When and for whom is enacted stigma associated with alcohol and cannabis use at the event-level among sexual and gender minority older adults?

Abstract

Objective:

Sexual and gender minorities (SGM) are at elevated risk for alcohol and cannabis use disorders compared to cisgender, heterosexual individuals. This has been attributed to the unique stressors that SGM experience (SGM stress); however, recent studies have found mixed evidence for a link between SGM stress and substance use. The current study tests an integrated theoretical model derived from minority stress theory and the multistage model of substance use to explain these mixed findings.

Method:

We used data from a 30-day ecological momentary assessment study of substance use among 109 SGM older adults (50+) to examine whether event-level associations between enacted stigma and alcohol and cannabis use, quantity consumed, and consequences are dependent on an individual’s typical pattern of substance use (e.g., frequency, coping motives, and substance use disorder symptoms).

Results:

Findings indicate that, at the event-level, enacted stigma was associated with increased likelihood of alcohol use among those who had a probable alcohol use disorder. Further, SGM with more AUD/CUD symptoms and who used alcohol to cope were more likely to engage in heavier cannabis use and experienced more alcohol consequences when they experienced enacted stigma. However, several other moderations were not significant, contrary to hypotheses.

Conclusions:

Findings provide partial support for an integrated theoretical model incorporating minority stress theory and the multistage model of substance use. Findings suggest that alcohol and cannabis use disorder interventions for SGM would benefit from addressing minority stress.

Introduction

Sexual minority (i.e., individuals who identify as lesbian, gay, bisexual or with another non-heterosexual identity) and gender minority individuals (i.e., individuals’ whose gender identity does not match their sex assigned at birth) are at elevated risk for alcohol and cannabis use disorders compared to their cisgender, heterosexual counterparts (Kerridge et al., 2017; Krueger et al., 2020). These disparities have been attributed to sexual and gender minority stress, the unique stress experienced by sexual and gender minorities (SGM) as a result of the stigmatization of non-heterosexuality and gender diversity (Meyer, 2003). However, ecological momentary assessment (EMA) studies have found mixed evidence for an event-level association between sexual and gender minority stress and alcohol and cannabis use (e.g., minority stress being associated with substance use several hours later; Dyar et al., 2021; Dyar et al., 2023; Ehlke et al., 2022; Lewis et al., 2021). A recent theoretical model integrating SGM stress theory (Hatzenbuehler, 2009; Hendricks & Testa, 2012; Meyer, 2003) and the multistage model of substance use (Koob, 2013; Koob & Volkow, 2010) has been proposed to explain these mixed findings (Dyar et al., 2023). According to this theory, risk factors for substance use differ based on an individual’s frequency of substance use and their level of substance use disorder symptoms. It is theorized that factors like SGM stress only becoming risk factors for substance use among individuals who already use frequently and have more substance use disorder symptoms (Dyar et al., 2023). This integrated theoretical model has received initial support from one EMA study with young adult sexual minority women and gender diverse individuals assigned female at birth (SMWGD; Dyar et al., 2023). However, there is a need to determine whether this framework also applies to the broader population of SGM. Therefore, this study aims to determine whether these event-level associations between enacted stigma and substance use are moderated by a set of individual-level characteristics identified by this integrated theoretical framework among a sample of SGM older adults (50+ years of age). We focus on SGM older adults as they continue to experience elevated rates of substance use disorders (Capistrant & Nakash, 2019; Fredriksen-Goldsen et al., 2013; Peralta et al., 2019; Schuler et al., 2018; Schuler et al., 2019) but are largely excluded from SGM research on substance use.

Event-level associations between enacted stigma and substance use

At the event-level, EMA studies have provided mixed evidence of an association between SGM enacted stigma (i.e., experiences of biased treatment by others) and substance use (Dyar et al., 2021; Dyar et al., 2023; Ehlke et al., 2022; Lewis et al., 2021; Livingston et al., 2017; Wolford-Clevenger et al., 2021). Studies often find some support for a concurrent association between sexual minority enacted stigma and at least one substance use outcome, but evidence for prospective effects is very limited. Ehlke et al. (2022) and Lewis et al. (2021) found evidence for a concurrent association between sexual minority enacted stigma and elevated likelihood of drinking, while Wolford-Clevenger et al. (2021) found evidence for a concurrent association between gender minority enacted stigma and elevated likelihood of drug use. However, Dyar et al. (2023) did not find a significant association between sexual minority enacted stigma and likelihood of either alcohol or cannabis use. Ehlke (2020); Ehlke et al. (2022) also linked sexual minority enacted stigma with heavier concurrent drinking and experiencing more consequences, but Lewis et al. (2021) did not find support for these associations. Dyar et al. (2021) found that sexual minority enacted stigma was associated with experiencing more consequences of alcohol use but not with heavier consumption.

Evidence for prospective associations is far sparser. Neither Dyar et al. (2021) nor Dyar et al. (2023) found evidence of prospective associations between sexual minority enacted stigma and substance use outcomes (with 24 hours and 12-hours lags, respectively), while Lewis et al. (2021) linked sexual minority enacted stigma with heavier next-day drinking, but not with a higher likelihood of drinking or consequences. Livingston et al. (2017) also found evidence for a prospective association between enacted stigma based on a range of identities (sexual identity, gender identity, race/ethnicity, disability, or health status) and substance use at the event-level. Taken together, these mixed findings suggest that there may be moderators that explain variation in findings across samples and individuals. Further, given that most studies have only examined lags from one day to the next, these non-significant prospective associations may also be due to the use of a lag that is too long to detect prospective effects with substance use. To address this limitation, the current study utilized two surveys per day to allow for the examination of within-day lagged effects (e.g., enacted stigma in the morning predicting substance use in the evening).

Differences in associations between enacted stigma and specific alcohol use outcomes (e.g., likelihood of drinking, quantity consumed, consequences) across studies are notable. These findings suggest that enacted stigma may be associated with different patterns of alcohol use. Ehlke et al. (2022)’s findings indicate that SGM are more likely to drink, drink more heavily, and experience more consequences on days when they experience enacted stigma, while Lewis et al. (2021)’s results suggest that SGM have a higher likelihood of drinking, but do not drink more heavily or experience more consequences on days when they experience enacted stigma. In contrast, Dyar et al. (2021)’s findings indicate that SGM are more likely to experience consequences on days when they experience enacted stigma, even when they do not drink more heavily. The latter suggests that SGM may engage in patterns of drinking (e.g., fast-paced drinking) that are higher risk or may be less likely to use protective behavioral strategies to reduce consequences following experiences of enacted stigma. Given these diverse patterns of results across studies, it is necessary to examine at least three outcomes in association with enacted stigma (i.e., likelihood of use, quantity consumed, and consequences) in order to determine which components of substance use are associated with enacted stigma and what this indicates about SGM individuals’ patterns of substance use following experiences of stigma.

An integrated theoretical model of minority stress and substance use

Dyar et al. (2023) proposed an integrated theoretical model to explain when and for whom enacted stigma may contribute to substance use. This theory integrates minority stress theory (Hatzenbuehler, 2009; Hendricks & Testa, 2012; Meyer, 2003) and the multistage model of substance use (Koob, 2013; Koob & Volkow, 2010). Minority stress theory posits that SGM enacted stigma and other types of SGM stress deplete SGM individuals’ coping resources, because SGM stress is experienced in addition to the general stressors experienced by everyone (Hatzenbuehler, 2009; Hendricks & Testa, 2012; Meyer, 2003). This depletion of coping resources, in turn, is theorized to lead to the use of substances to cope with negative affect arising from minority stress (Hatzenbuehler, 2009). However, this model does not specify which SGM are most likely to respond to experiences of minority stress with substance use. Potential moderating factors are identified by the multistage model of substance use.

The multistage model posits that factors contributing to substance use at the event-level change as individuals progress from infrequent use with few substance use disorder (SUD) symptoms to more frequent use with more SUD symptoms. When use is infrequent and SUD symptoms are low/absent, substance use is theorized to be driven by the desire to increase positive affect and sociability (i.e., positive reinforcement). However, when use is frequent and more SUD symptoms are present, substance use transitions to being motivated by the desire to reduce negative affect (i.e., negative reinforcement), leading coping motives to be a dominant driver of substance use during this stage.

Notably, numerous animal models and neuropsychological studies have provided support for the multistage model (Koob, 2013; Koob & Volkow, 2010; Kwako & Koob, 2017). Studies demonstrate that increases in substance use dysregulate neurological reward processes (Koob et al., 2004). Initially, the rewarding neurochemical effects of substances (e.g., increased dopamine; decreased stress neurochemicals) drive positive reinforcement of substance use (Koob et al., 2004). This is attenuated as substance use becomes more frequent and heavier and the brain adapts to the effects of the substance (Koob et al., 2004). However, when individuals’ substance use is disrupted, individuals who use frequently can experience withdrawal, which includes negative affect and a desire to use the substance to reduce the negative affect (negative reinforcement; Koob et al., 2004).

Some of the neurochemical pathways that contribute to negative affect during withdrawal are also activated in response to psychosocial stressors (Kwako & Koob, 2017; Lijffijt et al., 2014). This is theorized to result in cross-sensitization of these pathways, with stress increasing desire for substance use to reduce negative affect due to stress and increasing reactivity to stress (Lijffijt et al., 2014). Further, as substance use becomes more frequent, noradrenergic systems are sensitized, contributing to dysregulation of emotion regulation and problem-solving coping (Lijffijt et al., 2014).

Integrating this multistage model with minority stress theory, the transition to negative reinforcement of substance use, cross-sensitization of stress and substance use, and interference with other coping processes are theorized contribute to an increased likelihood of using substances to cope with enacted stigma among SGM who use substances frequently or have an SUD (Koob, 2013; Koob et al., 2004; Kwako & Koob, 2017; Lijffijt et al., 2014). This integrated model results in several hypotheses. First, individual-level frequency of substance use and SUD symptoms should moderate event-level associations between minority stress and substance use. Among SGM who use substances frequently and have more SUD symptoms, minority stress should show an association with increased substance use at the event-level as a result of the coping motives and negative reinforcement processes that the multistage model posits to be the main drivers of substance use in this group. However, among SGM who use substances infrequently and have few/no SUD symptoms, minority stress would not be expected to lead to substance use at the event-level because coping motives and negative reinforcement processes are not proposed drivers of substance use in this group. Second, coping motives should similarly act as an individual-level moderator of the association between minority stress and substance use because of their higher prevalence and role as a driving force for substance use among those who use substances more frequently and have more SUD symptoms. Specifically, at the event-level, minority stress should be associated with more substance use among those with higher endorsement of coping motives for use. Given that there are three potential moderators that may contribute to differences in event-level associations between minority stress and substance use based on the multistage model, it is necessary to examine each of these moderators to determine which best operationalizes differences based on the multistage model in real-world settings.

Providing initial support for this integrated theoretical model of minority stress and substance use, Dyar et al. (2023) found that individual-level frequency of alcohol/cannabis use, alcohol use disorder (AUD) symptoms, and cannabis use disorder (CUD) symptoms moderated event-level associations between sexual minority stress and alcohol and cannabis use. Specifically, among SMWGD who used alcohol more frequently and had more alcohol use disorder symptoms (AUD), enacted stigma concurrently and prospectively predicted an increased likelihood of alcohol use. The same pattern was present for the moderation of the association between sexual minority enacted stigma and cannabis use by cannabis use frequency and cannabis use disorder symptoms. Individual-level coping motives also moderated the association between sexual minority enacted stigma and cannabis use, but not alcohol use. Among those with higher coping motives for cannabis use, sexual minority enacted stigma was associated with a higher likelihood of cannabis use. While this study provides an important initial test of this integrated theoretical model, the study was limited by four factors: 1) focus on young adults; 2) inclusion of only SMWGD, 3) use of only likelihood of alcohol/cannabis use as outcomes (neglecting quantity consumed and consequences of use), and 4) exclusive focus on sexual minority stress.

Current Study

The current study aimed to further our understanding of when and for whom experiences of enacted stigma contribute to substance use among SGM. We did so among a sample of older adult (50+) SGM because this group continues to experience profound disparities in SUDs but remains underrepresented in SGM research on substance use. Utilizing a different subpopulation of SGM also helps to address concerns that the only prior study to test the integrated theoretical model of minority stress and substance use focused on young adult SMWGD. We also examined moderation of event-level associations between enacted stigma and likelihood of substance use as well as heaviness of use (i.e., quantity consumed and frequency of consumption per day) and consequences of use. Further, we included enacted stigma based on a range of identities, including sexual identity, gender identity, race/ethnicity, and age. Although we were not adequately powered to test each of these types of enacted stigma separately, prior research suggests that including a range of experiences of enacted stigma may help to reduce error associated with excluding experiences of enacted stigma based on marginalized identities not held by the entire sample (Dyar et al., 2025). Using data from a baseline assessment and 30 days of twice daily EMA surveys, we tested the following sets of hypotheses:

1. Individual-level frequency of alcohol use, AUD symptoms, and coping motives for alcohol use would moderate event-level associations between experiences of enacted stigma and likelihood of alcohol use, quantity of drinks consumed on drinking days, and consequences experienced on drinking days.

   We expected that among individuals who drank more frequently, had more AUD symptoms, and reported higher coping motives for drinking, experiencing enacted stigma would be concurrently and prospectively associated with a higher likelihood of alcohol use, more drinks consumed on drinking days, and more consequences on drinking days.

2. Individual-level frequency of cannabis use, CUD symptoms, and coping motives for cannabis use would moderate event-level associations between enacted stigma and likelihood of cannabis use, duration of intoxication on cannabis use days, number of cannabis use sessions on cannabis use days, and consequences experienced on cannabis use days.

   We expected that among individuals who used cannabis more frequency, had more CUD symptoms, and reported higher coping motives for cannabis use, experiencing enacted stigma would be concurrently and prospectively associated with a higher likelihood of cannabis use, longer duration of intoxication on cannabis use days, more sessions of cannabis use on cannabis use days, and more consequences on cannabis use days.

Methods

Participants and Procedures

The current analyses used data from a longitudinal study of mental health and substance use among older adult (50+) sexual and gender minorities (SGM). Participants were recruited via online advertisements on social media (e.g., Facebook) between January and May 2023. The study included a baseline assessment (day 0), a 30-day EMA study (days 1-30), and a follow-up assessment (completed within two weeks of day 30). This study used data from the 30-day EMA study. During the EMA period, participants completed one survey in the morning (8:00am-1:00pm in their time zone) and one in the evening (6:00pm-12:00am in their time zone). The study received IRB approval at Ohio State University.

Eligible participants were U.S. residents, age 50 or older; identified as a sexual minority; and met alcohol or cannabis use criteria (i.e., used cannabis at least three times in the past month, consumed at least two drinks on four days in the past month, or drank 4/5 or more drinks on two occasions for females/males in the past month).¹ Transgender and gender diverse individuals were included in the study if they also identified as a sexual minority. Participants were paid up to $140 based on completion rates: $15 for baseline, $15 for follow-up, $1 for each EMA survey, and $5 bonus for each 6 surveys completed in a row.

Participants who appeared eligible based on their responses to the eligibility survey were text messaged by study team members to verify their eligibility and their access to a mobile phone with text message capabilities. Only participants who provided cell phone numbers from valid cell phone carriers, not VOIP (voice over internet phone) carriers, were eligible. To verify their eligibility, participants were asked to text demographic information (i.e., age, state of residence, email address), which was cross-checked with their responses in the eligibility survey.

There were 109 participants. The sample was diverse in race/ethnicity, with 62.8% of the sample identifying exclusively as non-Latine White, 16.5% identifying as Black, and 21.1% identifying with another racial/ethnic identity. The sample include 21.1% gender minority participants, 40.4% cisgender women, and 38.5% cisgender men. Participants identified primarily as lesbian or gay (63.3%), followed by bisexual or pansexual (25.7%), and other sexual minority identities (11.0%). The average age for the sample was 60.31 (SD = 6.26).

Measures

EMA Measures

Enacted stigma was assessed for a range of identities, including sexual identity, age, race/ethnicity, TGD identity, and at the intersection of age and sexual identity as well as the intersection of sexual identity and race/ethnicity. All participants were asked about experiences related to their age, sexual identity, and the intersection of these two identities. Only people of color were asked questions about race/ethnicity and the intersection of sexual identity and race/ethnicity, and only TGD participants were asked about stigma related to their TGD identity.

Two items were asked for each identity or set of identities and they followed the same pattern. First, participants were asked an item adapted from Mohr and Sarno (2016): “Did you experience anything stressful or negative related to your [piped in identity] since the last survey? This could be something that was relatively minor (e.g., [feeling that your sexual identity was not respected]) or major (e.g., [being physically attacked because of your sexual orientation]).” Examples in parentheses differed across identities. Participants were asked to indicate yes or no. Regardless of their response, participants were asked to “indicate which of the following events you have experienced since the last survey because of your [piped in identity]” and provided with a list of common experiences of enacted stigma. The sexual identity version of this EMA measure of enacted stigma has been used in several prior studies and has demonstrated reliability and validity in samples of young adult sexual minorities (Dyar, 2023; Dyar et al., 2022; Dyar et al., 2023). Specific examples for the first item and common experiences of enacted stigma included in the checklists for age, race/ethnicity, intersectional age by sexual identity, and intersection race/ethnicity by sexual identity were adapted from existing measures of enacted stigma for these identities (Allen et al., 2022; Balsam et al., 2011; Maki, 2018; Smith et al., 2022; Wight et al., 2015; Zimmerman, 2017). Given that the endorsement of multiple types of enacted stigma on the same day was uncommon, we created a binary indicating whether the participant experienced enacted stigma based on any of their marginalized identities. See Dyar and Morgan (in press) for specific items and psychometric properties.

Substance Use was assessed by asking participants, “which of the following have you used since the last survey?” Participants could select alcohol and/or marijuana or “none of the above.” Two binary variables were created indicating no use (0) or use (1) of each substance.

Cannabis Use Outcomes.

The following items were assessed when participants indicated having used cannabis. Frequency and duration of intoxication items were adapted from the Cannabis Use Inventory and participants were asked to answer them thinking about their use since the last survey (Cuttler & Spradlin, 2017).

Frequency of Cannabis Use was assessed by the item “How many separate occasions/sessions of marijuana use did you have?” Response options ranged from 0 to 50+.

Duration of Intoxication was measured by asking “How many hours were you high?” Responses were provided in increments of one hour from 0 to 11 hours and participants could also indicate being high for 12 or more hours. Duration of intoxication acts as a parsimonious proxy for cannabis quantity that functions across modes of use at the event level (Calhoun et al., 2022).

Cannabis Consequences were measured only during morning assessments (and asked about the past 24 hours) by using six selected items from two existing measures of marijuana consequences (Lee et al., 2021; Simons et al., 2012). Participants were asked which “of the following things happened to you as a result of your marijuana use yesterday?” (e.g., “I felt dizzy or sick”). Participants could select multiple items. A sum of consequences endorsed (ranging from 0 to 6) was calculated.

Alcohol Use Outcomes.

The following items were assessed when participants indicated drinking.

Number of Drinks consumed was assessed by asking participants “How many drinks did you have since the last survey?” Participants could indicate the specific number of drinks they consumed from 0 to 24 or indicate that they consumed 25 or more drinks.

Alcohol Consequences were measured only during morning assessments (and asked about the past 24 hours) using a five item adapted version of a measure of consequences used in a previous EMA study of alcohol use among sexual minority women (Dyar et al., 2021). Participants were asked “Did any of the following things happen to you yesterday as a result of drinking?” and could indicate 1 (yes) or 0 (no) to five negative consequences (e.g., “I did something that embarrassed me”; “I had a hangover”). A sum of consequences endorsed was calculated.

Baseline Measures

Drinking Moderators

Drinking frequency were assessed using the first item of the Alcohol Use Disorders Identification Test (Saunders et al., 1993). “How often do you have a drink containing alcohol?” on a scale of 0 (never) to 4 (4 or more times a week).

Coping motives for drinking were assessed using the coping subscale of the Drinking Motives Measure (Cooper, 1994; Grant et al., 2007). Participants were asked to indicate how often they drank to cope (5 items; α = .88; “to forget my worries). Responses were provided on a scale of 1 (almost never/never) to 5 (almost always/always) and were averaged.

Alcohol use disorder symptoms were assessed using the Alcohol Use Disorder Identification Test (AUDIT; Saunders et al., 1993). The AUDIT includes 10 items rated on different scales. For example, the item “How often do you have a drink containing alcohol?” was rated from 0 (never) to 4 (4 or more times a week). Items also assess problems arising from alcohol use (e.g., “How often during the past 6 months have you found that you failed to do what was normally expected of you because of drinking?”). Responses were summed. Total scores ranged from 0 to 40 (a = .86), with scores of 7-15 indicating mild DSM-5 AUD and 16+ indicating the probable presence of moderate to severe AUD (Ingesson-Hammarberg et al., 2024; Saunders et al., 1993).

Cannabis Use Moderators

Cannabis use frequency was assessed by the item “In the past month, how many days did you use marijuana?” Participants could enter an integer value between 0 and 31.

Cannabis use motives were assessed using a brief version of the coping subscale of the Comprehensive Marijuana Motives Questionnaire (Lee et al., 2009), in which the two highest loading items were selected. Participants were asked to indicate how often they used marijuana to cope (α = .73; “to forget your problems”). Responses were provided on a scale of 1 (almost never/never) to 5 (almost always/always) and were averaged.

Cannabis use disorder symptoms were assessed using the Cannabis Use Disorder Identification Test Revised (CUDIT-R; Adamson et al., 2010). The CUDIT-R includes 8 items rated on different scales. For example, the item “How often during the past 6 months did you fail to do what was normally expected from you because of using marijuana?” was rated from 0 (never) to 4 (daily or almost daily). Responses were summed. Total scores ranged from 0 to 32 (a = .79), with scores of greater than 10+ indicating a DSM-5 cannabis use disorder (Bonn-Miller et al., 2016) and 13+ indicating probable DSM-IV cannabis dependence (Adamson et al., 2010).

Analytic Plan

Analyses were conducted in Mplus version 8.8. There were a total of 5,309 completed surveys from 109 participants. The median completion rate was 91.7% (M = 80.7%, SD = 25.9%). Within completed surveys, 2% of data were missing. Missing data were handled using Bayesian methods (Asparouhov & Muthén, 2010). Analyses of cannabis use outcomes at the event-level (e.g., frequency of use, duration of intoxication, and consequences) included observations during which cannabis use was reported (1,419 observations) by individuals who reported using cannabis at least once in the past month at baseline (n = 71). Analyses of alcohol use outcomes at the event-level (e.g., number of drinks and consequences) included observations during which drinking was reported (1,306 observations) among individuals who reported drinking at least once in the past month at baseline (n = 102). A subset of participants did not report any experiences of enacted stigma (n = 23) or did not report alcohol (n = 11) or cannabis use (n = 7) during the EMA period despite meeting criteria for inclusion in analyses. These participants were included in the models but did not contribute to within-person estimates.

Bayesian multilevel structural equation modeling (MSEM) with diffuse (non-informative) priors was used. MSEM utilizes latent variables, rather than group- and grand-mean centering, to separate within- from between-person variance (Ludtke et al., 2008). By removing the between-person variance from the event-level variance, the event-level variables indicate the extent to which an individual was experiencing more/less of a construct than usual (above/below their person mean) on a particular day (e.g., experiencing more/less minority stress than usual). We used Markov Chain Monte Carlo (MCMC) algorithms to generate a series of 10,000 random draws from the multivariate posterior distribution of our sample for each model. Trace plots and the Gelman-Rubin potential scaling reduction (PSR) were used to determine whether convergence was achieved (Depaoli & Clifton, 2015; Muthen, 2010).

We examined concurrent and prospective effects models. In the first set of models, we examined within-person (i.e., event-level) and between-person associations between experiencing enacted stigma and the likelihood of alcohol or cannabis use. In concurrent effects models, event-level enacted stigma during one observation predicted event-level alcohol or cannabis use during the same observation. In prospective models, event-level enacted stigma during one observation (t) predicted event-level alcohol or cannabis use during the next observation (t+1) and an autocorrelation for alcohol or cannabis use was included (which effectively controls for alcohol or cannabis use at t). Prospective associations included both within-day (stigma reported in the morning survey predicting cannabis use during the evening survey) and across-day lags (stigma reported in the evening survey predicting cannabis use during the morning survey) because the results were similar when across-day lags were included and when they were excluded. A probit link was used for binary outcomes (e.g., cannabis use) as logistic regression is not available in Bayesian MSEM. Probit regression coefficients represent the variance shared by the predictor and the latent continuous response variables underlying each binary observed item (Agresti, 2003). In all models, we controlled for day of assessment and assessment type (weekend/weekday; morning/evening) at the within-person level. Within-person associations among enacted stigma and substance use outcomes as well as autocorrelations were allowed to vary across individuals. Age, sexual identity, sex/gender identity (i.e., cisgender women, cisgender men, transgender and gender diverse), and race/ethnicity were included as covariates at the between-person level.

In the second set of models, which followed the same pattern described above, we examined concurrent and prospective effects of enacted stigma on outcomes of alcohol use (i.e., drinks consumed, consequences) during observations when drinking was reported. Similarly, we examined concurrent and prospective effects of enacted stigma on outcomes of cannabis use (i.e., number of sessions of cannabis use, duration of intoxication, consequences) during observations when cannabis use was reported. As consequences were only assessed during the morning survey and asked about experiences over the last 24 hours, day-level aggregate measures of enacted stigma (i.e., coded as 1 if enacted stigma experienced during the last 24 hours, coded as 0 if enacted stigma not experienced during the last 24 hours) were utilized in analyses of consequences. In concurrent models, enacted stigma reported during an evening survey (t) and the next morning survey (t+1) were aggregated and its association with consequences as reported during the same morning survey were examined. In prospective models, enacted stigma during one 24 hour period predicted consequences of cannabis/alcohol use during the next 24 hours.

In the third set of models, we added individual-level moderators to each of the event-level associations between enacted stigma and alcohol/cannabis use and outcomes. Alcohol use moderators (i.e., drinking frequency, coping motives for drinking, alcohol use disorder symptoms) were added as predictors of the event-level associations between enacted stigma and alcohol use and related outcomes. Similarly, cannabis use moderators (i.e., cannabis use frequency, coping motives for cannabis use, cannabis use disorder symptoms) were added as predictors of the event-level associations between enacted stigma and cannabis use and related outcomes. The Johnson-Neyman technique was also used to determine at which precise values of the moderator the association became significantly positive or negative for significant interactions.

A correction for multiple testing was not made as such corrections are only appropriate when there are omnibus hypotheses upon which subsequent hypotheses are dependent (García-Pérez, 2023; Rubin, 2021, 2024). It is not appropriate when individual hypotheses’ significance is not interdependent. The current manuscript tests independent hypotheses. This is evidenced by the fact that one association can be significant regardless of the significance of any other association. For example, enacted stigma may predict duration of intoxication even if there is not an association between enacted stigma and likelihood of cannabis use. This would indicate that enacted stigma is not associated with an elevated likelihood of cannabis use, but when individuals use cannabis on days when enacted stigma is experienced, they tend to be intoxicated for longer.

Sensitivity Analyses

Given that AUDIT and CUDIT-R scores include both items about frequency/quantity of use and AUD/CUD symptoms, we also conducted sensitivity analyses in which only AUDIT and CUDIT-R items that captured AUD/CUD symptoms (AUDIT items 4-10; CUDIT-R items 3-8) were used in the sum score and tested whether the moderation results were the same when the full AUDIT/CUDIT-R scores were used compared to scores that only included symptoms.

We also conducted sensitivity analyses in which only enacted stigma based on sexual identity and TGD identity were included. Of note, this resulted in fewer observations of enacted stigma (n = 1048 for all types of enacted stigma; n = 676 for sexual identity or TGD enacted stigma only) and lower power.

Transparency and Openness

Given the risk of deductive disclose when data from minoritized populations are made publicly available, deidentified data are only available from the study team after the development of a data sharing agreement. Data analysis code is available from the authors upon request. This study was not preregistered.

Results

Participants reported enacted stigma on 20% of days, drinking on 25% of days, and using cannabis on 42% of days (among those who met cannabis use criteria). The most common type of enacted stigma reported was based on one’s sexual identity (n = 554 observations), followed by age (n = 523 observations), the intersection of sexual identity and age (n = 168 observations), race/ethnicity (n = 135 observations), and the intersection of sexual identity and race/ethnicity (n = 105 observations). The majority of participants were below the AUDIT/CUDIT thresholds for a probable substance use disorder (n = 84 for alcohol; n = 59 for cannabis), while 15-16% of those who used alcohol or cannabis met criteria for probably mild AUD or CUD (n = 17 for alcohol; n = 11 for cannabis) and 8-14% met criteria for probable moderate to severe AUD or cannabis dependence (n = 8 for alcohol; n = 10 for cannabis). Table 1 provides means, standard deviations, and intraclass correlations.

Table 1.

Correlations, Means, Variances, and Intraclass Correlations

	Mean	Standard Deviation	Range	Intraclass Correlation
Enacted Stigma	.20	.40	0-1	.27
Alcohol Use	.25	.43	0-1	.31
Number of Drinks	2.77	2.20	1-20	.59
Alcohol Consequences	.09	.35	0-4	.30
Cannabis Use	.42	.49	0-1	.28
Duration Cannabis Intoxication	2.93	2.28	0-12	.70
Cannabis Use Sessions	1.72	1.10	0-8	.42
Cannabis Consequences	.25	.61	0-4	.53
Baseline Drinking Frequency	2.50	1.30	0-4	-
Baseline Alcohol Coping Motives	1.68	.78	1-5	-
Baseline Drinking Consequences (AUDIT)	5.67	5.75	0-40	-
Baseline Cannabis Use Frequency	19.73	10.75	0-30	-
Baseline Cannabis Coping Motives	1.56	.75	1-5	-
Baseline Cannabis Use Consequences (CUDIT)	8.46	5.66	1-32	-

Unmoderated event-level associations

At the event-level, enacted stigma was not associated with the likelihood of using alcohol or cannabis either concurrently (Table 2) or prospectively (Table 3). However, during observations when participants drank and experienced enacted stigma, they experienced more consequences than when they did not experience enacted stigma. Johnson Neyman figures for significant moderations are presented in Figures 1 and 2 and simple slopes for a subset of values of the moderator are presented in Table 4. Concurrent associations between enacted stigma and drinks consumed, frequency of cannabis use, duration of cannabis intoxication, and number of cannabis use consequences were not significant. No prospective event-level associations between enacted stigma and alcohol/cannabis use and related outcomes were significant.

Table 2.

Concurrent within-person direct effect estimates and moderation

Model	Moderator	Path	Any Substance Use			Quantity Consumed			Frequency of Use			Consequences
			b	95% CI	p	b	95% CI	p	b	95% CI	p	b	95% CI	p
Enacted Stigma → Drinking Outcome	Unmoderated	Enacted Stigma → Drinking Outcome	.09	−.10, .27	.34	.13	−.09, .36	.26	-	-	-	.15	.03, .26	.01
	Drinking Frequency	Enacted Stigma → Drinking Outcome
		Average Slope	.22	−.39, .78	.47	−.30	−1.31, .70	.56	-	-	-	.08	−.36, .52	.70
		Drinking Frequency → Slope	−.04	−.23, .15	.64	.13	−.16, .42	.40	-	-	-	.01	−.12, .15	.81
	Drinking Motives	Enacted Stigma → Drinking Outcome
		Average Slope	−.29	−.72, .15	.19	−.003	−.58, .57	.99	-	-	-	−.26	−.54, .01	.06
		Coping Motives → Slope	.21	−.01, .41	.06	.07	−.20, .34	.62	-	-	-	.21	.07, .35	.004
	AUDIT	Enacted Stigma → Drinking Outcome
		Average Slope	−.13	−.41, .13	.32	−.004	−.36, .33	.98	-	-	-	−.08	−.23, .06	.26
		AUDIT → Slope	.03	.01, .07	.02	.02	−.01, .05	.29	-	-	-	.03	.01, .04	< .001
Enacted Stigma → Cannabis Outcome	Unmoderated	Enacted Stigma → Cannabis Use	−.17	−.39, .04	.12	.01	−.33, .37	.93	.16	−.002, .32	.05	.01	−.08, .11	.78
	Cannabis Use Frequency	Enacted Stigma → Cannabis Use
		Average Slope	−.12	−.60, .34	.60	−.37	−1.49, .79	.53	−.06	−.59, .50	.81	−.02	−.39, .35	.91
		Cannabis Use Frequency → Slope	−.001	−.02, .02	.89	.02	−.03, .06	.49	.01	−.01, .03	.41	.001	−.01, .02	.85
	Cannabis Use Motives	Enacted Stigma → Cannabis Use
		Average Slope	−.35	−.77, .08	.10	.36	−.41, 1.15	.35	.44	.10, .76	.01	−.03	−.25, .18	.76
		Coping Motives → Slope	.11	−.09, .31	.28	−.20	−.60, .20	.31	−.16	−.33, .01	.08	.03	−.08, .14	.64
	CUDIT	Enacted Stigma → Cannabis Use
		Average Slope	−.48	−.87, −.09	.02	.21	−.55, .92	.56	.29	−.04, .63	.09	.02	−.20, .23	.87
		CUDIT → Slope	.03	.000, .07	.05	−.02	−.08, .04	.55	−.01	−.04, .01	.37	−.001	−.02, .02	.95

Covariates included: within-person (day of EMA period, weekend vs weekday, and morning vs. evening survey); between-person (race/ethnicity, sexual identity, gender identity, age). Associations among variables at the between-person level were also estimated but are not included in tables. Significant effects are presented in bold. Quantity consumed for cannabis use refers to duration of intoxication, a proxy for cannabis quantity that is effective across modes of use.

Table 3.

Prospective within-person direct effect estimates and moderation

Model	Moderator	Path	Any Substance Use			Quantity Consumed			Frequency of Use			Consequences
			b	95% CI	p	b	95% CI	p	b	95% CI	p	b	95% CI	p
Enacted Stigma → Drinking Outcome	Unmoderated	Enacted Stigma → Drinking Outcome	.11	−.07, .29	.23	.04	−.26, .30	.80	-	-	-	−.10	−.22, .03	.14
	Drinking Frequency	Enacted Stigma → Drinking Outcome
		Average Slope	.04	−.58, .60	.90	−.05	−1.28, 1.16	.94	-	-	-	.29	−.22, .84	.26
		Drinking Frequency → Slope	.03	−.15, .22	.79	.03	−.33, .37	.88	-	-	-	−.12	−.28, .03	.12
	Drinking Motives	Enacted Stigma → Drinking Outcome
		Average Slope	.12	−.33, .58	.60	−.47	−1.24, .31	.22	-	-	-	.46	.19, .74	< .001
		Coping Motives → Slope	−.004	−.22, .21	.97	.26	−.13, .63	.18	-	-	-	−.30	−.44, −.17	< .001
	AUDIT	Enacted Stigma → Drinking Outcome
		Average Slope	.17	−.10, .44	.21	−.16	−.56, .25	.43	-	-	-	.18	.03, .34	.02
		AUDIT → Slope	−.01	−.04, .02	.56	.02	−.01, .06	.19	-	-	-	−.03	−.05, −.02	< .001
Enacted Stigma → Cannabis Use	Unmoderated	Enacted Stigma → Cannabis Use	.04	−.20, .30	.74	.23	−.08, .54	.13	.16	−.01, .34	.07	−.004	−.11, .10	.94
	Cannabis Use Frequency	Enacted Stigma → Cannabis Use
		Average Slope	−.25	−.81, .28	.36	−.28	−1.27, .71	.56	−.11	−.68, .46	.71	.37	−.03, .74	.07
		Cannabis Use Frequency → Slope	.01	−.01, .04	.26	.02	−.02, .06	.27	.01	−.01, .03	.32	−.01	−.03, .001	.06
	Cannabis Use Motives	Enacted Stigma → Cannabis Use
		Average Slope	.08	−.43, .60	.76	.07	−.60, .75	.83	.11	−.31, .49	.60	.04	−.18, .28	.71
		Coping Motives → Slope	−.03	−.28, .22	.80	.09	−.24, .43	.60	.04	−.16, .24	.74	−.03	−.15, .10	.65
	CUDIT	Enacted Stigma → Cannabis Use
		Average Slope	.04	−.42, .50	.86	−.55	−1.17, .05	.08	−.19	−.58, .18	.31	.06	−.25, .32	.70
		CUDIT → Slope	−.002	−.04, .04	.92	.08	.03, .13	.004	.03	.003, .07	.03	−.01	−.04, .02	.52

Covariates included: within-person (day of EMA period, weekend vs weekday, and morning vs. evening survey); between-person (race/ethnicity, sexual identity, gender identity, age). Autocorrelations for the outcome were also included in prospective models. Associations among variables at the between-person level were also estimated but are not included in tables. Significant effects are presented in bold. Quantity consumed for cannabis use refers to duration of intoxication, a proxy for cannabis quantity that is effective across modes of use.

Figure 1.

Johnson-Neyman plots of significance of simple slopes for significant interactions involving concurrent associations. Solid lines represent the slope of the within-person effect of interest as a function of the between-person AUD (alcohol use disorder) symptoms or coping motives for alcohol use. Dashed lines represent the 95% CI around the simple slope line. The simple slope is significant at values of x, when the simple slope estimate and both 95% credibility interval lines are on the same size of line indicating 0 on the y axis (grey line).

Figure 2.

Johnson-Neyman plots of significance of simple slopes for significant interactions involving prospective associations. Solid lines represent the slope of the within-person effect of interest as a function of the between-person AUD (alcohol use disorder) symptoms or coping motives for alcohol use. Dashed lines represent the 95% CI around the simple slope line. The simple slope is significant at values of x, when the simple slope estimate and both 95% credibility interval lines are on the same size of line indicating 0 on the y axis (grey line).

Table 4.

Within-Person Simple Slopes for Significant Interactions

Outcome	Moderator	Level of Moderator	Concurrent		Prospective
			b	95% CI	b	95% CI
Alcohol Use	AUDIT	7 (cut point for DSM-5 mild AUD)	.11	−.07, .29	-	-
		16 (cut point for DSM-5 moderate/severe AUD)	.42	.07, .78	-	-
Alcohol Consequences	Coping Motives	Almost Never/Never	−.05	−.21, .10	.16	.01, .31
		Half of the time	.36	.18, .56	−.45	−.63, −.27
		Almost Always/Always	.77	.35, 1.23	−1.05	−1.50, −.63
	AUDIT	7 (cut point for DSM-5 mild AUD)	.11	.02, .21	−.07	−.17, .03
		16 (cut point for DSM-5 moderate/severe AUD)	.37	.22, .53	−.38	−.54, −.23
Cannabis Use	CUDIT	10 (cut point for DSM-5 CUD)	−.15	−.34, .03	-	-
		13 (cut point for DSM-IV cannabis dependence)	−.05	−.25, .15	-	-
Quantity Cannabis Consumed	CUDIT	10 (cut point for DSM-5 CUD)	-	-	.22	−.07, .50
		13 (cut point for DSM-IV cannabis dependence)	-	-	.44	.13, .76
Frequency of Cannabis Use	CUDIT	10 (cut point for DSM-5 CUD)	-	-	.16	−.02, .33
		13 (cut point for DSM-IV cannabis dependence)	-	-	.27	.07, .47

"-" indicates that moderation was not significant. Simple slopes were calculated at meaningful levels of the moderator when possible (e.g., coping motives half the time).

Moderated concurrent event-level associations

Individual-level alcohol use disorder symptoms moderated event-level associations between enacted stigma, likelihood of alcohol use, and alcohol consequences. Johnson-Neyman decomposition indicates that experiencing enacted stigma was significantly associated with a higher likelihood of alcohol consumption among participants who had a score of 11 or higher on the AUDIT, but this association was not significant among individuals with lower AUDIT scores. Similarly, experiencing enacted stigma was associated with more alcohol consequences among individuals with AUDIT scores of 7 or higher, with the association being non-significant for those with lower AUDIT scores. No variables significantly moderated the event-level association between enacted stigma and quantity of alcohol consumed, and individual-level drinking frequency did not significantly moderate any associations.

Individual-level coping motives for alcohol use moderated the event-level association between enacted stigma and alcohol consequences and (marginally) likelihood of alcohol use (p = .06). Johnson-Neyman decomposition indicates that experiencing enacted stigma was associated with more alcohol consequences on days when alcohol was consumed among individuals who had scores of 1.8 or higher on coping motives for drinking (1 = almost never/never; 2 = some of the time). This association was non-significant for those who reported lower coping motives for drinking.

Only one concurrent moderation effect was significant for analyses of cannabis. Individual-level cannabis use disorder symptoms moderated event-level associations between enacted stigma and likelihood of cannabis use. However, the pattern of simple slopes was different for this moderation effect. Among individuals with a CUDIT score of 9 or lower, experiencing enacted stigma was associated with a lower likelihood of using cannabis during the same observation. This association was not significant for individuals with CUDIT scores of 10 or higher.

Moderated prospective event-level associations

A different moderation pattern emerged for prospective event-level associations between enacted stigma and alcohol related outcomes. Specifically, coping motives for drinking and alcohol use disorder symptoms moderated the associations between enacted stigma and alcohol consequences on observations when alcohol was consumed. The simple slope pattern was similar for both of these moderations. Individuals with low AUDIT scores (0 or 1) and who reported never/almost never drinking to cope (1.0 on drinking to cope measure) experienced more consequences during observations when they experienced enacted stigma. These associations were non-significant for individuals with scores of 2 to 7 on the AUDIT or 1.1 to 1.8 on coping motives for drinking. Among individuals with higher AUDIT scores (8 or higher) and those who endorsed more coping motives for drinking (1.9 or higher), experiencing enacted stigma was associated with experiencing fewer consequences. Notably, the same pattern of moderation effects remained when the autocorrelation for alcohol consequences was dropped, suggesting that this is not an artifact of a high correlation between current and prior alcohol consequences. No other prospective associations with alcohol use or related outcomes were significantly moderated.

A pattern similar to the concurrent moderation analyses was present for cannabis-related outcomes. Specifically, cannabis use disorder symptoms moderated prospective event-level associations between enacted stigma and duration of intoxication as well as number of cannabis use sessions. Among individuals with CUDIT scores of 11 or higher, experiencing enacted stigma was associated with longer duration of intoxication and more sessions of cannabis use, while these associations were not significant among those with lower CUDIT scores. No other moderations were significant for cannabis use or related outcomes.

Sensitivity Analyses

As AUDIT and CUDIT-R scores include items assessing quantity/frequency of use as well as AUD/CUD symptoms, we conducted sensitivity analyses in which quantity/frequency items were excluded from AUDIT and CUDIT-R scores and only items assessing AUD/CUD symptoms were included. The moderation analyses for these symptom-only scores followed the same pattern as those for the full AUDIT and CUDIT-R scores.

Analyses in which only sexual identity and TGD enacted stigma were included produced largely the same pattern of results as those in which all types of enacted stigma were included. Two moderation effects (i.e., AUD symptoms moderating concurrent association between enacted stigma and likelihood of alcohol use; CUD symptoms moderating concurrent association between enacted stigma and likelihood of cannabis use) became marginally significant, but the coefficients for the moderation effects were the same as those for analyses including all types of enacted stigma. This suggests that these two moderation effects became marginally significant as a result of reduced power in these sensitivity analyses. Only one moderation that was significant in the main analyses became nonsignificant in sensitivity analyses and did not have a similar coefficient size, the moderation of the concurrent association between enacted stigma and alcohol consequences by AUD symptoms (b = .01, p = .32, 95% CI: −.01, .04).

Two other moderations became significant in sensitivity analyses: 1) frequency of cannabis use moderating the concurrent association between enacted stigma and cannabis consequences and 2) CUD symptoms moderating prospective associations between enacted stigma and cannabis consequences. Among individuals who used cannabis on 11 or fewer days in the past month at baseline, experiencing SGM enacted stigma was associated with experiencing fewer consequences of cannabis use on the same day, while these associations were not significant among those who used cannabis more frequently. Among individuals with a CUDIT score or 19 or higher, experiencing SGM enacted stigma predicted experiencing more consequences of cannabis use on the next day, while these associations were not significant among those with lower CUDIT scores.

Discussion

The current study extends research testing the integrated model of minority stress and substance use by 1) examining individual-level moderators of event-level associations between enacted stigma and substance use as well as quantity consumed and consequences of use, 2) including a range of diverse enacted stigma experiences, and 3) utilizing a sample of SGM older adults, who have been vastly underrepresented in research on substance use. Results indicate that enacted stigma was only concurrently associated with an elevated likelihood of alcohol use among SGM with more AUD symptoms, partially consistent with Dyar et al. (2023) and the integrated model of minority stress and substance use. However, prospective associations between enacted stigma and likelihood of alcohol and cannabis use were not moderated by AUD/CUD symptoms, and neither coping motives for substance use nor frequency of substance use were moderators of associations between enacted stigma and likelihood of alcohol or cannabis use, counter to our hypotheses. Some associations between enacted stigma and quantity and consequences of substance use were moderated by coping motives for alcohol use and AUD/CUD symptoms, but the majority of these moderations were not significant. Overall, findings provide partial support for the integrated model of minority stress and substance use, but results were not consistent across moderators or different substance use outcomes.

AUD and CUD symptoms were the most consistent moderators of concurrent event-level association between enacted stigma and likelihood of substance use. In most cases, enacted stigma was only associated with an increased likelihood of substance use among those with more AUD/CUD symptoms. This suggests that rather than enacted stigma acting as a risk factor for substance use for all SGM (as posited by minority stress theory), enacted stigma may only be associated with likelihood of substance use among those who already have some problems with substance use. This is consistent with findings from Dyar et al. (2023), who proposed an integrated model incorporating both minority stress and the multistage model of substance use to explain mixed findings for the association between enacted stigma and substance use in EMA and daily diary studies. This integrated model proposes that negative reinforcement drives substance use among those with substance use disorders and who use substances more frequently, and thus, risk factors linked with negative reinforcement (e.g., stressors like minority stress) would only be expected to be risk factors for substance use among those with a substance use disorder or those who use frequently. Among individuals with fewer substance use disorder symptoms or who use less frequently, substance use is theorized to be driven by positive reinforcement (e.g., socializing) rather than negative reinforcement. Thus, enacted stigma would not be expected to predict substance use among this group. Our moderation analyses of AUD/CUD symptom were consistent with this theory, providing support in a new subpopulation of SGM (SGM older adults) and broadening this support to include enacted stigma based on a range of identities (i.e., sexual identity, gender identity, race/ethnicity, and age). Notably, prospective associations between enacted stigma and likelihood of substance use were not moderated by AUD/CUD symptoms, which is inconsistent with results from Dyar et al. (2023) and our hypotheses. One potential explanation for these divergent findings may be differences in sample size for these two studies. The current study had 109 participants and Dyar et al. (2023) had 429, resulting in substantially higher power for the prior study than the current study.

The current study is the first we are aware of to test whether heavier use and more consequences are also more likely following enacted stigma experiences among SGM with more AUD/CUD symptoms. Results provide some support for this hypothesis. Enacted stigma was concurrently associated with more alcohol consequences and prospectively predicted heavier cannabis use (i.e., longer duration of intoxication and more sessions of use) among those with probable AUD/CUD, but not among those with fewer AUD/CUD symptoms. This suggests that not only are SGM older adults with more AUD/CUD symptoms more likely to consume alcohol or cannabis when they experience enacted stigma, but this use is also heavier (for cannabis) and associated with more consequences (for alcohol). This highlights the importance of determining the effects of enacted stigma on substance use in this population to identify those who may benefit most from interventions focused on enacted stigma as a risk factor for substance use.

For alcohol, coping motives for drinking and AUD symptoms had similar moderating effects on associations with alcohol consequences. Endorsing even slight coping motives for drinking was associated with experiencing more consequences of alcohol use during observations when enacted stigma was experienced, while individuals with mild to moderate AUD symptoms experienced more consequences in association with enacted stigma. This is inconsistent with Dyar et al. (2023) who did not find support for coping motives as a moderator of the association between enacted stigma and likelihood of alcohol use. However, this may be because Dyar et al. (2023) only examined the likelihood of alcohol/cannabis use and not consequences of use. Why might enacted stigma be associated with more consequences of alcohol use, but not higher quantity consumed? While counterintuitive, this finding is consistent with some other studies (Dyar et al., 2021; Wilson et al., 2016). One potential explanation for this finding may be that SGM who use alcohol to cope or have mild to moderate AUD symptoms are more likely to respond to enacted stigma by drinking in ways that increase the likelihood of consequences without increasing the quantity of alcohol consumed (e.g., faster consumption of alcohol) or that they are less likely to engage in protective behavioral strategies to reduce the likelihood of alcohol consequences when they experience enacted stigma. Future research should explore more nuanced patterns of alcohol consumption and the use of protective behavioral strategies to test these potential explanations.

For cannabis, only CUD symptoms moderated prospective associations between enacted stigma, duration of intoxication, and number of cannabis use sessions, but CUD symptoms did not moderate prospective associations between enacted stigma, likelihood of cannabis use, and consequences. This suggests that SGM with severe CUD symptoms were not more likely to use cannabis following experiences of enacted stigma, but if they did use cannabis following an experience of enacted stigma, they used more heavily but did not experience more consequences. This is nearly the opposite pattern found for alcohol use and suggests that at least among older adult SGM, patterns of alcohol and cannabis use may differ following experiences of enacted stigma. Further research should explore potential reasons for these different findings for alcohol and cannabis use.

In another divergent pattern, Dyar et al. (2023) found that frequency of substance use was a consistent moderator of the association between enacted stigma and substance use, with those who used substances more frequently being more likely use substances following enacted stigma. However, frequency of use did not moderate any associations in the current study. This may be due to sample differences, as the current study focused on older adult SGM, while Dyar et al. (2023) focused on young adult sexual minority women and gender diverse individuals assigned female at birth. Specifically, AUD/CUD symptoms may be a more important predictor of substance use in response to stigma than frequency of alcohol or cannabis use among older adult SGM, while both may be important predictors for young adult SGM. Future research should explore whether there are age differences in these moderation effects to test this potential explanation.

Two prospective moderation effects were in the opposite direction expected. Experiencing enacted stigma was associated with experiencing fewer consequences on the next day among individuals who endorsed more coping motives for drinking and had more AUD symptoms. It is interesting that these are the same individuals who were more likely to experience consequences on the same day that they experienced enacted stigma. Consequences were also the only variable that was assessed once per day (instead of twice), doubling the lag for prospective associations with this variable. We posit that this unexpected pattern may be explained by the same-day increase in consequences experienced by these individuals, which may in turn make them less likely to drink enough to experience elevated consequences on the next day. This would be consistent with the results of some daily diary studies examining the effects of experiencing alcohol consequences on next-day alcohol use and cognitions. These studies suggest that experiencing alcohol consequences on one day is associated with increased perceptions of the likelihood of experiencing negative consequences (Lee et al., 2018; LoParco et al., 2021) and lower willingness to drink (LoParco et al., 2021) on the subsequent day. Interestingly, the reversal of the association between enacted stigma and substance use was only present for alcohol consequences (not likelihood of drinking or quantity of drinks consumed), which may be a result of the longer lag for prospective effects, or it may suggest that this group of individuals isn’t necessarily reducing the amount they drink on the next day but may instead be changing the way they drink to reduce the likelihood of consequences (e.g., drinking more slowly). Notably, this is also consistent with prior research as studies have typically failed to find an association between experiencing negative consequences and actual next-day drinking behavior (i.e., likelihood of drinking and number of drinks consumed) (Lee et al., 2018; Merrill et al., 2021).

Clinical Implications

Given disparities in alcohol and cannabis use disorders affecting SGM (Kerridge et al., 2017; Krueger et al., 2020) and evidence that these disparities are maintained throughout older adulthood (Capistrant & Nakash, 2019; Fredriksen-Goldsen et al., 2013; Peralta et al., 2019; Schuler et al., 2018; Schuler et al., 2019), there is a need for interventions that reduce substance use disorders in this population. The current study suggests that experiences of enacted stigma may increase the likelihood of substance use and contribute to heavier use of cannabis and more consequences arising from drinking. Interventions that aim to reduce substance use disorders in this population should include components that teach alternative skills for coping with enacted stigma based on SGM and other marginalized identities held by clients. Minority stress theory posits that the depletion of alternative coping resources contributes to reliance on substances to cope. Therefore, the results of the current study and theory suggest that preventive interventions may also benefit from focusing on building skills and resources for coping with minority stress. Interventions like ESTEEM and EQuIP include components that address minority stress, building skills for coping with minority stress, and reducing problematic alcohol use (Pachankis et al., 2015; Pachankis et al., 2020). Thus, these interventions (with modifications to address cannabis use) may be effective in reducing AUD and CUD among SGM. Other interventions that address coping motives for alcohol use may also be effective, such as motivational enhancement and brief interventions targeting drinking motives (Blevins & Stephens, 2016; LaBrie et al., 2008). Such interventions would need to be adapted to address minority stress processes among SGM, but this avenue is promising given that coping motives for drinking emerged as a significant risk factor for experiencing more alcohol consequences when enacted stigma is experienced.

Limitations

The results of the current study should be considered in light of its limitations. First, only SGM older adults who used alcohol or cannabis regularly and lived in the US were included in this study. As a result, it is unclear whether similar patterns will be found among SGM individuals who live outside the US or those who use alcohol or marijuana less frequently. Second, only two surveys were administered per day. This approximately 12-hour gap between surveys may have been too long to detect some prospective effects. Further, the longer gaps for prospective effects of enacted stigma on consequences (~24 hours) may have especially limited our ability to detect prospective effects on consequences. Future research with more frequent surveys should determine the most appropriate lag for associations between enacted stigma and substance use using methods like dynamic time varying equation models (DTVEM) (Jacobson et al., 2019). Third, our concurrent analyses cannot determine the temporality or directionality of the association between enacted stigma and substance use. It is possible that enacted stigma may lead to substance use, that substance use may be more likely to occur in contexts in which enacted stigma is more likely, or that a third variable may be contributing to the co-occurrence of enacted stigma and substance use. Further research with shorter lags between associations are necessary to determine the directionality and temporality. Fourth, only a subset of participants were included in analyses of cannabis use (n = 67). While we were able to detect several significant moderation effects for cannabis use, we had lower power for these moderation analyses than we did for moderation analyses of alcohol use. Fifth, we did not examine whether participants used cannabis for medical purposes, recreational purposes, or a combination of both. These diverse reasons for cannabis use may affect associations between enacted stigma and cannabis use outcomes, and this should be explored by future research. Sixth, the sample included a relatively small number of participants above the AUDIT threshold for probable DSM-5 moderate to severe AUD (n = 8) and the CUDIT threshold for probable DSM-IV cannabis dependence (n = 10). This may have affected our ability to detect some moderation effects based on AUDIT and CUDIT scores, and future studies should determine if study findings replicate in samples with a higher number of individuals with moderate to severe AUD/CUD.

Conclusions

This study was the second to test whether individuals’ patterns of alcohol and cannabis use moderated the likelihood of alcohol or cannabis use following experiences of enacted stigma. Findings partially confirm prior results, indicating that AUD/CUD symptoms are a consistent moderator of event-level concurrent effects of enacted stigma on substance use. Providing additional support for an integrated model of minority stress theory and the multistage model of substance use, results indicate that when SGM older adults who had more AUD/CUD symptoms experienced enacted stigma, they were more likely to drink or use cannabis and experienced more drinking consequences during the same observation. They also reported heavier cannabis use during the next observation. Results also suggest that coping motives for alcohol use may be associated with more consequences of alcohol use when enacted stigma is experienced. Notably, a number of hypothesized moderation results were not significant, suggesting that the integrated theory may only partially explain differences in associations between enacted stigma and substance use outcomes across studies. Findings suggest that interventions designed to address AUD and CUD use among SGM older adults would benefit from addressing minority stress and coping skill-building.

Public Significance Statement.

The results of this study suggest that discrimination and microaggressions are risk factors for same/next-day alcohol and cannabis use among LGBTQ+ individuals who have a substance use disorder. However, discrimination and microaggressions do not appear to be risk factors for same/next-day alcohol and cannabis use among LGBTQ+ individuals who do not have a substance use disorder.