Harnessing mental health and neuroscience evidence for tobacco and cannabis health messaging

: (Alternative Format Thesis)

Abstract

Tobacco use is a global epidemic and significant public health concern. Although prevalence of tobacco smoking has declined globally, 1.25 billion people still smoke and are exposed to the harms of tobacco. Tobacco smoking is also disproportionately higher in disadvantaged groups, such as people with mental health problems. Targeted interventions for tobacco smoking and mental health are needed to reduce health inequalities from tobacco.

Cannabis is often co-used with tobacco and is used by around 4.4% of the global adult population. Cannabis use is highest in adolescents, who could be more vulnerable to the harms of cannabis due to high levels of use and being in a critical period of development. Potency of cannabis, level of delta9-tetrahydrocannabinol (THC), has increased in recent years, and cannabis products with higher potency carry an increased risk for mental health problems and addiction, compared to lower potency products. The standard THC unit has been proposed as a novel method to estimate potency and provide a more accurate measurement of cannabis use. Standard THC units have been recommended as a strategy to communicate potential harms of cannabis, like what is implemented for alcohol.

Both tobacco and cannabis public health policies use health messaging to communicate the harms of these substances and reduce risk, however, there are important gaps in these health messaging approaches which this thesis aims to address. An extensive body of evidence suggests that smoking tobacco increases risks of mental health problems, and that quitting smoking can improve mental health. Tobacco warning labels depicting this evidence for smoking and mental health is a major gap in public health policy for tobacco control, and research is needed to develop and test effective messaging communicating links between smoking and mental health.

Additionally, considering the neurobiological effects of tobacco and cannabis, and concerns of mental health messaging for stigma, brain-based messaging could be another strategy to communicate the harms of these substances. However, there are important gaps in the neuroimaging literature which first need to be addressed, such as a comprehensive synthesis of studies investigating tobacco, cannabis and co-use and brain structure across different study designs. And for cannabis specifically, longitudinal evidence investigating associations between cannabis use, and brain structure, using standard THC units as a novel measurement of cannabis use, and directly comparing adolescents and adults to investigate adolescent vulnerability.

Therefore, this thesis had two aims: the first, to develop and evaluate novel health messages communicating the links between smoking and mental health; and the second to examine associations between tobacco use and cannabis use and brain structure, addressing key gaps in the literature, which could be used to inform the development of health messaging focused on the brain.

To address the first aim, in chapter 1, in consultation with a patient and public (PPI) group I developed six novel health messages for tobacco packaging, communicating evidence of the negative impact of smoking on mental health. I then tested perceptions of these warnings where 687 participants rated the labels on measures of perceived effectiveness, believability, arousal, valence, acceptability, reactance, novelty of information. I found that the negatively framed mental health labels were perceived as low to moderately effective, and less effective compared to current physical health warning labels, there was also a concern for the loss-framed mental health warning labels to stigmatise people with mental health conditions. To build on these findings, in chapter 2, I developed four gain-framed message labels communicating the mental health benefits of stopping smoking in consultation with PPI group. In a sample of 631 people who smoked tobacco weekly, I tested whether these messages were more effective at increasing motivation to stop smoking, compared to blank labels and labels of the physical health benefits of quitting smoking. I found that positive messaging for tobacco packaging could motivate people to stop smoking, similarly, to gain framed physical health labels.

To address the second aim of my thesis, chapter 3 examined associations between tobacco use, cannabis use and co-use and brain structure in the largest, most comprehensive systematic review and meta-analysis. This chapter reviewed 103 studies: 57 for cannabis use, 45 for tobacco use, and one for tobacco and cannabis co-use, and through random effects meta-analyses provided evidence that cannabis use was associated with smaller volume in the amygdala and tobacco use was associated with smaller amygdala, insula, pallidum and total grey matter volume. I also highlighted significant gaps in the literature, including a lack of studies using longitudinal and causal inference designs, as well as a lack of research on cannabis and tobacco co-use.

Finally, in chapter 4, I address a current gap in the literature of longitudinal evidence directly comparing the associations between cannabis use and brain structure in adults versus adolescents, using standard THC units as a novel measurement of cannabis use, and inform the second aim of my thesis. I used data from the CannTeen study, and using linear mixed effect models to examine associations between cannabis use and brain structure change over one year, measured as cortical thickness, surface area and subcortical volume, in adolescents aged 16-17 (n = 34) and adults aged826-29 (n = 35), who used cannabis and matched controls (adolescents = 35, adults = 36). I found little evidence for associations between cannabis use status or average weekly THC consumption and structural brain changes over one year, and this did not differ between adults and adolescents. There were no significant 3-way interactions for Cme by age group by cannabis use group for any outcome measure, in either model, where cannabis was modelled as status or as mean weekly THC units. There was some evidence for time by cannabis use interaction for thickness of the paracentral lobule (β = 0.28 [0.13 - 0.44]), and postcentral gyrus (β = 0.18 [0.07 - 0.29]). There were no other significant time by cannabis use status, or time by mean weekly THC units interactions. There was evidence for main effects of cannabis use status and average weekly THC consumption: cannabis use was associated with lower volume of the nucleus accumbens and average weekly THC consumption with greater volume of the caudate, and week evidence for smaller amygdala volume. There was evidence for age related changes, with reduced thickness in adolescents compared to adults over time, across multiple regions, consistent with the developmental literature.

Overall, this thesis has generated important new evidence into the use of novel loss-framed and gain-framed mental health messaging for tobacco. Additionally, this thesis advances knowledge, giving a more comprehensive understanding of associations between tobacco use, cannabis use and brain structure, addressing important gaps in the literature. Taken together, these findings can be used to inform modern public health messaging for cannabis and tobacco, based on mental health and brain outcomes, which can address key public health aims to inform people of the risks of using these substances and reduce harm.

Date of Award	20 May 2026
Original language	English
Awarding Institution	University of Bath
Supervisor	Tom Freeman (Supervisor), Gemma Taylor (Supervisor), Esther Walton (Supervisor) & Tom Lancaster (Supervisor)