Marijuana use and high-risk health behaviors among diverse college students post- legalization of recreational marijuana use

Objectives This study examined high-risk health behaviors in marijuana-users among a diverse college population in Southern California, post legalization of marijuana for recreational use. Study design A cross-sectional research design was employed utilizing existing data via the 2018 National College Health Assessment (NCHA) from a large Minority-Serving Institution (MSI) population [n = 1345 (Hispanic/Latino/a, n = 456; White, n = 353; Asian Pacific Islander (API), n = 288; Multiracial/Biracial, n = 195; Other, n = 53)]. Methods Chi square and t-tests assessed differences in descriptive characteristics (age, gender, race/ethnicity and GPA) and high-risk behaviors (alcohol, tobacco and sexual behaviors) among marijuana users and non-users. Logistic regression analyses examined the relationship between race/ethnicity and high-risk behaviors with marijuana use (dependent variable). Results Among marijuana-users, significant (p = 0.004) differences were observed between race/ethnicity with Whites reporting using most (32.7%), followed by Hispanics (27.6%) and then APIs (17.8%). Marijuana-users compared with non-users consistently reported high-risk alcohol behaviors (p < 0.0001), were more likely to smoke tobacco (p < 0.0001) and engaged in more high-risk sexual behaviors (p < 0.0001). Logistic regression showed after adjusting for demographic characteristics and high-risk behaviors, race/ethnicity was borderline significantly associated with marijuana use, specifically for Whites (OR = 1.53; 95% CI: (−0.01, 0.86), p = 0.06) and the Other race/ethnicity category (OR = 2.32; 95% CI: (0.12, 1.56), p = 0.02) compared with APIs. Conclusion Our findings clearly demonstrate deleterious high-risk behaviors such as alcohol use, tobacco use, and certain sexual behaviors occur more among marijuana-users compared to non-users, post legalization of marijuana for recreational use. Further, race-ethnic differences were observed. Therefore, continued examination of marijuana use trends and high-risk behaviors is critical in monitoring the implications of marijuana policy changes, specifically in diverse populations.


Introduction
As of April 2021, 36 states and the District of Columbia have passed laws that allow for the medical use of marijuana and 16 states and the District of Columbia have passed laws that legalize recreational use of marijuana; more states appear to be heading toward similar legislation [1]. Even though the debate on marijuana's safety and benefits continues to be fiercely debated [2], public opinion on the legalization of marijuana has grown more favorable. In 1969 only 12% of the adult population supported legalization compared to 66% in 2019. Among the 18-34-year-old subgroup, support was even higher with 81% favoring legalization in 2019 [3]. With this data, it is not surprising that marijuana use has also gradually increased with college age adults, 18-25 years, experiencing the greatest growth from 17.3% in 2002 to 22.1% in 2018 [4]. Furthermore, in their examination of national survey results on drug use from 1975 to 2016, Schulenberg and colleagues [5] noted historical trends of marijuana use increased at different levels and for different lengths of time across younger ages, 19-20 years through 29-30 years, with almost all age groups reporting increasing prevalence from 2010 onward.
Marijuana use has been shown to have a negative impact on the lives of college students [6][7][8][9]. Earlier studies have shown numerous negative associations with marijuana use and academic success indicators such as lower GPA, reduced studying time, late assignments [6,8], missing class and discontinuous enrollment in college [7,9]. Furthermore, the heavier the marijuana use, the more likely negative outcomes were reported [10]. Importantly, the majority of the studies investigating negative impacts and marijuana use in college students are prior to legalization of recreational marijuana [6][7][8][9], and even less of this data are available on a diverse student population.
Adverse high-risk behaviors, including tobacco use, binge drinking and use of other illicit drugs have also been associated with marijuana use in college students [11,12]. Using marijuana and tobacco at the same time may lead to increased exposure to harmful chemicals, causing greater risks to the lungs, and the cardiovascular system [13]. According to the National Survey on Drug Use and Health, alcohol and marijuana were substances most frequently used by college students (18-22 years old) [4,14]. In addition, among users of both substances, alcohol and marijuana have been shown to be more concurrently used rather than alone [15].
Another area of concern is sexual risk behaviors. Marijuana use, among adolescents and young adults, has been associated with sexual risk behaviors and outcomes, including inconsistent condom use, multiple sexual partners, and STI diagnoses [16][17][18]. Bryan and colleagues [17] showed that marijuana use increased the likelihood of intercourse due to reduced inhibitions and effects on cognitive ability. In addition, they noted a decreasing ability to negotiate and carry out condom use. Furthermore, Metrik and colleagues [19] found that both alcohol and marijuana use were independently correlated with greater odds of having casual sexual intercourse.
Data from the National Survey on Drug Use and Health also showed that among 18-24-year-old respondents, Blacks and Hispanics experienced more marijuana use disorder than other ethnic groups [13]. Keyes and her colleagues [20] found similar results when examining race/ethnicity in Monitoring the Future data from 2006 to 2015, pre-legalization. However, they note the importance of a deeper look into diverse populations as they also found confounding factors of class size for Black high school seniors and urban setting for Hispanic seniors [20].
With legalization legislation in the United States beginning only a decade ago and in California as recently as 2016, there is limited data available on high-risk health behaviors and marijuana use postlegalization in college-aged students. Even more limited are those that focus on ethnically diverse college groups [21][22][23]. Therefore, this study examined health behaviors in marijuana-users post legalization in a diverse, urban college population in Southern California. Specifically, we investigated alcohol and tobacco use, as well as sexual behaviors among marijuana-users compared with non-users post legalization in a racially/ethnically diverse college group. We hypothesize that marijuana-users will partake in increased high-risk behaviors compared with non-users, and that use will vary by race/ethnicity.

Methods
The National College Health Assessment (NCHA) is a nationally recognized research survey designed to collect data on students' health habits, behaviors, and perceptions. NCHA data collected from a large Minority-Serving Institution (MSI) in Southern California (approximately 40,000 students), during the spring semester of 2018 was utilized for the present study. The assessment consists of 60 questions and approximately 300 items addressing a range of student health topics such as alcohol, tobacco and other drug use; sexual health; nutrition and exercise; mental health; and personal safety and violence. For the present study, a computer-generated random sample of 10,000 students were invited, via email, to complete the online survey from March 5, 2018 to March 23, 2018 with periodic reminders being sent to those yet to complete it. The response rate was 14.28% equating to n = 1428 participants completing the survey. Of the 1428 respondents, 1345 had complete data on demographics (Hispanic/Latino/a, n = 456; White, n = 353; Asian/Pacific Islander n = 288; Multiracial/Biracial, n = 195; Black, n = 32; Native American, n = 21), marijuana use, and other health-risk behaviors, and were included in the present study. Approval was secured by the campus' Institutional Review Board, IRB #17-18-80.

Student demographics
Participants provided demographic information such as age, gender, and race/ethnicity. In addition, academic performance, via grade point average (GPA) was also ascertained. Race/ethnicity categories were as follows: White, Black, Hispanic or Latino/a (abbr. Hispanic for this study), Asian/Pacific Islander (API), American Indian, Alaskan Native, Multiracial/Biracial (abbr. Multiracial for this study), and Other.

Marijuana and tobacco use
Questions related to tobacco and marijuana use were asked as follows: "Within the last 30 days, on how many days did you use the following …" Response options included: Never used; Have used, but not in last 30 days; 1-2 days; 3-5 days; 6-9 days; 10-19 days; 20-29 days; and used daily.

Alcohol consumption
Three questions assessed alcohol consumption. Prior to answering questions related to alcohol, the instrument provided a clear definition/ standardization pertaining to the term "drink" as follows: One drink of alcohol is defined as a 12 oz. can or bottle of beer or wine cooler, a 4 oz. glass of wine, or a shot of liquor straight or in a mixed drink. To assess alcohol consumption, participants were queried as follows, the last time you "partied"/socialized, how many drinks of alcohol did you have? A blank space was provided for subjects to type in a number; directions advised subjects to enter 0 (zero) if they did not drink alcohol. Additionally, to assess high-risk behaviors/binge drinking related to alcohol consumption, an item was included with two parts and queried participants to refer to the last 30 days when answering the following: Did you drive after drinking any alcohol? and Did you drive after drinking five or more drinks of alcohol? The options provided for both were: N/A don't drive; N/A don't drink; No; or Yes.

Sexual behavior and contraception
Questions related to sexual behavior were as follows: Within the last 12 months, with how many partners have you had oral sex, vaginal intercourse, or anal intercourse? A blank space was provided for subjects to type in a number; directions advised subjects to enter 0 (zero) if they did not have a sex partner within the last 12 months. Other questions were as follows: Within the last 30 days, how often did you or your partner(s) use a condom or other protective barrier (e.g., male condom, female condom, dam, glove) during vaginal sex? The response options included two nonapplicable selections reflective of one who had never did this sexual activity or one who had not done this sexual activity during the last 30 days. The other response options listed were: Never; Rarely; Sometimes; Most of the time; and Always.

Statistical analyses
The data were analyzed using both descriptive and inferential statistical methods. Descriptive statistics were calculated for the demographic data including age, gender, ethnicity, and GPA. Chi-square tests were conducted to study differences in gender, ethnicity and GPA between marijuana-users and non-users, and t-test assessed difference in age.
The response variable, marijuana use was dichotomized as follows: users were categorized as those indicating any use in the past 30 days and non-users or never used as those indicating no use within the past 30 days. For the race/ethnicity category, the variable was categorized as: White, Hispanic/Latino/a, Asian/Pacific Islander, Biracial/Multiracial and Other. Black, and American Indian/Alaskan Native were included in the "Other" category due to small sample sizes. All analyses were performed using SPSS version 25.
We initially conducted univariate analyses, t-test and chi-square test, respectively, to assess differences in various high-risk behaviors between marijuana-users and non-users. First, mean number of drinks consumed while socializing was calculated and compared between the two groups. Participants were also asked about alcohol consumption (binge drinking) during the past two weeks, and whether they consumed five or more drinks at a sitting. Responses were on an ordinal 8-point Likert scale with 1 = N/A don't drink through to 8 = 6 times (engaged in binge drinking). Mean differences in binge drinking were calculated on the ordinal scale (as a continuous variable) between marijuana-users and non-users. Also, we conducted chi-square analyses to assess differences in drinking after driving both for drinking any alcohol at all, as well as binge drinking (five or more drinks) between marijuana-users and nonusers. Second, differences in tobacco use between the two groups were assessed. Tobacco use was categorized as no or yes, with no indicating never use or no use in the last 30 days and yes indicating any use of tobacco within the past 30 days. Last, we measured differences in sexual behaviors, including number of sexual partners (continuous) and condom use (for vaginal sex) in the last 30 days via t-test and chi-square, respectively, between marijuana-users and non-users.
Logistic regression was used to assess the association between race/ ethnicity and high-risk behaviors with marijuana use compared to no use. The analysis population included only respondents with nonmissing data on marijuana use, and race/ethnicity and high-risk behavior data. The dependent variable, marijuana use was dichotomized as use/no use, while independent variables included the following: Age (years, continuous), gender (male vs. females), race/ ethnicity [White, Hispanic, Multiracial, Other, API (ref)], GPA (continuous), # of drinks at last party (continuous), # of sexual partners (continuous) and tobacco use (yes vs. no).

Results
Descriptive statistics on demographic data and GPA on the entire sample are presented in Table 1. Sample mean age was 23.10 years (±5.66 years) with majority being female (74.5%). Race/ethnicity was predominantly equally distributed between White (26.2%), Hispanic/ Latino/a (33.9%) and API (21.4%) followed by Multiracial/Biracial (14.5%) and Other (4.0%). Table 2 compares demographic and academic performance data between and within marijuana-users and non-users. Of the 1345 survey respondents, 275 (20.4%) were marijuana-users. Demographic data showed that marijuana-users were approximately one year younger (22.27 years) than non-users (23.19 years), indicating that younger college students were significantly more likely to use marijuana (p = 0.016). The proportion of users were more likely to be male, although not statistically significant. Also, significant (p = 0.004) differences were observed for marijuana-users compared with non-users for race/ ethnicity with Whites reporting using most (32.7%), followed by Hispanics/Latino/a (27.6%) and then API (17.8%). Regarding grade point average (GPA), marijuana-users had a significantly (p = 0.009) lower GPA than non-users (p = 0.009), with a higher proportion of non-users reporting A's (31.1%) compared with non-users (20.7%).
Sexual risk behaviors also yielded statistically significant differences among marijuana-users and non-users (p < 0.0001 for all behaviors). Marijuana-users reported 2.6 times as many sexual partners on average compared to non-users. Moreover, the reported use of a condom during vaginal intercourse was significantly (p < 0.0001) less likely among marijuana-users with 25.1% indicating never use, in relation to only 14.2% of non-users.

Discussion
Our findings post legalization of marijuana for recreational use clearly demonstrated a higher prevalence of high-risk behaviors such as alcohol use, tobacco use, drinking and driving and sexual activities among college students who use marijuana compared to those who do not. Specifically, adjusting for all covariates in the logistic regression model, younger respondents were found to be more likely to be marijuana users; similarly, students earning D grades were 1.68 times more likely to use marijuana compared to those earning A grades (although not significant). Compared to APIs, Whites were 53% more likely and the Other race/ethnicity category were twice as likely to be marijuana users. These findings are consistent with similar studies examining risktaking behaviors among marijuana-users for recreational use pre legalization [9][10][11][15][16][17][18][19]29,[33][34][35][36][37][38][39][40][41][42][43][44].
This study is one of the first to assess marijuana use in an ethnically diverse student population. We found marijuana use was higher among White college students which mirrors most recent national data for 18-25-year-olds [4]. Interestingly, when examining marijuana-use disorders Pacek and colleagues [24] found African Americans twice as likely to be diagnosed with this condition versus both Whites and Hispanics. Due to small sample size of certain ethnic groups we did not specifically examine marijuana use among Blacks and Native American students. However, we found a strong relationship between the Other race/ethnicity category, (which predominantly comprised Blacks and Native Americans), and marijuana use, suggesting higher use of marijuana in this respective population. Nonetheless, further studies are needed with larger samples of African Americans and Native Americans.
While we found marijuana use slightly higher among males than females, it was not a statistically significant difference. National data shows men age 18 to 25 are more likely to use marijuana at least once a month compared to women [13,25,26]. However, when analyzing a 15-year trend, Johnson and colleagues [27] found that the male-female differences in marijuana use decreased over time; with the most recent SAMSA data indicating among 18-25 year-olds, males are only slightly higher in reported marijuana use than females [4].
In addition, we found an inverse relationship between academic performance and marijuana use; specifically, adjusting for all covariates in the logistic regression model, younger respondents were found to be more likely to use marijuana; similarly, students earning D grades were 1.68 times more likely to use marijuana compared to those earning A grades. This is similar to the findings from Arria and colleagues [8]. Their research found that not only did marijuana-users have lower grade point averages, but also skipped classes more and took longer to graduate college. Whether these negative outcomes are related to poor academic behaviors or poor cognitive function or both continues to be researched. In a review of the literature, Crane and colleagues [28] found that numerous studies continue to demonstrate the negative effect of cannabis on learning and memory, in addition to deficits in attention, concentration, and abstract reasoning.
With regard to alcohol use, our findings indicate that increased alcohol use is also associated with an increase in marijuana use. Recent studies exploring the role of marijuana policies on alcohol and marijuana use, show varying results. Some hypothesize that alcohol use may decrease as marijuana becomes legalized for recreation use as it would substitute for alcohol [29,30]. Others posit that with more liberal marijuana policies, both marijuana and alcohol use will increase as they may complement each other [31,32]. A critical review of the literature provides some evidence of both. The researchers of these studies note the issue is complex, suggesting likely factors such as how long the policy has been in place, how it is implemented as well as the age of users, play a role [33].
Tobacco was another substance we found marijuana smokers more likely to use. This was a significant finding as tobacco users were four times more likely to be marijuana users compared with non-tobacco users. Interestingly, some researchers have found that using nicotine Table 3 High-risk behaviors among marijuana users and non-users.   with cannabis together can intensify the effects of cannabis [34][35][36]. In their study Ream and colleagues [37] found participants reported smoking cigarettes directly after using marijuana to enhance the intoxication. This may be due to the close overlapping in the distribution of brain receptors for nicotine and cannabis [38]. The relationship of smoking and marijuana use is supported by findings that show marijuana-users who also smoke cigarettes have increased risk of marijuana relapse when they are trying to quit [39].
As is the case with many substances, marijuana impairs judgement; therefore, we found a positive association between high-risk sexual behavior (# of sexual partners) and marijuana use [40,41]. Our univariate analyses also showed that marijuana-users in this study indicated they never use a condom during vaginal intercourse 25% of the time compared to only 14% of the time for non-users. These findings are similar to data showing marijuana use associated with non-use of condoms and having a higher number of sexual partners [16,[42][43][44]. Some researchers assert this association may be due to marijuana use potentially increasing sexual desire or sensation [45,46] while others believe decreased cognition [47,48] and increased disinhibition [49] are contributing factors.

Strengths and limitations
Strengths and limitations of this study should be recognized. Although we assessed several behavioral factors among those who use marijuana and those who do not, we were limited by the number and types of items that addressed these behaviors. The National College Health Assessment is a comprehensive instrument designed to asses general student health status and engagement in a variety of health behaviors. In addition, because our population of college students was from one university, the findings are not generalizable to other age groups or college students outside of Southern California. With that said, it is worth noting the sample was large and was representative of the surrounding diverse metropolitan area. One other limitation includes the cross-sectional design which does not allow for causal inference regarding marijuana use and other health behaviors. Despite these limitations, this study is one of the first to provide insight into marijuana use post legalization for recreational purposes.

Implications and conclusions
Overall, this study provides data indicating high-risk health behaviors are more prevalent among college student marijuana-users than non-users post legalization. While propositions for the legalization of marijuana for recreational use expand across the United States, there is still much debate on the health and social impacts of marijuana use. Continued examination of marijuana use trends as well as health behavior trends is critical in monitoring the implications of policy changes. In addition, it will be important to be vigilant in the surveillance of vulnerable populations as we are seeing an increase in marijuana use among females and ethnic/racial groups. More specifically, college health educators and student affairs personnel should be aware of the results from studies like this and work to address marijuana use in college students and the other high-risk behaviors that likely accompany it.

Funding
This study was part of the Big Data Discovery and Diversity through Research Education Advancement and Partnerships (BD3-REAP) Project funded by National Institutes of Health (NIH), NIMHHD-R25; # 1R25MD010397-01.
All authors confirm there are no relevant financial or non-financial competing interests to report.
Ethical approval was granted by the appropriate Institutional Review Board, IRB #17-18-80.

Declaration of competing interest
None.