Examining Disparities in Excessive Alcohol Use Among Black and Hispanic Lesbian and Bisexual Women in the United States: An Intersectional Analysis

Abstract

Objective:

Sexual minority (i.e., lesbian, bisexual) women and racial-ethnic minority groups in the United States are disproportionately harmed by excessive alcohol use. This study examined disparities in excessive alcohol use at the intersection of race-ethnicity and sexual identity for non-Hispanic Black and Hispanic sexual minority women.

Method:

Using data from the 2015 National Survey on Drug Use and Health, we compared the age-adjusted prevalence of binge drinking and heavy alcohol use among sexual minority women of color, sexual minority White women, and heterosexual women of color with that of White heterosexual women. The joint disparity is the difference in the prevalence of excessive alcohol use between sexual minority women of color and White heterosexual women. The excess intersectional disparity is the portion of the joint disparity that is due to being both a racial-ethnic minority and a sexual minority woman.

Results:

Black and Hispanic sexual minority women reported the highest prevalence of binge drinking (45.4% and 43.4%, respectively), followed by White sexual minority women (35.7%) and White heterosexual women (23%). Black and Hispanic heterosexual women reported the lowest prevalence of binge drinking (20.8% and 20.2%, respectively). The joint disparity in binge drinking between Black sexual minority women and White heterosexual women was 21.2%, and the excess intersectional disparity was 17.7%. The joint disparity in binge drinking between Hispanic sexual minority women and White heterosexual women was 16.8%, and the excess intersectional disparity was 10.8%.

Conclusions:

Disparities in excessive alcohol consumption for Black and Hispanic sexual minority women, compared with White heterosexual women, were larger than what would be expected when considering differences by race or sexual identity individually.

Excessive alcohol consumption is the third leading cause of death attributable to modifiable behaviors in the United States (Mokdad et al., 2004). Excessive alcohol consumption is defined as binge drinking (4+/5+ drinks on an occasion for women/men), heavy alcohol use (8+/15+ drinks weekly for women/men), or any drinking by those under 21 or pregnant people (Centers for Disease Control and Prevention, n.d.; National Institute on Alcohol Abuse and Alcoholism, n.d.). Excessive alcohol consumption contributes to acute and chronic health issues including motor vehicle fatalities, coronary heart disease, and cancer (Blimcoe et al., 2015; Connor, 2017; Zhao et al., 2017). Besides population-level consequences, excessive alcohol use is associated with a higher risk for depression (Bellos et al., 2016) and contributes to firearm violence, homicide, and suicide death (Branas et al., 2016; Cherpitel et al., 2004; Fowler et al., 2018). Notwithstanding the mortality and morbidity, excessive alcohol consumption costs the United States approximately $223.5 billion in lost productivity, health care costs, and criminal justice costs (Bouchery et al., 2011).

Excessive alcohol consumption and alcohol-related harms severely affect members of marginalized groups, specifically sexual minority populations. Sexual minority populations include individuals who identify as gay, lesbian, or bisexual. Lesbian and bisexual (LB) women are more likely to engage in excessive alcohol consumption compared with their heterosexual peers (Bauer et al., 2010; Fish et al., 2018; Gonzales & Henning-Smith, 2017; Hughes, 2003; Medley et al., 2016; Veldhuis et al., 2017). Using data from the Behavioral Risk Factor Surveillance System in 27 states, Gonzales and Henning-Smith (2017) found that LB women had 49% and 84% higher odds of binge drinking compared with heterosexual women. Data from the National Survey of Family Growth show that an estimated 52% of women with female sex partners reported binge drinking compared with 11% of exclusively heterosexual women (Bauer et al., 2010).

Black and Hispanic populations are less likely to consume alcohol compared with White populations in the United States (Delker et al., 2016; Mulia et al., 2009). But these populations are more likely to experience alcohol dependence and negative consequences from drinking (Mulia et al., 2009). A systematic review found that, for both sexual minority and racial-ethnic minority populations, the increased prevalence of harmful alcohol use and alcoholrelated harms may be due to experiences of prejudice and discrimination through the excess stress produced by these experiences (Gilbert & Zemore, 2016). Despite increased attention to disparities in excessive alcohol consumption and related harms by sexual identity or race-ethnicity, few studies have contrasted disparities across the intersections of these identities. Such studies may shed light on the relative contributions of racism, homophobia, and their potential synergistic effects on alcohol use.

Intersectionality is a social theory that posits that individual social categories of difference cannot be fully understood apart from other categories (Bowleg, 2012). Sexual identity, a facet of sexual orientation, cannot be understood separately from sex, gender, gender identity, class, or race, because the mechanisms of homophobia work through sexism, heterosexism, classism, and racism to produce inequalities among individuals and populations. Using an intersectional framework challenges the current paradigm in which researchers prioritize one part of an individual’s identity, making it the focal point around which all other identities are based (Hankivsky, 2012). Instead, intersectionality allows for the complexity of intersecting social locations as the starting point for understanding health disparities (Bowleg, 2012). Bowleg asserts that, although all persons have multiple identities, intersectional frameworks are first and foremost concerned with those persons and groups that have been “historically oppressed and marginalized” (Bowleg, 2012). This is because the mechanisms that drive health disparities are often intricately linked with historical and political processes that have not valued diversity or equity.

Previous studies have used a variety of statistical methods when applying intersectionality theory to health outcomes. One approach has been to construct logistic regression models with interaction terms (Agénor et al., 2014; Mereish & Bradford, 2014). Other approaches use intersecting social identities to explain variation in population-wide outcomes (Evans et al., 2018). Another approach, concerned with documenting and understanding intersectional inequalities, is to define intersectional disparity measures and compare their magnitude (Jackson, 2017; Jackson et al., 2016). Additive measures are preferred for assessing public health burden (Rothman et al., 2015). We sought to describe and understand absolute disparities in specific patterns of excessive alcohol consumption, specifically binge drinking and heavy alcohol use among women in the United States living at the intersection of sexual minority and racial-ethnic minority identity because they experience both homophobia and racism.

Method

Study design

The 2015 National Survey on Drug Use and Health (NSDUH) provided the data for this analysis. NSDUH is a nationally representative survey conducted annually by the Substance Abuse and Mental Health Services Administration (2015). Using complex sampling methods, NSDUH collects data on approximately 70,000 U.S. residents 12 years and older sampled from the noninstitutionalized civilian population. For this analysis, we included participants who identified as women and were age 18 years and older. Our analytic sampled consisted of 21,024 individuals. NSDUH asks participants several questions about alcohol, tobacco, and illicit drug use as well as mental health. Interviews are conducted using an audio computer-assisted self-interview to ensure complete privacy and anonymity, given the sensitive nature of the survey questions. Participants are provided $30 for their time (Substance Abuse and Mental Health Services Administration, 2015). (All monetary amounts are in U.S. dollars.) The weighted response rate for 2015 was 69.25% (Center for Behavioral Health Statistics and Quality, 2016). These survey weights account for nonresponse bias and nonsampling error (Center for Behavioral Health Statistics and Quality, 2016).

Measures

In 2015, NSDUH began measuring sexual identity (Medley et al., 2016) with the question: “Which one of the following do you consider yourself to be?” Response options included, “heterosexual, that is straight,” “lesbian or gay,” or “bisexual.” Participants were asked several questions about their drinking behaviors. Among women, current binge drinking was defined as four or more drinks in about 2 hours within the past 30 days, in accordance with the definition from the National Institute on Alcohol Abuse and Alcoholism (n.d.). Heavy alcohol use was defined as binge drinking on five or more occasions within the past 30 days, in accordance with the Substance Abuse and Mental Health Services Administration (National Institute on Alcohol Abuse and Alcoholism, n.d.). Race-ethnicity was assessed with two questions. The first question asked participants whether they identify as “Hispanic, Latino, or of Spanish origin.” The second question asked participants to identify their race as “White,” “Black or African American,” “American Indian or Alaska Native,” “Native Hawaiian or other Pacific Islander,” “Asian,” or “other.” Participants who selected multiple races were grouped into one “multiple races” category. Among women who identified as American Indian/Alaska Native, Native Hawaiian, or other Pacific Islander, there were fewer than 100 people who identified as lesbian or bisexual in each racial-ethnic group. Because of these constraints in the data, this analysis was limited to women who identified as Hispanic, non-Hispanic White, or non-Hispanic Black. Age, education level, and family income were also collected as nominal categorical variables.

Statistical analysis

We calculated the population-level prevalence of age, education level, and family income by sexual minority identity and race-ethnicity. To examine the intersection of sexual minority identity and race-ethnicity on binge drinking and heavy alcohol use, we conducted parallel analyses comparing sexual minority women of color, heterosexual women of color, sexual minority non-Hispanic White women, and heterosexual non-Hispanic White women. The first group of analyses included non-Hispanic Black women and non-Hispanic White women. The second group of analyses included Hispanic women and non-Hispanic White women. For both sets of analyses, non-Hispanic White heterosexual women were the reference group.

The methods used for the intersectionality analysis were developed by Jackson et al. (2016). In brief, Jackson et al. propose examining the intersection of multiple marginalization using statistical decomposition techniques. An important assumption of this technique is that disparities in health outcomes attributed to social categorizations are a representation of the historical oppression faced by particular groups. For example, disparities among African Americans are the result of the historical legacy of structural racism, racial-ethnic discrimination, and prejudice faced by this group in the United States (Williams et al., 2019). Similarly, health disparities among Hispanic/Latinx populations are the result of racism, xenophobia, discriminatory immigration policies, and prejudice faced by this group. Underlying this quantitative technique is the understanding of relativity of oppression such that racism or homophobia “play out differently,” depending on levels of the other and on other social categorizations (Jackson, 2017; Jackson et al., 2016).

For this analysis, we report the age-adjusted joint disparity, referent disparities, and excess intersectional disparity for binge drinking and heavy alcohol use (Jackson, 2017; Jackson et al., 2016). The primary measure of interest is the joint disparity (µ₁₁ – µ₀₀), which compares the prevalence of binge drinking/heavy alcohol use between LB women of color (µ₁₁) and non-Hispanic White heterosexual women (µ₀₀). This difference represents the excess prevalence of binge drinking/heavy alcohol use that we attribute to the racism and homophobia experienced by LB women of color. The referent race disparity (µ₁₀ – µ₀₀) measures the excess prevalence of excessive alcohol use among heterosexual women of color attributable to racial-ethnic marginalization (racism) in the absence of homophobia. The referent sexual identity disparity (µ₀₁ – µ₀₀) measures the excess prevalence of excessive alcohol use among LB White women attributable to homophobia in the presence of racial-ethnic privilege. If the joint disparity exceeds the sum of the referent race disparity and referent sexual identity disparity, it suggests that the joint action of racism and homophobia may synergistically impact the prevalence of excessive alcohol use, which is captured by the excess intersectional disparity (µ₁₁ − µ₁₀ − µ₀₁ + µ₀₀). Because the referent disparities and excess intersectional disparity always sum to the joint disparity, their relative magnitude gives a sense of whether racism alone or homophobia alone or their interlocking presence could produce the observed prevalence for LB women of color (Jackson, 2017) In general linear model/analysis of variance parlance, the referent disparities are termed “main effects” and the intersectional disparity is an interaction.

To compute the age-adjusted joint disparity, referent disparities, and excess intersectional disparity, we used unweighted logistic regression to model the probability of (a) sexual identity, (b) race-ethnicity given sexual identity, (c) sexual identity given age, and (d) race-ethnicity given sexual identity and age. The predicted probabilities from these models were used to construct standardized inverse probability weights (Jackson et al., 2016). These weights standardize the age distribution for each racial-ethnic and sexual identity group according to the entire study population. The inverse probability weights were multiplied by the design-based survey weights, which adjust for nonresponse and nonsampling error, to generalize to the U.S. population. All analyses were conducted in R 3.4.2 (R Core Team, 2017).

Results

Demographic characteristics

Sociodemographic characteristics and health behaviors differed significantly across racial-ethnic and sexual identity axes. Compared with White heterosexual women (n = 12,860), Black (n = 281) and Hispanic (n = 305) LB women were more likely to be 18–25 years old, to not have graduated from high school, and to have a family income of less than $10,000. Black (n = 2,862) and Hispanic (n = 3,719) heterosexual women had a similar age distribution as White heterosexual women; however, they were more likely to not have graduated from high school and to have a family income of less than $10,000. Although White LB women were just as likely to have graduated from high school as White heterosexual women, they were significantly more likely to have a family income of less than $10,000 (Table 1).

Table 1.

Characteristics of women in the National Survey on Drug Use and Health by race-ethnicity and sexual identity group (N = 21,024)

	Non-Hispanic Black lesbian/bisexual (n = 281)		Non-Hispanic Black heterosexual (n = 2 862)		Non-Hispanic Black lesbian/bisexual (n = 997)		Non-Hispanic Black heterosexual (n = 12.860)		Hispanic lesbian/bisexual (n = 305)		Hispanic heterosexual (n = 3,719)
Variable	na	%b	na	%b	na	%b	na	%b	na	%b	na	%b
Age group
18–25 years	171	46.4	968	14.8	527	31.1	3,504	10.8	194	41.5	1,402	18.2
26–34 years	63	29.7	597	15.9	230	23.1	2,549	13.1	62	25.1	861	19.7
35–49 years	41	19.8	762	27.8	162	19.9	3,472	22.6	33	15.1	1,007	30.6
50–64 years	4	2.1	358	25.1	55	17.8	1,839	28.4	11	12.5	308	20.0
≥65 years	2	1.9	177	16.4	23	8.0	1,496	25.2	5	5.9	141	11.5
Education
Less than high school	50	17.0	445	17.8	115	9.5	1,026	8.0	85	30.7	994	31.1
High school	99	34.4	846	28.4	285	23.4	3,009	24.0	81	27.0	1,043	25.5
Some college	101	35.6	1,089	35.0	377	36.3	4,651	33.5	100	26.0	1,207	28.9
College graduate	31	13.0	482	18.9	220	30.8	4,174	34.4	39	16.3	475	14.5
Family incomec
<$10,000	66	17.1	514	13.1	127	11.3	1,007	5.1	49	11.8	402	9.8
$10,000–$19,999	61	22.0	610	21.9	161	16.3	1,339	9.9	68	25.3	710	17.7
$20,000–$29,999	38	14.6	388	12.7	138	9.3	1,258	9.6	37	9.6	584	15.1
$30,000–$39,999	35	16.5	344	11.6	116	9.9	1,216	9.6	28	6.0	466	11.4
$40,000–$49,999	30	10.4	277	9.6	97	8.7	1,281	9.1	36	17.8	453	11.9
$50,000–$74,999	22	7.3	316	12.9	137	16.7	2,181	17.5	39	12.0	521	15.3
≥$75,000	29	12.1	413	18.2	221	27.9	4,578	39.2	48	17.4	583	18.8

^aNotes: Unweighted counts;

^bweighted prevalence;

^cin U.S. dollars.

Prevalence of excessive alcohol consumption

Overall, one fifth of women who identify as Black, White, or Hispanic in the United States binge drank within the past 30 days, and approximately 5% binge drank on five or more occasions within the past 30 days. The prevalence of binge drinking was highest among Black LB women (45.4%) and Hispanic LB women (43.4%), followed by White LB women (35.7%). Black and Hispanic heterosexual women had the lowest prevalence of binge drinking (20.8% and 20.2%, respectively). This pattern was similar for heavy alcohol use (i.e., five or more occasions of binge drinking within the past 30 days). Black LB (11.8%) and Hispanic LB (8.4%) women had the highest prevalence of heavy alcohol use, followed by White LB women (8.2%). Black and Hispanic heterosexual women were least likely to engage in heavy alcohol use (3.3% and 2.9%, respectively) (Table 2).

Table 2.

Prevalence of current binge drinking and heavy alcohol use among non-Hispanic Black, non-Hispanic White, and Hispanic sexual identity groups (N = 21,024)

	Non-Hispanic Black lesbian/bisexual (n = 281)		Non-Hispanic Black heterosexual (n = 2,862)		Non-Hispanic White lesbian/bisexual (n = 997)		Non-Hispanic White heterosexual (n = 12,860)		Hispanic lesbian/bisexual (n = 305)		Hispanic heterosexual (n = 3,719)		All groups
Variable	na	% [95% CI]b	na	% [95% CI]b	na	% [95% CI]b	na	% [95% CI]b	na	% [95% CI]b	na	% [95% CI]b	% [95% CI]b
Binge drinking	135	45.4 [39.1, 51.6]	683	20.8 [18.8, 22.8]	406	35.7 [31.8, 39.5]	3,739	23.0 [21.8, 24.1]	150	43.4 [35.1, 51.7]	913	20.2 [18.3, 22.0]	22.3 [21.5, 23.1]
Heavy alcohol use	31	11.8 [7.1, 16.5]	99	3.3 [2.3, 4.3]	109	8.2 [6.0, 10.5]	853	5.2 [4.7, 5.7]	41	8.4 [4.8, 11.9]	137	2.9 [2.1, 3.8]	4.6 [4.2, 4.9]

Notes: Binge drinking is defined as four or more drinks on at least one occasion (about 2 hours) in the past 30 days; heavy alcohol use is defined as four or more drinks on the same occasion (about 2 hours) on 5 or more days in the past 30 days. CI = confidence interval.

^aUnweighted counts;

^bweighted prevalence (used survey weights to account for nonresponse).

Intersectional disparities in age-adjusted prevalence of excessive alcohol consumption

The joint disparity is the total difference in the prevalence of binge drinking/heavy drinking between multiply marginalized women and those who have racial and sexual identity privilege. The excess intersectional disparity is the amount of the total difference that is attributable to being both a racial-ethnic and a sexual minority woman. Whereas the intersectional quantities describing disparities among women of color and White heterosexual and LB women are qualitatively similar for binge drinking and heavy alcohol use, differences in binge drinking are larger. The joint disparity in binge drinking comparing Black LB women and White heterosexual women was 21.2% (95% CI [12.4, 29.9]). The referent race disparity was -0.8% (95% CI [-2.8, 1.1]), and the referent sexual identity disparity was 4.3% (95% CI [-1.5, 10.1]). The estimated excess intersectional disparity for binge drinking across race and sexual identity for these groups was 17.7% (95% CI [6.9, 28.5]). The joint disparity in heavy alcohol use comparing Black LB women and White heterosexual women was 6.5% (95% CI [0.6, 12.5]).

The referent race disparity was -1.8% (95% CI [-2.9, -0.7]), and the referent sexual identity disparity was 0.0% (95% CI [-2.5, 2.4]). The estimated excess intersectional disparity for heavy alcohol use across race and sexual identity for these groups was 8.4% (95% CI [2.0, 14.7]) (Table 3).

Table 3.

Joint disparity, referent race disparity, referent sexual identity disparity for binge drinking and heavy alcohol use adjusted for age among Black and White sexual identity groups

Variable	Joint disparity	Referent race disparity	Referent sexual identity disparity	Excess intersectional disparity
% [95% CI]	µ11 − µ00	µ10 − µ00	µ01 − µ00	µ11 − µ10 − µ01 + µ00
Binge drinking	21.2 % [12.4, 29.9]	-0.8% [-2.8, 1.1]	4.3% [-1.5, 10.1]	17.7% [6.9, 28.5]
Heavy alcohol use	6.5% [0.6, 12.5]	-1.8% [-2.9, -0.7]	0.0% [-2.5, 2.4]	8.4% [2.0, 14.7]

Notes: Binge drinking is defined as four or more drinks on at least one occasion (about 2 hours) in the past 30 days; heavy alcohol use is defined as four or more drinks on the same occasion (about 2 hours) on 5 or more days in the past 30 days. CI = confidence interval; µ₁₁ = prevalence of binge drinking or heavy alcohol use among non-Hispanic Black sexual minority women adjusted for age; µ₁₀ = prevalence of binge drinking or heavy alcohol use among non-Hispanic Black non-sexual minority women adjusted for age; µ₀₁ = prevalence of binge drinking or heavy alcohol use among White sexual minority women adjusted for age; µ₀₀ = prevalence of binge drinking or heavy alcohol use among White non–sexual minority women adjusted for age.

The joint disparity in binge drinking comparing Hispanic LB women and White heterosexual women was 16.8% (95% CI [6.9, 26.8]). The referent race disparity was 0.7% (95% CI [-1.6, 3.0]), and the referent sexual identity disparity was 5.3% (95% CI [-0.3, 10.9]). The estimated excess intersectional disparity for binge drinking across ethnicity and sexual identity for these groups was 10.8% (95% CI [-0.9, 22.6]). The joint disparity in heavy alcohol use comparing Hispanic LB women and White heterosexual women was 2.1% (95% CI [-1.4, 5.6]). The referent ethnic disparity was -1.6% (95% CI [-2.6, -0.5]), and the referent sexual identity disparity was 0.3% (95% CI [-2.2, 2.8]). The estimated excess intersectional disparity for heavy alcohol use across ethnicity and sexual identity for these groups was 3.3% (95% CI [-1.5, 8.2]) (Table 4).

Table 4.

Joint disparity, referent race disparity, and referent sexual identity disparity for binge drinking and heavy alcohol use adjusted for age among Hispanic and White sexual identity groups

Variable	Joint disparity	Referent race disparity	Referent sexual identity disparity	Excess intersectional disparity
% [95% CI]	µ11 − µ00	µ10 − µ00	µ01 − µ00	µ11 − µ10 − µ01 + µ00
Binge drinking	16.8% [6.9, 26.8]	0.7% [-1.6, 3.0]	5.3% [-0.3, 10.9]	10.8% [-0.9, 22.6]
Heavy alcohol use	2.1% [-1.4, 5.6]	-1.6% [-2.6, -0.5]	0.3% [-2.2, 2.8]	3.3% [-1.5, 8.2]

Notes: Binge drinking is defined as four or more drinks on at least one occasion (about 2 hours) in the past 30 days; heavy alcohol use is defined as four or more drinks on the same occasion (about 2 hours) on 5 or more days in the past 30 days. CI = confidence interval; µ₁₁ = prevalence of binge drinking or heavy alcohol use among Hispanic sexual minority women adjusted for age; µ₁₀ = prevalence of binge drinking or heavy alcohol use among Hispanic non–sexual minority women adjusted for age; µ₀₁ = prevalence of binge drinking or heavy alcohol use among White sexual minority women adjusted for age; µ₀₀ = prevalence of binge drinking or heavy alcohol use among White non-sexual minority women adjusted for age.

Discussion

This study examined differences in binge drinking and heavy alcohol use (i.e., binge drinking on five or more occasions) across sexual identity and race-ethnicity among women in the United States using nationally representative survey data. Results from this analysis align with previous studies showing that sexual minority women have a higher prevalence of excessive alcohol consumption compared with heterosexual women (Gonzales & Henning-Smith, 2017; Gonzales et al., 2016; Medley et al., 2016). One fifth of women who identify as Black, White, or Hispanic in the United States reported binge drinking, and 5% reported heavy alcohol use in the past 30 days. Sexual minority women of color (Black and Hispanic) had more than double the prevalence of current binge drinking and heavy alcohol use in comparison with heterosexual White women. When decomposing differences in the prevalence of excessive alcohol use, we found large excess intersectional disparities relative to disparities based on race-ethnicity or sexual identity alone—which may suggest that the higher prevalence of excessive alcohol consumption among sexual minority women of color may be attributable to combined experiences of racism and homophobia. Moreover, the negative race disparity between heterosexual women of color and heterosexual White women suggests that the impact of racism on heavy alcohol use and binge drinking may work differently for heterosexual women of color and sexual minority women of color.

Few quantitative studies have examined disparities in excessive alcohol consumption specifically among sexual minority women (Bauer et al., 2010; Fish et al., 2018; Hughes, 2003). In a cohort study conducted by Hughes, lesbian women were more likely to develop problematic alcohol use compared with age- and race-matched heterosexual women (Hughes, 2003). Fish et al. (2018) found that sexual minority women had higher odds of consuming eight or more drinks per binge episode compared with heterosexual women. Bauer and colleagues found that heterosexual-identifying women with female sex partners had a similar 12-month prevalence of binge drinking as lesbian-identified women (Bauer et al., 2010). Although these studies use diverse samples, none has specifically applied an intersectional lens to understand intragroup differences in problematic alcohol use among this population. Our findings advance beyond previously published work to suggest that disparities in binge drinking and heavy alcohol use between LB women of color and non-Hispanic heterosexual White women may be driven by structural racism and homophobia among those living at the intersection of sexual minority and racial-ethnic minority identity.

The literature supports several possible hypotheses for the higher prevalence of excessive alcohol use among sexual minority women of color and the very low prevalence of excessive alcohol use among heterosexual women of color. First, sexual minority women of color face stressors based on their gender, racial minority identity, and sexual minority identity that may influence increased substance use. For example, Calabrese et al. (2015) found that Black sexual minority women reported more occurrences of discrimination compared with White sexual minority women and Black sexual minority men. These differences were associated with more depressive symptoms and poorer psychological well-being among Black sexual minority women (Calabrese et al., 2015). Previous research has shown a causal link between excessive alcohol use and incident depression and poorer mental health (Bellos et al., 2016). Like Black sexual minority women, Latina sexual minority women face minority stress because of multiple marginalized identities, and this stress is related to both illicit and licit substance use in this population, including heavy drinking (Matthews et al., 2014). The results of a previous study suggest that acculturation, when measured as nativity and languages spoken, may be linked to substance use and mediated by experiences of discrimination among Latina sexual minority women (Matthews et al., 2014). However, acculturation is a complex cultural, social, and psychological phenomenon that occurs when there is contact between multiple cultural groups (Berry & Sam, 2016). Although this contact can serve individuals and groups in positive ways, stressful negative reactions (i.e., acculturative stress) can also arise (Sam, 2015). Acculturative stress, rather than acculturation more broadly, has been linked with substance use among Latino adolescents (Buchanan et al., 2009). Additional research is needed to understand pathways between acculturative stress and substance use for Latina sexual minority women.

Second, religiosity influences drinking behavior, as some religions stress that individuals should limit or avoid alcoholic beverages. Drabble et al. (2016) found that in the National Alcohol Survey greater religiosity was associated with reduced odds of hazardous drinking for both heterosexual and sexual minority women. However, a greater proportion of heterosexual women, compared with sexual minority women, reported that religion or spirituality was “very important” to them (Drabble et al., 2016). In addition, religious norms around alcohol consumption were protective for heterosexual women but not for sexual minority women (Drabble et al., 2016). The authors treat race-ethnicity as a covariate rather than stratifying their analyses by race-ethnicity and sexual identity. Therefore, it is unknown how religiosity affects hazardous alcohol use at the intersection of race-ethnicity and sexual identity. Future studies could elucidate these potential differences.

Third, differences in excessive alcohol use between sexual minority and heterosexual women may be related to variations in social norms around alcohol use behavior between these populations. Several qualitative studies have explored social norms and beliefs about alcohol among sexual minority women. One study examined the role that alcohol plays in the lives of sexual minority women in comparison with heterosexual women (Drabble & Trocki, 2014). Drabble and Trocki (2014) found that sexual minority women commonly discussed alcohol use as a coping mechanism for the trauma they experienced in their lives. Another qualitative study explored how alcohol factors into stress and coping among a diverse sample of LB women (Condit et al., 2011). Those researchers found that sexual minority women discussed excessive alcohol use as a normative behavior within the LGBT community because it provided a means of escape from stressful life experiences, particularly when women drank in gay bars and clubs. Emslie et al. (2017) found that drinking behavior and choice of drink factored heavily into how lesbian, gay, bisexual, and transgender individuals chose to display their sexual and gender identities in public drinking spaces. Some sexual minority women discussed choosing “manly drinks” such as pints of beer and drinking excessively to subvert traditional feminine stereotypes and reinforce their sexual minority identity (Emslie et al., 2017). These studies were limited in exploring how race-ethnicity may affect stress and social norms associated with alcohol use. Future qualitative research may consider building interview guides that specifically incorporate intersectional theory to more fully understand alcohol-related disparities for multiply marginalized groups and to explore the lived experiences of sexual minority women of color.

This analysis has several important limitations. First, a common limitation in sexual minority health research is small sample sizes—which often leads to combining heterogenous groups. This study used a national sample to attempt to mitigate this common challenge, yet issues of small numbers in some strata led to combining LB women into one group. However, literature suggests that bisexuals face bi-specific stressors that can lead to excessive alcohol consumption, depression, and other health challenges (Friedman et al., 2014; Molina et al., 2015). LB Asian women, Native American women, and women of multiple races were excluded from the analysis due to small numbers as well. However, future research should consider oversampling these populations to provide stable estimates to assess health disparities (Jackson et al., 2016). Moreover, NSDUH does not capture specific Latinx identities. Instead, anyone who identifies as Hispanic, Latinx, or of Spanish origin is grouped into the Hispanic category despite important differences in health status among specific Latinx ethnic groups (Schur et al., 1987; Weinick et al., 2004). The cross-sectional nature of the NSDUH survey precludes mechanistic decomposition analyses that can be used to unpack how the differential distribution and effects of risk factors contribute to negative alcohol outcomes (Jackson, 2018; Jackson & VanderWeele, 2018, 2019). Future studies could empirically test for whether disparities arise from distress using longitudinal data and mediation analyses (Jackson, 2017; Jackson & VanderWeele, 2019).

Second, geographic context plays an important role in adult binge drinking that may be obscured in national-level data. Alcohol policy is determined at the state and local level. Previous work has shown that states with stronger alcohol policy environments that increase the price and limit the availability of alcohol also have a lower prevalence of adult binge drinking (Naimi et al., 2014; Xuan et al., 2015). There may be important interactions between the alcohol policy environment and structural barriers based on race-ethnicity or sexual orientation that could explain health disparities among racial-ethnic minority LB women. Future research should explore the relationship between alcohol policy and disparities in hazardous alcohol use among sexual minority women of color.

Despite these limitations, this study adds to a growing body of literature on the health and well-being of sexual minority populations. Quantitative methods for applying intersectionality can reveal important health disparities among heterogenous populations. Additionally, by quantifying these intersectional health disparities, public health professionals can more readily identify groups with substantially high risk (Jackson et al., 2016). The combination of quantitative and qualitative methods driven by social and behavioral theory may provide the strongest evidence for understanding health disparities and building interventions to achieve health equity. More research is needed to further elucidate the mechanisms that drive excessive alcohol use among sexual minority women of color and determine the impact of current population-level strategies on this population.

Our results have implications for how policy and programmatic interventions to reduce alcohol abuse should be tailored to specific populations. Public health efforts to reduce excessive alcohol consumption should bring a health equity perspective and focus on multiply marginalized populations whose drinking behaviors may partially be driven by structural stigma, prejudice, and discrimination. Current strategies to decrease population-level excessive alcohol consumption—including reducing alcohol outlet density (Campbell et al., 2009), increasing the price of alcohol through excise taxes (Elder et al., 2010), and implementing alcohol screening and brief intervention in health care settings (Tansil et al., 2016)—are universal primary prevention strategies. However, research has yet to determine whether these interventions can reduce disparities in excessive alcohol consumption or what impact they have across population subgroups. Recent studies on the alcohol policy environment provide some evidence that legislation that regulates the availability, affordability, and accessibility of alcohol products reduces binge drinking among the general adult population, but not among Black and Hispanic adults (Xuan et al., 2015). Furthermore, little is known about how these policies affect drinking among sexual minority populations. Interventions implemented in the healthcare setting, such as alcohol screening and brief intervention, may never reach LB women of color who may face barriers to healthcare because of their social position as both sexual minorities and racial-ethnic minorities (Barbara et al., 2001; Smedley et al., 2003). Therefore, multilevel interventions are needed at the individual and structural level that target sexual minority women, particularly women of color. These interventions should be theory driven and designed in consultation with these populations.

Acknowledgments

The authors thank Dr. Tonia Poteat and Dr. Danielle German for their comments on earlier versions of the manuscript.