Substance-Specific Risk Factors Among Young Adults: Potential Prevention Targets Across Cannabis-Permissive Environments

Abstract

Purpose:

This study examined levels of substance-specific risk factors such as perception of harm from substance use among young adults in a range of cannabis-permissive environments. The main objective was to inform future preventive interventions aimed at reducing cannabis use in the context of increasingly permissive environments.

Methods:

Data came from the Community Youth Development Study (CYDS) collected in 2016 when participants were about 23 years old (n = 1,722 participants residing in 46 U.S. states). Young adults self-reported their perceptions about the harms related to cannabis, alcohol, and cigarette use; attitudes about and ease of access to cannabis and other substances; and perceived wrongfulness and social acceptability of cannabis, alcohol, and cigarette use and of selling of cannabis and other illegal drugs.

Results:

Young adults in more permissive cannabis contexts reported higher levels of all cannabis-specific risk factors (e.g., greater access to and more favorable attitudes about cannabis use), except for perception of harm from regular cannabis use. However, permissiveness of the cannabis environment was not associated with heightened levels of risk factors for other substance use (such as alcohol, cigarettes, and opioids).

Conclusions:

Future preventive interventions for young adults living in more permissive cannabis contexts may need to focus on cannabis-specific risk factors in particular and go beyond considerations of harm from regular use. Future studies should replicate these findings with other samples.

Introduction

Cannabis use generally peaks in young adulthood, and the past two decades have seen an increase in use in this age group in the United States. Studies have shown that a typical developmental pattern for both cannabis and other substance use follows an arc that increases throughout adolescence, with a peak in the early to mid-20s and a subsequent decline in the 30s (Palmer et al., 2009; Schulenberg et al., 2019; Schuler et al., 2015). Data from national studies in the United States indicate increases in the prevalence of young adult cannabis use over the past two decades; according to the data from the National Survey of Drug Use and Health, the prevalence of past-year cannabis use among young adults ages 18 - 25 has increased by over 5% between 2002 and 2019, with more than a third (35.4%) of young adults reporting use in 2019 (Substance Abuse and Mental Health Administration, 2019). Prevention and intervention research has suggested that reducing availability and targeting malleable cannabis-specific risk factors such as perceived harm from use and favorable attitudes and norms may help to reduce use (e.g., Catalano et al., 2012). Reducing cannabis use in young adulthood can prevent serious short- and long-term health, economic, and social consequences that include addiction, respiratory and cardiovascular health complications, accidents and motor vehicle crashes, stunted academic achievement and educational attainment, and occupational underperformance (Hall et al., 2019; Volkow et al., 2014). Thus, the focus on malleable risk factors represents a public health and safety priority.

Indeed, targeting factors that increase the likelihood of health risk behaviors (i.e., risk factors) is one of the main tools of preventive efforts focused on health and well-being in general and efforts to prevent or reduce harms related to substance use in particular (Catalano et al., 2012). These efforts can be bolstered by understanding what substance-specific risk factors are elevated in what contexts, including the increasingly diverse legal contexts related to adult cannabis use. Young adults in the United States live in cannabis environments that vary substantially depending on how individual state legislatures approach restrictions on cannabis. These restrictions range from being consistent with the federal law and prohibiting and penalizing cannabis use for any purpose, to permitting medical use, decriminalizing nonmedical use, and legalizing use for nonmedical or “recreational” purposes for adults age 21 and over (for review see, e.g., Lancione et al., 2020; Pacula & Smart, 2017). The first state to legalize medical cannabis use was California in 1996, and many other states soon followed. In 2012, legalization of nonmedical cannabis use started at the state level when Washington State and Colorado voters approved statewide initiatives for the sale and use of cannabis for adults over the age of 21. Currently, more than two thirds of the U.S. population live in states where cannabis has been legalized for medical or nonmedical use (Johnson & Guttmannova, 2019), and all states that had legalized nonmedical use by 2016—when this study took place—had previously legalized medical use and decriminalized nonmedical use (Klieger et al., 2017). These changes in the legal context change access to cannabis and also shape perceptions about the acceptability of use and other beliefs, attitudes, and norms about cannabis, all of which are risk factors for use (e.g., Guttmannova, Jones, et al., 2019; Johnson & Guttmannova, 2019; Pacula & Smart, 2017).

The gradual loosening of legal restrictions on cannabis occurring in many U.S. states has sparked concerns not only about potential increases in cannabis use but also in other, often comorbid, substance use (e.g., Guttmannova, Jones, et al., 2019; Guttmannova et al., 2016; Johnson & Guttmannova, 2019; Pacula et al., 2015). Past research provides strong evidence that young adults who use cannabis are more likely to engage in alcohol, tobacco, or other substance use when compared to young adults who do not use cannabis (e.g., Palmer et al., 2009; Yurasek et al., 2017). This suggests that more permissive cannabis contexts could be associated with higher levels of risk factors not just specifically for cannabis, but for other substances as well, and effective prevention and intervention efforts aimed at reducing harmful behavior and improving health may need to include an additional focus on risk factors specific for other substances. Thus, the following two main research questions guided this study: (a) Are cannabis-specific risk factors higher among young adults who live in legally more permissive cannabis contexts? (b) Are other substance-specific risk factors (e.g., perceived harm of smoking cigarettes or drinking alcohol) also elevated in more permissive cannabis contexts?

Specifically, we examined perceptions about the harms related to cannabis; attitudes about and ease of access to cannabis; and perceived wrongfulness and social acceptability of cannabis, alcohol, and cigarette use and of selling of cannabis and other illegal drugs among young adults living in different legal cannabis contexts. Because past experience with cannabis use could influence beliefs, attitudes, and norms about cannabis use in a reciprocal manner (e.g., Guttmannova, Skinner, et al., 2019), and cannabis use may already be elevated in more permissive contexts (e.g., Pacula & Smart, 2017; Wen et al., 2015), we accounted for the prevalence of past-year use in the models assessing the association between permissiveness of the legal context and the risk factors. Other controls included sociodemographic factors associated with elevated prevalence of young adult substance use, such as sex, race/ethnicity, and education level (e.g., Banks & Zapolski, 2018; Townsend et al., 2007). We hypothesized that greater legal permissiveness toward cannabis use would predict elevated levels of cannabis-specific risk factors as well as risk factors specific to other substance use.

Methods

Sample

Data came from the Community Youth Development Study (CYDS; for a description of the study, original recruitment and sample, see Hawkins et al., 2008), a community-randomized controlled trial of the Communities That Care (CTC) prevention system originally conducted in 24 rural and small-town communities in seven states (Colorado, Illinois, Kansas, Maine, Oregon, Utah, and Washington) with first data collection occurring in 2004. The present study included young adults from the control sample only. The intervention sample was excluded because the trial showed that CTC reduced levels of risk factors and drug use in adolescence (e.g., Hawkins et al., 2014), with some intervention effects on drug use persisting into young adulthood (Oesterle et al., 2018) and the effects of the CTC intervention were not the focus of the present study. The sample is sex-balanced (52% female), with 26.6% self-identifying as Latino/a or Hispanic. In terms of race, 60.9%% self-identified as White, 2.7% as Black, 1.6% as Asian, 3.7% as Native American, 0.4% as Pacific Islander, 5.0% as multiracial, 24.7% as “other” (76.1% of participants who chose this category self-identified as Latino/a or Hispanic), and 0.9% did not answer the race question.

The data for the present study were collected in 2016 (n = 1,700; 85.3% of the active, still-living control sample) when the average age of participants was 23 years (M = 22.70; SD = 0.57) and all participants were over the age of 21, the legal age to consume alcohol and, in states where cannabis had been legalized for nonmedical or “recreational” purposes, to use cannabis.

Data were collected using the Youth Development Survey (based on the Communities That Care Youth Survey; Arthur et al., 2002) that was offered online or as a mailed paper questionnaire. Participants received completion incentives of $45 and the vast majority of participants (77%) completed the survey online. The study protocol was reviewed and approved by the University of Washington’s Human Subjects Review Committee.

Measures

The survey used the term “marijuana” in the assessment of risk factors and use without defining the term. Because participants could have interpreted the term in different ways, we use the more general and encompassing term “cannabis” in this paper.

Substance-specific risk factors.

Participants were asked about their perception of harm from using cannabis regularly and their perception of harm to children from seeing adults using cannabis in public. Response options ranged from 0 = high risk to 3 = no risk. Participants were also asked how wrong they think it is for someone their age to use cannabis and sell cannabis (0 = very wrong to 3 = not wrong at all); what the chances are that they would be seen as “cool” if they used cannabis (0 = no chance to 3 = very good chance); and how easy it would be to get some cannabis if they wanted to get it (0 = very hard to 3 = very easy). Participants were also asked about their perception of harm from smoking one or more packs of cigarettes a day, and harm from having five or more drinks of an alcoholic beverage in one occasion at least once or twice each week; how wrong it would be for someone their age to engage in these behaviors, and what were the chances they would be seen as “cool” if they engaged in these behaviors. Because alcohol and cigarette use were legal for this age group in all states, we did not assess ease of access to these two substances but did assess ease of access to prescription opioids (e.g., OxyContin, Vicodin, Percocet) and other prescription drugs (e.g., Adderall, Ritalin, Valium, Xanax) not prescribed to them by their doctor, and access to other drugs “like cocaine, LSD, or amphetamines.” We also assessed how wrong respondents thought it would be to sell drugs (other than cannabis). The response options for these questions about other substances were analogous to those asked about cannabis.

Legal permissiveness of cannabis context.

The legal permissiveness of young adults’ cannabis contexts was coded according to the legality of cannabis use in their state of residence at the time of the interview (February-October, 2016). The permissiveness of the cannabis legal context was coded on a 4-point scale ranging from 1 = not legal, 2 = legal medical, 3 = legal medical and decriminalized nonmedical, and 4 = legal medical as well as nonmedical (“recreational”). The majority (57.3%, n = 1,025) of participants resided where cannabis use was not legal, 5.4% (n = 93) where only medical cannabis use was legal, 8.2% (n = 141) in states that had legalized cannabis use for medical use and decriminalized nonmedical use, and 26.8% (n = 463) where medical and nonmedical use was legal. Six participants resided outside of the United States and were excluded from the analyses.

Covariates.

Cannabis use was based on a question about the frequency of past-year use and coded as any past-year use of cannabis (no use = 0; some use = 1). Analyses also adjusted for variation in demographic characteristics of participants and included: participant’s sex (male = 1, female = 0); race (White = 1; other = 0); whether the participant self-identified as Hispanic or Latino(a) (yes = 1, no = 0); and their highest level of education, coded as three mutually exclusive indicator variables of dropped out of high school (referent), high school only, and more than high school.

Analysis

All analyses were conducted in STATA 14.0. We estimated a series of ordered logistic regression models to examine the association between the permissiveness of the cannabis legal context and substance use-specific risk factors (outcomes). The predictors were entered in a series of steps. First, we entered the permissiveness of the context variable (Model 1), then we included the background and demographic covariates (Model 2). Past-year cannabis use was entered in the final model (Model 3) to control for the possibility that risk factors are higher due to more prevalent use in more permissive contexts. At the time of the Age 23 survey, participants were dispersed across 46 different U.S. states with little clustering within states beyond the original study states. Analyses adjusted for clustering in the original 12 study communities to reflect the study design. Specifically, all models included community indicator variables in order to control for clustering of participants in their original communities.

Results

Table 1 provides descriptive statistics of all study variables. Supplemental Tables 1 – 4 show full regression results for all three models and all outcomes. Figures 1 – 4 focus on the key association of interest by displaying adjusted odds ratios with their 95% confidence intervals from Model 3, which controlled for all covariates, including prior cannabis use. Results suggest that greater permissiveness of the cannabis legal context was significantly associated with higher levels of all cannabis-specific risk factors except for perception of harm from regular cannabis use, but not with any risk factors pertaining to other substance use.

Table 1.

Descriptive statistics for all study variables

Variables	Response options	%
Legal permissiveness of cannabis context	not legal	59.52
	legal medical	5.4
	legal medical and decriminalized	8.19
	legal medicl and nonmedical	26.89
Demographic covariates	Male	48.03
	White	61.76
	Hispanic	26.1
	No high school diploma	8.93
	High school diploma or GED	57.04
	Beyond high school diploma	34.03
Prevalence of past year cannabis use	Used cannabis in past year	36.42
Perception of harm from using cannabis regularly	No risk	20.09
	Slight risk	28.42
	Moderate risk	23.59
	Great risk	27.9
Perception of harm to children from seeing adults using cannabis in public	No!	8.14
	no	20.88
	yes	25.42
	Yes!	45.55
How wrong they think it is for someone their age to use cannabis	Very wrong	36.28
	Wrong	13.84
	Little bit wrong	17.5
	Not wrong at all	32.38
How wrong they think it is for someone their age to sell cannabis	Very wrong	55.78
	Wrong	15.34
	Little bit wrong	15.46
	Not wrong at all	13.42
Chances that they would be seen as “cool” if they used cannabis	No chance	47.78
	Little chance	21.88
	Some chance	16.92
	Pretty good chance	7.93
	Very good chance	5.48
How easy would it be to get some cannabis	Very hard	13.7
	Sort of hard	9.29
	Sort of easy	20.38
	Very easy	56.62
Perception of harm from smoking one or more packs of cigarettes a day	No risk	2.45
	Slight risk	3.9
	Moderate risk	16.78
	Great risk	76.86
Perception of harm from heavy and regular alcohol use 1	No risk	2.85
	Slight risk	17.58
	Moderate risk	36.96
	Great risk	42.61
How wrong it would be for someone their age to smoke one or more packs of cigarettes a day	Very wrong	50.84
	Wrong	22.2
	Little bit wrong	15.46
	Not wrong at all	11.5
How wrong it would be for someone their age to engage in heavy and regular alcohol use 1	Very wrong	39.3
	Wrong	19.65
	Little bit wrong	23.49
	Not wrong at all	17.56
Chances they would be seen as “cool” if they smoked one or more packs of cigarettes a day	No chance	65.13
	Little chance	25.36
	Some chance	6.82
	Pretty good chance	1.63
	Very good chance	1.05
Chances they would be seen as “cool” if they engaged in heavy and regular alcohol use 1	No chance	42.77
	Little chance	25.93
	Some chance	19.46
	Pretty good chance	8.92
	Very good chance	2.91
Ease of access to prescription opioids (e.g., OxyContin, Vicodin, Percocet) not prescribed by their doctor	Very hard	35.62
	Sort of hard	27.37
	Sort of easy	20.63
	Very easy	16.39
Ease of access to other prescription drugs not prescribed by their doctor	Very hard	34.07
	Sort of hard	25.58
	Sort of easy	21.8
	Very easy	18.55
Ease of access to other drugs “like cocaine, LSD, or methamphetamines”	Very hard	35.74
	Sort of hard	25.9
	Sort of easy	22.41
	Very easy	15.95
How wrong it would be to sell drugs (other than cannabis)	Very wrong	77.89
	Wrong	13.55
	Little bit wrong	6.81
	Not wrong at all	1.75

Notes: GED=General Educational Development diploma;

¹five or more drinks of alcohol in one occasion weekly or more often.

Figure 1.

Adjusted odds ratios and 95% confidence intervals around the odds ratios from ordered logistic regressions estimating associations between legal permissiveness of cannabis context and perception of easy access to different substances.

Figure 4.

Adjusted odds ratios and 95% confidence intervals around the odds ratios from ordered logistic regressions estimating associations between legal permissiveness of cannabis context and perception of harm from using different substances.

More specifically, as indicated in Figure 1 and supplemental Table 1, controlling for the aforementioned covariates, there was a significant positive association between permissiveness of context and perceived availability of cannabis (OR = 1.31; 95% CI = [1.11; 1.54]; p = 0.002), but not other illegal drugs, prescription opioids, or other prescription drugs. Similarly, there was a positive association between permissiveness of cannabis legal context and (a) the belief that cannabis use would make them “look cool” (OR = 1.18; 95% CI = [1.03; 1.36]; p = 0.021; see Figure 2 and supplemental Table 2); (b) the belief it is not wrong to use cannabis (OR = 1.16; 95% CI = [1.00; 1.36]; p = 0.049; see Figure 3 and supplemental Table 3); (c) the belief it is not wrong to sell cannabis (OR = 1.28; 95% CI = [1.10; 1.50]; p = 0.002; see Figure 3 and supplemental Table 3); and (d) perception of harm to children from seeing adults using cannabis in public (OR = 1.22; 95% CI = [1.05; 1.40]; p = 0.007; see Figure 4 and supplemental Table 4). However, and contrary to our hypotheses, the association between the permissiveness of cannabis context and the perception of harmfulness of regular substance use was not statistically significant for cannabis use nor the use of alcohol and cigarettes (see Figure 4 and supplemental Table 4).

Figure 2.

Adjusted odds ratios and 95% confidence intervals around the odds ratios from ordered logistic regressions estimating associations between legal permissiveness of cannabis context and favorable attitudes toward substance use (seen as “cool” if engaging in…).

Figure 3.

Adjusted odds ratios and 95% confidence intervals around the odds ratios from ordered logistic regressions estimating associations between legal permissiveness of cannabis context and favorable attitudes toward substance use (“How wrong would it be for someone your age to…”).

Discussion

The legal and social environment has become increasingly permissive toward cannabis use over the last two decades. This study examined associations between beliefs, attitudes, and perceptions about cannabis as well as other substances and the permissiveness of the cannabis legal context in which young adults resided in order to inform future preventive interventions for substance use among young adults. The main findings of this study are that young adults who were living in legally more permissive cannabis contexts, indeed, were more likely to report more approving beliefs, attitudes, and perceptions towards cannabis, but not toward other substances, including alcohol and tobacco. These findings are consistent with reports of increased presence of cannabis-related advertising and pro-cannabis public messaging in more permissive cannabis contexts (e.g., Fiala et al., 2018) and studies that have reported a positive association between pro-cannabis media presence and cannabis-specific risk factors and use among adolescents (e.g., D’Amico et al., 2015; Fiala et al., 2020) as well as young adults (e.g., Krauss et al., 2017).

While perception of harm to children from seeing adults using cannabis in public was positively associated with the permissiveness of the legal context, perception of harm from using cannabis regularly was not. This finding is unexpected given reports of increasing likelihood that young adults perceive no or low health risk related to cannabis use in more permissive cannabis environments, such as where medical cannabis has been legalized (Wen et al., 2019). The lack of a statistically significant association between the permissiveness of the legal context and perception of harm from regular use in our study could reflect a general trend towards decreased perception of harm from cannabis use nationwide, even where cannabis laws have not become more permissive (Schulenberg et al., 2020). Continuing to focus prevention and intervention efforts on perceptions of harm from regular cannabis use no matter the legal context may, therefore, be advised. The prevalence of alcohol and cigarette use among young adults has been declining nationally in recent years (Substance Abuse and Mental Health Administration, 2019) and may be a consequence of longstanding public health campaigns that clearly communicate harms from cigarettes (e.g., Bala et al., 2017; Dilley et al., 2012) and regular alcohol use (e.g., Young et al., 2018).

Although the results of this study suggest that greater permissiveness of cannabis legal context is associated with higher levels of cannabis-specific risk factors among young adults, the good news is that this increase does not extend to risk factors specific to other substances, such as favorable attitudes toward alcohol and tobacco use. However, an increase in cannabis-specific risk factors may impact other substance use indirectly as studies have documented complementary relationships between cannabis, alcohol, and tobacco use as well as positive associations between substance-specific risk factors for different substances (e.g., Jones et al., 2019). Such associations suggest that increases in cannabis use could be associated with increases in other substance use (e.g., Guttmannova et al., 2016). In the context of increasing legal permissiveness of cannabis across the nation, the findings of this study point to the importance of prevention and intervention efforts for young adults that focus specifically on reducing cannabis-specific risk factors.

Limitations and Directions for Future Research

The present study has some limitations and the findings should be interpreted with these in mind. First, while we controlled for past-year cannabis use in our models of association between cannabis legal context and cannabis-specific risk factors, the reports are cross-sectional and future studies could incorporate both concurrent and reciprocal associations between risk and use in order to understand their dynamic interplay. Next, this study did not evaluate the impact of state-level cannabis policies at a broad, population-level, aggregate impacts, but instead examined the association between self-reported substance use risk factors and the permissiveness of that person’s particular context. This constitutes a novel way of examining the role of broad societal changes in individuals’ lives but precludes making conclusions about the effects of cannabis policy changes on outcomes. Similarly, there was temporal variability in the length of period when cannabis was legalized in different jurisdictions and this study was not designed to capture it. Future studies, particularly those focusing on the effects of cannabis policy changes on outcomes, could zoom in on the effect that length of legality in different jurisdictions may have on young adults’ attitudes and related factors. Along those lines, studies that account for the possibility that cannabis-related risk factors such as favorable attitudes toward cannabis might already be elevated in jurisdictions that eventually implement cannabis policy change would be particularly informative. Enforcement of cannabis laws, which was not included in this analysis, may vary by jurisdiction (Ward et al., 2018), and has been shown to disproportionately impact people and communities of color and those with socioeconomic disadvantage (Stohr et al., 2020). These factors could also contribute to perceptions and beliefs about cannabis. Future studies, particularly those designed to evaluate the effect of changes in cannabis policy on risk factors and use, could focus specifically on the implementation of legalization with more targeted samples. Next, the study’s sample consisted of young adults who grew up in rural and small towns in seven states in the West, Midwest, and Northeast regions of the United States. Although they lived in 46 out of 50 U.S. states by the time they reached young adulthood, the original locations from which the sample was drawn may limit generalizability of the findings.

Conclusions

The findings from this study suggest that future preventive interventions for cannabis use for young adults who live in legally more permissive cannabis contexts should (a) be cannabis-specific; and (b) go beyond targeting harm of regular cannabis use to also include other perceptions, norms, and beliefs about cannabis use. For example, more permissive cannabis contexts provide opportunities for education and prevention and intervention efforts focused on risk factors such as perception of harm to children from seeing adults using cannabis in public. While risk factors specific to other substances were not elevated in more legally permissive cannabis contexts—an encouraging finding—epidemiologic studies should continue to monitor their levels to ensure continued appropriate tailoring of preventive interventions aimed at reducing harm from substance use.