Subjective responses to simultaneous alcohol and cannabis use relative to alcohol‐only use and cannabis‐only use: An ecological momentary assessment investigation
Subjective responses to simultaneous use
Coelho et al.
Department of Psychology York University Toronto Ontario Canada
Department of Population and Public Health Sciences University of Southern California Los Angeles California USA
Institute for Addiction Science University of Southern California Los Angeles California USA
Institute for Mental Health Policy Research Centre for Addiction and Mental Health Toronto Ontario Canada
Department of Psychiatry University of Toronto Toronto Ontario Canada
* CorrespondenceJeffrey D. Wardell, Department of Psychology, York University, 277 Behavioural Sciences Building, 4700 Keele St., Toronto, ON M3J 1P0, Canada.
Email: jwardell@yorku.ca
Abstract
Background
Few studies have examined subjective responses to simultaneous alcohol and cannabis use in naturalistic settings. The current study used ecological momentary assessment (EMA) to compare subjective responses between simultaneous use and both alcohol‐only use and cannabis‐only use sessions, while also examining the moderating role of quantities of alcohol and cannabis consumed at both the session and person levels.
Methods
Young adults (N = 149, 59.73% women) reporting recent simultaneous use completed three 21‐day EMA bursts, spaced 6 months apart. Participants completed a survey each time they initiated a new session of alcohol or cannabis use, along with two hourly follow‐up surveys. Surveys assessed alcohol use (quantities), cannabis use (quantities, forms of cannabis), and current acute subjective responses.
Results
At the session level, simultaneous use (vs. alcohol‐only use) was associated with greater peak sedation and intoxication, with the latter association strengthened during sessions involving relatively lighter drinking. Simultaneous use sessions also involved greater peak liking ratings relative to alcohol‐only use sessions, though only among participants who reported relatively lower average alcohol consumption. In addition, relative to cannabis‐only use sessions, simultaneous use sessions were associated with greater peak energized and liking ratings, with the former association strengthened during sessions involving relatively heavier cannabis concentrate use. Simultaneous use sessions also involved lower peak sedated and anxious ratings relative to cannabis‐only use sessions, though only among participants who reported relatively lighter average cannabis consumption.
Conclusions
Overall, simultaneous use was experienced as more impairing (i.e., greater intoxication and sedation) than alcohol‐only use and as more reinforcing (i.e., greater stimulation and liking) than cannabis‐only use. Some differences in subjective responses between simultaneous use and single‐substance use sessions depended on session‐level or person‐level consumption amounts, which may inform tailored interventions for simultaneous use.
Graphical
This ecological momentary assessment study found that when young adults were engaging in the simultaneous use of alcohol and cannabis, they tended to report greater intoxication and sedation relative to alcohol‐only use sessions and greater stimulation and liking relative to cannabis‐only use sessions. Some differences in subjective responses between simultaneous use and single‐substance use sessions depended on the amounts of alcohol and cannabis used in the sessions and on typical levels of alcohol and cannabis use, which may inform tailored interventions for simultaneous use.
Boxed Text
Article notes
Coelho, S.G. , Hendershot, C.S. & Wardell, J.D. (2025) Subjective responses to simultaneous alcohol and cannabis use relative to alcohol‐only use and cannabis‐only use: An ecological momentary assessment investigation. Alcohol: Clinical and Experimental Research, 49, 923–940. Available from: 10.1111/acer.70017 PMC1201286840059034
INTRODUCTION
Young adults are increasingly engaging in simultaneous use of alcohol and cannabis (hereinafter called simultaneous use) (Lee et al., 2022; McCabe et al., 2021; Patrick et al., 2019), which refers to the use of alcohol and cannabis with overlapping effects or within the same narrow time window (Sokolovsky et al., 2020; Subbaraman & Kerr, 2015). Mounting, albeit varied, evidence suggests that young adults report greater negative consequences during simultaneous use days relative to single‐substance use days (e.g., Boyle et al., 2024; Drohan et al., 2023; Lee et al., 2022; Linden‐Carmichael et al., 2020; Sokolovsky et al., 2020; Wardell et al., 2024), underscoring the importance of understanding factors that reinforce simultaneous use in the moment to inform targeted interventions. Simultaneous use may be directly reinforced or punished through individuals' acute subjective responses to the combined effects of the substances. Indeed, much research has shown that individuals who report heightened stimulant or reward responses and blunted sedative responses to alcohol are at risk for heavier drinking (Berey et al., 2017; King et al., 2014, 2021; Morean & Corbin, 2010; Quinn & Fromme, 2011b; Wardell et al., 2015). Similarly, greater stimulant/arousal responses, positive subjective responses, and subjective high following cannabis use are related to greater cannabis use frequency and risk for cannabis use disorder (Davidson & Schenk, 1994; Fergusson et al., 2003; Le Strat et al., 2009; Treloar Padovano & Miranda Jr., 2018). Subjective responses to substance use are also important in understanding risk for harms during use; for example, heightened subjective intoxication predicts alcohol‐related harms such as impaired driving, sexual risk‐taking, and accidental injury (Stevely et al., 2020). Thus, subjective responses may be relevant for understanding simultaneous use behavior and outcomes.
Although subjective responses to each of alcohol use and cannabis use, individually, have been widely studied, relatively few studies have investigated subjective responses to simultaneous alcohol and cannabis use. Some human laboratory studies have observed greater euphoria following co‐administration of alcohol and cannabis relative to cannabis‐only administration (Lukas & Orozco, 2001), and have found alcohol to potentiate the duration of subjective responses, including euphoria (Lukas & Orozco, 2001) and feeling high, stimulated, stoned, and sedated (Hartman et al., 2016). In addition, a recent laboratory study found that participants reported greater elevations in subjective effects related to euphoria specifically following co‐administration of alcohol and cannabis relative to both alcohol‐only and cannabis‐only administration (Wickens et al., 2022). Although these studies preliminarily evince differences in subjective responses between simultaneous use and single‐substance use, the cannabis products administered in laboratory settings often differ in potency, form, and dose from the cannabis products more typically used by young adults (Yurasek et al., 2017), potentially truncating the range of subjective responses elicited. Further, the administration of alcohol and cannabis in a laboratory setting limits ecological validity—an important gap given that subjective responses are sensitive to the context of substance use events (Corbin et al., 2015, 2021; Sher, 1985).
Ecological momentary assessment (EMA) methods can assess real‐time subjective responses to substance use in naturalistic settings, complementing human laboratory studies by providing more ecologically valid data. Although EMA has been used extensively to assess subjective responses to both alcohol use and cannabis use in naturalistic settings (Acuff et al., 2024; Fridberg et al., 2021; Piasecki et al., 2012; Trela et al., 2016; Treloar Padovano & Miranda Jr, 2018), few studies to date have used EMA to compare subjective responses to simultaneous use relative to single‐substance use (for a review, see Bedillion et al., 2024), with those that have yielding mixed results (Linden‐Carmichael et al., 2020; Sokolovsky et al., 2020; Waddell et al., 2023). Sokolovsky et al. (2020) observed greater daily peak subjective intoxication, assessed via a single intoxication index (i.e., capturing feeling both ‘drunk’ and ‘high’), on simultaneous use days relative to both alcohol‐only use and cannabis‐only use days. By contrast, Linden‐Carmichael et al. (2020) did not observe significant differences between simultaneous use days and alcohol‐only use or cannabis‐only use days in subjective intoxication from alcohol (‘drunk’) or cannabis (‘high’), respectively, although daily subjective intoxication was assessed retrospectively on the next day's scheduled morning survey. More recently, Waddell et al. (2023) extended this prior work by examining momentary, in addition to daily, associations of simultaneous use with subjective responses and by distinguishing between high‐arousal (stimulation) and low‐arousal (sedation) positive and negative subjective response indices. At both the momentary and daily levels, while controlling for social context, simultaneous use was associated with greater high‐arousal positive, low‐arousal positive, and low‐arousal negative subjective responses relative to alcohol‐only use, whereas differences between simultaneous use and alcohol‐only use in high‐arousal negative subjective responses were not observed (Waddell et al., 2023). Though extending prior research by identifying momentary differences in subjective responses to simultaneous use relative to alcohol‐only use, this study did not compare simultaneous use events to cannabis‐only use events. Further, this study did not assess subjective intoxication and liking—the former an indicator of impairment and risk for substance‐use related harms (Stevely et al., 2020), and the latter an indicator of the degree to which substance effects are experienced as positive, thus with relevance to positive reinforcement of use. Consequently, there remains a need to compare simultaneous use to both alcohol‐only use and cannabis‐only use at the event level, across a range of subjective responses.
In addition, limited prior EMA research has examined whether differences in subjective responses between simultaneous use and single‐substance use events depend on the quantities of alcohol and cannabis consumed. Sokolovsky et al. (2020) observed a significant day‐level interaction between session type (simultaneous use vs. alcohol‐only use) and the number of drinks consumed in the prediction of subjective intoxication, with more pronounced differences in subjective intoxication on days involving lighter relative to heavier drinking. Perhaps because heavy drinking is already associated with high intoxication, the addition of cannabis to a heavy drinking episode may have less of an incremental effect on intoxication than the addition of cannabis to a light drinking episode. Differences in subjective responses between simultaneous use and cannabis‐only use events may also depend on the quantities of cannabis used, although this has not been examined in prior EMA studies to our knowledge. Further, whether alcohol or cannabis quantities moderate associations of simultaneous use (relative to single‐substance use) with a range of subjective response indices (along arousal [i.e., stimulant, sedative] and valence [i.e., positive, negative] dimensions) at the event level has yet to be empirically examined.
The present study
This study used EMA data to examine differences between simultaneous use sessions and both alcohol‐only use and cannabis‐only use sessions on a range of subjective responses, in addition to liking and subjective intoxication. Consistent with prior research suggesting greater reinforcement from simultaneous use relative to single‐substance use (Waddell et al., 2023), we hypothesized that simultaneous use sessions would involve greater positive subjective responses (e.g., stimulation, calmness) and greater liking relative to single‐substance use sessions. Further, given evidence suggesting elevated impairment during simultaneous use relative to single‐substance use (Lee et al., 2022; Linden‐Carmichael et al., 2020; Sokolovsky et al., 2020; Waddell et al., 2023), we hypothesized that participants would report greater subjective intoxication during simultaneous use relative to single‐substance use sessions. We also examined whether session‐level alcohol and cannabis quantities moderated session‐level differences in subjective response between simultaneous use and single‐substance use. We hypothesized that at the session level, lighter (relative to heavier) alcohol use would potentiate differences in subjective responses between simultaneous use and alcohol‐only use sessions, whereas our examination of session‐level interactions between session type and cannabis quantities was exploratory given the dearth of prior research. Further, as an exploratory aim, we tested participants' average quantities of alcohol and cannabis consumed across the EMA period as person‐level moderators of session‐level associations between simultaneous use (vs. single‐substance use) and subjective responses.
MATERIALS AND METHODS
Participants and recruitment
Data were drawn from a broader EMA study of young adult simultaneous use (see Wardell et al., 2024). The current analysis utilized data from three 21‐day EMA bursts that involved event‐contingent surveys administered during alcohol and cannabis use events. Eligible participants were young adults ages 19–25 years residing in Ontario, Canada, who reported minimum weekly use of both alcohol and cannabis in the past month, reported simultaneous use at least twice in the past month, and owned a compatible smartphone (Android or iOS). Exclusion criteria were current treatment for (or efforts to reduce) cannabis or alcohol use, monthly or more frequent illegal drug use, and exclusive medical cannabis use.
Young adults were recruited from across Ontario using fliers, online advertisements, and social media posts, which directed interested individuals to complete an online eligibility screener. An initial 177 young adults enrolled in the study; however, those who were found to be ineligible during the baseline visit (n = 9) or who completed less than 10% of the prompted EMA surveys (n = 17) were withdrawn, resulting in a sample of N = 151. Two of these participants did not provide data on any of the surveys included in analyses (see EMA Protocol and Data Analysis for types of surveys that were included). Consequently, N = 149 (59.73% women) participants comprised the analytic sample for the current study. Sample characteristics are provided in Table 1.
| n (%) | |
|---|---|
| Sex a | |
| Male | 53 (35.57) |
| Female | 95 (63.76) |
| Gender b | |
| Man | 56 (37.58) |
| Woman | 89 (59.73) |
| Transgender | 2 (1.34) |
| Nonbinary | 5 (3.36) |
| Race/ethnicity b | |
| White/Caucasian | 94 (63.09) |
| Black/African Descent/African | 13 (8.72) |
| Asian | 23 (15.44) |
| Pacific Islander | 2 (1.34) |
| Indigenous/Native North American | 1 (0.67) |
| East Indian | 14 (9.40) |
| Middle Eastern | 5 (3.36) |
| Hispanic/Latinx | 7 (4.70) |
| Not listed | 6 (4.03) |
| Prefer not to respond | 2 (1.34) |
| Income c | |
| $0–$19,999 | 32 (21.48) |
| $20,000–$49,999 | 47 (31.54) |
| $50,000–$99,999 | 34 (22.82) |
| $100,000–$159,999 | 14 (9.40) |
| $160,000+ | 20 (13.42) |
| Highest level of education | |
| Less than high school, GED, or high school diploma | 53 (35.57) |
| Some college, technical certification, or associates degree | 41 (27.52) |
| Bachelors degree | 52 (34.90) |
| Masters or doctoral degree | 3 (2.01) |
Procedures
All study procedures were approved by the research ethics boards at York University and the Centre for Addiction and Mental Health. Upon enrolling in the study, participants attended a baseline visit with a research assistant (either in person or via secure videoconferencing), during which informed consent was obtained. Participants were then assisted in installing the EMA mobile application (MetricWire, Inc., Waterloo, ON) and were oriented to the EMA protocol. The orientation included training in reporting alcohol quantities in standard drinks, using a standard drink conversion chart, and in reporting cannabis quantities in grams, using a cannabis flower infographic depicting various amounts of cannabis flower in grams from the Daily Sessions, Frequency, Age of Onset, and Quantity of Cannabis Use Inventory (DFAQ‐CU) (Cuttler & Spradlin, 2017). The baseline visit also involved interview and questionnaire assessments, the latter administered via online surveys either in‐person or remotely, immediately following the visit (see Measures below for those assessments relevant to the current analyses). The next day, participants began completing a 21‐day EMA protocol (see EMA Protocol). Study visit and EMA procedures were repeated at 6‐ and 12‐month follow‐ups.
EMA protocol
The EMA protocol included event‐contingent, daily, and randomly timed surveys, although only event‐contingent surveys assessed subjective response and were used in analyses. Participants were asked to complete a 2‐min new‐use survey to report each new session of alcohol or cannabis use, defined as alcohol or cannabis use occurring two or more hours after the most recent use of alcohol or cannabis (i.e., after a minimum 2‐h break from use); this 2‐h break between sessions was selected to align with our study's operationalization of simultaneous use as using both cannabis and alcohol within a 2‐h period. Participants were asked to complete the new‐use survey immediately after finishing their first drink or first “unit” of cannabis (e.g., joint, bowl, edible) of the session; asking participants to complete a new‐use survey after finishing their first drink during an alcohol use session was consistent with prior EMA research on subjective response to alcohol (e.g., Carpenter et al., 2017; Miranda et al., 2018; Piasecki et al., 2011; Trela et al., 2016; Treloar et al., 2015), and we extended this approach to cannabis use sessions for consistency. If participants forgot to complete a new‐use survey at the start of their session, they could complete a new‐use survey late, within 2 h of using alcohol or cannabis. Participants were sent notifications to complete brief follow‐up surveys at 60 and 120 min after submitting a new‐use survey. Each follow‐up survey remained available in the mobile application for 40 min.
When additional alcohol or cannabis use was reported in a follow‐up survey, further follow‐up surveys were administered; however, as a minority of the sessions included in analyses involved extended follow‐up surveys (25.82% of alcohol sessions, 20.14% of cannabis sessions), and to achieve a relatively standardized time period during which subjective responses were assessed across sessions, analyses focused on follow‐up surveys submitted within the initial follow‐up time period (i.e., within 175 min of the submission of the new‐use survey).
Measures
Online questionnaire
Demographic characteristics
Participants reported their age, sex assigned at birth, gender, race/ethnicity, annual household income, and highest level of education.
Alcohol use
At each burst, items from the National Institute on Alcohol Abuse and Alcoholism (NIAAA) Recommended Alcohol Screening Questions (NIAAA Task Force on Recommended Questions, 2003) were administered for descriptive purposes, including items assessing the average number of alcohol use days in the past 3 months and the average number of standard drinks per alcohol use day in the past 3 months. At baseline, participants also completed the Alcohol Use Disorder Identification Test (AUDIT) (Saunders et al., 1993), which was used to characterize levels of hazardous alcohol use in our sample.
Cannabis use
Items from the DFAQ‐CU (Cuttler & Spradlin, 2017) were administered at each burst for descriptive purposes. Participants reported the number of days in the past month during which they used cannabis (any form) and how many grams of cannabis flower they use in a typical cannabis use session, day, and week. Participants were shown images of different quantities of cannabis flower in grams alongside a one‐dollar bill to provide relative scale. At baseline, the eight‐item Cannabis Use Disorder Identification Test – Revised (CUDIT‐R) (Adamson et al., 2010) was also administered to characterize levels of hazardous cannabis use.
EMA surveys
Alcohol and cannabis use
In each new‐use survey, participants were first asked whether they had used each of alcohol and cannabis in the past 2 h, and they reported how long ago (in minutes) their substance use session began; time since beginning use of each substance was assessed using an ordinal item with six response options (0–10, 11–30, 31–60, 61–90, 91–120 min, more than 2 h), the midpoints of which were used in analyses. This allowed us to capture all uses of both alcohol and cannabis within the past 2 h (which would be considered part of the same session), even if the participant did not initiate a new‐use survey immediately upon finishing their first drink or cannabis product, as instructed. In new‐use surveys in which alcohol use was reported, participants reported the number of standard drinks they had consumed in the past 2 h. In new‐use surveys in which cannabis use was reported, participants reported the forms of cannabis used in the past 2 h on items adapted from the DFAQ‐CU (Cuttler & Spradlin, 2017); response options included marijuana (bud/leaf), concentrates (e.g., oil, wax, shatter, butane hash oil, dabs), edibles, cannabis beverages, and other. Participants reported the amount of each form of cannabis used in the past 2 h in grams of cannabis flower, the number of hits of cannabis concentrates, and the number of servings of cannabis edibles or beverages. All items from new‐use surveys (except the item assessing time since the session began) were re‐administered in follow‐up surveys; in these surveys, participants reported on their use since the last survey, rather than in the past 2 h.
Subjective response
In each new‐use and follow‐up survey, participants rated the extent to which they were currently experiencing each of four subjective responses, selected to reflect the four arousal‐valence quadrants by which subjective responses are commonly categorized (Morean et al., 2013): high‐arousal positive (“energized”), high‐arousal negative (“anxious”), low‐arousal positive (“calm”), and low‐arousal negative (“sedated”). The high‐arousal positive and low‐arousal negative items were respectively drawn from the stimulation and sedation subscales of the Brief Biphasic Alcohol Effects Scale (B‐BAES) (Rueger & King, 2013); although from a measure of subjective responses to alcohol, these items were similar to items used to assess subjective responses to cannabis in prior research (Green et al., 2009). The high‐arousal negative and low‐arousal positive items were drawn from prior research on subjective responses to cannabis (Green et al., 2009; Hunault et al., 2014), though they were expected also to apply to subjective responses to alcohol or simultaneous use. Participants were instructed to respond to each item based on how they felt “right now”, with response options ranging from 0 = not at all to 4 = extremely.
Each new‐use and follow‐up survey also included items adapted from the Drug Effects Questionnaire (DEQ) (Hamilton et al., 2011; Morean et al., 2021) and prior EMA and human laboratory research (Heinz et al., 2013; Marczinski & Fillmore, 2009; Wray et al., 2016), in which participants reported whether they liked the effects they were currently feeling, how “high” they were (in new‐use surveys in which cannabis use was reported and in all follow‐up surveys), and how “intoxicated” they were (in new‐use surveys in which alcohol use was reported and in all follow‐up surveys), with response options ranging from 0 = not at all to 4 = extremely.
Data analysis
All analyses included sessions from all three bursts (baseline, 6‐month follow‐up, 12‐month follow‐up). Thus, to account for the clustering of sessions within bursts, data were analyzed using three‐level multilevel models with sessions (level 1) nested within bursts (level 2) and bursts nested within participants (level 3). Session‐level variables were aggregated across all surveys in a session (defined as a new‐use survey and its associated follow‐up surveys). Two sets of multilevel models were constructed: one that included all alcohol use sessions and compared simultaneous use to alcohol‐only use sessions, and another that included all cannabis use sessions and compared simultaneous use to cannabis‐only use sessions. Examining alcohol use and cannabis use sessions in separate models permitted the inclusion of alcohol quantities and cannabis quantities in the models, as these were only relevant during alcohol use and cannabis use sessions, respectively, and consequently would have been redundant with the session type variable if models included all sessions (i.e., due to alcohol‐only use sessions always involving zero cannabis consumption and vice versa).
Separate models were specified with each of the following subjective responses as the dependent variable: energetic, sedated, anxious, calm, liking, intoxicated (for alcohol use sessions models), and high (for cannabis use sessions models). Subjective response items were specified as dependent variables in separate models, given that each was selected as a single‐item indicator of a domain validated in prior subjective response research (i.e., each arousal‐valence quadrant, liking, and subjective intoxication/high). In each model, the dependent variable was the peak subjective response reported across all surveys in a session; we elected to model peak subjective response, rather than average subjective response, to provide indices of maximum reinforcement or impairment that were comparable across substances and routes of administration with varying pharmacokinetic and pharmacodynamic profiles. Peak subjective response variables were treated as linear outcomes and modeled using a gaussian distribution with an identity link function. The primary independent variable in each model was session type (simultaneous use vs. alcohol‐only use or cannabis‐only use), defined according to the substances that were reported on the new‐use survey of the session. Analyses excluded sessions in which only alcohol or cannabis use was reported in the new‐use survey but use of the other substance was subsequently reported in a follow‐up survey (7.89% of sessions); thus, simultaneous use sessions included in analyses were homogeneous with respect to both alcohol and cannabis use being reported prior to all subjective response assessments during the session.
In the alcohol use sessions models, session‐level (i.e., level 1) independent variables of interest were session type (0 = alcohol‐only use, 1 = simultaneous use) and number of standard drinks consumed. At level 1, we controlled for whether the session took place on a weekend (Friday–Sunday, coded 1) versus a weekday (Monday–Thursday, coded 0); the social context of the session, or whether the participant was with others at any point during the session (coded 1) versus being alone during the entire session (coded 0); how long before completing the new use survey (in minutes) the participant reported initiating substance use; and the number of surveys (new‐use and follow‐ups) completed during the session. At level 2 (i.e., the burst level), burst timepoint (0 = baseline, 1 = 6‐month follow‐up, 2 = 12‐month follow‐up) was specified as an independent variable. At level 3 (i.e., the person level), person‐level means of all level 1 independent variables were included as independent variables. Additional level‐3 covariates were age (in years), sex (0 = male, 1 = female), and baseline CUDIT and AUDIT total scores (as these measures were not administered at all bursts). Further, to examine whether alcohol quantities moderated associations of session type with peak subjective response, models included a level 1 interaction between session type (level 1) and standard drinks (level 1), and a cross‐level interaction between session type (level 1) and person‐mean standard drinks (level 3).
In the cannabis use sessions models, level 1 independent variables of interest were session type (0 = cannabis‐only use, 1 = simultaneous use), grams of cannabis flower used, number of cannabis concentrate hits used, and number of cannabis edible/beverage servings consumed; the person‐level means of each of these variables were included as independent variables at level 3. All other independent variables were the same as those included in alcohol sessions models, except for standard drinks. To examine whether cannabis quantities moderated associations of session type with peak subjective response, models included level 1 interactions between session type and quantities of each form of cannabis (grams of cannabis flower, number of concentrate hits, number of edible/beverage servings), and cross‐level interactions between session type and person‐mean quantities of each form of cannabis.
Prior to estimating models, session‐level alcohol and cannabis quantity variables were inspected for extreme outliers, defined as values with z scores greater than 3.29 that were disconnected from the distribution (Tabachnick & Fidell, 2007), and extreme outliers were winsorized to one unit higher than the next non‐outlying value. To disaggregate variance in predictors across levels, all level 1 independent variables were person‐mean‐centered within each burst, and the burst timepoint variable at level 2 was person‐mean centered (Enders & Tofighi, 2007; Yaremych et al., 2021). We also grand‐mean centered level 3 independent variables to simplify the interpretation of parameter estimates. Models were specified with a random intercept and fixed slopes, as given the complexity of models, convergence issues arose when including random slopes. All models were fit using maximum likelihood estimation and Laplace approximation to integrate random effects using the glmmTMB package in R (Brooks et al., 2017; R Core Team, 2022).
Statistically significant interactions were probed by conditioning the association of session type with peak subjective response on high (80th percentile) and low (20th percentile) quantities of alcohol or cannabis; percentiles were used due to skewness of several person‐level quantity variables. Interaction terms that were not statistically significant suggested that the effect of session type—the lower‐order coefficient of the interaction—was not conditional on quantities of alcohol or cannabis consumed. Thus, statistically non‐significant interaction terms were trimmed from models such that the effect of session type was no longer conditioned on quantities of alcohol or cannabis consumed and could be interpreted as a main effect (Aiken & West, 1991).
RESULTS
Descriptive statistics
Of the N = 149 participants who comprised the analytic sample, n = 110 (73.83%) completed the 6‐month follow‐up, and n = 98 (65.77%) completed the 12‐month follow‐up. Participants who completed both follow‐ups (n = 90) were significantly older than participants who missed one or more follow‐ups but did not differ significantly on any other demographic or baseline alcohol or cannabis use variables. Characteristics of participants who did and did not complete both follow‐ups are summarized in Table S1. Person‐level descriptive statistics on the number of cannabis‐only use, alcohol‐only use, and simultaneous use sessions that participants contributed at each burst timepoint are shown in Table S2.
At the baseline, 6‐month follow‐up, and 12‐month follow‐up, a respective 148, 102, and 92 participants contributed one or more sessions to analyses. Across all participants and all bursts, 3431 separate sessions were included in analyses, of which 1297 (37.80%) included only a new‐use survey, 2134 (62.20%) included one or more follow‐up surveys, and 1245 (36.29%) included two or more follow‐up surveys. In 90 sessions (2.62%), a technical issue with the EMA mobile application resulted in more than the expected two follow‐up surveys being completed within the initial follow‐up window; in these cases, all follow‐up surveys submitted within the initial follow‐up window were retained for analyses to capture all available data on alcohol and cannabis use and subjective responses occurring within the 2‐h time period following the new use survey. Of the 3431 sessions included in analyses, 2242 (65.35%) involved only cannabis use, 809 (23.58%) involved only alcohol use, and 380 (11.08%) involved simultaneous use. On average, participants reported 15.05 (SD = 18.23) cannabis‐only use sessions, 5.43 (SD = 6.22) alcohol‐only use sessions, and 2.55 (SD = 3.98) simultaneous use sessions during the study; repeated measures analyses of variances indicated significant within‐person differences in the number of sessions reported across session types (F(1.18, 174.63) = 52.85, p < 0.001), with participants tending to report significantly more cannabis‐only use sessions relative to both alcohol‐only use sessions (p < 0.001) and simultaneous use sessions (p < 0.001), and significantly more alcohol‐only use sessions relative to simultaneous use sessions (p < 0.001). Person‐level and session‐level descriptive statistics on session characteristics are shown in Tables 2 and 3, respectively.
| All sessions | Cannabis‐only use sessions a | Alcohol‐only use sessions b | Simultaneous use sessions c | |
|---|---|---|---|---|
| M (SD) | M (SD) | M (SD) | M (SD) | |
| Number of simultaneous use sessions | 2.55 (3.98) | – | – | – |
| Number of cannabis‐only use sessions | 15.05 (18.23) | – | – | – |
| Number of alcohol‐only use sessions | 5.43 (6.22) | – | – | – |
| Average peak energized rating across sessions | 1.71 (0.78) | 1.46 (0.91) | 1.88 (0.87) | 2.03 (1.15) |
| Average peak sedated rating across sessions | 1.80 (0.79) | 2.03 (0.86) | 1.39 (0.89) | 1.92 (1.07) |
| Average peak anxious rating across sessions | 0.79 (0.67) | 0.84 (0.69) | 0.70 (0.72) | 0.81 (0.96) |
| Average peak calm rating across sessions | 2.67 (0.69) | 2.72 (0.73) | 2.51 (0.82) | 2.68 (0.84) |
| Average peak liking rating across sessions | 2.86 (0.60) | 2.91 (0.64) | 2.62 (0.77) | 3.00 (0.71) |
| Average peak high rating across sessions | 2.08 (0.79) d | 2.08 (0.79) | – | 2.07 (0.95) |
| Average peak intoxicated rating across sessions | 1.22 (0.81) e | – | 1.15 (0.82) | 1.52 (1.07) |
| Average grams of cannabis flower across sessions | 0.25 (0.28) d | 0.24 (0.27) | – | 0.32 (0.40) |
| Average concentrate hits across sessions | 0.82 (1.70) d | 0.81 (1.67) | – | 1.02 (3.02) |
| Average edible/beverage servings across sessions | 0.25 (0.45) d | 0.25 (0.45) | – | 0.15 (0.39) |
| Average standard drinks across sessions | 2.41 (1.15) e | – | 2.33 (1.30) | 2.56 (1.45) |
| Proportion of sessions using with other people (vs. alone) | 0.61 (0.28) | 0.48 (0.34) | 0.83 (0.25) | 0.84 (0.29) |
| Proportion of sessions on weekend days (vs. weekdays) | 0.45 (0.18) | 0.41 (0.26) | 0.5 (0.30) | 0.5 (0.36) |
| Average number of surveys across sessions | 1.91 (0.47) | 1.82 (0.53) | 2 (0.62) | 1.84 (0.70) |
| Average time between starting use and initiating new use survey (in minutes) across sessions | 31.42 (16.28) | 27.62 (18.31) | 35.32 (21.24) | 50.15 (30.84) |
| All sessions (n = 3431) | Cannabis‐only use sessions (n = 2242) | Alcohol‐only use sessions (n = 809) | Simultaneous use sessions (n = 380) | |
|---|---|---|---|---|
| M (SD) | M (SD) | M (SD) | M (SD) | |
| Peak energized rating | 1.81 (1.27) | 1.66 (1.29) | 2.00 (1.15) | 2.23 (1.23) |
| Peak sedated rating | 1.84 (1.23) | 1.97 (1.23) | 1.43 (1.16) | 1.94 (1.20) |
| Peak anxious rating | 0.78 (0.98) | 0.78 (0.95) | 0.72 (1.00) | 0.86 (1.11) |
| Peak calm rating | 2.76 (0.99) | 2.87 (0.93) | 2.50 (1.04) | 2.65 (1.06) |
| Peak liking rating | 2.93 (0.92) | 3.00 (0.90) | 2.71 (0.96) | 2.97 (0.91) |
| Peak high rating a | 2.28 (1.15) | 2.31 (1.14) | – | 2.12 (1.19) |
| Peak intoxication rating b | 1.30 (1.15) | – | 1.11 (1.09) | 1.69 (1.20) |
| Grams of cannabis flower a | 0.38 (0.51) | 0.35 (0.48) | – | 0.53 (0.63) |
| Concentrate hits a | 0.80 (2.51) | 0.80 (2.55) | – | 0.80 (2.27) |
| Edible/beverage servings a | 0.20 (0.67) | 0.21 (0.68) | – | 0.16 (0.65) |
| Standard drinks b | 2.51 (2.05) | – | 2.37 (1.92) | 2.81 (2.28) |
| Number of surveys in session | 2.02 (0.94) | 2.00 (0.94) | 2.06 (0.92) | 2.05 (1.02) |
| Time since started use (minutes) | 32.15 (30.24) | 26.56 (26.12) | 38.18 (32.03) | 52.30 (37.18) |
Models predicting peak subjective responses
Peak subjective responses during simultaneous use versus alcohol‐only use sessions
Results of multilevel models predicting peak subjective responses during alcohol use sessions are shown in Table 4. In the models predicting peak energetic, sedated, anxious, and calm ratings, all interactions were statistically non‐significant (ps ≥0.162) and were trimmed from the final models. In the final trimmed model predicting peak sedated ratings, simultaneous use was significantly associated with greater peak sedated ratings relative to alcohol‐only use at the session level. However, in the final trimmed models predicting peak energetic, anxious, and calm ratings, there were no statistically significant differences between simultaneous use and alcohol‐only use sessions in subjective responses.
| Energized | Sedated | Anxious | Calm | Liking | Intoxication | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| b | SE | p | b | SE | p | b | SE | p | b | SE | p | b | SE | p | b | SE | p | |
| Level 1 | ||||||||||||||||||
| Simultaneous use (vs. alcohol‐only use) | <0.01 | 0.08 | 0.979 | 0.42 | 0.09 | <0.001 | 0.11 | 0.07 | 0.149 | 0.08 | 0.08 | 0.301 | 0.19 | 0.07 | 0.006 | 0.44 | 0.08 | <0.001 |
| Standard drinks | 0.09 | 0.02 | <0.001 | 0.03 | 0.02 | 0.139 | 0.01 | 0.02 | 0.751 | −0.01 | 0.02 | 0.658 | 0.08 | 0.01 | <0.001 | 0.23 | 0.02 | <0.001 |
| Simultaneous use (vs. alcohol‐only use) X standard drinks | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – | −0.09 | 0.04 | 0.034 |
| Level 2 | ||||||||||||||||||
| Burst timepoint | −0.03 | 0.05 | 0.556 | 0.14 | 0.05 | 0.012 | −0.01 | 0.04 | 0.765 | 0.04 | 0.04 | 0.348 | −0.02 | 0.04 | 0.578 | 0.08 | 0.05 | 0.083 |
| Level 3 | ||||||||||||||||||
| Proportion of simultaneous use (vs. alcohol‐only use) sessions | 0.41 | 0.21 | 0.051 | 0.33 | 0.23 | 0.148 | 0.06 | 0.2 | 0.784 | 0.24 | 0.21 | 0.242 | 0.59 | 0.19 | 0.002 | 0.38 | 0.2 | 0.057 |
| Person‐mean standard drinks | 0.15 | 0.06 | 0.021 | −0.07 | 0.07 | 0.351 | −0.06 | 0.06 | 0.328 | <0.01 | 0.06 | 0.938 | 0.06 | 0.06 | 0.309 | 0.1 | 0.06 | 0.118 |
| Cross‐level interactions | ||||||||||||||||||
| Simultaneous use (vs. alcohol‐only use) X person‐mean standard drinks | – | – | – | – | – | – | – | – | – | – | – | – | −0.17 | 0.07 | 0.012 | – | – | – |
In the model predicting peak liking, the only statistically significant interaction was the cross‐level interaction between session type and person‐mean standard drinks (p = 0.010), which was retained in the final model. In the final trimmed model, simple slope analyses showed that for participants who consumed relatively high numbers of standard drinks on average, the session‐level association between session type and peak liking was not statistically significant (b = 0.05, SE = 0.08, p = 0.552), whereas for participants who consumed relatively low numbers of standard drinks on average, simultaneous use was significantly associated with greater peak liking ratings relative to alcohol‐only use at the session level (b = 0.34, SE = 0.10, p < 0.001); this interaction is shown in Figure 1. At the person level, participants with a greater proportion of simultaneous use (relative to alcohol‐only use) sessions aggregated across sessions reported significantly greater peak liking overall across sessions.
In the model predicting peak intoxication ratings, the only statistically significant interaction was the session‐level interaction between session type and standard drinks (p = 0.034), which was retained in the final trimmed model. Simple slope analyses showed that for sessions involving both high and low levels of standard drinks, simultaneous use was significantly associated with greater peak intoxication ratings relative to alcohol‐only use at the session level (high standard drinks: b = 0.38, SE = 0.08, p < 0.001; low standard drinks: b = 0.54, SE = 0.09, p < 0.001), and this association was stronger for sessions involving lower versus higher numbers of standard drinks (see Figure 2).
Peak subjective responses during simultaneous use versus cannabis‐only use sessions
Results of trimmed multilevel models predicting peak subjective responses during cannabis use sessions are shown in Table 5. In the models predicting peak calm, liking, and high ratings, all interactions were statistically non‐significant (ps ≥0.064) and were trimmed from the final models. In the final trimmed model predicting peak liking, at the session level, participants reported significantly greater peak liking during simultaneous use sessions relative to cannabis‐only use sessions, whereas at the person level, participants who had a greater proportion of simultaneous use relative to cannabis‐only use sessions tended to report significantly lower peak liking ratings overall across sessions. In the final trimmed model predicting peak high ratings, the session‐level association of session type with peak high ratings was not statistically significant, but at the person level, participants who had a greater proportion of simultaneous use relative to cannabis‐only use sessions tended to report significantly lower peak high ratings overall across sessions. In the final trimmed model predicting peak calm ratings, session‐level and person‐level associations of session type with peak calm ratings were not statistically significant.
| Energized | Sedated | Anxious | Calm | Liking | High | |||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| b | SE | p | b | SE | p | b | SE | p | b | SE | p | b | SE | p | b | SE | p | |
| Level 1 | ||||||||||||||||||
| Simultaneous use (vs. cannabis‐only use) | 0.33 | 0.06 | <0.001 | −0.1 | 0.07 | 0.116 | −0.1 | 0.06 | 0.076 | −0.08 | 0.05 | 0.105 | 0.12 | 0.05 | 0.018 | −0.07 | 0.07 | 0.281 |
| Grams of cannabis flower | −0.03 | 0.05 | 0.628 | 0.25 | 0.06 | <0.001 | −0.05 | 0.05 | 0.292 | 0.1 | 0.04 | 0.023 | 0.18 | 0.04 | <0.001 | 0.5 | 0.06 | <0.001 |
| Concentrate hits | −0.02 | 0.01 | 0.079 | 0.02 | 0.01 | 0.013 | −0.01 | 0.01 | 0.373 | <0.01 | 0.01 | 0.841 | <0.01 | 0.01 | 0.509 | 0.03 | 0.01 | 0.001 |
| Edible/beverage servings | 0.01 | 0.03 | 0.717 | 0.07 | 0.03 | 0.039 | −0.05 | 0.03 | 0.084 | 0.03 | 0.02 | 0.21 | 0.04 | 0.02 | 0.078 | 0.14 | 0.03 | <0.001 |
| Simultaneous use (vs. cannabis‐only use) X concentrate hits | 0.11 | 0.03 | 0.001 | – | – | – | – | – | – | – | – | – | – | – | – | – | – | – |
| Level 2 | ||||||||||||||||||
| Burst timepoint | −0.04 | 0.04 | 0.31 | 0.04 | 0.04 | 0.363 | −0.02 | 0.03 | 0.45 | −0.01 | 0.03 | 0.824 | −0.08 | 0.03 | 0.014 | −0.02 | 0.04 | 0.671 |
| Level 3 | ||||||||||||||||||
| Proportion of simultaneous use (vs. cannabis‐only use) sessions | 0.31 | 0.31 | 0.328 | −0.14 | 0.35 | 0.684 | 0.14 | 0.28 | 0.624 | −0.35 | 0.28 | 0.22 | −0.73 | 0.24 | 0.002 | −0.86 | 0.29 | 0.004 |
| Person‐mean grams of cannabis flower | 0.83 | 0.26 | 0.002 | −0.5 | 0.29 | 0.089 | −0.48 | 0.23 | 0.04 | 0.44 | 0.24 | 0.068 | 0.22 | 0.19 | 0.251 | 0.15 | 0.24 | 0.546 |
| Person‐mean concentrate hits | 0.03 | 0.04 | 0.518 | 0.06 | 0.05 | 0.194 | −0.03 | 0.04 | 0.44 | 0.09 | 0.04 | 0.014 | 0.08 | 0.03 | 0.009 | 0.05 | 0.04 | 0.176 |
| Person‐mean edible/beverage servings | 0.22 | 0.15 | 0.163 | −0.08 | 0.17 | 0.651 | −0.02 | 0.14 | 0.878 | 0.01 | 0.14 | 0.959 | −0.03 | 0.12 | 0.81 | 0.01 | 0.15 | 0.966 |
| Cross‐level interactions | ||||||||||||||||||
| Simultaneous use (vs. cannabis‐only use) X person‐mean grams of cannabis flower | – | – | – | 0.61 | 0.13 | <0.001 | 0.26 | 0.12 | 0.033 | – | – | – | – | – | – | – | – | – |
| Simultaneous use (vs. cannabis‐only use) X person‐mean edible/beverage servings | – | – | – | – | – | – | 0.38 | 0.15 | 0.014 | – | – | – | – | – | – | – | – | – |
In the model predicting peak energetic ratings, only the session‐level interaction between session type and concentrate hits was statistically significant (p = 0.001) and retained in the final model. In the final trimmed model, simple slopes analyses showed that for sessions involving both high and low quantities of cannabis concentrates (i.e., number of hits), simultaneous use was significantly associated with greater peak energetic ratings relative to cannabis‐only use at the session level (high concentrate hits: b = 0.33, SE = 0.06, p < 0.001; low concentrate hits: b = 0.31, SE = 0.06, p < 0.001), and this association was stronger at high relative to low levels of concentrate hits.
In the model predicting peak sedation, the cross‐level interaction between session type and person‐mean grams of cannabis flower was statistically significant (p < 0.001) and was retained in the final model. In the final trimmed model, simple slopes analyses showed that for participants who consumed relatively more grams of cannabis flower on average, the session‐level association of session type with peak sedation was not statistically significant (b = −0.04, SE = 0.06, p = 0.554), whereas for participants who consumed relatively fewer grams of cannabis flower on average, cannabis‐only use was significantly associated with greater peak sedated ratings relative to simultaneous use at the session level (b = −0.27, SE = 0.08, p = 0.001); this interaction is shown in Figure 3.
In the model predicting peak anxious ratings, the cross‐level interactions between session type and person‐mean grams of cannabis flower and between session type and person‐mean edible/beverage servings were statistically significant (ps ≤0.026) and retained in the final model. Simple slope analyses showed that for participants who consumed relatively high average quantities of cannabis flower (b = −0.08, SE = 0.06, p = 0.183) or edible/beverage servings (b = −0.08, SE = 0.06, p = 0.152), the session‐level association of session type with peak anxious ratings was not statistically significant. However, for participants who consumed relatively low average quantities of cannabis flower (b = −0.17, SE = 0.07, p = 0.018) or edible/beverage servings (b = −0.18, SE = 0.07, p = 0.009), cannabis‐only use was significantly associated with greater peak anxious ratings relative to simultaneous use at the session level. The interaction with person‐mean grams of cannabis flower is shown in Figure 4.
DISCUSSION
Subjective responses may be important risk factors for simultaneous alcohol and cannabis use and associated risk for harms. Yet, few studies have examined event‐level differences in subjective responses between simultaneous use and single‐substance use in naturalistic settings. The current EMA study extends prior research by examining differences in subjective responses between simultaneous use and both alcohol‐only and cannabis‐only sessions, and by examining quantities of alcohol and cannabis consumption (at both the session and person levels) as moderators of these session‐level differences. Results provide novel insight into the subjective responses that uniquely characterize simultaneous use events among young adults.
As hypothesized, participants reported elevated sedation and intoxication during their simultaneous use sessions relative to their alcohol‐only use sessions. These results align with EMA studies demonstrating greater low‐arousal negative subjective responses (Waddell et al., 2023) and intoxication (Sokolovsky et al., 2020) during simultaneous use relative to alcohol‐only use events or days. Elevations in sedation and intoxication during simultaneous use sessions suggest that simultaneous use may produce greater impairment relative to alcohol‐only use, perhaps contributing to increased risk for negative consequences linked to simultaneous use relative to single‐substance use (Lee et al., 2022). However, diverging from our hypotheses and from prior EMA research (Waddell et al., 2023), we generally did not find evidence to support differences between simultaneous use and alcohol‐only use sessions in positive arousal responses (i.e., energized ratings). One potential reason for the divergence of our findings from prior research may be that all young adults in our sample were required to report weekly alcohol and cannabis use as well as twice‐monthly simultaneous use, whereas young adults in Waddell et al.'s (2023) sample were only required to report one or more instances of same‐day alcohol and cannabis use during the 21‐day EMA protocol. Perhaps adding cannabis to alcohol does not increase stimulation among young adults who regularly combine alcohol and cannabis to the extent that it might among young adults who typically use alcohol on its own. However, as null findings cannot confirm the absence of effects and few EMA studies have compared positive arousal subjective responses between simultaneous use and alcohol‐only use events, further research is needed to reconcile mixed findings.
Notably, as hypothesized, elevations in intoxication during simultaneous use relative to alcohol‐only use sessions were more pronounced when comparing sessions involving lighter drinking (relative to heavier drinking). This finding replicates the day‐level interaction observed by Sokolovsky et al. (2020) at the session level, suggesting that consuming greater quantities of alcohol can attenuate within‐person elevations in intoxication during simultaneous use relative to alcohol‐only use. It is possible that during heavy alcohol consumption, changes in intoxication due to the addition of cannabis are overshadowed by the already‐high intoxication associated with heavy drinking, whereas the addition of cannabis to a light‐drinking session more noticeably increases intoxication. This finding may also shed light on previous research demonstrating more pronounced elevations in negative consequences on simultaneous use days compared to alcohol‐only use days when comparing days involving light (vs. heavy) drinking (Boyle et al., 2024), given that high intoxication may underlie risk for certain acute negative consequences (Quinn & Fromme, 2011a).
Interestingly, we found that simultaneous use sessions were associated with greater liking relative to alcohol‐only use sessions, but only among participants reporting relatively low (and not high) average levels of alcohol consumption. This finding may reflect the tendency for young adults who drink more heavily on average to report high liking of alcohol on its own (King et al., 2021; Morean & Corbin, 2010; Quinn & Fromme, 2011b), which may attenuate the effect of adding cannabis to alcohol on their subjective liking. However, inspection of Figure 1 does not show notable differences in liking ratings between those with relatively higher versus lower average levels of alcohol consumption in our sample, perhaps reflecting a restriction in the range of alcohol use in our sample—that is, our sample had low representation both of young adults reporting very infrequent alcohol use and of young adults reporting very heavy alcohol use or alcohol use disorder, and consequently, differences in acquired tolerance to the effects of alcohol at either extreme of the drinking spectrum may not be represented in the observed interactions. Thus, additional research in samples with a wider range of drinking patterns is needed to further investigate the moderating role of alcohol consumption levels on the relation between simultaneous use and subjective liking.
Extending prior EMA research on differences in subjective responses between simultaneous use and alcohol‐only use events, we also examined differences in subjective responses between simultaneous use and cannabis‐only use sessions. Consistent with hypotheses, participants reported feeling more energetic and liking effects more during simultaneous use sessions relative to cannabis‐only use sessions. These differences may in part reflect pharmacological interactions between alcohol and cannabis; indeed, select human laboratory studies have found that co‐administering alcohol with cannabis increases plasma levels of delta‐9‐tetrahydrocannabinol (THC) (Downey et al., 2013; Lukas & Orozco, 2001), and that participants report greater euphoria during alcohol‐cannabis co‐administration relative to cannabis‐only administration (Lukas & Orozco, 2001; Wickens et al., 2022). Taken together, across methodologies, simultaneous use appears to be associated with greater positive arousal and rewarding effects relative to cannabis‐only use. This pattern aligns with alcohol's well‐documented stimulant effects on the ascending limb of the blood alcohol concentration (BAC) curve, which are often experienced as rewarding (Hendler et al., 2013; Morean & Corbin, 2010). Further, that participants reported greater positive arousal and rewarding effects during simultaneous use sessions relative to cannabis‐only use sessions sheds light on young adults' frequent endorsement of using alcohol and cannabis simultaneously to “cross‐fade” (Boyle et al., 2021; Patrick et al., 2020), or to enhance the effects of alcohol or cannabis by using both substances simultaneously. Thus, young adults' simultaneous use behaviors may be positively reinforced through heightened rewarding subjective responses relative to cannabis‐only use.
At the session level, quantities of cannabis consumed generally did not moderate differences in subjective responses between simultaneous use and cannabis‐only use. Still, one statistically significant session‐level interaction emerged, wherein elevations in peak energetic ratings during simultaneous use sessions (relative to cannabis‐only use sessions) were heightened when comparing simultaneous use and cannabis‐only use sessions involving more (vs. fewer) hits of cannabis concentrates. This interaction may reflect cannabis concentrates intensifying the arousal effects of simultaneous use when used in large quantities, perhaps owing to the high potencies of many cannabis concentrate products (Spindle et al., 2019). However, this interaction should be interpreted with caution given the small number of participants in our sample who used concentrates and the imprecision of estimating cannabis concentrate quantities in hits. Thus, further research is needed to confirm the observed interaction and explore potential mechanisms.
We also found that cannabis‐only use sessions were associated with greater negative subjective responses, including sedation and anxiety, relative to simultaneous use sessions, but that these associations were stronger and were only statistically significant among participants reporting relatively low (and not high) consumption of cannabis flower on average across sessions (for both sedation and anxiety) and relatively low edible or beverage consumption on average (for anxiety). The specificity of these associations to participants reporting relatively lighter cannabis use overall may reflect lower acquired tolerance to the sedative and anxiogenic acute effects of cannabis among these participants (Colizzi & Bhattacharyya, 2018a, 2018b; LaFrance et al., 2020); that is, young adults reporting relatively lighter cannabis use may experience heightened sedation and anxiety in response to cannabis use, which may be dampened by the addition of alcohol to cannabis, given alcohol's acute stimulant and anxiolytic effects (Hendler et al., 2013; Morean & Corbin, 2010). In further support of this interpretation, Figures 3 and 4 show that in our sample, anxiety and sedation were lower across both cannabis‐only use and simultaneous use sessions among participants reporting relatively heavier average cannabis use relative to those reporting relatively lighter average cannabis use. Although speculative, these findings suggest that young adults reporting relatively lighter cannabis use could find simultaneous use reinforcing insofar as alcohol may lessen adverse subjective effects of cannabis for which they have not acquired a tolerance.
Limitations
Results of this study should be interpreted with consideration of several limitations. First, in many sessions, participants missed one or more follow‐up surveys; although we controlled for the number of follow‐up surveys completed as a session‐level covariate in all models, results may predominantly pertain to a narrow (e.g., 1‐h) time window following use. Relatedly, it is possible that participants missed new‐use surveys at times when they initiated alcohol or cannabis use, limiting inference to the substance use sessions that were captured in the EMA surveys. Another limitation is that subjective responses for a given session were first assessed upon completion of the participant's first standard drink or first “unit” of cannabis (e.g., joint, bowl, edible) for that session. Although the approach of asking participants to initiate a survey indicating the start of a new session following the first drink is consistent with prior EMA research on subjective responses to alcohol (e.g., Carpenter et al., 2017; Miranda et al., 2018; Trela et al., 2016; Treloar et al., 2015), there are some limitations in extending this approach to cannabis. For example, it is possible that peak subjective responses may have occurred prior to the point of initiating a new‐use survey, such as when using routes of cannabis administration with rapid onsets of action, and there may have been heterogeneity in the quantities of cannabis consumed by the time of initiating a new‐use survey across different forms of cannabis (e.g., a full joint of cannabis flower versus the first hit of a cannabis concentrate). Intoxication may also have been higher after finishing a whole joint or bowl of cannabis relative to a single alcoholic drink, although this would depend on factors such as acquired tolerance. Moreover, subjective responses were assessed only immediately after finishing the first drink or unit of cannabis of a session, and then 60 and 120 min later. Although we selected this frequency and duration of follow‐up assessments to minimize participant burden while still capturing initial subjective responses to substance use, subjective responses occurring at later points during extended substance use sessions were not examined. Future research should consider administering more frequent follow‐up assessments spanning a longer period after use to permit the modeling of detailed trajectories of subjective responses to alcohol‐only use, cannabis‐only use, and simultaneous use, consistent with recent high‐resolution EMA studies on subjective responses to alcohol use (e.g., Fridberg et al., 2021, 2023). For sessions involving alcohol use, it should also be noted that whether subjective responses reported were during the ascending or descending limb of the BAC curve could not be definitively ascertained.
Some limitations also pertain to the measures in our EMA surveys. First, we used a single item to assess each subjective response domain to reduce participant burden, which may have limited the reliability of assessments. In addition, several items used to assess subjective responses were drawn from measures designed only to assess subjective responses to either alcohol or cannabis (and not simultaneous use), and items selected reflect only a subset of subjective responses to these substances. Further research extending our comparisons of simultaneous use and single‐substance use sessions to a wider range of subjective responses, including those specific to simultaneous use, is needed. Moreover, estimates of cannabis quantities consumed did not consider potency, and units of consumption for cannabis concentrates (i.e., hits) and edibles or beverages (i.e., servings) were imprecise; future studies should consider assessing potency and perhaps measuring cannabis consumption in standard THC units (Freeman & Lorenzetti, 2021), which would also permit aggregation across different forms of cannabis.
Finally, the results of this study may have limited generalizability to clinical samples of young adults and to young adults who use alcohol or cannabis less frequently than required by our eligibility criteria, use illegal drugs frequently, or reside in jurisdictions where non‐medical cannabis is illegal or where the legal drinking or cannabis use age is higher than it is in Ontario, Canada. Participants in our sample also reported simultaneous use relatively less frequently than alcohol‐only and cannabis‐only use, potentially limiting generalizability to young adults who more frequently use alcohol and cannabis simultaneously. Still, that we observed many more cannabis‐only use days than alcohol‐only use or simultaneous use days may reflect the high likelihood of daily or near‐daily cannabis use among those who use cannabis frequently (Health Canada, 2024)—a pattern that differs from alcohol use, which tends to be more episodic (e.g., drinking on weekends) among young adults (Joyner et al., 2019). Although not required by our eligibility criteria, participants in our study also reported relatively light drinking, and thus replication of this research with heavier‐drinking samples is an important direction for future research.
CONCLUSIONS
In sum, the results of this study suggest that subjective responses differentiate simultaneous use sessions from both alcohol‐only use and cannabis‐only use sessions, though in different ways; generally, young adults in our study appeared to experience simultaneous use as more impairing than alcohol‐only use and as more reinforcing than cannabis‐only use. The results also indicate that differences between simultaneous use and single‐substance use sessions in some subjective responses may depend on session‐level quantities of alcohol or cannabis consumed or individual differences in average levels of alcohol or cannabis consumption. Future research should examine whether harm reduction interventions for simultaneous use among young adults may benefit from psychoeducation on risks associated with heightened subjective intoxication during simultaneous use relative to alcohol‐only use and whether there may be utility in tailoring interventions based on levels of alcohol and cannabis consumption.
FUNDING INFORMATION
This research was supported by a grant from the Canadian Institutes of Health Research (159,754; PIs: Jeffrey D. Wardell and Christian S. Hendershot). The views expressed herein do not necessarily represent the official policy of the Canadian Institutes of Health Research.
CONFLICT OF INTEREST STATEMENT
The authors declare no conflicts of interest.
Supporting information
ACKNOWLEDGEMENTS
We wish to thank Dinat Khan, Brian Periera, Marina Charalampopoulou, Korina Taguba, Dennis Padilla, and all members of the Behavioral Alcohol and Cannabis Lab at York University for their assistance with data collection.
DATA AVAILABILITY STATEMENT
The data that support the findings of this study are available from the corresponding author upon request.