Highlight
- Most adults in US recreational cannabis states exhibit maladaptive responses to health warnings.
- Maladaptive responses are linked to higher reactance, less accurate beliefs about cannabis harms, and reduced intentions to minimize use-related harm.
- BERTopic modeling and EPPM coding provide nuanced insights into consumer resistance and acceptance of warning messages.
- Findings suggest a need for tailored, evidence-based messaging to improve harm reduction in cannabis consumers.
Background
The legalization of recreational cannabis across multiple US states has brought renewed attention to the effectiveness of public health communications targeting cannabis use. As prevalence of use increases, understanding how adults perceive and respond to health warnings becomes vital for mitigating potential adverse health outcomes. Cannabis use is associated with a spectrum of health risks, including impaired cognitive function, increased risk of psychiatric disorders, and potential for dependence, especially among frequent users and those at higher risk for problematic cannabis use (PCU) [1,2]. Public health authorities have implemented warning labels to inform consumers, but the real-world impact of these warnings—particularly in environments where cannabis is normalized—remains uncertain.
Study Overview and Methodological Design
Massey et al. (2025) conducted a cross-sectional, mixed-methods study guided by the Extended Parallel Process Model (EPPM), a behavioral framework positing that individuals respond to threat-based messages with either adaptive (risk acknowledgment and behavior change) or maladaptive (denial, reactance, or avoidance) responses [3]. In 2022, 1,078 adults aged 21 or older, residing in US states permitting recreational cannabis use and who reported cannabis use in the past year, were randomly assigned to view either text-only or pictorial health warnings on cannabis products.
Participants provided free-text responses to the warnings. These qualitative reactions were analyzed using BERTopic, a machine learning-based topical modeling approach, to identify thematic clusters. Manual coding subsequently classified responses according to EPPM categories (adaptive or maladaptive). Three linear regression models were performed, with the primary outcomes being levels of reactance to warnings, accuracy of health beliefs regarding cannabis harms, and intentions to adopt harm-preventing behaviors. Predictors included cannabis use frequency, PCU risk, risk perceptions, and EPPM response, with warning type as a controlling variable.
Key Findings
BERTopic modeling revealed 11 main subtopics in participant responses. Of 1,078 participants, 57.9% (n=624) exhibited maladaptive responses, while 42.1% (n=454) showed adaptive responses. Notably, maladaptive responses were significantly associated with:
- Higher reactance to health warnings (regression coefficient β = 0.38)
- Lower accuracy in health beliefs about cannabis-related harms (β = -0.32)
- Reduced intentions to prevent cannabis-associated harms (β = -0.31)
These associations remained significant after adjusting for frequency of use, PCU risk, and warning condition. The findings underscore a substantial resistance among cannabis consumers to established health warnings, with implications for the design and delivery of future public health interventions.
Mechanistic Insights and Pathophysiological Context
The EPPM posits that warning messages must balance threat (severity and susceptibility) with efficacy (response and self-efficacy) to motivate adaptive behaviors [3]. When perceived threat outweighs perceived efficacy, individuals may react defensively—denying risk, minimizing message credibility, or dismissing the warning entirely (maladaptive response). In this study, maladaptive responders demonstrated greater psychological reactance, a phenomenon wherein individuals resist perceived attempts to restrict their autonomy. This is particularly salient in contexts where cannabis is legal and culturally accepted, potentially undermining message credibility and perceived relevance.
Clinical Implications
For clinicians and public health strategists, these findings highlight both the limitations and opportunities inherent in current cannabis risk communication. The predominance of maladaptive responses suggests that existing warning strategies may insufficiently address consumer beliefs or may inadvertently trigger reactance, particularly among frequent or high-risk users. Integrating behavioral science principles—such as enhancing message efficacy, tailoring content to address common misperceptions, and employing non-judgmental, harm reduction-oriented messaging—may improve engagement and promote safer use behaviors [4,5].
Case vignette: John, a 29-year-old regular cannabis user from Colorado, reports routinely ignoring warning labels, perceiving them as exaggerated and irrelevant. He expresses skepticism about the evidence and little intention to alter his patterns of use. This scenario reflects the study’s findings and underscores the need for innovative approaches to risk messaging.
Limitations and Controversies
Several limitations should be considered. First, the cross-sectional design precludes causal inference regarding the relationship between warning responses and behavioral intentions. The reliance on self-reported data introduces potential biases, including social desirability and recall bias. Additionally, the study population—adults in recreational cannabis states who are current users—may not reflect attitudes among non-users, medical cannabis users, or populations in non-legalized states. The analysis did not explore potential demographic moderators (e.g., age, socioeconomic status, ethnicity) that may influence message receptivity. Finally, while the EPPM is a robust behavioral model, its application to cannabis risk messaging requires ongoing validation.
Controversy persists regarding the optimal balance of threat and efficacy messaging for cannabis, particularly as the legal and cultural landscape evolves. Some experts advocate for stronger, pictorial warnings modeled after tobacco control, while others caution against approaches that may further alienate or stigmatize users [6].
Expert Commentary or Guideline Positioning
Current guidelines, such as those from the Centers for Disease Control and Prevention (CDC), recommend plain-language warnings describing known cannabis risks, especially for vulnerable populations (youth, pregnant women, individuals with psychiatric comorbidities) [7]. However, translation into practice remains inconsistent, and evidence regarding warning effectiveness is still emerging. As Dr. Ziva Cooper, director of the UCLA Cannabis Research Initiative, recently noted, “Messaging must be evidence-based, nonjudgmental, and responsive to the lived experiences of users if we are to impact health outcomes.”
Conclusion
This study offers valuable, data-driven insights into the complex landscape of adult responses to cannabis health warnings in legalized US states. The predominance of maladaptive reactions signals a critical need for reimagined public health messaging, leveraging behavioral science and user-centered design. Future research should explore longitudinal impacts of tailored warnings, demographic moderators, and integration with broader harm reduction frameworks to maximize public health benefit.
References
1. Volkow ND, Han B, Compton WM, McCance-Katz EF. Self-reported medical and nonmedical cannabis use among pregnant women in the United States. JAMA. 2019;322(2):167-169.
2. Hasin DS, Saha TD, Kerridge BT, et al. Prevalence of marijuana use disorders in the United States between 2001-2002 and 2012-2013. JAMA Psychiatry. 2015;72(12):1235-1242.
3. Massey ZB, Tong C, Zhang T, Wexell KH, Li Y, Zhao J. Assessing Responses to Cannabis Health Warnings among Adults in US Recreational States. Am J Prev Med. 2025 Jul 25:107998. doi: 10.1016/j.amepre.2025.107998
4. Noar SM, Francis DB, Bridges C, et al. The impact of strengthening cigarette pack warnings: Systematic review of longitudinal observational studies. Soc Sci Med. 2016;164:118-129.
5. Hall W, Stjepanović D, Caulkins J, et al. Public health implications of legalising the production and sale of cannabis for medicinal and recreational use. Lancet. 2019;394(10208):1580-1590.
6. Barry RA, Glantz SA. A public health framework for legalized retail marijuana based on the US experience: Avoiding a new tobacco industry. PLoS Med. 2016;13(9):e1002131.
7. Centers for Disease Control and Prevention. Marijuana and public health: Health effects. https://www.cdc.gov/marijuana/health-effects/index.html. Accessed June 2024.
