Abstract
Background:
Poor sleep is common in the context of cannabis use, but experimental and daily-life studies provide mixed evidence for whether cannabis use helps or disturbs same-night sleep. Despite a high prevalence of co-use of alcohol and cannabis, most studies of cannabis use and sleep do not consider alcohol use. We sought to add to this literature by examining cannabis-sleep associations in the context of alcohol co-use and to examine these associations bidirectionally using ecological momentary assessment.
Methods:
Participants were 88 adults reporting cannabis use at least 3x/week and simultaneous alcohol and cannabis use at least 2x/week. For 14 days, participants completed morning surveys assessing hours slept and perceived sleep quality from the night before. In morning surveys, random surveys, and substance-specific surveys, participants also reported on alcohol and cannabis craving, use, and amounts.
Results:
Primary results from multilevel models demonstrated that cannabis use was not independently associated with sleep (duration or quality). However, cannabis use attenuated alcohol’s negative effects on sleep quality.
Conclusions:
Results question the utility of cannabis use to improve sleep but highlight the attenuated negative effects of alcohol as a potential reinforcer of alcohol-cannabis co-use. Future work should continue to consider polysubstance use and integrate additional self-report and objective measures of sleep health to further clarify how cannabis use affects sleep.
Keywords: sleep, alcohol, cannabis, simultaneous use, ecological momentary assessment
Introduction
Alcohol and cannabis are the most used substances among young adults in the United States (Miech et al., 2024), and 20–25% of young adults report using alcohol and cannabis together (Lee et al., 2022). Unfortunately, sleep problems are also prevalent in the United States (Wang et al., 2023), with increasing rates of clinically significant sleep disturbance among US adults (Ford et al., 2015). Sleep is among the most commonly-cited motives for cannabis use, with 50–74% of adults who use cannabis indicating that they do so specifically to help with sleep (Bachhuber et al., 2019, Metrik et al., 2018, Belendiuk et al., 2015, Altman et al., 2019). Indeed, up to 84% of those who use cannabis for sleep rate it as very or extremely helpful, and ~85% report reducing or stopping use of sleep medications as a result (Bachhuber et al., 2019). Such strong beliefs about the sleep benefits of cannabis imply day-level associations between cannabis use and sleep that could be tested empirically.
Consistent with these perceived benefits of cannabis, experimental and daily assessment studies find that use of cannabis before bedtime improves perceived sleep health (broadly defined) among those who use regularly. We use the term “sleep health” (Buysse, 2014) to characterize the varied dimensions of sleep: subjective sleep quality, daytime impairment, circadian alignment, and sleep efficiency (which encompasses sleep onset latency, wake after sleep onset, sleep duration, and time in bed). In general, those who use cannabis tend to report shorter time to sleep onset and (in some studies) less wake after sleep onset, longer sleep duration, and better sleep quality following days of cannabis use (Gorelick et al., 2013, Goodhines et al., 2019, Graupensperger et al., 2021b, Sznitman et al., 2020). Some studies also indicate more next-day fatigue and daytime naps (Gorelick et al., 2013, Goodhines et al., 2019), which might imply cannabis-induced changes in sleep architecture that lead to less restorative sleep. This is consistent with early work indicating that chronic use of cannabis leads to tolerance of its sleep-enhancing properties (Barratt et al., 1974, Freemon, 1982), as well as more recent studies indicating objective impairments in sleep during acute cannabis withdrawal (Bolla et al., 2008, Cohen-Zion et al., 2009, Vandrey et al., 2011). Among young adults with a history but no past-month use of cannabis, high-dose cannabis administration also reduced both slow-wave “deep” sleep and sleep duration (Nicholson et al., 2004), indicating potentially negative effects of high-dose cannabis on sleep architecture. However, adults who use cannabis tend to self-report benefits of cannabis on sleep at the daily (within-person) level (Gorelick et al., 2013, Goodhines et al., 2019, Graupensperger et al., 2021b, Sznitman et al., 2020).
Despite the growth of research on cannabis and sleep, most studies to date share at least two limitations. First, a growing number of adults are using cannabis in conjunction with other substances, particularly alcohol (McCabe et al., 2021); and previous studies have not consistently accounted for other substance use in statistical models. Moreover, few studies account for the potential impact of sleep on subsequent cannabis use (bidirectional associations). In retrospective longitudinal studies, poor sleep health among those who are substance-naïve has been linked to increased subsequent likelihood of cannabis use (Pasch et al., 2012, Miller et al., 2017, Wong et al., 2009), providing some evidence that poor sleep health may precede (and confer risk for) cannabis use. However, few studies examine these associations at the momentary or daily level.
Ecological momentary assessment (EMA) is particularly well-suited to characterize bidirectional, day-level associations between cannabis and alcohol use and sleep health because it minimizes recall bias and allows us to separate between- from within-person associations. Although retrospective designs are important for documenting change over time, they are limited in that participants may have difficulty recalling how long they slept, how well they slept, or what substances they used several days (or weeks or months) ago. Momentary studies circumvent this issue by asking participants about their behaviors over the past few hours. EMA also allows investigators to separate between- from within-person effects (Curran and Bauer, 2011). That is, rather than estimating the extent to which people who use cannabis frequently differ in sleep from those who use cannabis infrequently (between-person effects), intensive longitudinal designs like EMA allow investigators to determine the extent to which one’s sleep changes on nights of substance use versus nights without use (within-person effects).
Only one EMA study to date has examined associations between sleep and substance use among those reporting simultaneous use of both alcohol and cannabis (such that drug effects overlap). In one analysis of these data, Graupensperger and colleagues (2021b) found that young adults reporting simultaneous alcohol/cannabis use reported better sleep quality and fewer symptoms of insomnia (less difficulty falling/staying asleep) following days of cannabis use, but worse sleep quality and more daytime impairment following days of alcohol use. Interestingly, use of cannabis on drinking days also seemed to attenuate alcohol’s negative effects on subjective sleep quality. In a subsequent analysis exploring effects in the opposite direction (sleep effects on substance use; Graupensperger et al., 2022), they found that less time in bed (which might imply shorter sleep duration) was associated with stronger craving for – but not use of – alcohol and cannabis at the daily level. They also found more alcohol use during assessment bursts of shorter time in bed (Graupensperger et al., 2022). Together, these analyses indicate that poor sleep is linked to stronger craving for (if not use of) alcohol and cannabis, while alcohol and cannabis use interact in their effects on sleep.
This study aimed to extend previous research by examining bidirectional, day-level associations between alcohol/cannabis and two critical components of sleep health: sleep quality and sleep duration (Buysse, 2014). Consistent with our pre-registration (https://osf.io/9hj63), we conducted secondary data analysis of a 14-day EMA study within a community sample of adults reporting weekly simultaneous alcohol/cannabis use. Aim 1 tested sleep associations with next-day alcohol and cannabis variables, including craving (Aim 1a), yes/no use (Aim 1b), quantity of use (Aim 1c), and co-use (Aim 1d). We hypothesized that shorter self-reported sleep duration and worse sleep quality would be associated with higher next-day alcohol/cannabis craving, greater odds and quantity of next-day alcohol/cannabis use, and greater odds of next-day co-use (compared to only using alcohol or cannabis). Aim 2 tested the inverse associations: alcohol and cannabis use on that night’s sleep, specifically yes/no use (Aim 2a), quantity of use (Aim 2b), and co-use (Aim 2c). Consistent with previous studies (Graupensperger et al., 2021b, Sznitman et al., 2020), we hypothesized that alcohol/cannabis use (yes/no and quantity) would be associated with longer self-reported sleep duration that night, that alcohol use would be associated with worse sleep quality, and that cannabis use would be associated with better sleep quality. Moreover, we hypothesized that co-use would attenuate associations with sleep quality; specifically, cannabis use would attenuate the negative association between alcohol use and sleep quality, while alcohol use would attenuate the positive association between cannabis and sleep quality. We did not make a similar hypothesis regarding co-use predicting sleep duration compared to single use, as those models are exploratory.
Materials and Methods
Participants
Participants in the present study were all 88 adults who participated in the parent study and met the following inclusion criteria (1) between 18 and 45 years old, (2) owned a smartphone, (3) reported cannabis use at least three times per week, and (4) reported simultaneous alcohol and cannabis use, such that their effects overlap, at least twice per week. Exclusion criteria included past-month use of illicit substances other than cannabis; being pregnant or planning to become pregnant in the next month; having a history of head trauma that resulted in changes to their mood, concentration, or memory; reporting past-year physiological withdrawal symptoms for alcohol or cannabis; being in or seeking treatment for alcohol- or cannabis-related problems; and past-year unsuccessful efforts to quit or cut down on use of alcohol or cannabis. Because the parent study design involved alcohol administration, we also excluded participants if they denied drinking at least 4 drinks (for women) or 5 drinks (for men) at least once in the past six months, reported any medical contraindications for drinking alcohol, or reported a history of flushing while drinking alcohol. The resulting 88 participants were 25.22 years old on average, and most participants (79.6%) were under the age of 30. Participants were 60% female, 85% white, 8% Black or African American, 7% Hispanic or Latino, 88% single or never married, and 77% employed. Although we did not assess current college student status, over half of our sample was likely in college based on endorsement of completing “some post high school education” (including currently enrolled). Table 1 presents additional demographic details, and Table 2 presents baseline substance use information.
Table 1.
Sample characteristics (N = 88)
| Demographic and substance use categories | Mean | SD | range | n | % |
|---|---|---|---|---|---|
| Age (years) | 25.2 | 6.9 | 18 to 44 | ||
| Gender identity | |||||
| Female | 53 | 60.2 | |||
| Male | 34 | 38.6 | |||
| Nonbinary | 1 | 1.1 | |||
| Racea | |||||
| White | 75 | 85.2 | |||
| Black or African American | 7 | 8.0 | |||
| Asian or Asian American | 4 | 4.6 | |||
| Other (write in)b | 4 | 4.6 | |||
| American Indian or Alaska native | 3 | 3.4 | |||
| Native Hawaiian or other Pacific Islander | 1 | 1.1 | |||
| Ethnicity | |||||
| Hispanic or Latino | 6 | 6.8 | |||
| Marital status | |||||
| Single or never married | 77 | 87.5 | |||
| Married | 6 | 6.8 | |||
| Divorced | 3 | 3.4 | |||
| Living with someone as married | 1 | 1.1 | |||
| Separated | 1 | 1.1 | |||
| Highest level of education completed | |||||
| High school or equivalent | 12 | 13.6 | |||
| Some post high school education (includes currently enrolled) | 45 | 51.1 | |||
| Vocational or technical program | 2 | 2.3 | |||
| Undergraduate degree | 10 | 11.4 | |||
| Some post graduate education (includes currently enrolled) | 10 | 11.4 | |||
| Graduate degree | 9 | 10.2 | |||
| Annual household income | |||||
| $0 to $25,000 | 44 | 50.0 | |||
| $25,001 to $50,000 | 23 | 26.1 | |||
| $50,001 to $75,000 | 5 | 5.7 | |||
| $75,001 to $100,000 | 5 | 5.7 | |||
| Above $100,000 | 11 | 12.5 | |||
| Employed | 68 | 77.3 | |||
| Substance use | |||||
| AUDIT | 10.5 | 4.9 | 3 to 26 | ||
| CUDIT-R | 10.9 | 5.7 | 3 to 25 | ||
| Age at first alcohol use | 14.6 | 2.6 | 6 to 22 | ||
| Age at first cannabis use | 16.3 | 3.5 | 8 to 31 | ||
| Alcohol use frequency (TLFB past 30 days) | 14.2 | 6.0 | 0 to 30 | ||
| Cannabis use frequency (TLFB past 30 days) | 23.5 | 7.2 | 10 to 30 | ||
Note. SD = standard deviation. AUDIT = Alcohol Use Disorders Identification Test. CUDIT-R = Cannabis Use Disorder Identification Test—Revised. TLFB = Timeline Follow Back.
Participants could select all that apply.
Write-in responses included Hispanic or Latino (2 participants), Mexican (1 participant), and Sri Lankan (1 participant).
Table 2.
Descriptive statistics and correlations for analyzed measures (N = 88)
| Correlations | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Measure | M | SD | range | 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 |
| 1. Daily sleep duration | 7.3 | 1.8 | 0 to 14 | −.01 | .15*** | .04 | −.23*** | .07 | ||||
| 2. Daily sleep quality | 2.9 | 0.8 | 1 to 4 | .44*** | .01 | .14*** | .04 | −.14* | .04 | |||
| 3. Alcohol craving | 1.2 | 0.6 | 1 to 5 | −.04 | .02 | |||||||
| 4. Cannabis craving | 1.5 | 0.9 | 1 to 5 | .05 | −.00 | .71*** | ||||||
| 5. Alcohol use day | 0.4 | 0.5 | 0 to 1 | −.00 | .02 | .28*** | .09** | |||||
| 6. Cannabis use day | 0.6 | 0.5 | 0 to 1 | .13*** | .06* | .08** | .20*** | .37*** | ||||
| 7. Co-use day | 0.3 | 0.5 | 0 to 1 | .04 | .04 | .22*** | .13*** | -- | -- | |||
| 8. Daily drink number | 4.9 | 5.3 | 1 to 49 | −.09 | .00 | .11* | .07 | -- | .07 | .07 | ||
| 9. Daily high | 5.5 | 4.0 | 1 to 29 | .09* | .02 | −.00 | .05 | .20*** | -- | .20*** | .21*** | |
p < .05,
p < .01,
p < .001
Note. SD = standard deviation. Variables presented here are in their raw (uncentered) form. All variables are day-level variables except for alcohol and cannabis craving. Daily drink number and daily high are reported only for alcohol and cannabis use days, respectively. Because we predicted craving at the momentary level, we report their momentary-level means, standard deviations, and ranges. For correlations between craving variables and day-level variables, we aggregated craving to the day-level. All correlations are Pearson correlations except for those between alcohol use day and cannabis use day, which is a polychoric correlation. We report the correlation between alcohol and cannabis craving at the momentary level. Below the diagonal, correlations between sleep and substance use variables are for sleep and next-day substance craving and use. Above the diagonal, we report correlations between sleep and prior day substance use, given that this temporal direction is relevant to Aim 2 analyses.
Procedures
The Institutional Review Board of the University of Missouri approved all study procedures (IRB Protocol Number 2016077). We collected data between August 2020 and June 2021. Participant recruitment occurred via convenience sampling, specifically advertisements in a university mailing list and social media. Interested individuals contacted researchers by email to express interest and provide contact information. Research staff contacted individuals by phone to complete eligibility screening and provide more information about study participation. Eligible participants completed a one-hour orientation session, during which they completed self-report questionnaires, downloaded the EMA app (TigerAware; Morrison et al., 2018), and were trained on the EMA protocol. Participants were compensated $10 for their time.
Participants completed 14 days of EMA starting the day after orientation. Each day, the EMA protocol included one morning survey (available 0700–1200, with reminders at 0915 and 1130) and up to five random surveys (between 1200–2200, available for 20min with one reminder for each survey). We also instructed participants to complete a user-initiated survey any time they initiated a new episode using alcohol, cannabis, or both. After reporting substance use in any survey, participants were sent two follow-up surveys at one-hour intervals. If participants reported additional use in either follow-up survey, they continued to receive follow-ups hourly until they reported no substance use in two consecutive surveys. Follow-ups would result in cancellation of any random surveys that were previously scheduled during the same timeframe to limit participant burden. Participants were able to suspend surveys during any specified periods of time if they would be unable to use their phone (e.g., in an important meeting or while driving).
Study staff monitored compliance every weekday and contacted participants by phone to troubleshoot and encourage compliance if it fell below 80%. At the end of the 14-day EMA period, participants were scheduled for a brief Zoom meeting to debrief and coordinate compensation. Participants were compensated $80 for 80% or higher compliance overall to morning, random, and follow-up surveys. Compensation was prorated $10 for every 10% drop in compliance below 80%, with no compensation for less than 50% compliance. On average, morning report compliance was 87% (SD = 17.5, range = 25–100%), random prompt compliance was 74% (SD = 17.0, range = 12–98%), substance use follow-up compliance was 76% (SD = 19.5, range = 22–100%), and total compliance across response types was 76% (SD = 15.6, range = 20–97%). On average, participants completed 4.9 user-initiated substance use surveys (SD = 4.2, range = 0–19). Of the 88 participants included in the present analyses, 83 had ≥ 50% compliance. We conducted sensitivity analyses with the individuals who had at least 50% compliance, and the pattern of results largely stayed the same. We report analyses with all 88 participants here and note in the results section which findings differ when we restrict the sample to only those participants with 50%+ compliance.
Measures
Baseline Demographic and Substance Use Variables
Participants completed baseline questionnaires at the orientation session. Demographic variables included age in years, gender identity, race, ethnicity, marital status, education, annual household income, and employment. Participants completed the Alcohol Use Disorders Identification Test (AUDIT; Saunders et al., 1993) and the Cannabis Use Disorders Identification Test—Revised (CUDIT-R; Adamson et al., 2010) to indicate alcohol- and cannabis-related problems. Participants also reported age at first alcohol and cannabis use and alcohol and cannabis use frequency via 30-day Timeline Follow-Back (Sobell and Sobell, 1992).
EMA Sleep
In the morning report, participants indicated their sleep duration and perceived sleep quality. Sleep duration was assessed by the question, “How many hours did you sleep last night?” which participants indicated on a sliding scale from 0 to 18 hours. Sleep quality was assessed using the consensus sleep diary (Carney et al., 2012) item, “How would you rate the quality of your sleep?” Response options were (1) very bad, (2) fairly bad, (3) fairly good, and (4) very good.
EMA Substance Craving
At every type of survey, participants were asked, “In the past 15 minutes, how much have you craved alcohol?” and, “In the past 15 minutes, how much have you craved cannabis?” Response options ranged from 1 to 5, where 1 = very slightly or not at all, 2 = a little, 3 = moderately, 4 = quite a bit, and 5 = extremely. We used this item as our primary craving variable of interest. However, participants were also asked at every survey, “In the past 15 minutes, how hard was it to stop thinking about drinking alcohol?” and “In the past 15 minutes, how hard was it to stop thinking about using cannabis?” which they also rated on the same 1 to 5 scale. We conducted sensitivity analyses to determine whether the pattern of results changed if we used these craving items instead of the items reported in the main text. The pattern of results did not change, so we report results for the first set of craving items.
EMA Substance Use
At every type of survey, participants indicated whether they used alcohol, cannabis, or both, since the last survey. When participants reported cannabis use, we asked them to select their mode of use: dry leaf, concentrate, or edible. When participants reported using dry leaf cannabis, we asked them to estimate number of grams used (including fractions) and THC content (%). When participants reported using cannabis concentrates, we asked them to estimate THC content (%) and number of hits. When participants reported using edible cannabis, we asked them to estimate amount consumed in milligrams. We coded days with any alcohol use as “alcohol use days” and days with any cannabis use as “cannabis use days.”
EMA Substance Use Quantity
When participants reported alcohol use, they were asked to indicate how many standard drinks they had consumed since their last survey. When participants reported cannabis use, they were asked to indicate how high they felt “right now” on the 1 to 5 scale where 1 = very slightly or not at all, 2 = a little, 3 = moderately, 4 = quite a bit, and 5 = extremely. To create a day-level drinking amount variable, we aggregated the number of standard drinks reported within each day. To create a day-level cannabis amount variable, we took a similar approach to that by Graupensperger et al. (2021b), which assessed each morning how high participants felt when they used cannabis the day before. To capitalize on our momentary assessments of how high participants felt each time they reported cannabis use, we averaged momentary ratings of “high” within each day and then multiplied that by the number of cannabis use moments that day to create a “daily high” variable that reflects both the intensity of their high and number of instances of cannabis use each day. We ideally would have computed standard THC units (5mg THC) for each moment of use and then summed those within each day but were unable to compute standard THC units when participants used cannabis concentrates. To examine the reliability and validity of our daily high variable, we computed standard THC units for when our participants used dry leaf (flower) cannabis from participants’ self-reported % THC content and grams used, and when participants used edible cannabis from their self-reported number of milligrams used. Excluding days on which participants used cannabis concentrates, as that use would contribute to daily high but not our estimate of standard THC units, our daily high variable (which reflects both intensity of high and frequency of use) was strongly correlated with standard THC units used that day (r = .63, p < .001). This correlation was larger than the correlation between standard THC units and mean high each day (r = .24, p < .001), supporting our consideration of number of uses in addition to momentary ratings of high.
EMA Substance Co-Use
We coded days with use of both alcohol and cannabis as “co-use days.”
EMA Covariates
The EMA app recorded date and timestamps of all completed surveys. Dates were used to derive “day of the week” and “day in the study” for use as covariates. We also subtracted each day’s morning report timestamp from every survey’s timestamp from the same day to derive an “hour after wake” variable for use as a covariate in models predicting momentary outcomes.
Analytic Method
We used SAS 9.4 to test all hypotheses.
Aim 1: Sleep Associations with Next-Day Alcohol and Cannabis Craving and Use
To test sleep associations with next-day momentary alcohol and cannabis craving (Aim 1a), we used three-level multilevel models with moments nested within days and days nested within person. To test hypotheses regarding sleep associations with next-day alcohol and cannabis use (Aim 1b), next-day alcohol and cannabis use quantity (Aim 1c), and next-day co-use (Aim 1d), we used two-level multilevel models with days nested within person. Predictors of interest for all Aim 1 models (regardless of three- versus two-level models) were the prior night’s sleep duration and sleep quality (both person-mean centered). We used three-level models to predict craving despite our predictors of interest remaining at the day-level (sleep) so that we could adjust for recent alcohol and cannabis use in our prediction of alcohol and cannabis craving, while keeping their temporal precedence intact. We included person means of sleep duration and quality (both sample-mean centered) as covariates (Curran and Bauer, 2011). We used linear mixed models in proc mixed to predict alcohol craving and daily high. We re-scaled momentary alcohol and cannabis craving from 1-to-5 to 0-to-4 and then log-transformed them to reduce their skewness and kurtosis to levels appropriate for use as normally distributed continuous outcomes. The model predicting daily high was restricted to cannabis use days. We used generalized linear mixed models in proc glimmix to predict nonlinear outcomes. Specifically, we predicted alcohol day and cannabis day as binary outcome variables with logit link functions. We predicted daily drink number as a negative binomial outcome variable with a log link function, using a dataset that was restricted to alcohol use days. We predicted day-level co-use (versus alcohol only or cannabis only day) as a multinomial outcome variable with a glogit link function, using a dataset that was restricted to days with any substance use. Depending on the model, covariates included momentary alcohol and cannabis use to adjust for their potential associations with subsequent craving, hour after wake due to its potential associations with substance use and craving, same-day alcohol and cannabis use due to their potential influences on each other, sample-centered AUDIT total score, sample-centered CUDIT-R total score, age (sample-centered), gender (male = 1, female or non-binary = 0), day of the week to adjust for potential differences in substance use across days of the week, and day in study to adjust for potential reactivity over the course of the study. We specified a random intercept for person in all models and a random intercept for days within person for three-level models. We also included random slopes for day-level sleep duration and quality within person, which we retained if they demonstrated significant variability around the mean slope.
Aim 2: Alcohol and Cannabis Use Associations with Same-Night Sleep
We used two-level multilevel models in proc mixed with days nested within person to test hypotheses that alcohol and cannabis use (Aim 2a), alcohol and cannabis use quantity (Aim 2b), and co-use (Aim 2c) would predict nighttime sleep duration and quality. First, using a dataset of all study days, we ran a set of models specifying prior-day alcohol use (yes/no), cannabis use (yes/no), and their interaction (co-use) as predictors of interest, while adjusting for person-level proportions of alcohol, cannabis, and co-use days. Second, using a dataset comprising only alcohol use days, we specified prior-day drink number, prior-day co-use of cannabis (yes/no), and their interaction as predictors of interest, while adjusting for person-level daily drink number (sample-centered) and proportion of co-use days. Finally, using a dataset comprising only cannabis use days, a third set of models specified prior-day high, prior-day co-use of alcohol (yes/no), and their interaction as predictors of interest, while adjusting for person-level daily high (sample-centered) and person-level proportion of co-use days. Covariates included age (sample-centered), gender (male = 1, female or non-binary = 0), day of the week to adjust for potential differences in sleep on different days of the week, and day in study to adjust for potential reactivity over the course of the study. We specified a random intercept for person in all models. We also included random slopes for day-level predictors of interest, but none demonstrated significant variability around the mean slope, so we did not retain them in Aim 2 models.
Results
Descriptive Results
Participants completed on average 13.3 days of EMA (SD = 2.8, range = 2 to 16), with 24 participants completing 15 or 16 days due to completing user-initiated surveys on the days immediately preceding and following the 14-day protocol. We analyzed a total of 1174 days across 88 participants. Participants responded to an average of 5.4 surveys per day (SD = 1.4). On average, participants reported alcohol use on 5.2 days (SD = 3.2), cannabis use on 8.6 days (SD = 4.2), and co-use on 4.1 days (SD = 3.1). Participants reported using dry leaf cannabis at 68% of cannabis-use moments, cannabis concentrate at 26% of cannabis-use moments, and edible cannabis at 6% of cannabis-use moments. Participants reported on average 7.3 hours of sleep each night (SD = 1.8) and rated their average sleep quality as 2.9 (SD = 0.8, range = 1–4). Over one quarter (28.1%) of our participants reported a nightly average of less than 7 hours per sleep per night. There were 29 instances (2% of days) across 15 different participants where they reported less than 4 hours of sleep, and four instances (0.3% of days) across four different participants where they reported zero hours of sleep. AUDIT total scores categorized 31% of the sample at low risk consumption, 50% of the sample at hazardous or harmful consumption, and 19% of the sample as likely having moderate or severe alcohol use disorder. CUDIT-R total scores categorized 27% of the sample at hazardous use and 40% of the sample at possible cannabis use disorder. Additional descriptive statistics and correlations are reported in Tables 1–2. Power simulations by Arend & Schäfer (2019) demonstrate that we should be adequately powered to detect relatively small effects (as low as 0.18) for our least powered models, which are those restricted to days with alcohol use.
Aim 1
Sleep Predicting Next-Day Alcohol Craving and Use
Table 3 reports results from models using sleep duration and quality to predict next-day alcohol craving and use. Neither within-person sleep duration nor quality were associated with next-day alcohol craving, odds of alcohol use, or number of drinks.
Table 3.
Aim 1 models: sleep predicting next-day momentary substance craving and day-level substance use (N = 88)
| Momentary alcohol craving (n moments = 5358) |
Alcohol day (yes/no) (n days = 933) |
Alcohol drink number (n days = 365) |
Momentary cannabis craving (n moments = 5359) |
Cannabis day (yes/no) (n days = 933) |
Cannabis high (n days = 584) |
|||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Effect | b | SE | OR | 95% CI | IRR | 95% CI | b | SE | OR | 95% CI | b | SE |
| Intercept | 0.11 | .07 | 0.85 | [0.46, 1.58] | 4.43 *** | [3.25, 6.03] | 0.30 *** | 0.05 | 4.85 *** | [2.28, 10.31] | 5.99 *** | 0.52 |
| Predictors of interest | ||||||||||||
| Prior night sleep duration | −0.00 | 0.01 | 0.95 | [0.84, 1.06] | 1.04 | [0.98, 1.12] | −0.01 | 0.01 | 1.17 * | [1.02, 1.34] | −0.09 | 0.10 |
| Prior night sleep quality | 0.05 | 0.03 | 0.83 | [0.63, 1.10] | 1.00 | [0.86, 1.16] | −0.01 | 0.02 | 0.81 | [0.60, 1.11] | −0.04 | 0.24 |
| Covariates | ||||||||||||
| Betw-person sleep duration | −0.05* | 0.02 | 0.98 | [0.79, 1.21] | 0.96 | [0.87, 1.06] | 0.02 | 0.02 | 1.75 *** | [1.30, 2.36] | 0.48 * | 0.21 |
| Betw-person sleep quality | 0.03 | 0.06 | 1.90 * | [1.06, 3.43] | 0.98 | [0.75, 1.29] | 0.02 | 0.06 | 1.14 | [0.53, 2.47] | −0.15 | 0.57 |
| Last-moment alcohol use | 0.04 | 0.03 | -- | -- | -- | -- | 0.02 | 0.03 | -- | -- | -- | -- |
| Last-moment cannabis use | 0.01 | 0.03 | -- | -- | -- | -- | −0.01 | 0.02 | -- | -- | -- | -- |
| Same-day cannabis use | -- | -- | 2.85 *** | [1.96, 4.14] | 1.15 | [0.94, 1.41] | -- | -- | -- | -- | -- | -- |
| Same-day alcohol use | -- | -- | -- | -- | -- | -- | -- | -- | 3.17 *** | [2.11, 4.77] | 0.95 *** | 0.29 |
| Age | −0.01* | 0.00 | 1.03 | [1.00, 1.07] | 1.00 | [0.98, 1.02] | 0.00 | 0.00 | 1.10 *** | [1.04, 1.16] | 0.06 | 0.04 |
| Male gender | −0.07 | 0.06 | 0.99 | [0.60, 1.61] | 0.99 | [0.73, 1.33] | −0.11 | 0.06 | 0.84 | [0.42, 1.67] | 0.46 | 0.54 |
| Day of the week (Sat = ref) | ||||||||||||
| Sunday | −0.02 | 0.06 | 0.47 ** | [0.26, 0.82] | 0.94 | [0.73, 1.22] | −0.04 | 0.04 | 0.68 | [0.34, 1.35] | 0.74 | 0.47 |
| Monday | 0.05 | 0.06 | 0.31 *** | [0.17, 0.57] | 0.80 | [0.59, 1.08] | −0.06 | 0.04 | 0.65 | [0.33, 1.28] | −0.49 | 0.49 |
| Tuesday | −0.06 | 0.06 | 0.50 * | [0.29, 0.88] | 0.72 * | [0.56, 0.93] | 0.05 | 0.04 | 0.55 | [0.28, 1.06] | −0.45 | 0.47 |
| Wednesday | −0.03 | 0.06 | 0.29 *** | [0.16, 0.52] | 0.64 ** | [0.48, 0.85] | 0.06 | 0.04 | 0.54 | [0.27, 1.05] | −1.55** | 0.47 |
| Thursday | 0.01 | 0.06 | 0.53 * | [0.30, 0.94] | 0.72 * | [0.55, 0.94] | −0.01 | 0.04 | 0.35 ** | [0.18, 0.70] | −0.97* | 0.49 |
| Friday | 0.01 | 0.05 | 0.97 | [0.57, 1.67] | 0.98 | [0.79, 1.23] | 0.04 | 0.04 | 0.70 | 0.36, 1.36] | −1.44** | 0.46 |
| Hour after wake | 0.01 ** | 0.00 | -- | -- | -- | -- | 0.00 | 0.00 | -- | -- | -- | -- |
| Study day | 0.01 | 0.00 | 0.95 * | [0.92, 0.99] | 1.00 | [0.98, 1.02] | 0.00 | 0.00 | 0.93 ** | [0.89, 0.97] | −0.13*** | 0.03 |
| AUDIT total score | 0.01 | 0.01 | 1.02 | [0.97, 1.07] | 1.03 | [0.99, 1.06] | -- | -- | -- | -- | -- | -- |
| CUDIT-R total score | -- | -- | -- | -- | -- | -- | 0.00 | 0.00 | 1.21 *** | [1.13, 1.29] | 0.12 * | 0.05 |
Note. SE = standard error. OR = odds ratio. CI = confidence interval. IRR = incidence rate ratio. Significant effects at p < .05 are bolded. -- not in the model. Craving models are three-level models. The only momentary-level predictors in the craving models are covariates: last-moment alcohol and cannabis use and hour after wake. All other predictors and covariates in the craving models are day- and person-level variables, identical to those in the rest of the (two-level) models. Depending on the model, some participants had no variability on the outcome variable: 10 for momentary alcohol craving, 4 for alcohol day, 3 for alcohol drink number, 5 for momentary cannabis craving, 14 for cannabis day, and 2 for daily cannabis high. A sensitivity analysis for the cannabis day model removing the 14 participants with no variability (due to 13/14 of them reporting cannabis use every day) resulted in the same pattern of findings with almost identical effect sizes as reported here, so we left all participants in the model.
p < .05,
p < .01,
p < .001
At the between-person level, a one-unit increase in average hours slept during the study was associated with a small decrease in momentary alcohol craving on any given day (−0.05 on a five-item scale). In addition, a one-unit increase in average sleep quality throughout the study (on the four-item scale) was associated with almost twice the odds of reporting alcohol use on any given day.
Sleep Predicting Next-Day Cannabis Craving and Use
Table 3 reports results from models using sleep duration and quality to predict next-day cannabis craving and use. Neither sleep duration nor quality were associated with next-day cannabis craving. Sleep duration was associated with cannabis use at the within-person level, such that a one-hour increase in hours slept was associated with 17% higher odds of cannabis use. Sleep duration was not related to daily high. Sleep quality was not related to next-day cannabis use or daily high within persons.
At the between-person level, participants reporting one more hour of sleep throughout the study had 75% higher odds of cannabis use on any given day, and a 0.48 increase in daily high on cannabis use days.
Sleep Predicting Next-Day Co-use of Alcohol and Cannabis
Table 4 reports results from models using sleep duration and quality to predict next-day co-use of alcohol and cannabis. Neither sleep duration nor sleep quality were associated with co-use at the within-person level.
Table 4.
Aim 1 models: sleep predicting next-day substance co-use (N = 88, n days = 672)
| Predicting co-use (compared to using alcohol alone) | Predicting co-use (compared to using cannabis alone) | |||
|---|---|---|---|---|
| Effect | OR | 95% CI | OR | 95% CI |
| Intercept | 5.64 *** | [2.36, 13.49] | 1.63 | [0.88, 3.01] |
| Predictors of interest | ||||
| Prior night sleep duration | 1.09 | [0.87, 1.36] | 0.95 | [0.82, 1.10] |
| Prior night sleep quality | 0.64 | [0.38, 1.09] | 0.74 | [0.52, 1.10] |
| Covariates | ||||
| Person-level sleep duration | 1.52 ** | [1.17, 1.98] | 0.96 | [0.76, 1.21] |
| Person-level sleep quality | 1.25 | [0.63, 2.51] | 2.02 * | [1.10, 3.71] |
| Age | 1.11 *** | [1.05, 1.18] | 1.05 * | [1.01, 1.09] |
| Male gender | 0.91 | [0.47, 1.77] | 1.04 | [0.61, 1.76] |
| Day of the week (Saturday = reference) | ||||
| Sunday | 1.11 | [0.38, 3.23] | 0.61 | [0.31, 1.20] |
| Monday | 0.48 | [0.16, 1.38] | 0.28 *** | [0.14, 0.58] |
| Tuesday | 0.58 | [0.22, 1.51] | 0.48 * | [0.25, 0.94] |
| Wednesday | 0.84 | [0.28, 2.45] | 0.34 ** | [0.17, 0.68] |
| Thursday | 0.36 * | [0.14, 0.94] | 0.49 * | [0.24, 0.98] |
| Friday | 1.05 | [0.43, 2.57] | 1.17 | [0.61, 2.25] |
| Study day | 1.00 | [0.93, 1.07] | 0.99 | [0.94, 1.03] |
| AUDIT total score | 0.94 | [0.87, 1.01] | 1.01 | [0.95, 1.07] |
| CUDIT-R total score | 1.22 *** | [1.14, 1.31] | 1.01 | [0.97, 1.06] |
Note. The outcome variable, co-use, was a multinomial variable with three levels: co-use, alcohol use only, or cannabis use only. Thus, the columns reflect estimates from the same model predicting co-use compared to alcohol use alone or cannabis use alone. OR = odds ratio. CI = confidence interval. Significant effects at p < .05 are bolded. Three participants had no variability on the outcome variable, co-use.
p < .05,
p < .01,
p < .001
At the between person level, participants reporting one more hour of sleep throughout the study had 52% higher odds of co-use compared to alcohol use only on any given day. Sleep quality was associated with co-use as well, such that participants reporting better sleep on average (by one point on the four-point scale) were twice as likely to report co-use compared to cannabis use alone on any given day.
Aim 2
Alcohol Use Predicting Same-Night Sleep
Table 5 reports results from models using alcohol use to predict same-night sleep duration and quality. In contrast to our hypotheses, alcohol use and quantity were not associated with sleep duration that night. Consistent with our hypotheses, however, alcohol-only days (compared to days with no substance use) were associated with a decrease in sleep quality that night (a third of a point lower sleep quality on a four-point scale). Similarly, consuming one extra drink on alcohol-only days was associated with a very small decrease in sleep quality that night (0.07 decrease on a four-point scale).
Table 5.
Aim 2 results: substance use predicting same-night sleep (N = 88)
| Use (Yes/No) Models | Drinking Amount Models | Cannabis Amount Models | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Sleep duration (n days = 876) |
Sleep quality (n days = 876) |
Sleep duration (n days = 340) |
Sleep quality (n days = 340) |
Sleep duration (n days = 540) |
Sleep quality (n days = 539) |
|||||||
| Effect | b | SE | b | SE | b | SE | b | SE | b | SE | b | SE |
| Intercept | 6.53 *** | 0.40 | 2.62 *** | 0.16 | 6.01 *** | 0.41 | 2.42 *** | 0.16 | 7.45 *** | 0.36 | 2.96 *** | 0.13 |
| Predictors of interest | ||||||||||||
| Alcohol day | −0.37 | 0.21 | −0.34*** | 0.09 | -- | -- | -- | -- | −0.25 | 0.15 | −0.04 | 0.06 |
| Cannabis day | 0.15 | 0.16 | 0.07 | 0.07 | 0.25 | 0.26 | 0.34 ** | 0.11 | -- | -- | -- | -- |
| Alcohol day × cannabis day | 0.16 | 0.25 | 0.28 ** | 0.11 | -- | -- | -- | -- | -- | -- | -- | -- |
| Number of drinks | -- | -- | -- | -- | −0.08 | 0.05 | −0.07** | 0.02 | -- | -- | -- | -- |
| Number of drinks × cannabis day | -- | -- | -- | -- | 0.03 | 0.06 | 0.06 * | 0.03 | -- | -- | -- | -- |
| Cannabis high | -- | -- | -- | -- | -- | -- | -- | -- | −0.00 | 0.03 | 0.00 | 0.01 |
| Cannabis high × alcohol day | -- | -- | -- | -- | -- | -- | -- | -- | −0.03 | 0.04 | −0.01 | 0.02 |
| Covariates | ||||||||||||
| Betw-person alcohol day proportion | 0.03 | 0.60 | 0.62 ** | 0.23 | -- | -- | -- | -- | −0.21 | 0.61 | 0.12 | 0.20 |
| Betw-person cannabis day proportion | 0.96 | 0.49 | 0.06 | 0.19 | 1.25 ** | 0.45 | 0.22 | 0.18 | -- | -- | -- | -- |
| Betw-person number of drinks | -- | -- | -- | -- | −0.13** | 0.04 | −0.02 | 0.02 | -- | -- | -- | -- |
| Betw-person cannabis high | -- | -- | -- | -- | -- | -- | -- | -- | 0.02 | 0.05 | 0.01 | 0.02 |
| Age | 0.01 | 0.02 | 0.00 | 0.01 | 0.00 | 0.02 | 0.01 | 0.01 | 0.03 | 0.02 | 0.01 | 0.01 |
| Male gender | 0.10 | 0.25 | 0.12 | 0.10 | −0.22 | 0.28 | 0.12 | 0.11 | −0.08 | 0.30 | 0.05 | 0.10 |
| Day of the week (Saturday = reference) | ||||||||||||
| Sunday | 0.04 | 0.20 | 0.02 | 0.08 | 0.06 | 0.28 | −0.15 | 0.12 | −0.09 | 0.23 | −0.06 | 0.10 |
| Monday | 0.21 | 0.20 | −0.01 | 0.08 | 0.29 | 0.33 | −0.05 | 0.14 | 0.03 | 0.24 | −0.10 | 0.10 |
| Tuesday | −0.10 | 0.20 | 0.02 | 0.09 | −0.50 | 0.34 | −0.03 | 0.14 | −0.51* | 0.24 | −0.03 | 0.10 |
| Wednesday | 0.18 | 0.20 | 0.12 | 0.09 | 0.36 | 0.32 | 0.09 | 0.13 | −0.08 | 0.25 | −0.02 | 0.10 |
| Thursday | 0.17 | 0.20 | −0.02 | 0.08 | 0.15 | 0.34 | −0.09 | 0.14 | 0.03 | 0.24 | −0.03 | 0.10 |
| Friday | −0.24 | 0.20 | 0.00 | 0.08 | 0.01 | 0.32 | 0.04 | 0.13 | −0.46 | 0.25 | −0.19 | 0.10 |
| Study day | 0.00 | 0.01 | −0.01 | 0.01 | 0.02 | 0.02 | 0.00 | 0.01 | 0.03 | 0.02 | −0.00 | 0.01 |
Note. SE = standard error. Significant effects at p < .05 are bolded. -- not in the model. One participant had no variability on sleep duration, and 10 participants had no variability on sleep quality.
p < .05,
p < .01,
p < .001
At the between-person level, participants who consumed alcohol daily (compared to never) during the study reported better sleep quality (0.62 higher on a four-point scale), and those reporting an average of one more drink on drinking days reported ~8 minutes shorter sleep duration.
Cannabis Use Predicting Same-Night Sleep
Table 5 reports results from models using cannabis use to predict same-night sleep duration and quality. Using cannabis (compared to not) was not associated with sleep duration or sleep quality that night. Using cannabis on drinking days was not associated with sleep duration but was associated with a third of a point improvement in sleep quality on a four-point scale compared to not using cannabis on drinking days. Daily high was not significantly associated with sleep duration or quality that night.
At the between-person level, participants who consumed cannabis daily (compared to never) reported approximately an hour and a half longer sleep duration.
Co-Use of Alcohol and Cannabis Predicting Same-Night Sleep
Table 5 reports results from models interacting alcohol and cannabis use (representing co-use) to predict same-night sleep duration and quality. Cannabis use attenuated the association between alcohol use and sleep quality, such that the effect of alcohol use on sleep quality was not significant on cannabis use days (simple slope for alcohol use on cannabis use days: b = −0.06, SE = 0.06, p = .304). Cannabis use also attenuated the association between number of drinks and nighttime sleep quality, such that it was not significant on co-use days (simple slope for number of drinks on cannabis use days: b = −0.01, SE = 0.01, p = .480). However, the interaction of cannabis use and daily drink number was no longer significant (b = 0.04, SE = 0.03, p = .128) in a sensitivity model restricting the sample to only participants with at least 50% EMA compliance; though the main effects of daily drink number (b = −0.07, SE = 0.02, p = .003) and cannabis use (b = 0.33, SE = 0.11, p = .003) remained significant. Alcohol use did not interact with daily high to predict sleep duration or quality.
Discussion
We sought to add to a growing body of literature examining associations between cannabis use and sleep by studying a sample of individuals not only with frequent cannabis use but also regular simultaneous use of cannabis and alcohol. Given the prevalence of alcohol-cannabis co-use (McCabe et al., 2021), better understanding how cannabis use and sleep are related in the context of alcohol co-use may clarify treatment considerations for individuals who use both substances. We used EMA to examine these associations bidirectionally, given evidence that substance use and sleep both confer risk to each other. Results partially supported our hypotheses, which we derived based on previous work (Graupensperger et al., 2022, Graupensperger et al., 2021b, Sznitman et al., 2023).
Consistent with our hypotheses, we found that alcohol use on a given day and greater number of drinks on drinking days were associated with worse sleep quality that night. This finding is in line with work demonstrating the negative associations between alcohol use and sleep quality among healthy adults (Graupensperger et al., 2021a, Lydon et al., 2016, Geoghegan et al., 2012, Galambos et al., 2009, Patrick et al., 2018) and extends to our sample of individuals with frequent cannabis use as well. We also found that co-using cannabis on drinking days attenuated those associations, such that the negative associations of alcohol use with sleep quality were essentially cancelled out on cannabis use days. Cannabis use attenuating alcohol’s negative associations with sleep replicates and extends findings by Graupensperger et al. (2021b) and Sznitman et al. (2023). However, our data lie somewhat counter to the expectations of many individuals that cannabis use helps their sleep (Altman et al., 2019, Bachhuber et al., 2019), as cannabis only buffered the deleterious associations of alcohol and was not linked independently to better sleep quality or longer sleep duration. Extending research on cannabis use and sleep to the context of its co-use with alcohol provides a fuller characterization of cannabis’s relationship to sleep. That is, the possibility that cannabis may mitigate alcohol’s negative associations with sleep could reinforce and maintain cannabis co-use among people who drink alcohol and notice worse sleep quality after drinking. This interpretation aligns with qualitative work identifying improved sleep as a positive perceived consequence of using alcohol and cannabis simultaneously as opposed to only using alcohol (Boyle et al., 2021).
In contrast to our hypotheses and some prior work (Sznitman et al., 2023), alcohol and cannabis use were not related to sleep duration. Also in contrast to our hypotheses, longer (not shorter) sleep duration predicted increased odds of next-day cannabis use. As noted above, studies examining associations between sleep and next-day substance use are rare, and some have found no significant associations (Graupensperger et al., 2022, Galambos et al., 2009). One study found that longer sleep duration was linked to lower next-day drinking quantity; however, this association was not significant in multivariate models, and better sleep quality was linked to more (not less) drinking in their sample (Fucito et al., 2018). Similarly, two studies in young adults found heavier drinking after nights of longer sleep duration (Miller et al., 2021) and less drinking on days of higher fatigue (Hamilton et al., 2023). Based on these studies, we speculate that longer sleep duration may lead to less fatigue and more positive mood/enhancement motives (Tomaso et al., 2021, Rea et al., 2022, Wescott et al., 2023, Kenney et al., 2013), which may in turn be linked to alcohol or, as in our study, cannabis use (Dora et al., 2023, Bonar et al., 2017, Stevenson et al., 2023).
This study had several limitations. Because this study was a secondary data analysis, our assessments of sleep included only sleep quality and sleep duration. As noted above, sleep is a multidimensional construct (Buysse, 2014), and different components of sleep (e.g., quality vs quantity vs timing) may have different associations with substance use outcomes. Additional self-reported measures of sleep such as sleep onset latency, number of nighttime awakenings, early awakenings, and next-day fatigue attributed to sleep are encouraged in future research to clarify how cannabis use acts on processes linking alcohol and sleep. Further, physiological measures of sleep health using actigraphy, for example, provide objective sleep data that are not fully captured by self-report (Aili et al., 2017). We did not recruit participants on the basis of experiencing sleep problems. Although prior studies of alcohol, cannabis, and sleep health also used relatively healthy sleep samples (Graupensperger et al., 2021b, 2022, Sznitman et al., 2023), associations between sleep and substance use may differ among those with sleep and substance use disorders (Miller et al., 2021, Roehrs et al., 1999, Brower and Hall, 2001), in which case findings may not generalize to these groups. Acute and chronic cannabis use also seem to confer different risks for pathophysiology (Cservenka et al., 2018), so assessing history of cannabis use may also improve understanding of the point along the cannabis use trajectory that specific associations emerge between sleep, cannabis use, and its co-use with alcohol.
The generalizability of our results is also limited by the demographic makeup of our sample. Participants in our study were predominantly non-Hispanic white, single, young adults, with many participants likely in college. Although our participants’ AUDIT and CUDIT-R scores demonstrate a range of alcohol and cannabis use severity, we excluded participants who experienced past-year physiological withdrawal symptoms from alcohol or cannabis, who were in or seeking alcohol or cannabis use treatment, and who had past-year unsuccessful efforts to cut down or quit their use of either substance. Additional work is needed to examine associations between sleep and alcohol and cannabis (co-)use among more diverse samples and among individuals with higher severity of substance use and related problems.
We used EMA to examine the bidirectional associations between sleep and cannabis, alcohol, and their co-use. Whereas much of what we know about cannabis associations with sleep come from studies of cannabis use alone, many people who use cannabis also use alcohol. Primary findings included that cannabis use did not predict better sleep in the absence of alcohol use, which contrasts individuals’ belief that cannabis use is beneficial for sleep. However, cannabis use attenuated the negative associations that alcohol use had with sleep quality, highlighting mitigated sleep problems as one potential contributor or reinforcer or patterns of co-use of cannabis among people who consume alcohol. Further research on alcohol-cannabis co-use will improve our understanding of how sleep relates to substance use processes among people who use multiple substances.
Figure 1.
Cannabis co-use attenuates the association between alcohol use and worse sleep quality
Figure 2.
On drinking days, cannabis co-use attenuates the association between total daily drink number and worse sleep quality
Acknowledgments
The National Institutes of Health supported the researchers’ effort (F31 AA027958, PI: Wycoff; K23 AA026895, PI: Miller).
Footnotes
Conflict of interest: Timothy J. Trull is a co-founder of TigerAware LLC which developed the software described in this study. He receives no compensation or royalties for the use of the software.
References
- AILI K, ÅSTRÖM-PAULSSON S, STOETZER U, SVARTENGREN M & HILLERT L 2017. Reliability of actigraphy and subjective sleep measurements in adults: The design of sleep assessments. Journal of Clinical Sleep Medicine, 13, 39–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- ALTMAN BR, MIAN MN, SLAVIN M & EARLEYWINE M 2019. Cannabis expectancies for sleep. Journal of Psychoactive Drugs, 51, 405–412. [DOI] [PMC free article] [PubMed] [Google Scholar]
- AREND MG & SCHÄFER T 2019. Statistical power in two-level models: A tutorial based on Monte Carlo simulation. Psychological methods, 24, 1–19. [DOI] [PubMed] [Google Scholar]
- BACHHUBER M, ARNSTEN JH & WURM G 2019. Use of cannabis to relieve pain and promote sleep by customers at an adult use dispensary. Journal of Psychoactive Drugs, 51, 400–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
- BARRATT ES, BEAVER W & WHITE R 1974. The effects of marijuana on human sleep pattersn. Biological Psychiatry, 8, 47–54. [PubMed] [Google Scholar]
- BELENDIUK KA, BABSON KA, VANDREY R & BONN-MILLER MO 2015. Cannabis species and cannabinoid concentration preference among sleep-disturbed medicinal cannabis users. Addictive Behaviors, 50, 178–181. [DOI] [PubMed] [Google Scholar]
- BOLLA KI, LESAGE SR, GAMALDO CE, NEUBAUER DN, FUNDERBURK FR, CADET JL, DAVID PM, VERDEJO-GARCIA A & BENBROOK AR 2008. Sleep disturbance in heavy marijuana users. Sleep: Journal of Sleep and Sleep Disorders Research, 31, 901–908. [DOI] [PMC free article] [PubMed] [Google Scholar]
- BONAR EE, GOLDSTICK JE, COLLINS RL, CRANFORD JA, CUNNINGHAM RM, CHERMACK ST, BLOW FC & WALTON MA 2017. Daily associations between cannabis motives and consumption in emerging adults. Drug and Alcohol Dependence, 178, 136–142. [DOI] [PMC free article] [PubMed] [Google Scholar]
- BOYLE HK, GUNN RL, LOPEZ G, FOX OS & MERRILL JE 2021. Qualitative examination of simultaneous alcohol and cannabis use reasons, evaluations, and patterns among heavy drinking young adults. Psychology of Addictive Behaviors, 35, 638–649. [DOI] [PMC free article] [PubMed] [Google Scholar]
- BROWER KJ & HALL JM 2001. Effects of age and alcoholism on sleep: A controlled study. Journal of Studies on Alcohol, 62, 335–343. [DOI] [PMC free article] [PubMed] [Google Scholar]
- BUYSSE DJ 2014. Sleep health: Can we define it? Does it matter? SLEEP, 37, 9–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- CARNEY CE, BUYSSE DJ, ANCOLI-ISRAEL S, EDINGER JD, KRYSTAL AD, LICHSTEIN KL & MORIN CM 2012. The consensus sleep diary: Standardizing prospective sleep self-monitoring. Sleep: Journal of Sleep and Sleep Disorders Research, 35, 287–302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- COHEN-ZION M, DRUMMOND SPA, PADULA CB, WINWARD J, KANADY J, MEDINA KL & TAPERT SF 2009. Sleep architecture in adolescent marijuana and alcohol users during acute and extended abstinence. Addictive Behaviors, 34, 976–979. [DOI] [PMC free article] [PubMed] [Google Scholar]
- CSERVENKA A, LAHANAS S & DOTSON-BOSSERT J 2018. Marijuana use and hypothalamic-pituitary-adrenal axis functioning in humans. Frontiers in Psychiatry, 9, 1–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- CURRAN PJ & BAUER DJ 2011. The disaggregation of within-person and between-person effects in longitudinal models of change. Annual Review of Psychology, 62, 583–619. [DOI] [PMC free article] [PubMed] [Google Scholar]
- DORA J, PICCIRILLO M, FOSTER KT, ARBEAU K, ARMELI S, AURIACOMBE M, BARTHOLOW BD, BELTZ AM, BLUMENSTOCK SM, BOLD K, BONAR EE & AL. E 2023. The daily association between affect and alcohol use: A meta-analysis of individual participant data. Psychological Bulletin, 149, 1–24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- FORD ES, CUNNINGHAM TJ, GILES WH & CROFT JB 2015. Trends in insomnia and excessive daytime sleepiness among US adults from 2002 to 2012. Sleep Medicine, 16, 372–378. [DOI] [PMC free article] [PubMed] [Google Scholar]
- FREEMON FR 1982. The effect of chronically administered delta-9-tetrahydrocannabinol upon the polygraphically monitored sleep of normal volunteers. Drug and Alcohol Dependence, 10, 345–353. [DOI] [PubMed] [Google Scholar]
- FUCITO LM, BOLD KW, VAN REEN E, REDEKER NS, O’MALLEY SS, HANRAHAN TH & DEMARTINI KS 2018. Reciprocal variations in sleep and drinking over time among heavy-drinking young adults. Journal of Abnormal Psychology, 127, 92–103. [DOI] [PMC free article] [PubMed] [Google Scholar]
- GALAMBOS NL, DALTON AL & MAGGS JL 2009. Losing sleep over it: Daily variation in sleep quantity and quality in Canadian students’ first semester of university. Journal of Research on Adolescence, 19, 741–761. [Google Scholar]
- GEOGHEGAN P, O’DONOVAN MT & LAWLOR BA 2012. Investigation of the effects of alcohol on sleep using actigraphy. Alcohol and Alcoholism, 47, 538–544. [DOI] [PubMed] [Google Scholar]
- GOODHINES PA, GELLIS LA, ANSELL EB & PARK A 2019. Cannabis and alcohol use for sleep aid: A daily diary investigation. Health Psychology, 38, 1036–1047. [DOI] [PMC free article] [PubMed] [Google Scholar]
- GORELICK DA, GOODWIN RS, SCHWILKE E, SCHROEDER JR, SCHWOPE DM, KELLY DL, ORTEMANN-RENON C, BONNET D & HUESTIS MA 2013. Around-the-clock oral THC effects on sleep in male chronic daily cannabis smokers. Am J Addict, 22, 510–4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- GRAUPENSPERGER S, CADIGAN JM, EINBERGER C & LEE CM 2021a. Multifaceted covid-19-related stressors and associations with indices of mental health, well-being, and substance use among young adults. International Journal of Mental Health and Addiction. [DOI] [PMC free article] [PubMed] [Google Scholar]
- GRAUPENSPERGER S, FAIRLIE AM, RAMIREZ JJ, CALHOUN BH, PATRICK ME & LEE CM 2022. Daily-level associations between sleep duration and next-day alcohol and cannabis craving and use in young adults. Addictive Behaviors, 132. [DOI] [PMC free article] [PubMed] [Google Scholar]
- GRAUPENSPERGER S, FAIRLIE AM, VITIELLO MV, KILMER JR, LARIMER ME, PATRICK ME & LEE CM 2021b. Daily-level effects of alcohol, marijuana, and simultaneous use on young adults’ perceived sleep health. Sleep: Journal of Sleep and Sleep Disorders Research, 44, 1–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- HAMILTON AR, ARMELI S & TENNEN H 2023. Too tired to drink? Daily associations of sleep duration and fatigue with own and others’ alcohol consumption. Psychology of Addictive Behaviors, 37, 267–274. [DOI] [PMC free article] [PubMed] [Google Scholar]
- KENNEY SR, LAC A, LABRIE JW, HUMMER JF & PHAM A 2013. Mental health, sleep quality, drinking motives, and alcohol-related consequences: A path-analytic model. Journal of Studies on Alcohol and Drugs, 74, 841–851. [DOI] [PMC free article] [PubMed] [Google Scholar]
- LEE CM, CALHOUN BH, ABDALLAH DA, BLAYNEY JA, SCHULTZ NR, BRUNNER M & PATRICK ME 2022. Simultaneous alcohol and marijuana use among young adults: A scoping review of prevalence, patterns, psychosocial correlates, and consequences. Alcohol Research: Current Reviews, 42. [DOI] [PMC free article] [PubMed] [Google Scholar]
- LYDON DM, RAM N, CONROY DE, PINCUS AL, GEIER CF & MAGGS JL 2016. The within-person association between alcohol use and sleep duration and quality in situ: An experience sampling study. Addictive Behaviors, 61, 68–73. [DOI] [PMC free article] [PubMed] [Google Scholar]
- MCCABE SE, ARTERBERRY BJ, DICKINSON K, EVANS-POLCE RJ, FORD JA, RYAN JE & SCHEPIS TS 2021. Assessment of changes in alcohol and marijuana abstinence, co-use, and use disorders among US young adults from 2002 to 2018. JAMA Pediatrics, 175. [DOI] [PMC free article] [PubMed] [Google Scholar]
- METRIK J, BASSETT SS, ASTON ER, JACKSON KM & BORSARI B 2018. Medicinal versus recreational cannabis use among returning veterans. Translational Issues in Psychological Science, 4, 6–20. [DOI] [PMC free article] [PubMed] [Google Scholar]
- MIECH RA, JOHNSTON LD, PATRICK ME, O’MALLEY PM & BACHMAN JG 2024. Monitoring the Future national survey results on drug use, 1975–2023: Overview and detailed results for secondary school students. Monitoring the Future Monograph Series. Ann Arbor, MI: Institute for Social Research, University of Michigan. [Google Scholar]
- MILLER MB, FREEMAN LK, DEROCHE CB, PARK CJ, HALL NA & MCCRAE CS 2021. Sleep and alcohol use among young adult drinkers with insomnia: A daily process model. Addictive Behaviors, 119, 1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
- MILLER MB, JANSSEN T & JACKSON KM 2017. The prospective association between sleep and initiation of substance use in young adolescents. Journal of Adolescent Health, 60, 154–160. [DOI] [PMC free article] [PubMed] [Google Scholar]
- MORRISON W, GUERDAN L, KANUGO J, SHANG Y & TRULL T 2018. TigerAware: An innovative mobile survey and sensor data collection and analytics system. International Conference on Data Science in Cyberspace, 115–122. [Google Scholar]
- NICHOLSON AN, TURNER C, STONE BM & ROBSON PJ 2004. Effect of Δ−9-Tetrahydrocannabinol and Cannabidiol on Nocturnal Sleep and Early-Morning Behavior in Young Adults. Journal of Clinical Psychopharmacology, 24, 305–313. [DOI] [PubMed] [Google Scholar]
- PASCH KE, LATIMER LA, CANCE JD, MOE SG & LYTLE LA 2012. Longitudinal bi-directional relationships between sleep and youth substance use. Journal of Youth and Adolescence, 41, 1184–1196. [DOI] [PMC free article] [PubMed] [Google Scholar]
- PATRICK ME, GRIFFIN J, HUNTLEY ED & MAGGS JL 2018. Energy drinks and binge drinking predict college students’ sleep quantity, quality, and tiredness. Behavioral Sleep Medicine, 16, 92–105. [DOI] [PMC free article] [PubMed] [Google Scholar]
- REA EM, DECARLO SANTIAGO C, NICHOLSON L, HEARD EGBERT A & BOHNERT AM 2022. Sleep, affect, and emotion reactivity in first-year college students: A daily diary study. International Journal of Behavioral Medicine. [DOI] [PubMed] [Google Scholar]
- ROEHRS T, PAPINEAU K, ROSENTHAL L & ROTH T 1999. Ethanol as a hypnotic in insomniacs: Self-administration and effects on sleep and mood. Neuropsychopharmacology, 20, 279–286. [DOI] [PubMed] [Google Scholar]
- SAUNDERS JB, AASLAND OG, BABOR TF, DE LA FUENTE JR & GRANT M 1993. Development of the Alcohol Use Disorders Identification Test (AUDIT): WHO collaborative project on early detection of persons with harmful alcohol consumption. II. Addiction, 88, 791–804. [DOI] [PubMed] [Google Scholar]
- SOBELL LC & SOBELL MB 1992. Timeline follow-back: A technique for assessing self-reported alcohol consumption. Measuring alcohol consumption: Psychosocial and biochemical methods. Totowa, NJ: Humana Press. [Google Scholar]
- STEVENSON BL, PARKS MJ & PATRICK ME 2023. Daily associations between affect, drinking motives, and drinking intensity among U.S. young adults. Psychology of Addictive Behaviors, 37, 275–284. [DOI] [PMC free article] [PubMed] [Google Scholar]
- SZNITMAN SR, SHOCHAT T & GREENE T 2020. Is time elapsed between cannabis use and sleep start time associated with sleep continuity? An experience sampling method. Drug and Alcohol Dependence, 208. [DOI] [PubMed] [Google Scholar]
- SZNITMAN SR, SHOCHAT T, VAN RIJSWIJK L, GREENE T & COUSIJN J 2023. Cannabis and alcohol use and their associations with sleep: A daily diary investigation of single-use and co-use in college students. Cannabis and Cannabinoid Research, 8, 527–536. [DOI] [PubMed] [Google Scholar]
- TOMASO CC, JOHNSON AB & NELSON TD 2021. The effect of sleep deprivation and restriction on mood, emotion, and emotion regulation: Three meta-analyses in one. Sleep, 44, 1–30. [DOI] [PMC free article] [PubMed] [Google Scholar]
- VANDREY R, SMITH MT, MCCANN UD, BUDNEY AJ & CURRAN EM 2011. Sleep disturbance and the effects of extended-release zolpidem during cannabis withdrawal. Drug Alcohol Depend, 117, 38–44. [DOI] [PMC free article] [PubMed] [Google Scholar]
- WANG S, ROSSHEIM ME & NANDY RR 2023. Trends in prevalence of short sleep duration and trouble sleeping among US adults, 2005–2018. Sleep, 46, 1–13. [DOI] [PubMed] [Google Scholar]
- WESCOTT DL, TAYLOR ML, KLEVENS AM, FRANZEN PL & ROECKLEIN KA 2023. Waking up on the wrong side of the bed: Depression severity moderates daily associations between sleep duration and morning affect. Journal of Sleep Research, 25. [DOI] [PMC free article] [PubMed] [Google Scholar]
- WONG MM, BROWER KJ & ZUCKER RA 2009. Childhood sleep problems, early onset of substance use, and behavioral problems in adolescence. Sleep Medicine, 10, 787–796. [DOI] [PMC free article] [PubMed] [Google Scholar]