0.35-0.89); males (pp = 0.69) and females (pp = 0.40); Hispanic (pp = 0.27), non-Hispanic White (pp = 0.68), and non-Hispanic Black (pp = 0.90) adults; across all educational attainment (pp range = 0.35-0.76), annual household income (pp range = 0.49-0.63), employment (pp range = 0.44-0.60) statuses; metropolitan status subgroups (pp range = 0.52-0.63; and among residents of states with and without MCL coverage (p = 0.54 for both).

CONCLUSIONS: CUD prevalence was higher in recent years, with consistent upward trends observed across sociodemographic subgroups, including historically disadvantaged populations. These findings underscore the need for continued policy attention to address rising CUD risk.

PMID:42617538 | DOI:10.1016/j.addbeh.2026.108835

Document this CPD