Daily Fluctuations in Drinking Intensity: Links with Vaping and Combustible Use of Nicotine and Marijuana

Abstract

Background:

Whether alcohol use intensity on a given day is linked with nicotine or marijuana use that same day is not well known, nor are links of drinking intensity with different modes of nicotine and marijuana use. This study examined these within-person links across 14 days in a national sample of young adults (YAs).

Methods:

Past 30-day drinkers participating in the U.S. nationally-representative Monitoring the Future study of 12^th graders in 2018, who also reported alcohol use during a 14-day data collection one year later in the Young Adult Daily Life Study in 2019, were included (N=487). Weighted multilevel modeling estimated within- and between-person associations of drinking intensity with cigarette smoking, nicotine vaping, marijuana smoking, and marijuana vaping.

Results:

Within-person fluctuations in drinking intensity on a given day were associated with cigarette smoking, nicotine vaping, and marijuana smoking, but not marijuana vaping. There were significant between-person associations of means of drinking intensity and each outcome, except for cigarette smoking.

Conclusion:

Drinking intensity on a given day was associated with multiple modes of nicotine use and marijuana smoking that day. Nicotine and marijuana use remain critical areas of concern for public health, and future research and interventions should consider the comorbidity of drinking intensity and multiple modes of nicotine and marijuana use. Focusing on the same-day use of alcohol may provide a tailored avenue for preventing and reducing nicotine and marijuana emerging trends among YAs.

1. Introduction

1.1. Overlapping substance use

Young adults (YAs) who drink alcohol are more likely to use nicotine and marijuana, as documented by many cross-sectional and longitudinal studies. The prevalence of using both alcohol and nicotine is high, as many smokers also can be classified as heavy drinkers (Dawson, 2000; McKee and Weinberger, 2013; Roche et al., 2019), and many YA marijuana users also engage in alcohol use (Patrick et al., 2019b, 2020; Sokolovsky et al., 2020). However, when use is measured over long periods of time (e.g., months or years), it cannot be easily determined whether alcohol and nicotine or marijuana are used simultaneously or on the same day. Documented between-person associations do not provide evidence about use on the same occasions, which is critical for understanding overlapping patterns of substance use among YAs.

Simultaneous or same-day use of alcohol with other substances is receiving increasing attention (Patrick et al., 2020, 2019b; Roche et al., 2019, 2016; Sokolovsky et al., 2020). Many YAs engage in simultaneous use of alcohol and marijuana (Patrick et al., 2020, 2019b; Sokolovsky et al., 2020; Subbaraman and Kerr, 2015), which increases risk of detrimental consequences of both substances (Lee et al., 2020; Sokolovsky et al., 2020). However, little work has examined whether days with greater drinking intensity are more likely to be days when nicotine or marijuana is used. More is known about between-person associations of alcohol use measured retrospectively with nicotine or marijuana use across longer time frames (e.g., 30 days), but less about how within-person differences in alcohol use relate to nicotine or marijuana use. Notable exceptions have examined simultaneous use of alcohol and nicotine or marijuana with data across days (Jackson et al., 2010; Metrik et al., 2018; Patrick et al., 2019a; Witkiewitz et al., 2012). These studies have shown that days with increased alcohol use are associated with greater likelihood of marijuana smoking and light/moderate cigarette smoking. We extend this work by specifically examining how within-person fluctuations in drinking intensity are related to smoking and vaping of nicotine and marijuana in a sample of YA drinkers.

1.2. High-intensity drinking

Examining any alcohol use or binge drinking (e.g., consuming 4 or more drinks for females and 5 or more drinks for males [National Institute on Alcohol Abuse and Alcoholism, 2010; Courtney and Polich, 2009]) is common in research linking drinking with nicotine or marijuana use. Alternative, graded measures are needed to differentiate higher-risk groups and situations associated with greater likelihoods of nicotine and marijuana use (Linden-Carmichael et al., 2019; Patrick, 2016; Patrick et al., 2017a, 2016). High-intensity drinking (HID) measures capture higher-risk drinking occasions, defined as consuming 8 or more drinks for females and 10 or more drinks for males (Patrick, 2016). People who engage in HID experience more harmful outcomes (Linden-Carmichael et al., 2017; Patrick et al., 2017b), and occasions when people engage in HID result in more consequences (Patrick et al., 2016; Patrick and Terry-McElrath, 2021).

Some characteristics differentiate those who engage in HID from moderate and/or binge drinkers: males, cigarette smokers, and those with depression symptoms are more likely to engage in HID (Patrick et al., 2021). Other risk factors do not differentiate binge and high-intensity drinkers, including drinking motives and nicotine vaping frequency (Patrick et al., 2021). The between-persons link between HID and marijuana use is unclear (Patrick et al., 2021), and there is a dearth of research on how drinking intensity relates to same-day use of different modes of nicotine and marijuana use.

HID has some unique risk factors and consequences, indicating that binge and HID are distinct. We contend that a categorical measure that distinguishes levels of drinking intensity across days can more precisely depict links with modes of nicotine and marijuana use on those days.

1.3. Modes of nicotine and marijuana use

Separately examining smoking and vaping is increasingly important because prevalence of nicotine and marijuana vaping are rapidly increasing among YAs. Combustible tobacco use has declined among adolescents and YAs since the 1980s, with dramatic declines since 2010 (U.S. Department of Health and Human Services, 2012; Schulenberg et al., 2020). Yet, the introduction of highly marketed novel delivery systems (e.g., JUUL, Puff Bar) has presaged a nicotine vaping epidemic among adolescents and YAs (King et al., 2020; Miech et al., 2020; Schulenberg et al., 2020). Moreover, since approximately 2012, cigarette smoking initiation and escalation is increasingly occurring in young adulthood (Hair et al., 2017; Perry et al., 2018; Terry-McElrath et al., 2017; Villanti et al., 2019), and marijuana vaping prevalence increased in 2018 and 2019 (Schulenberg et al., 2020; Seaman et al., 2020). This is important because e-cigarette, or vaping, product use-associated lung injury (EVALI) has been linked to vitamin E acetate and marijuana vaping (EVALI; King et al., 2020; Ellington. et al., 2020).

The current project explicitly acknowledges tobacco and marijuana products currently used by YAs by separately examining cigarette smoking, nicotine vaping, marijuana smoking, and marijuana vaping. Moreover, research on alcohol and nicotine or marijuana use tends not to use national samples of YAs, and our project addresses this gap by utilizing a national sample of YAs across the US. Finally, we conduct within-person analyses that examine same-day associations between alcohol and nicotine or marijuana use, providing same-day associations that rule out stable confounders (i.e., both measured and unmeasured).

1.4. Current study

Understanding whether within-person fluctuations in drinking intensity co-occur with lower or higher likelihoods of other substance use and mode of use on those same days is key to documenting patterns and harms as they occur in real life. Intensive data on the same YAs across multiple days can decompose more stable between-person differences (people who use one substance may use another, in general) from within-person fluctuations (days with greater drinking intensity may be linked with more use or different modes of use). Analyses on within-person fluctuations in drinking intensity (relative to people’s mean level of drinking) on a given day and different modes of nicotine and marijuana use (i.e., cigarette use, nicotine vaping, combustible marijuana use, and marijuana vaping) on that day is also critical because more is known about the link between alcohol and more traditional forms of use, but less is known about emerging forms of nicotine or marijuana use. Therefore, our results would apply to patterns and harms within the changing landscape of nicotine and marijuana product use.

In this project, we use a 14-day survey design in a national sample of YAs who reported recent drinking and past 30-day drinking at age 18 to examine the relationships between levels of drinking intensity (moderate, binge, and high-intensity drinking) and cigarette smoking, nicotine vaping, marijuana smoking, and marijuana vaping. We consider how within-person fluctuations in drinking intensity across days relate to smoking and vaping of both substances, while accounting for between-person differences in alcohol use intensity. We hypothesized that greater drinking intensity would be positively associated with cigarette smoking, nicotine vaping, marijuana smoking, and marijuana vaping.

2. Methods

2.1. Data and sample

Data came from the first wave (2019) of the Young Adult Daily Life (YADL) Study (Patrick & Terry-McElrath, 2021). YADL participants were drawn from a nationally-representative sample of 12^th grade students in the U.S. who participated in the Monitoring the Future (MTF) study in Spring 2018 (see Miech et al., 2020 for detailed methodology on the MTF 12^th grade study). MTF participants (N=14,502) were eligible for YADL participation if they reported using alcohol in the past 30 days in the MTF 12th grade survey (n=4,240), were not randomly selected for the MTF longitudinal study (n=828 excluded; for detailed MTF longitudinal study methods, see Schulenberg et al., 2020), and provided contact information necessary for follow-up (n=1,208 excluded; for more information on YADL see Patrick & Terry-McElrath, 2021). In total, after removing these 2,036 participants, there remained 2,204 eligible participants. In Year 1 of YADL, there was an annual online survey followed by 14 consecutive daily online surveys in Spring 2019. Respondents received up to a $100 incentive. The study was approved by an Institutional Review Board. Of the 2,204 eligible individuals, 911 (41.3%) consented and responded to the initial wave of the YADL.

The YADL sample had an average age of 19.3 years (SD=0.40), and was 56.8% male. Non-Hispanic White YAs were 67.5% of the sample, followed by Hispanic YAs (20.2%) and non-Hispanic Other YAs at 12.3%. Using MTF 12^th grade data, participation in YADL among eligible students was more likely among participants who were female (vs. male; p<0.001); had 2 parents in the household (vs. fewer; p<0.001); lived in the Northeast (vs. other regions; p<0.05); had high school grades of B- or above (vs. lower; p<0.01); had definitive plans to graduate from a 4-year college (vs. not; p<0.001); did not engage in binge drinking (vs. did p<0.05); and had lower past 30-day drinking frequency in 12^th grade (p<0.001). Race/ethnicity, high school religiosity, high school truancy, and parental education did not predict YADL participation.

The present study included only participants who reported any alcohol use during the 14-day daily YADL survey period (63.4%, n=487) due to the focus on within-person fluctuations in drinking intensity at the day level. The analytic sample of 487 individuals was used in all analyses, and pairwise deletion was used for missing data. From the 487 individuals in analysis, there were a total of 6,019 and 6,023 days for nicotine and marijuana outcomes, respectively. Among the 487 individuals, over 80% provided a full 14 days of data.

2.2. Measures

2.2.1. Substance use outcomes

Each day, respondents were asked to report on their substance use the day before, from the time they woke up to the time they went to sleep. Respondents were asked on each day, “Did you do any of the following on this day?”, and we examined the two response options of “use nicotine or tobacco” and “use marijuana.” Follow-up questions asked: “How did you use nicotine or tobacco?” or “How did you use marijuana?” The response options examined were: “smoked cigarettes,” “vaped nicotine,” “smoked marijuana,” and “vaped marijuana.” Four dichotomous measures captured any (1) cigarette smoking, (2) nicotine vaping, (3) marijuana smoking, and (4) marijuana vaping (all coded 1=yes, 0=no) on a given day. These measures were worded to ensure participants could report any use of each product; questions did not ask if products included both tobacco and marijuana (e.g., blunts).

2.2.2. Drinking intensity

Respondents were asked about any alcohol use, and on drinking days respondents were asked how many total drinks they consumed (1–25+ drinks). Drinking intensity was coded using sex-specific thresholds: 0=no drinking; 1=moderate drinking (1–3 drinks for women/1–4 for men); 2=binge drinking (4–7 drinks for women/5–9 for men); and 3=high-intensity drinking (8+ drinks for women/10+ for men [Patrick, 2016]). Our drinking measure did not include a time component. The modal time spent drinking was 6–7 hours (43.9%) on HID days, 3–4 hours (28.6%) on binge drinking-only days, and <1 hour (45.3%) on moderate drinking days. There is no clear indication that timeframes (e.g., a two-hour window) should be imposed on ordinal measures that capture drinking intensity on a given day; for instance, HID involves a large amount of alcohol consumption on a given day regardless of a two-hour window (Patrick, 2016).

Covariates

At the day level, we included two controls: Weekend days vs. weekdays (1=Thursday, Friday, or Saturday, 0=other days; Maggs et al., 2011), and survey day number (0–13; Patrick and Terry-McElrath, 2021]).

At the person level, covariates included, race/ethnicity (recoded dichotomously to indicate non-Hispanic White=1 versus Other=0 due to small sample sizes among drinkers), sex (1=male, 0=female), and college student status (currently attending 4-year college or university full-time=1 vs. 0=other).

2.3. Analytic strategy

We used two-level multilevel models (MLM) to examine within- and between-person associations across days between alcohol use and cigarette smoking, nicotine vaping, marijuana smoking, and marijuana vaping. The YADL’s data structure nests days within individuals, which results in correlated residuals. MLM accounts for this dependence in error terms by analyzing days and individuals as separate levels of data and including random effects (Raudenbush and Bryk, 2002). Our MLMs were logistic models due to binary outcomes. Weights were used to adjust for sampling and nonresponse.

MLMs estimate two levels of regression equations simultaneously (Raudenbush and Bryk, 2002). At the day level, a within-person equation modeled the time-varying relationships between drinking and the four outcomes across days. Time-varying covariates (i.e., day of week, day in study) were included and within-person residuals estimated. At the person level, between-person equations modeled the average log-odds of each of the four outcomes and allowed for the inclusion of time-invariant variables (i.e., person means of the drinking intensity measure across drinking days, race/ethnicity, sex, and student status). Regarding between-person residuals, slopes between drinking intensity and 3 of the 4 outcomes varied significantly across persons (not included for cigarette smoking), thus an error term allowing slopes to vary randomly was included in each of these models.

To isolate the within-person associations between drinking intensity and different modes of nicotine or marijuana use, drinking intensity was group-mean centered at the day level and the person means of drinking intensity were included in the individual-level equation (Raudenbush and Bryk, 2002). We conducted supplemental analyses to support the use of the ordinal drinking measure (full details in Appendix A). All analyses were conducted in Stata v.17, using the “melogit” command with probability weights and unstructured covariance for models with random slopes.

3. Results

Across drinking days among the analytic sample of YAs who reported any alcohol use during the 14-day survey period (N=487), nicotine vaping was the most prevalent outcome (23.6% of days), followed by marijuana smoking (17.2%), marijuana vaping (10.0%), and cigarette smoking (3.3%). Across persons, 40.8% reported nicotine vaping, 38.7% reported marijuana smoking, 31.1% reported marijuana vaping, and 8.2% reported cigarette smoking at least once in the 14 reported days. Supplementary analyses showed that that the percent of days cigarette smoking, nicotine vaping, marijuana smoking, and marijuana vaping occurred was high on days with HID even relative to binge drinking; for instance, nicotine vaping occurred on 37% of HID days compared to 35% and 27% of binge and moderate drinking days, respectively.

Approximately 9% of variance in cigarette smoking and 6% of variance in nicotine vaping was within persons (intraclass correlation coefficients = 0.91 and 0.94, respectively). For marijuana use; 17% of variance in marijuana smoking and 25% of variance in marijuana vaping was within persons (intraclass correlation coefficients = 0.83 and 0.75, respectively).

3.1. MLM Result

3.1.1. Cigarette smoking

Table 2 shows MLM results. Drinking intensity on a given day was positively associated with cigarette smoking that day (AOR=2.82; 95% CI=1.57, 5.80). At the day level, each step up the 0 to 3 scale of drinking intensity corresponded with a 182% increase in the odds of cigarette smoking. Cigarette smoking was less likely to occur on weekdays than weekends (AOR=0.61; 95% CI=0.40, 0.91).

Table 2.

Multilevel Logistic Regression Results for Day- and Individual-Level Associations between Drinking Intensity and Multiple Smoking and Vaping Measures among Young Adult Drinkers^a

	Nicotine				Marijuana
	Cigarette Smoking		Vaping		Smoking		Vaping
Variables	Odds Ratio	95% C.I.	Odds Ratio	95% C.I.	Odds Ratio	95% C.I.	Odds Ratio	95% C.I.
Level 1: Day
Drinking intensity (range 0 to 3)	2.83 **	(1.57, 5.08)	2.57 **	(1.49, 4.43)	1.99**	(1.27, 3.12)	1.10	(0.67, 1.81)
Weekend (vs. weekday)	0.61 *	(0.40, 0.91)	1.24	(0.91, 1.70)	0.90	(0.67, 1.22)	0.95	(0.72, 1.27)
Survey day number (0 to 13)	0.93	(0.79, 1.08)	0.97	(0.90, 1.04)	0.97	(0.92, 1.02)	0.96	(0.90, 1.02)
Level 2: Individual
Person mean drinking intensity	4.30	(0.73, 25.45)	3.80 *	(1.24, 11.66)	2.58 *	(1.16, 5.73)	2.21 *	(1.17, 4.17)
Hispanic/Other (vs. non- Hispanic White)	0.12	(0.00, 4.17)	0.08 *	(0.01, 0.98)	1.57	(0.47, 5.26)	0.96	(0.29, 3.24)
Male sex (vs female)	8.39	(0.62, 114.39)	1.01	(0.16, 6.20)	7.67**	(1.92, 30.67)	6.76 ***	(2.63, 17.36)
Full-time student at 4- year college (vs other)	0.03 *	(0.00, 0.69)	0.03 **	(0.00, 0.28)	0.09 ***	(0.02, 0.33)	0.67	(0.21, 2.15)
Level-2 variance components
Intercept	35.74 ***		### ***		15.76 ***		9.93 ***
Drinking intensity slope	~		0.82 *		0.24 *		0.24 **

^aNotes. Individual N = 487 young adults who reported any drinking on at least 1 day out of 14 daily surveys

Day N for cigarette smoking = 6019; nicotine vaping = 6019; marijuana smoking = 6023; marijuana vaping = 6023

^~Random slopes were not included/needed for cigarette smoking

Significance tests for variance components were derived from unweighted models

^***=p<.001;

^**=p<.01;

^*=p<.05

For between-person covariates, there were no differences in cigarette smoking by person means of drinking level, sex, or race/ethnicity. Full-time college students had lower odds of cigarette smoking compared to other YAs (AOR=0.03; 95% CI=0.01, 0.69).

3.1.2. Nicotine vaping

Drinking intensity on a given day was positively related to nicotine vaping that day (AOR=2.57; 95% CI=1.49, 4.43). At the day level, each step up the 0 to 3 scale of drinking intensity corresponded with a 157% increase in the odds of nicotine vaping.

Person means of drinking intensity were positively related to nicotine vaping. There were no statistically significant differences by sex. Other race/ethnicity YAs had lower odds of nicotine vaping compared to White YAs (AOR=0.08; 95% CI=0.01, 0.98). Full-time college students had lower odds of nicotine vaping (AOR=0.03; 95% CI=0.01, 0.28).

3.1.3. Marijuana smoking

Drinking intensity on a given day was positively associated with marijuana smoking (AOR=1.99; 95% CI=1.27, 3.12). Each step up the 0 to 3 scale of drinking intensity on a given day corresponded with a 99% increase in the odds of marijuana smoking that day.

Person means of drinking intensity were positively related to marijuana smoking. Females compared to males (AOR=0.13; 95% CI=0.03, 0.52) and full-time college students versus others (AOR=0.09; 95% CI=0.02, 0.33) had lower odds of marijuana smoking.

3.1.4. Marijuana vaping

Drinking intensity on a given day was not related to marijuana vaping that day (AOR=1.10; 95% CI=0.67, 1.81).

Person means of drinking intensity were positively related to marijuana vaping (AOR=2.21; 95% CI=1.17, 4.17). Females had lower odds of marijuana vaping compared to males (AOR=0.15; 95% CI=0.06, 0.38).

4. Discussion

Drinking intensity on a given day was a key predictor of different modes of nicotine and marijuana among YA drinkers. That is, on days when individuals drank more intensely than usual, they were more likely to smoke cigarettes, vape nicotine, and smoke marijuana on those days. We did not find a significant association between drinking intensity and marijuana vaping across days in this sample of YAs who engaged in alcohol use at ages 18 and 19.

4.1. Within-person drinking intensity

Understanding within-person fluctuations in drinking intensity can provide context for concurrent risks for traditional and emerging modes of nicotine and marijuana use. YAs fluctuate greatly in their drinking intensity across occasions (Maggs et al., 2011; Patrick & Terry-McElrath, 2021), and therefore drinking intensity is not a static state that is easily or well captured with aggregate reports of past 30-day use. Research on drinking should incorporate high-intensity levels (Patrick, 2016) across occasions. We found nicotine and marijuana use were especially likely on days when HID occurred, even relative to binge drinking. Understanding how this co-occuring substance use contributes to acute risks associated with days of HID is important for intervening to reduce harms.

4.2. Modes of nicotine and marijuana use

Cigarette smoking, nicotine and marijuana vaping, and marijuana smoking among YAs remain critical public health issues (Schulenberg et al., 2020; Villanti et al., 2019). We found clear patterns showing that nicotine and marijuana were more likely to be used on days with heavier alcohol consumption; yet there were differences across modes of use, as drinking intensity on a given day was associated with cigarette smoking, nicotine vaping, and marijuana smoking that day, but not with marijuana vaping.

4.1.1. Cigarette smoking

Cigarette smoking was the least common outcome, and the within-person relationship of drinking intensity with cigarette smoking was stronger than for other modes of nicotine and marijuana use. Within-person fluctuations in drinking intensity are clearly linked with same-day risk for cigarette smoking, but YAs who drink more intensely on average were not more likely to smoke cigarettes.

It is well known that YAs who use alcohol more often are more likely to smoke cigarettes (McKee and Weinberger, 2013; Roche et al., 2016), but less attention in tobacco control efforts has been devoted to whether YA smoke on days they drink more intensely. Tobacco control research has attempted to determine why certain populations continue to smoke despite dramatic declines in the prevalence of past 30-day cigarette use over the past several decades (U.S. National Cancer Institute., 2017). As cigarette smoking prevalence has declined, prevention efforts have been focused on reducing cigarette use among vulnerable populations that still smoke at disproportionately high rates. Understanding fluctuations in drinking intensity and their association with cigarette smoking on a given day, rather than between-person differences, could help prevention efforts by focusing on variables linked to smoking in young adulthood and not just focusing on populations of YAs who still smoke cigarettes.

Despite dramatic declines in smoking, achieving the endgame of zero tobacco use at the population level has become an elusive goal (Fairchild et al., 2014). More fine-grained analyses within groups of smokers could identify the situational triggers that could be targeted, including intensive drinking occasions and settings. While it is important to continue to tackle tobacco-related disparities by focusing on key stable predictors of cigarette use (e.g., socioeconomic status), our research indicates that tackling changes in drinking intensity on a given day is also important among those who continue to smoke cigarettes.

4.1.2. Nicotine vaping

Nicotine vaping was the most prevalent outcome, reflecting the epidemic of nicotine vaping among youth and YAs (Cullen et al., 2018; King et al., 2020). We found marked within- and between-person associations between drinking intensity and nicotine vaping. Vaping nicotine was most common on days when HID occurred (nicotine vaping occurred on 37% of HID days), indicating that interventions should acknowledge this co-occurring risk.

Our findings help to distinguish the extent to which nicotine vaping is associated with alcohol use, which has been a gap in research despite the call for more investigation by the Surgeon General (U.S. Department of Health and Human Services, 2016). Our findings identify differences both between people and across days, supporting a comprehensive approach to prevention (Cronce et al., 2018). To intervene on the day level and reduce nicotine vaping, alcohol-related messaging could be incorporated into health communication interventions designed to reach YAs in real time (Calabro et al., 2019). Vaping-specific messaging is needed to reduce nicotine use (Graham et al., 2021), but tailored interventions could also include alcohol-related messaging to prevent and reduce vaping on more intense drinking days.

4.1.3. Marijuana smoking

Marijuana smoking was common in our sample of YA drinkers (i.e., almost 40% of participants). Consistent with previous between-person research (Seaman et al., 2020; Terry-McElrath and Patrick, 2018; Looby et al., 2021), YAs who reported higher drinking intensity on average were more likely to smoke marijuana. Uniquely, this study also documented that YAs were more likely to smoke marijuana on days they drank more intensely. Effect sizes were somewhat smaller than for cigarette smoking and nicotine vaping. Because YAs reported drinking alcohol and smoking marijuana on many of the same days, future research and prevention efforts should target overlapping patterns of use. A key question is whether motivations for same-day use (Patrick et al., 2020) include efforts to complement or substitute the individual effects of each substance (Risso et al., 2020). Another important research question with prevention and clinical implications is whether enduring person-level characteristics such as depression or impulsivity may moderate same-day use patterns.

4.1.4. Marijuana vaping

In contrast to the three other outcomes, marijuana vaping on a given day did not co-occur with higher levels of drinking intensity. Comparing individuals to each other, however, YAs with higher drinking intensity on average were more likely to vape marijuana, similar to prior between-subjects research (Sokolovsky et al., 2020). There may not be an association between marijuana vaping and drinking on a given day among YAs in the sample, like there is for marijuana smoking. YAs who use alcohol and marijuana at the same time are more likely to smoke marijuana and use alcohol at parties (Looby et al., 2021), and therefore contexts of use appear to matter. For instance, at parties with HID, marijuana smoking may be more common than marijuana vaping, and perhaps marijuana vaping may occur more often when YAs are alone and not drinking intensely. Future research examining situational predictors and contexts of marijuana vaping is needed to understand if and how drinking relates to marijuana vaping. For instance, future research should consider how context moderates the drinking intensity-marijuana use relationship.

4.3. Limitations and strengths

These results provide a unique perspective on how drinking intensity on a given day is associated with nicotine and marijuana use. However, it is important to note that the data cannot decipher the sequence of use across given days, only associations and not causal relationships. The study had good compliance across the 14 study days, and sensitivity analyses (not reported) indicated noncompliance did not affect results. It is possible that participants combined modes of use—e.g., smoked tobacco while also smoking marijuana (a blunt)—although we did not capture dual nicotine and marijuana use. Future research should consider how drinking intensity across days relates to dual use of products and/or modes.

5. Conclusion

Drinking intensity on a given day was associated with cigarette smoking, nicotine vaping, and marijuana smoking, but not with marijuana vaping. Public health researchers and practitioners should consider the comorbidity of high-intensity drinking in prevention efforts geared toward preventing and reducing multiple modes of nicotine and marijuana use among YAs.

Table 1.

Descriptive Statistics for Sample of Young Adult Drinkers^a: Smoking and Vaping Outcomes and Day- and Individual-Level Variables

	%/Mean	S.E.
Outcomes at day level b
Cigarette smoking	3.3%	(0.3)
Nicotine vaping	23.6%	(0.7)
Marijuana smoking	17.2%	(0.7)
Marijuana vaping	10.0%	(0.6)
Day-level predictors
Day-level ordinal drinking intensity (on given day)c	0.26	(0.01)
Weekend day (Thurs/Fri/Sat; percent of days)	43.1%	(0.8)
Survey day number (0 to 13)	6.24	(0.07)
Individual-level predictors
Individual-level average ordinal drinking intensity	0.26	(0.02)
Race/ethnicity
Non-Hispanic White	67.5%	(2.8)
Hispanic/non-Hispanic Other	32.5%	(2.8)
Male sex (vs. female)	56.8%	(2.7)
Full-time student at 4-year college (vs. other)	64.5%	(3.0)

Notes. Days N varies based on missing data (range 6,028 – 6,089).

^aIndividual N = 487 young adult participants who reported any drinking on at least 1 day out of 14 daily surveys

^bOutcomes at the day level indicate the percent of days on which the outcome occurred.

^cFor alcohol measure: 0 = no drinking, 1 = moderate, 2 = binge, and 3 = high-intensity

Highlights.

• Limited extant research addresses within-person drinking intensity across days.

• Outcomes were different modes of nicotine and marijuana use in young adulthood.

• Drinking intensity on a given day is linked to cigarette use and nicotine vaping.

• Drinking intensity on a given day is linked to marijuana smoking, but not vaping.

• Tailored prevention should address same-day drinking and nicotine or marijuana use.