The George Washington University Committee on Human Research, Institutional Review Board determined that this study did not meet the definition of human subjects research. Table 3 presents the predicted probabilities of past-month marijuana use at each level of risk and by reasons for use. We also present marginal effects, which represent the change in the probability of each outcome as the risk factors change, while holding all other characteristics constant. Adjusted relative risk ratios from the multinomial logistic regression model are presented in supplemental Table S3 in the online version of this article. Overall, the average predicted probabilities of marijuana use for medical, recreational, and both reasons were 28.6%, 38.2%, and 33.1%, respectively. Several covariates were associated with marijuana users’ reported reasons for use including, but not limited to, gender, age, race, education status and employment. Associations between both health status and frequency of use and reasons for use were large in magnitude and statistically significant. For example, respondents who reported 14+ versus no days of poor mental health had significantly increased chances of reporting marijuana use for medical reasons or both reasons and a significantly decreased probability of reporting recreational reasons . This same pattern—that is, greater probabilities of reporting medical reasons or both reasons among those in poor health—was also evident for physical health. Marijuana users who reported daily use had a 6.3% increased probability of reporting medical reasons and a 15.6% increased probability of reporting both reasons. Daily users had a significantly reduced likelihood of reporting recreational reasons for use . Estimates from the logistic regression model showed similar patterns. See Supplemental Table S4 in the online version of this article.Among respondents in fully legal states, the chance of reporting recreational reasons was 5.5% lower than in illegal states ,grow table hydroponic but the chance of reporting both reasons was 7.0% higher . While the predicted probability of reporting medical reasons was lower with states’ liberalization of marijuana, differences were not statistically significant. Our user profiles confirmed these patterns .
We deliberately modeled illustrative profiles possessing characteristics associated with medical and recreational reasons for use based on our results: an older woman in poor mental and physical health who is a daily marijuana user and a young adult male who is an infrequent marijuana user and a binge drinker. In the first case, we found that the average predicted probabilities of reporting medical, recreational, and both reasons did not significantly vary by states’ legal environments. For the second user profile, we found that the predicted probability of reporting recreational reasons was 5.3% lower in a fully legal state than in a fully illegal state. Conversely, the predicted probability of reporting both reasons was 5.0% higher in a fully legal than in an illegal state. The difference in the probabilities of reporting medical reasons in legal versus illegal states was not statistically significant.Our study makes two, unique contributions. First, we estimated the change in probability associated with incremental changes in risk factors on each reason for use and created user profiles to illustrate these relationships. Second, we used three years of BRFSS data to examine marijuana users’ reasons for use, data which we do not believe has been previously used for this purpose. We found that the prevalence of past-month marijuana use in our sample was 11.2%, which was similar to, but not the same as, rates reported from other U.S. sample surveys. For example, two studies based on BRFSS data from 2016 and 2016-2017, respectively, found prevalence rates of 9.1% and 13.6% . The National Survey on Drug Use and Health reported rates of past-month marijuana use among persons aged 12 or older from 11.2% in 2017 to 13.0% in 2019 , 2018, 2019, 2020. Findings from a study that used 2005-2018 National Health and Nutrition Examination Survey data reported a 14.4% past-month marijuana prevalence rate . Differences in these estimates could be attributable to time trends, study inclusion criteria, and/or differences in each survey’s design and data collection procedures.
Like other studies , we found that adults were most likely to report recreational reasons for use followed by both reasons and medical reasons . The characteristics we identified as being associated with marijuana users’ reasons for use—gender, age, race, education, health status, and frequency of use—also comport with prior research . Consistent with prior studies , we found that medical marijuana users were less likely to report smoking marijuana and more likely to eat/drink, vape, and use other routes of administration. This pattern is consistent with evidence, which has found that vaporization is a commonly used mode of delivery among medical marijuana users because of its relative health advantages over smoking and the flexibility, portability, efficiency, and ease of use that accompanies vaporization devices . It is possible that eating/drinking offers similar benefits. Additional studies are needed to explore this phenomenon. We found that being a woman increased the odds of reporting medical reasons for past month marijuana use . While prior studies have found that women were more likely to report using marijuana for medical reasons compared to recreational ones , in only one of these studies was the difference statistically significant . Our findings complement a recent study , which explored gender differences in medical marijuana use and found that women were more likely than men to use medical marijuana for a variety of symptoms including pain, anxiety, and nausea—conditions which commonly qualify patients for medical use . Finally, we found that past-month marijuana users who reported medical reasons for use were more likely to be older and in poorer health, and they were more likely to be daily users. Prior studies have produced similar findings . Notably, we also found that marijuana users who reported using for both medical and recreational reasons were the most likely to be daily users. Because the frequency and quantity of marijuana consumed have been associated with marijuana dependence and other adverse effects , persons who report using for medical and/or both reasons could bear greater risks. Studies that compare daily and intermittent marijuana users’ reasons for use and adverse outcomes, and which draw on diverse populations, are needed to better understand these relationships and optimize generalizability .
Our findings regarding the association between legalization and marijuana users’ reported reasons for use were unexpected. We found that in fully legal states , the predicted probability of reporting recreational use was significantly lower while the probability of reporting both reasons was significantly higher. While these findings are counter intuitive, restrictive recreational marijuana laws and higher tax rates incentivize medical use . These forces could drive recreational users, especially those who use marijuana to self-medicate , 2020c, towards using and reporting medical reasons or both reasons for use, which might explain our findings. Alternatively, states that have legalized marijuana for recreational use tend to hold liberal positions on other issues and attract residents who share those values. It is entirely possible that persons who hold such liberal values are more likely to view marijuana favorably, recognize its therapeutic benefits,grow table and attribute at least some of their use to a medical need. Our findings should be placed within the context of data and study limitations. BRFSS data are self-reported, which could introduce reporting bias. While we combined the three most recent years of BRFSS data, we effectively limited our sample to 20 states and 18,925 marijuana users. This diminishes the generalizability of our results. We know that marijuana prevalence varies by state , 2020c, at least in part, because of states’ policies regarding marijuana use, possession, and sales . If the subset of states that opted to include the BRFSS marijuana module also had different use rates or reasons for marijuana use, confounding could have been introduced. Further, we excluded 294,751 respondents from our sample . These respondents differed from those retained on many characteristics. For example, respondents in our study sample were more likely to be male, older, report using tobacco and binge drinking and less likely to be married. If this same pattern existed for the variables of greatest interest, our estimates could be biased.Additionally, the BRFSS’ asks respondents about their reasons with three responses: medical, recreational, or both. In reality, marijuana users’ reasons for use may be more nuanced; users might not fit “neatly” into these three categories. For example, some may view marijuana as having general health benefits—helping with relaxation or enhancing wellness. This is supported by studies, which have found users describing marijuana as a “natural” alternative to or substitute for other prescription medications and the marijuana industry’s marketing of it as a lifestyle product . In these cases, it is unclear how a user might answer the BRFSS question. Further, the BRFSS does not ask respondents about several potentially important predictors of marijuana users’ reasons for use, including quantity, duration of use or dose. Our coding of states’ policy environments as a categorical variable could be an imprecise reflection of how states’ policy environments influence consumption patterns . Given that local government entities have enacted additional policies that further regulate marijuana markets, studies that also account for local policies and other factors such as retail availability of marijuana are needed. Finally, our analysis offers insight into the correlates of reasons for marijuana use, but because of the BRFSS’ cross-sectional nature, causal inferences cannot be made.
Recent national data showed that 27.1% of high school students reported past 30-day use of any tobacco1 and 19.8% reported past 30-day use of any marijuana.Furthermore, several studies indicated that past 30-day co-use of two substances was higher than use of either tobacco or marijuana only among adolescents,and higher than use of marijuana only among young adults.The use of these substances during neurodevelopmental stages exposes AYA to numerous adverse health consequences and societal impacts .Understanding tobacco and marijuana co-use among AYA has become more important given the proliferation of new products. The tobacco landscape has shifted from conventional cigarettes to cigars/cigarillos, hookah, smokeless tobacco, and more recently, e-cigarettes.Marijuana is also available in a variety of combustible , vaporized , and edible products .Co-use can refer to use of both substances separately across the aforementioned products during the past 30 days, or at the same time or in the same product, such as with blunts .Recent studies have demonstrated a transformation of tobacco use patterns with the use of non-cigarette products now surpassing conventional cigarette use among adolescents.In contrast, despite the appeal of newer marijuana products, combustible marijuana remains the most common form used across all age groups.Therefore, it is critical to characterize tobacco and marijuana co-use in light of specific products. Past research on tobacco and marijuana co-use among AYA has predominantly focused on examining the overall relationship between tobacco and marijuana using “blanket terms” , or only co-use of combustible forms and blunts.Newer co-use research among AYA has taken into account specific products.However, these studies have examined only the co-use of individual tobacco products with “any marijuana,” which limits our understanding on specific marijuana products co-used with tobacco. Only two recent studies have assessed couse of specific products for both substances. One study among 1,420 high school students compared past 30-day use of cigarettes, cigars, hookah, and e-cigarettes between current blunt and combustible marijuana users.This study, however, focused on only combustible forms of marijuana; thus, it did not fill an important gap in the literature on alternative forms of marijuana .The other study among 2,668 adolescents assessed the relationship between previous use of tobacco products and subsequent use of marijuana products .Although this study provided more insights by including noncombustible marijuana products, it did not directly provide data on past 30-day co-use of tobacco and marijuana products, and more importantly, did not include a sample of young adults, a group with the highest risk of co-use relative to other age groups.National data indicated co-use prevalence among young adults in 2014–2015 was 21.3%,4 and nearly half of adult co-users were between 18–25 years old.To our knowledge, there has been no research among young adults considering specific products for both tobacco and marijuana. Given the current era of marijuana legislative reform nationwide as well as the proliferation of proposed national and state tobacco regulations, more research on co-use of tobacco and marijuana is needed to inform these actions.