Authors: Luisa Kcomt (1School of Social Work, Wayne State University, Detroit, MI; 2Center for the Study of Drugs, Alcohol, Smoking and Health, Department of Health Behavior and Biological Sciences, School of Nursing, University of Michigan, Ann Arbor, MI), Carol J. Boyd (2Center for the Study of Drugs, Alcohol, Smoking and Health, Department of Health Behavior and Biological Sciences, School of Nursing, University of Michigan, Ann Arbor, MI; 3Institute for Research on Women and Gender, University of Michigan, Ann Arbor, MI; 4Addiction Center, Department of Psychiatry, Rachel Upjohn Building, University of Michigan, Ann Arbor, MI; 7Center for Sexuality and Health Disparities, School of Nursing, University of Michigan, Ann Arbor, MI), Rebecca J. Evans-Polce (2Center for the Study of Drugs, Alcohol, Smoking and Health, Department of Health Behavior and Biological Sciences, School of Nursing, University of Michigan, Ann Arbor, MI), Phil Veliz (2Center for the Study of Drugs, Alcohol, Smoking and Health, Department of Health Behavior and Biological Sciences, School of Nursing, University of Michigan, Ann Arbor, MI; 3Institute for Research on Women and Gender, University of Michigan, Ann Arbor, MI; 5Institute for Social Research, University of Michigan, Ann Arbor, MI), Curtiss Engstrom (5Institute for Social Research, University of Michigan, Ann Arbor, MI), Brady T. West (2Center for the Study of Drugs, Alcohol, Smoking and Health, Department of Health Behavior and Biological Sciences, School of Nursing, University of Michigan, Ann Arbor, MI; 5Institute for Social Research, University of Michigan, Ann Arbor, MI; 6Institute for Healthcare Policy and Innovation, University of Michigan, Ann Arbor, MI), Sean Esteban McCabe (2Center for the Study of Drugs, Alcohol, Smoking and Health, Department of Health Behavior and Biological Sciences, School of Nursing, University of Michigan, Ann Arbor, MI; 3Institute for Research on Women and Gender, University of Michigan, Ann Arbor, MI; 5Institute for Social Research, University of Michigan, Ann Arbor, MI; 6Institute for Healthcare Policy and Innovation, University of Michigan, Ann Arbor, MI; 7Center for Sexuality and Health Disparities, School of Nursing, University of Michigan, Ann Arbor, MI; 8Rogel Cancer Center, University of Michigan, Ann Arbor, MI)
Categories: Article, discrimination, substance use, epidemiology, health disparities, LGBT health, prejudice, sexual orientation
Source: Journal of homosexuality
Authors: Luisa Kcomt, Carol J. Boyd, Rebecca J. Evans-Polce, Phil Veliz, Curtiss Engstrom, Brady T. West, Sean Esteban McCabe
We aimed to assess the probability of past-year DSM-5 alcohol use disorder (AUD) and past-year moderate-to-severe DSM-5 AUD as a function of past-year ethnic discrimination among U.S. Latino/Hispanic adults and as a function of past-year discrimination types among Latino/Hispanic sexual minorities (SM). We used data from the 2012-2013 National Epidemiological Survey on Alcohol and Related Conditions-III (n=36,309 U.S. adults aged ≥18 years). Our sample consisted of 6,954 Latino/Hispanic adults. Multivariable logistic regression analyses estimated the association of past-year ethnic discrimination with past-year AUD and past-year moderate-to-severe AUD among the overall Latino/Hispanic population. We tested the association between discrimination types (i.e., none, ethnic or sexual orientation discrimination, both) with AUD and moderate-to-severe AUD among Latino/Hispanic SM adults. Among Latino/Hispanic adults, 13.4% met criteria for past-year AUD and 6.4% met criteria for past-year moderate-to-severe AUD. Ethnic discrimination was significantly associated with AUD (AOR=1.09, 95% CI=1.07-1.12) and moderate-to-severe AUD (AOR=1.10, 95% CI=1.06-1.13). Latino/Hispanic SMs who experienced both ethnic and sexual orientation discrimination were more likely to meet criteria for AUD (AOR=3.44, 95% CI=1.97-6.03) and moderate-to-severe AUD (AOR=2.69, 95% CI=1.25-5.82) compared to those who did not experience discrimination. Discrimination is a risk factor for AUD and moderate-to-severe AUD among Latino/Hispanic overall and SM populations.
Harmful alcohol use contributes significantly to the global burden of disease and can result in substantial adverse health outcomes (Rehm et al., 2009; World Health Organization, 2018). Studies have demonstrated causal relationships between alcohol consumption and over 200 health conditions (World Health Organization, 2018). More than 65 million U.S. adults meet criteria for a lifetime alcohol use disorder (AUD; Grant et al., 2015b), with sexual minorities carrying a disproportionate burden of alcohol misuse and AUD across the life course (Fish et al., 2018; Gonzales et al., 2016; Gonzales & Henning-Smith, 2017; McCabe et al., 2019). Understanding the risk factors for AUD among sexual minorities who also identify as racial/ethnic minorities is a public health priority and can inform prevention and intervention services tailored to those communities (Institute of Medicine, 2011; National Academies of Sciences, Engineering, and Medicine, 2020).
Latino/Hispanic individuals are currently the largest minority group (18.5%) in the U.S., comprising of over 60 million people (United States Census Bureau, 2019), and this population share is projected to increase to 28.6% by the year 2060 (United States Census Bureau, 2015). They are among the fastest growing populations in the U.S., second to Asian Americans (Johnson & Lichter, 2016; Noe-Bustamante et al., 2020). Furthermore, although the number of U.S. individuals who identify as a sexual or gender minority have risen overall, Latino/Hispanics have seen the greatest increase since 2012. An estimated 6.1% of all Latino/Hispanic people identify as a sexual or gender minority, the highest prevalence by race/ethnicity (Newport, 2018).
Studies have shown that psychological, social, and cultural factors influence alcohol misuse, AUD, and substance use behaviors among U.S. Latino/Hispanic individuals (Blackson et al., 2015; Molina et al., 2016; Otiniano Verissimo et al., 2014; Savage & Mezuk, 2014). Acculturation—conceptualized as a multidimensional process of integrating the beliefs, values, and practices of the host society as a form of adaption (Siatkowski, 2007)—is positively associated with alcohol and substance use disorders in U.S. Latino/Hispanic populations (Blanco et al., 2013; Lui & Zamboanga, 2018; Savage & Mezuk, 2014; Unger et al., 2014). The risk of alcohol misuse and AUD increases with higher levels of acculturation to U.S. society (Lui & Zamboanga, 2018; Savage & Mezuk, 2014; Vaeth et al., 2012). English-language use/proficiency has been used as an indicator of acculturation (Siatkowski, 2007) and found to be positively associated with alcohol use behaviors and AUD among Latino/Hispanic people (Alvarez et al., 2017; Cano, 2020; Lui & Zamboanga, 2018) because of its utility as a key point of access into U.S. society.
Nativity is another commonly used demographic indicator of acculturation (Otiniano Verissimo et al., 2014; Pérez et al., 2008; Vaeth et al., 2012), with U.S.-born Latino/Hispanic individuals being more likely to adopt the practices of the U.S. culture relative to Latino/Hispanic immigrants (Schwartz et al., 2006). Thus, nativity may play a role in the risk for alcohol and substance use. The prevalence of lifetime alcohol use is lower among Latino/Hispanic immigrants (77%-82%) compared to their U.S.-born counterparts (90%-92%) (Borges et al., 2011). This is consistent with other research showing Latino/Hispanic immigrants having lower lifetime prevalence and lifetime risk for substance use disorders than U.S.-born respondents (Breslau et al., 2007). Research suggests that parental history of substance use problems in the country of origin increases young adult Latino/Hispanic immigrants’ risk for alcohol misuse during pre- and post-immigration (Blackson et al., 2015). These studies have produced important insights about alcohol use among Latino/Hispanic populations, but further work is needed to examine potential risk factors for greater AUD severity among Latino/Hispanic sexual minorities using DSM-5 criteria.
There is documented evidence that social forces such as discrimination can be a driver of alcohol use. A systematic review by Gilbert and Zemore (2016) revealed a positive association between experiences of discrimination and alcohol-related outcomes. Racial/ethnic discrimination increases the risk for mild, moderate, and severe AUD among racial/ethnic minority populations in the U.S. (Glass et al., 2020). Among U.S. Latino/Hispanic people, ethnic discrimination is associated with an elevated risk of alcohol and substance use (Cheng & Mallinckrodt, 2015; Tran et al., 2010; Unger et al., 2014; Vaeth et al., 2017), with variations based on sex (Molina et al., 2016; Otiniano Verissimo et al., 2013, 2014), as well as alcohol-related problem behaviors (e.g., physical fighting when or right after drinking; Salas-Wright et al., 2018).
Studies have also established links between sexual orientation discrimination and increased risk for alcohol-related outcomes among sexual minority individuals (Lehavot & Simoni, 2011; McCabe et al., 2019; Slater et al., 2017). Prior research has also established that individuals who were not sure about their sexual identity had increased risk for severe AUD (Boyd et al., 2019). These studies support the Minority Stress Model (Meyer, 1995, 2003), which posits that in addition to everyday life stress, individuals who are members of a marginalized social group may experience stigma, prejudice, and discrimination related to their minority identity. Minority stress processes include exposure to harassment, victimization and harassment; expectations for these events to occur, thereby producing vigilance; internalization of negative societal attitudes towards same-sex attraction or behavior; and identity concealment. Exposure to high levels of minority stress may contribute to alcohol misuse, which in turn can result in negative health outcomes.
Although there is documented evidence of the relationship between discrimination and alcohol-related outcomes among Latino/Hispanic and sexual minorities, respectively, researchers have yet to examine how ethnic and sexual orientation discrimination can influence, individually and collectively, the risk for AUD and AUD severity among Latino/Hispanic sexual minorities while accounting for acculturation factors. Thus, the aims of our study are (1) examine the differences between sexual minority and heterosexual Latino/Hispanic men and women in their prevalence of Latino/Hispanic ethnic discrimination; (2) compare Latino/Hispanic sexual minority men and women in their prevalence of sexual orientation discrimination; (3) assess the probability of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD as a function of Latino/Hispanic ethnic discrimination; and (4) assess the probability of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD as a function of past-year discrimination types experienced by Latino/Hispanic sexual minorities.
The 2012-2013 National Epidemiologic Survey on Alcohol-Related Conditions-III (NESARC-III) collected data from a nationally representative sample of the general civilian noninstitutionalized population of U.S. adults ages 18 years and older (n=36,309). The Alcohol Use Disorder and Associated Disabilities Interview Schedule-5 (AUDADIS-5) is a structured diagnostic interview and was used to conduct in-person interviews in households. The household response rate was 72%, the person response rate was 84%, resulting in an overall response rate was 60.1%. The NESARC-III study design is described in more detail elsewhere; all procedures received full human subjects review and institutional review board approval (Grant, Chu, et al., 2015). Our study sample was restricted to Latino/Hispanic adults (n=7,037).
DSM-5 alcohol use disorder (AUD) was assessed according to criteria of the DSM-5 using AUDADIS-5. Consistent with the DSM-5, a past-year AUD diagnosis was based on the presence of at least 2 of the 11 symptoms associated with DSM-5 AUD and excludes symptoms that are associated with mental illness and other substances. Past-year moderate-to-severe AUD was defined as the presence of 4 or more AUD symptoms (American Psychiatric Association, 2013; Hasin et al., 2015). Reliability and validity of the DSM-based diagnoses of AUD have been examined previously (Grant, Goldstein, et al., 2015; Hasin et al., 2015). DSM-5 AUD criteria scales demonstrated excellent reliability (intraclass correlation coefficient [ICC] = 0.9) in a large population sample (Grant, Goldstein, et al., 2015).
Discrimination was based on questions from the Experiences with Discrimination scale (Krieger et al., 2005; Krieger & Sidney, 1996). The past-year Latino/Hispanic ethnic discrimination measure assessed six different types of discrimination that respondents may have experienced because of their Latino/Hispanic ethnicity (e.g., obtaining health care, receiving health care, obtaining a job, applying to school, interacting with police, public locations, verbal or physical aggression). The range of responses for each item was never (0) to very often (4). A scale was created by summing responses to the six items for discrimination ranging from 0 to 24. The past-year Latino/Hispanic ethnic discrimination scale has good reliability (α=0.74; Ruan et al., 2008). A similar past-year sexual orientation discrimination scale was created based on six items assessing the same types of discrimination as the Latino/Hispanic ethnic discrimination scale (McCabe et al., 2010; Ruan et al., 2008). The past-year sexual orientation discrimination scale has excellent reliability (α=0.88; McCabe et al., 2019). We created a categorical past-year discrimination measure consisting of the following none, one type (experienced either ethnic discrimination or sexual orientation discrimination) or two types (experienced both ethnic and sexual orientation discrimination), consistent with prior work (McCabe et al., 2010). We also merged the two discrimination scales to form the past-year Latino/Hispanic ethnic and sexual orientation discrimination scale, ranging from 0 to 48, which had high reliability in this study (α=0.902).
Sexual Orientation consisted of three identity, attraction, and behavior. Respondents were asked which term best described heterosexual (straight), gay/lesbian, bisexual, or not sure. Sexual attraction was assessed by asking respondents to describe their attraction to other only attracted to females, mostly attracted to females, equally attracted to both sexes, mostly attracted to males, and only attracted to males. Respondents were asked about their sexual only with different sex, with both sexes, and never had sex. Sexual orientation subgroups were created based on prior research (Drabble et al., 2005; McCabe et al., 2019; Trocki et al., 2009) by combining the three sexual orientation dimensions into the following five mutually exclusive sexual orientation (1) lesbian/gay-identified, (2) bisexual-identified, (3) unsure of their sexual identity, (4) heterosexual-identified with same-sex attraction and/or behavior, and (5) heterosexual-identified with different-sex attraction or behavior. Sexual minorities were defined as lesbian/gay-, bisexual-, or heterosexual-identified individuals with same-sex attraction and/or behavior, and individuals who were not sure about their sexual identity.
Sociodemographic/background characteristics included sex (male/female), age (18-24, 25-44, 45-64, and ≥65), educational status (high school or less, some college, and college degree or higher), urbanicity (rural/urban), and geographic region (Northeast, Midwest, South, or West). Parental history of drug or alcohol problems was assessed by asking respondents whether a biological parent had any lifetime history of alcohol or drug problems (Ruan et al., 2008).
Acculturation factors consisted of English-language preference and nativity. The NESARC-III used the Bidimensional Acculturation Scale (BAS; Marin and Gamba, 1996) to measure respondents’ language use, proficiency, and media consumption. Respondents received 24 questions regarding their use of English and Spanish, respectively, with each item scored on a 4-point Likert scale. Consistent with prior research (Cano, 2020), we used the English portion of the BAS (hereafter referred to as the English-language preference scale) as a measure of acculturation. We summed the scores to the 12 items in the English-language preference scale and treated it as a continuous variable with a possible range of 0-36. Internal consistency was strong in this sample (α=0.976). Nativity was assessed by asking respondents the country in which they were born and was treated as a dichotomous variable (U.S.- or foreign-born), similar to prior research (Otiniano Verissimo et al., 2014; Pérez et al., 2008).
All statistical analyses performed in this study were design-based in nature, fully accounting for the NESARC-III survey weights in point estimation to ensure representative, unbiased estimation of parameters of interest in the finite NESARC-III target population, and fully accounting for the survey weights, stratum codes, and cluster codes in variance estimation (using Taylor Series Linearization) to ensure unbiased estimates of sampling variance for all point estimates. We employed commands in the Stata 16 software for data management and analysis, specifically using the various commands for the design-based analyses, and used appropriate options in these commands for unconditional subpopulation analyses throughout.
Initial analyses compared Latino/Hispanic men and women in their sexual orientation, sociodemographics, background characteristics, and past-year DSM-5 AUD. We next examined differences between Latino/Hispanic sexual minorities and heterosexuals in their experience of past-year discrimination based on Latino/Hispanic ethnicity. For the subgroup of sexual minority Latino/Hispanics, we compared men and women in their prevalence of experiencing past-year sexual orientation discrimination. These analyses employed design-adjusted Rao-Scott tests of the null hypothesis that no difference in distributions existed between Latino/Hispanic individuals based on their sexual orientation and between Latino/Hispanic sexual minority men and women, respectively (Rao & Scott, 1984; Rao & Thomas, 1988). Given the number of tests performed in these descriptive analyses, only test results significant at p < 0.001 were considered meaningful.
Multivariable analyses consisted of logistic regression models to assess the probability of a past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD as a function of the Latino/Hispanic ethnic discrimination scale adjusting for sex, age, educational attainment, urbanicity, geographic region, any parental history of drug or alcohol abuse, nativity, and English-language preference. For Latino/Hispanic sexual minorities, we fitted logistic regression models to assess the probability of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD as a function of past-year discrimination (none, 1 type, or 2 types). In this model, we also explored two-way interactions between sex and discrimination to examine whether the relationship of discrimination with each outcome varied by sex. We conducted additional analyses to determine the probability of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD as a function of the past-year Latino/Hispanic ethnic and sexual orientation discrimination scale. We explored two-way interactions between sex and past-year Latino/Hispanic ethnic and sexual orientation discrimination. We plotted marginal predicted probabilities and means (and their design-adjusted 95% CIs) on the outcome variables as a function of discrimination based on the fitted models using the “margins” and “marginsplot” commands in Stata 16. Lastly, we conducted supplemental analyses by assessing the probability of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD as a function of a 4-category past-year discrimination measure (none, Latino/Hispanic discrimination only, sexual orientation discrimination only, and Latino/Hispanic and sexual orientation combined) and explored two-way interactions between sex and the 4-category past-year discrimination measure. We also completed supplemental analyses using past-year Latino/Hispanic discrimination and past-year sexual orientation discrimination as separate continuous measures included into the models simultaneously and tested the interaction between these two types of discrimination in each supplemental model.
Table 1 presents estimated descriptive quantities for the sociodemographic features and key study measures for the overall Latino/Hispanic population and stratified by sex and sexual orientation, respectively. Sexual minority orientation was rare overall in the Latino/Hispanic population, with less than 4% of individuals identifying as lesbian/gay, bisexual, or being unsure of their status. Among the overall Latino/Hispanic population, an estimated 13.4% met criteria for past-year DSM-5 AUD and 6.4% met criteria for past-year moderate-to-severe DSM-5 AUD. Latino/Hispanic men had a higher prevalence of past-year DSM-5 AUD (17.4%) and past-year moderate-to-severe DSM-5 AUD (8.6%) compared to Latino/Hispanic women (9.0% and 4.1%, respectively). There were no significant differences between Latino/Hispanic men and women in their mean amount of past-year ethnic discrimination they experienced (1.8 and 1.7, respectively). The prevalence of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD was higher among Latino/Hispanic sexual minorities (20.9% and 10.2%, respectively) than Latino/Hispanic heterosexuals (12.4% and 6.0%, respectively). There were no significant differences between Latino/Hispanic sexual minorities and heterosexuals in their mean amount of past-year ethnic discrimination they experienced (1.9 and 1.7, respectively).
Four in ten Latino/Hispanic individuals experienced ethnic discrimination in the past year. Ethnic discrimination in public places was most prevalent in the overall Latino/Hispanic subgroup and for sexual minority men and women. Sexual minority women were more likely to report being called names (19.8%) compared to heterosexual women (12.6%) (P<0.001), whereas sexual minority men were more likely to be bullied or assaulted (17.5%) compared to heterosexual men (8.4%) (P<0.01). No other significant differences in individual experiences or frequency of discrimination situations were found between Latino/Hispanic sexual minority and heterosexual men and women (Supplemental Table 1).
Among Latino/Hispanic sexual minority adults, almost one in five experienced sexual orientation discrimination in the past year. And like ethnic discrimination, sexual orientation discrimination in public places was the most prevalent experience among the overall sexual minority subgroup. We found no differences between men and women in their distributions for individual experiences and frequency of sexual orientation discrimination situations (Supplemental Table 2). Almost four in ten (38.7%) Latino/Hispanic sexual minority adults experienced one type of past-year discrimination (either ethnic or sexual orientation discrimination) and one in ten (10.3%) experienced both types of discrimination (Supplemental Table 3). Among Latino/Hispanic sexual minorities who did not experience any ethnic or sexual orientation discrimination, 16.4% met criteria for a past-year DSM-5 AUD (data not shown).
Table 2 presents adjusted, marginal predicted probabilities of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD as a function of Latino/Hispanic ethnic discrimination for the overall Latino/Hispanic population and stratified by sexual orientation. These results show evidence of a positive relationship between Latino/Hispanic ethnic discrimination and AUD. Among the overall Latino/Hispanic population, a one unit increase in the Latino/Hispanic ethnic discrimination scale was associated with a 9% increase in the odds of past-year DSM-5 AUD (AOR=1.09, 95% CI=1.07, 1.12) and a 10% increase in odds of past-year moderate-to-severe DSM-5 AUD (AOR=1.10, 95% CI=1.06, 1.13). We tested for interactions between sexual orientation discrimination and Latino/Hispanic ethnic discrimination; results were not significant.
When stratifying by sexual orientation, we found the association of Latino/Hispanic ethnic discrimination with past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD was significant for both heterosexual and sexual minority Latino/Hispanic individuals. In each sexual orientation subgroup, individuals who experienced ethnic discrimination had increased odds of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD relative to counterparts who did not experience discrimination. Importantly, sexual minorities who experienced Latino/Hispanic ethnic discrimination had higher predicted probabilities of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD relative to their heterosexual counterparts (Figure 1). Lastly, nativity and English language preference as acculturation factors were both significantly associated with past-year DSM-5 AUD and moderate-to-severe DSM-5 AUD in the overall Latino/Hispanic population. Foreign-born Latino/Hispanic individuals had decreased odds of past-year DSM-5 AUD (AOR=0.73, 95% CI=0.57, 0.92) and past-year moderate-to-severe DSM-5 AUD (AOR=0.66, 95% CI=0.49, 0.90) relative to their U.S.-born counterparts. Individuals with an English language preference were more likely to have past-year DSM-5 AUD (AOR=1.04, 95% CI=1.03, 1.05) and past-year moderate-to-severe DSM-5 AUD (AOR=1.03, 95% CI=1.01, 1.05).
Table 3 presents adjusted predicted probabilities of past-year DSM-5 AUD and past-year moderate-to severe DSM-5 AUD among Latino/Hispanic sexual minorities, respectively, as a function of the number of past-year discrimination types experienced (none, either ethnic discrimination or sexual orientation discrimination, or both types of discrimination). Latino/Hispanic sexual minorities who experience both types of discrimination compared to those who do not experience discrimination were more likely to meet criteria for both past-year DSM-5 AUD (AOR=3.44, 95% CI=1.97, 6.03) and past-year moderate-to-severe DSM-5 AUD (AOR=2.69, 95% CI=1.25, 5.82) (see Supplemental Figure 1). We tested two-way interactions between sex and number of past-year discrimination types; results were not significant.
Table 4 presents adjusted, marginal predicted probabilities of past-year DSM-5 AUD and past-year moderate-to severe DSM-5 AUD, respectively, as a function of combined Latino/Hispanic ethnic and sexual orientation discrimination among Latino/Hispanic sexual minorities. The results show that a one unit increase in the Latino/Hispanic ethnic and sexual orientation discrimination scale increased the odds of past-year DSM-5 AUD by approximately 8% (AOR=1.08, 95% CI=1.04, 1.13) and the odds of a past-year moderate-to-severe DSM-5 AUD in the same fashion (AOR=1.08, 95% CI=1.02, 1.14) (See Supplemental Figure 2). We tested two-way interactions between sex and the combined discrimination scale and the interactions were not significant.
Results from our supplemental multivariable logistic regression models using the 4-category past-year discrimination measure showed that Latino/Hispanic sexual minorities who experienced Latino/Hispanic ethnic and sexual orientation discrimination had increased risk of past-year DSM-5 AUD (AOR=3.45, 95% CI=1.97, 6.05) and past-year moderate-to-severe AUD (AOR=2.69, 95% CI=1.24, 5.83) (Supplemental Table 4). We tested the two-way interaction between sex and the 4-category past-year discrimination measure and the interaction was not significant. Results from our supplemental multivariable analyses using the discrimination scales as separate continuous measures included simultaneously in the models showed that Latino/Hispanic ethnic discrimination was significantly associated with past-year DSM-5 AUD (AOR=1.09, 95% CI=1.02, 1.16) and past-year moderate-to-severe AUD (AOR=1.14, 95% CI=1.05, 1.25) (Supplemental Table 5). We tested the two-way interaction between past-year Latino/Hispanic discrimination and past-year sexual orientation discrimination in both models and this interaction was not significant.
Although ethnic discrimination is legally prohibited in the U.S., the prevalence of past-year ethnic discrimination among U.S. Latino/Hispanic adults was high (40%). Similar to prior research on alcohol use among Latino/Hispanic populations (Tran et al., 2010; Unger et al., 2014), we found that Latino/Hispanic people who experienced past-year ethnic discrimination had increased risk for past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD, thus providing support for the Minority Stress Model. Latino/Hispanic sexual minorities who experienced ethnic discrimination had higher risk for past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD relative to their heterosexual counterparts.
Our study revealed that one in five Latino/Hispanic sexual minority adults experienced past-year sexual orientation discrimination and one in ten experienced both ethnic and sexual orientation discrimination in the past year. We extend the current body of literature with the finding that discrimination based on ethnicity and sexual orientation can collectively influence Latino/Hispanic sexual minorities’ AUD risk. Latino/Hispanic sexual minority adults who did not experience any discrimination had a prevalence of past-year DSM-5 AUD (16.4%), which was similar to that of the U.S. general population (14%; Grant et al., 2015b). Higher levels of discrimination were associated with increased risk for past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD. Moreover, Latino/Hispanic sexual minority adults who experienced both types of discrimination were over three times more likely to meet criteria for past-year DSM-5 AUD and over two and a half times more likely to have past-year moderate-to-severe DSM-5 AUD, relative to those who did not experience discrimination. This suggests that individuals located at the intersection of multiple marginalized identities can experience distinct risks associated with discrimination (Bowleg, 2012).
Latino/Hispanic sexual minorities may experience compound marginalization resulting from complex, cumulative systems of disadvantage (National Academies of Sciences, Engineering, and Medicine, 2020). Their experiences of ethnic discrimination cannot be divorced from their experiences of sexual orientation discrimination. Most importantly, social forces such as discrimination drive their AUD disparities—not intrinsic personal characteristics related to their ethnicity or sexual orientation. Understanding these influencing factors and differential risks can inform AUD prevention and intervention strategies for Latino/Hispanic sexual minorities. Health providers and substance use treatment specialists should be cognizant of the minority stressors experienced by their clients and consider their impact before, during, and after treatment.
Discrimination in public places was the most prevalent situational context for both types of discrimination among the participants in our study. Importantly, public accommodation non-discrimination laws protecting sexual minorities are not universal at present (Movement Advancement Project, 2018). Thus, sexual minorities in many states remain vulnerable to receiving unfair treatment (e.g., refusal of service or entry) or discrimination in places accessible to the public. Prior research has demonstrated that exposure to sexual stigma (i.e., heterosexism) can produce negative affect and alcohol cravings (Mereish & Miranda, 2019). Structural and interpersonal discrimination can have deleterious effects on people in marginalized social groups (Hatzenbuehler, 2016). Our study underscores the ongoing need to create safe spaces for ethnic minorities, sexual minorities, and individuals who live at those intersections.
Our study found that higher levels of acculturation increased the risk of past-year DSM-5 AUD in our overall study population, consistent with prior research about acculturation and alcohol use among Latino/Hispanic populations (Lui & Zamboanga, 2018; Savage & Mezuk, 2014; Vaeth et al., 2012). A new contribution from our study is that higher levels of acculturation is also associated with past-year moderate-to-severe DSM-5 AUD. While our study used acculturation factors as control variables, other research found that acculturation factors can influence perceived ethnic discrimination among Latino/Hispanic populations (Pérez et al., 2008). Highly acculturated Latino/Hispanic individuals (i.e., those who were U.S.-born or emigrated at a younger age) were more likely to report perceived ethnic discrimination, whereas a strong ethnic identity had a protective effect against perceptions of discrimination. Future studies should examine whether acculturation moderates the relationship between discrimination and AUD among Latino/Hispanic sexual minorities.
Several limitations warrant mention. The cross-sectional nature of the study prevented us from making any causal inferences. The items related to alcohol use relied on self-report, which could influence the prevalence of AUD outcomes due to social desirability. We did not account for polysubstance use in this study. The discrimination measures did not assess the severity of the discrimination experienced and focused primarily on interpersonal but not structural discrimination. Protection against sexual orientation discrimination are not universal at the federal and state levels and thus, our findings may under-represent the scope of discrimination experienced by sexual minorities and its association with AUD. We did not account for other acculturation factors, such as subgroup differences by Latino/Hispanic ethnicity, in our study. Prior research has revealed that the association between discrimination and AUD (Otiniano Verissimo et al., 2014) and the reporting of discrimination (Pérez et al., 2008) varies across Latino/Hispanic ethnic subgroups. Future research should examine these cultural variations in the risk for AUD among Latino/Hispanic sexual minorities.
The NESARC-III does not assess other important measures, such as gender identity. Sexual identity labels are dynamic and has since evolved; the operationalization of sexual minorities in our study do not represent other sexual identities that exist (e.g., pansexual, asexual, etc.). As is common in large, population-based surveys, the prevalence of adults with sexual minority orientation was small and thus, the Latino/Hispanic sexual minority subgroups (lesbian/gay-, bisexual-, or heterosexual-identified individuals with same-sex attraction and/or behavior, and individuals who were not sure about their sexual identity) were combined into one category. Future research should examine whether there is heterogeneity in the association between experiences of discrimination and AUD among Latino/Hispanic sexual minority subgroups. In addition, policy changes have occurred at the structural level that may affect sexual minorities’ experiences of interpersonal discrimination. Given the current socio-political climate in the U.S., the past-year ethnic discrimination measure may not represent the frequency of ethnic discrimination experienced by Latino/Hispanic individuals in recent years. Further intersectional analyses of Latino/Hispanic populations using more recent data are warranted.
Our study highlighted the additivity of oppressions experienced by Latino/Hispanic sexual minority individuals through the use of the combined past-year Latino/Hispanic ethnic and sexual orientation discrimination scale. Although we report the reliability of this scale in our study, there are no data that affirm the validity of this combined measure. Furthermore, intersectionality scholars have cautioned against the assumption of additivity with the critique that the excess risk of negative health and social outcomes do not necessarily increase linearly with each additional marginalized social identity or position (Scheim & Bauer, 2019). Given the dearth of literature on Latino/Hispanic sexual minority populations, further analytic examination of the optimal intersectional discrimination measures and the impacts of intersectional discrimination is warranted.
Despite the limitations, the strengths of this study include the use of a nationally representative, probability-based sample of Hispanic adults in the U.S. The NESARC-III includes validated measures of AUD based on DSM-5 criteria as well as discrimination experiences. It also assesses multiple dimensions of sexual orientation, thus enabling us to include heterosexual-identified sexual minorities and individuals who were not sure about their sexual identity in our study.
This study found that ethnic discrimination is associated with higher risk of past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD in the overall Latino/Hispanic population. As individuals with multiple marginalized identities, Latino/Hispanic sexual minority adults experience distinct AUD risks associated with discrimination due to complex systems of disadvantage. Those who experienced both ethnic and sexual orientation discrimination were more likely to meet criteria for past-year DSM-5 AUD and past-year moderate-to-severe DSM-5 AUD, relative to those who did not experience discrimination. Continued efforts are needed to create safe spaces for minority populations and to eliminate discrimination in our social systems.