We placed flyers in marijuana dispensaries, vape shops, cafes, stores, on bulletin boards at community colleges, and on Craigslist and Facebook. We attempted to interview participants twice, in order to allow conversations to develop more deeply and to use the content of the first interview to inform questions in the second. Out of 32 enrolled participants, 24 completed the second interview. This study was approved by the University of California, San Francisco Committee on Human Research. Participants provided written informed consent. We used pseudonyms for this and all publications. Semi-structured qualitative interviews lasting 60-90 minutes were conducted between January and August 2015 by six trained interviewers. Discussion topics included definitions of smoking, experiences with tobacco, e-cigarettes, marijuana, marijuana vaporizers, and other products. To further generate discussion of comparative harms and benefits of products, participants were asked to arrange labeled pictures of various products from the least harmful to the most harmful and talk through their sort process.We audio recorded and professionally transcribed the interviews and coded transcripts using Dedoose software. LP and SS independently blind-coded and compared a sub-set of transcripts to develop the study’s coding guidelines. We created code definitions, developed a consistent coding scheme and discussed the coding results to ensure codes were applied consistently. SS coded the larger set of remaining transcripts. Given the emerging nature of the legalized marijuana market and the lack of existing research in this context, ebb and flow trays we adopted a thematic analysis approach that would allow us to discover emerging behaviors and meaningful categories for our participants and to generate themes iteratively during review of coded transcripts.

All authors reviewed memos with illustrative quotes summarizing each theme and discussed themes iteratively to reach consensus and theme saturation. Participants primarily evaluated products along five dimensions: whether or not the product was combusted, potency of the psychoactive agent delivered, presence of unnatural chemicals, addictiveness, and source of information about the product. Participants generally assessed harms in the context of perceived benefits, and frequently discussed alternative products delivering similar benefits, but with less harm. They gauged harms in nuanced ways, with criteria for judging harm differing between tobacco and marijuana products and comparing them with alcohol, illicit drugs, and pharmaceuticals. Participants commented on the addictive nature of tobacco compared to alcohol, opiates, and prescription drugs. In contrast, participants spoke about the habitual urges to use marijuana but rarely reported physiological withdrawal symptoms. Several participants also explicitly mentioned physiological changes in the body associated with addiction to marijuana. ‘Ben’, age 22, commented, “I feel that for a lot of people, especially in Colorado, it’s very much emotionally dependent, mentally dependent, on marijuana…Over a chronic period of [using] it, your body adjusts and lowers the blood flow to your cerebellum so that when you smoke, you have normal flow.” Some, like ‘Brad’, age 25, acknowledged that they “need [marijuana] to function.” Some participants discussed addiction to marijuana as affinity for euphoria or the positive feelings associated with getting high. Thus, it was the psychological benefits of marijuana that were viewed as addictive. When asked what the addictive component in marijuana is, ‘Owen’, age 20, said, “Euphoric feeling;” an observation confirmed among several other participants. Participants did not generally identify tolerance as a sign of addiction. Instead, some marijuana users discussed high potency products as a useful way to overcome tolerance and reclaim positive effects. Participants also discussed tolerance increasing with use; some temporarily stopped use for a few days or weeks to bring their tolerance level down in order to be able to better feel the effects of marijuana when using smaller doses or less concentrated products.

The ability to abstain from use without physical withdrawal validated for some participants that marijuana was not addictive. As ‘Erin’, age 24, commented, “I smoke [marijuana] because I like it, but, I don’t have to wake up in the morning and smoke a bowl like I have to wake up in the morning and smoke a cigarette. I’ve gone periods of time without smoking marijuana, and you can quit like that. I don’t feel like it has the addictive properties that cigarettes have.” Because of its perceived naturalness, marijuana was also believed to be superior for therapeutic functions to pharmaceutical “pills” , which were perceived as more addictive. Informational sources mentioned by participants can be broadly classified as external and internal . Participants described information from governmental agencies as exclusively concerning harms of products, and mostly related to tobacco use. Participants reported seeing warning labels on tobacco products, which was particularly salient to ‘Rashawn’, age 24, who had lost his grandmother to tobacco-related illness. As he explained, “my friends wanted to do [blunt wraps]…but I looked at them and it has this warning. … and I thought that was just for cigarettes. I wasn’t really prepared to have a warning of cancer on the blunt wrap. I was like, so you guys see this? … I’m not okay with it. I could possibly have cancer because my grandmother, she’s my favorite person in the world, she died of lung cancer.” While several participants explicitly referenced media campaigns when discussing tobacco related harms , it was less common for participants to cite health authorities as sources of knowledge on the harms of e-cigarettes or marijuana. Some participants noted the contradiction between government information on harms and their own experiences. ‘Sadie’, 24, shared, “My whole life I was taught that weed was really bad, and then when I smoked it, I was like, wow, this really makes me feel good. And I was depressed at the moment. And that was … the first time … I was okay.” Participants’ own bodies were a primary source of knowledge when discerning whether something was harmful or beneficial. ‘Timothy’, age 25, explained, “generally when things make you feel bad, they’re bad for you.” Similarly referencing bodily sensations, ‘Angela’, age 18, reflected, “Dabbing is really harsh on your lungs. It feels like someone is stabbing you in the lungs.” Participants often continued to use a product if their bodies did not communicate acute harms . ‘Timothy’, age 25, commented, “if I felt it affecting my lungs, then I would consider something else.” Few participants cited medical providers as a source of information about marijuana.

Employees of marijuana dispensaries and vape shops were an important source of guidance about benefits of particular products and strains. If a product yielded an unpleasant experience, retailers would recommend a different product or a lower dose rather than abstention. One of our participants, ‘Sadie’, age 24, worked at a marijuana dispensary and reported, “we have to find different strains out and what they do, how it affects people, and of course it affects people in a different way… if it’s an upper, a downer, if it’s a mellow one…” In sum, among our participants, retailers served as a source of information on product selection and customization, but not on potential harms. Our participants judged the harms and benefits of various tobacco and marijuana products along the dimensions of combustion, potency, presence of chemicals, potential addictiveness,4×8 flood tray and sources of knowledge. The narrative that non-combustible products are safer appears in the scientific literature, tobacco industry statements, and documents from regulatory agencies. While some participants recognized that smoking anything – tobacco or marijuana – was dangerous, more often combusted marijuana was perceived as safer because it was seen as having fewer chemicals and lower potency. Our findings complement the small literature on the comparative perceptions of harm of various tobacco and marijuana products, where perceptions of risk were lowest for marijuana flower. Our study enriches previous research by examining the reasons behind differential perceptions. Contrary to a previous hypothesis that marijuana flower was perceived as safer than marijuana concentrates because it was less addictive, few participants in our study considered the dependence risks of marijuana concentrates, but viewed them as a way to overcome increased tolerance. Perceived harm of marijuana concentrates was due in part to immediate adverse psychological effects, such as panic attacks. Chemicals were commonly perceived as harmful, particularly when used to manufacture tobacco or process marijuana into concentrates. Few participants associated harm with chemicals that naturally occur in tobacco or marijuana plants, thus perceiving “natural” tobacco products, such as Natural American Spirit cigarettes as less harmful, which may be unintentionally reinforced by national media campaigns educating youth about chemicals added to tobacco.

Participants utilized distinct sources of knowledge to evaluate the harms and benefits of tobacco and marijuana products. Anti-tobacco campaigns informed tobacco-related harm perceptions, whereas few public education campaigns informed perceptions of marijuana and e-cigarettes. Unpleasant personal experiences with marijuana intoxication were mentioned as potential harms more often than marijuana-related diseases or addiction, which might reflect a lack of information from government agencies on the health risks of marijuana and e-cigarettes. Even though the Colorado public awareness campaign ‘Good To Know’ publicized risks of highly potent marijuana edibles, our participants were experienced marijuana users, which may explain why no participants discussed how retailers had warned them to “go slow” with initial marijuana use. The findings from our study with a small convenience sample limited in age and geographically to young adults in Denver, CO, may not be generalizable. However, they point to issues worthy of future exploration and, together with other tobacco research, have implications for health education and product labeling. Educational campaigns about tobacco and marijuana products may be more relevant to young adults if they include diverse messages reflecting multiple dimensions of harms and benefits of tobacco and marijuana. Some aspects of tobacco harm might be relevant to marijuana, such as the dangers of marijuana combustion, as marijuana smoke impairs vascular function in ways similar to tobacco smoke. The harms of secondhand tobacco smoke exposure has been an effective theme in tobacco prevention and control media campaigns. However, educational messages that focus solely on dangers of combustion may be missing important nuances in how young people understand tobacco and marijuana related harms. Young adults in our study reported that they would rather smoke marijuana than use edibles, primarily because edibles were seen as more potent and more difficult to titrate dosage. While additional research is needed to ensure the generalizability of these findings, this study suggests that the availability of lower dose marijuana edibles may mitigate the tendency for users to choose combustible over edible products due to potency concerns. Our study also identified little awareness that there are harmful, naturally-occurring chemicals in tobacco in addition to chemicals added by manufacturers. This may be an opportunity for tobacco education that is also relevant to marijuana and e-cigarettes. These messages should be carefully tested to ensure they do not unintentionally encourage the belief that tobacco chemicals are harmless because they are natural. Studies are needed to determine how reduced risk perceptions of organic plant matter, whether tobacco or marijuana, impact use and prevalence among young people, and what education messages would address these misperceptions. Our findings suggest that messages using the words “addiction” or “dependence” may not resonate with young adults, particularly with respect to marijuana. Messages about tolerance or dose escalation may be the dimensions of addiction most relevant to marijuana users. Illustrating physical changes in the body or the brain as a result of long-term marijuana use may also aid understanding that marijuana dependence has physiological in addition to psychological risks. Stories from young users who have difficulty controlling their use would be worth exploring, similar to how the FDA’s Real Cost Campaign focused on “loss of control” as the relevant dimension of nicotine addiction. Future studies should evaluate these messages with larger representative samples. Additional research on the addictive qualities of marijuana and which demographics are particularly susceptible to these qualities is also needed. Medical professionals have important opportunities to educate patients about health risks of tobacco and marijuana use when screening for substance use. Recommendations will need to be updated continuously to address new products and delivery systems and provide guidance for medical professionals in this arena. Our finding that marketing describing products as “natural”, “pure”, “clean”, “additive-free”, or “organic” may increase product appeal or reduce perceived harm is in line with other studies. Taken together, these findings indicate the need for regulatory agencies to prohibit the use of such terms on tobacco and marijuana products .