Skip to main content
PMC home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Oct 21.
Published in final edited form as: Addiction. 2018 Apr 21:10.1111/add.14251. doi: 10.1111/add.14251

Sexual-orientation differences in alcohol use trajectories and disorders in emerging adulthood: results from a longitudinal cohort study in the United States

Robert W S Coulter 1,2,3,4, Hee-Jin Jun 5, Jerel P Calzo 5, Nhan L Truong 5, Christina Mair 1, Nina Markovic 2,6, Brittany M Charlton 7,8,9,10, Anthony J Silvestre 2,11, Ron Stall 1,2, Heather L Corliss 5,9
PMCID: PMC6667315  NIHMSID: NIHMS1038334  PMID: 29679419

Abstract

Aims

We estimated sexual-orientation differences in alcohol use trajectories during emerging adulthood, and tested whether alcohol use trajectories mediated sexual-orientation differences in alcohol use disorders (AUDs).

Design

Longitudinal self-reported survey data from the Growing Up Today Study.

Setting

United States.

Participants

A total of 12 493 participants aged 18–25 during the 2003, 2005, 2007 or 2010 surveys.

Measurements

Stratified by gender, longitudinal latent class analyses estimated alcohol use trajectories (using past-year frequency, quantity and binge drinking from 2003 to 2010). Multinomial logistic regression tested differences in trajectory class memberships by sexual orientation [comparing completely heterosexual (CH) participants with sexual-minority subgroups: mainly heterosexual (MH), bisexual (BI) and gay/lesbian (GL) participants]. Modified Poisson regression and mediation analyses tested whether trajectories explained sexual-orientation differences in AUDs (past-year DSM-IV abuse/dependence in 2010).

Findings

Six alcohol use trajectory classes emerged for women and five for men: these included heavy (23.5/36.9% of women/men), moderate (31.8/26.4% of women/men), escalation to moderately heavy (9.7/12.0% of women/men), light (17.0% for women only), legal (drinking onset at age 21; 11.1/15.7% of women/men) and non-drinkers (7.0/9.1% of women/men). Compared with CH women, MH and BI women had higher odds of being heavy, moderate, escalation to moderately heavy and light drinkers versus non-drinkers (odds ratios = 2.02–3.42; P-values < 0.01–0.04). Compared with CH men, MH men had higher odds of being heavy, moderate and legal drinkers versus non-drinkers (odds ratios = 2.24–3.34; P-values < 0.01–0.01). MH men and women, BI women and GLs had higher risk of AUDs in 2010 than their same-gender CH counterparts (risk ratios = 1.34–2.17; P-values < 0.01). Alcohol use trajectories mediated sexual-orientation differences in AUDs for MH and GL women (proportion of effect mediated = 30.8–31.1%; P-values < 0.01–0.02), but not for men.

Conclusions

In the United States, throughout emerging adulthood, several sexual-minority subgroups appear to have higher odds of belonging to heavier alcohol use trajectories than completely heterosexuals. These differences partially explained the higher risk of alcohol use disorders among mainly heterosexual and gay/lesbian women but not among sexual-minority men.

Keywords: Alcohol use disorder, alcohol use trajectories, cohort study, emerging adulthood, longitudinal latent class analysis, sexual orientation

INTRODUCTION

Alcohol use and binge drinking (i.e. consuming at least four drinks for women or five drinks for men on one occasion [1]) in the United States is highest during the emerging adulthood period (i.e. ages 18–25) [2–4]. Nevertheless, emerging adults have diverse alcohol use trajectories (AUTs). This diversity is often characterized using statistical approaches well-suited for identifying homogeneous subgroups of AUTs [5–7]. A recent review of longitudinal alcohol use throughout emerging adulthood found common trajectories, such as consistently heavy drinkers, escalation drinkers and non-drinkers [8]. Importantly, consistently heavy or escalating AUTs place emerging adults at greater risk for acquiring alcohol use disorders- [9–12], which are associated with morbidity, mortality and economic burden [13,14].

Alcohol use behaviors are not equal across all populations of emerging adults. For example, some sexual-minority subgroups (e.g. gay/lesbian, bisexual and those who describe their sexual orientation as ‘mostly heterosexual’) have higher alcohol use than their completely heterosexual peers, according to both cross-sectional [15–20] and longitudinal [21–26] research. Sexual-minority women are at greater risk of alcohol use and binge drinking than completely heterosexual women [15–24]. Contrastingly, findings among men are mixed: most studies found no sexual-orientation differences in the prevalence of alcohol use or binge drinking [15–18,21,22,26]; some studies found a higher prevalence of drinking and binge drinking for sexual-minority versus heterosexual men [20,22,23,26,27], while others found a lower prevalence of heavy drinking among sexual-minority men [15,18,19]. However, there is little information about how sexual orientation is related to membership in different AUT groups throughout emerging adulthood, which is a focus of the current study. Understanding sexual-orientation differences in AUTs throughout emerging adulthood can inform the need for and design of future alcohol use interventions, a sentiment echoed by researchers and the Institute of Medicine [28,29].

Sexual-minority adults are also at higher risk than heterosexual adults for having alcohol use disorders (AUDs; i.e. abuse or dependence) [30–32]. Scant research suggests that this is also true during emerging adulthood [33], when people in the United States are at greatest risk for AUDs [34]. Generally, the development of AUDs is associated positively with earlier or concurrent high AUTs [8,35,36]; therefore, sexual-orientation differences in AUTs may partially explain sexual-orientation differences in subsequent AUDs. Support for this hypothesis would suggest that interventions reducing sexual-orientation differences in AUTs may also decrease differences in AUDs. Additionally, theory and prior literature suggest that minority stressors and social stigma related to minority sexual orientations are contributors to elevated alcohol use and AUDs among sexual-minority populations [30,37–48].

This paper addressed gaps in knowledge about how sexual orientation is related to AUTs and AUDs using data from a prospective cohort study. First, we estimated longitudinal AUTs during the emerging adulthood period. Secondly, we tested for sexual-orientation differences in AUT memberships. Thirdly, we estimated sexual-orientation disparities in AUDs in late emerging adulthood and tested whether longitudinal AUTs throughout emerging adulthood mediated these disparities in AUDs. The results of this study can help to identify sexual-orientation subgroups at risk for having heavy AUTs and AUDs and can inform public health strategies aimed at reducing sexual-orientation disparities in AUTs and AUDs.

METHODS

Study design and population

We analysed data from participants in the Growing Up Today Study (GUTS), which began in 1996. GUTS initially enrolled 16 875 participants aged 9–14 years who were children of women participating in the Nurses’ Health Study II (NHS2)—a cohort study of 116 430 registered nurses from 14 US states begun in 1989. The GUTS sample is similar in racial/ethnic distribution as the NHS2 (93.4 and 93.8% non-Hispanic White, respectively). Additional information about GUTS is reported elsewhere [24,49]. The Brigham and Women’s Hospital Institutional Review Board approved original study procedures.

The current study included participants who were aged 18–25 years when they responded to at least one alcohol use item during the 2003, 2005, 2007 or 2010 survey waves, and who provided information on sexual orientation and did not select ‘unsure’ for their sexual orientation (11 participants were missing or ‘unsure’ of their sexual orientation). We selected these waves because they assessed alcohol use and contained the age range focus on in this analysis. Our analytical sample included 12 493 participants (7465 women; 5027 men), representing 74.0% of the cohort. Compared to cohort participants excluded from our analyses, participants in our analytical sample were: more likely to be female (59.8 versus 35.8%, respectively; P < 0.001); more likely to live in the western region of the United States at baseline (14.7 versus 11.5%, respectively; P < 0.001); more likely to have a mother who reported her household income as ≥ $100 000 in year 2001 (33.2 versus 27.4%, respectively); and less likely to have a mother who did not report her household income in 2001 (18.9 versus 29.1%, respectively; P < 0.001). Participants included versus excluded in our analytical sample did not differ by age at baseline (P = 0.094), their own race/ethnicity (P = 0.838) or their mother’s race/ethnicity (P = 0.5459).

Participants’ mean age in our analytical sample was 18.8 years in 2003 and 25.2 years in 2010. Among participants in our analytical sample, 84.2% responded to the 2003 survey, 83.3% in 2005, 78.1% in 2007 and 68.2% in 2010. Overall, 53.3% responded to all four waves, 19.7% responded to three waves, 14.6% responded to two waves and 12.5% responded to one wave. A higher number of completed surveys was associated with being female (P < 0.0001), younger age at baseline (P < 0.0001), living in the western region of the United States at baseline (P < 0.0001), having a mother who reported her household income as ≥ $100 000 in 2001 (P < 0.0001) and with having a minority sexual orientation (for most subgroups, see Table 1)—but was not associated with participant’s race/ethnicity (P = 0.6762) or mother’s race/ethnicity (P = 0.3943).

Table 1.

Characteristics of the sample by sexual orientation, stratified by gender: Growing Up Today Study, 2003–2010.

Sexual orientation
Completely heterosexual Mostly heterosexual Bisexual Gay/lesbian
n (%) n (%) P-value n (%) P-value n (%) P-value
Women
Total, row percentage 6057 (81.1) 1121 (15.0) 174 (2.3) 113 (1.5)
Race/ethnicity
 White 5691 (94.0) 1026 (91.5) 0.0065 156 (89.7) 0.0661 105 (92.9) 0.6707
 Non-white 366 (6.0) 95 (8.5) 18 (10.3) 8 (7.1)
Region
 West 898 (14.8) 247 (22.0) <0.0001 31 (17.8) 0.2348 20 (17.7) 0.1907
 Midwest 2039 (33.7) 296 (26.4) 46 (26.4) 27 (23.9)
 South 1092 (18.0) 179 (16.0) 32 (18.4) 21 (18.6)
 Northeast 2028 (33.5) 399 (35.6) 65 (37.4) 45 (39.8)
College attendance
 Never attended 119 (2.7) 36 (4.2) 0.0397 14 (9.6) 0.0059 2 (2.2) 0.7270
 Attended 4316 (97.3) 817 (95.8) 132 (90.4) 91 (97.8)
Life-time pregnancy
 No 3515 (77.5) 666 (75.8) 0.2603 104 (68.9) 0.0245 84 (88.4) 0.0020
 Yes 1018 (22.5) 213 (24.2) 47 (31.1) 11 (11.6)
Age in 2010, median (interquartile range) 25 (24–27) 25 (24–27) 0.5648 25 (24–27) 0.9518 26 (24–27) 0.1537
Mother’s household income in 2001
 <$50 000 613 (12.7) 105 (11.3) 0.7291 38 (26.2) 0.0002 13 (13.4) 0.1495
 $50000–74 999 1164 (24.1) 227 (24.5) 33 (22.8) 18 (18.6)
 $75 000–99 999 1078 (22.3) 209 (22.6) 23 (15.9) 16 (16.5)
≥ $100000 1976 (40.9) 386 (41.6) 51 (35.2) 50 (51.6)
Number of survey waves completed from 2003 to 2010
 1 597 (9.6) 79 (7.1) 0.0003 7 (4.0) 0.0050 7 (6.2) 0.5197
 2 769 (12.7) 124 (11.1) 16 (9.2) 17 (15.0)
 3 1060 (17.5) 244 (21.8) 41 (23.6) 18 (15.9)
 4 3649 (60.2) 674 (60.1) 110 (63.2) 71 (62.8)
Men
Total, row percentage 4546 (90.4) 301 (6.0) 34 (0.7) 146 (2.9)
Race/ethnicity
 White 4254 (93.6) 262 (87.0) 0.0017 33 (97.1) 0.2431 129 (88.4) 0.0538
 Non-white 292 (6.4) 39 (13.0) 1 (2.9) 17 (11.6)
Region
 West 751 (16.5) 69 (22.9) 0.0543 5 (14.7) 0.2664 24 (16.4) 0.3276
 Midwest 1532 (33.7) 94 (31.2) 10 (29.4) 41 (28.1)
 South 765 (16.8) 46 (15.3) 3 (8.8) 23 (15.8)
 Northeast 1498 (33.0) 92 (30.6) 16 (47.1) 58 (39.7)
College attendance
 Never attended 121 (4.7) 5 (2.2) 0.0251 1 (4.8) 0.9865 4 (3.5) 0.5239
 Attended 2462 (95.3) 218 (97.8) 20 (95.2) 109 (96.5)
Age in 2010, median 25 (24–26) 25 (24–26) 0.9189 25 (23–26) 0.2997 25 (24–27) 0.5756
(interquartile range)
Mother’s household income in 2001
 < $50000 475 (12.8) 28 (10.9) 0.5587 3 (9.7) 0.2043 14 (11.6) 0.9604
 $50000–74999 873 (23.4) 66 (25.8) 5 (16.1) 27 (22.3)
 $75 000–99 999 868 (23.3) 53 (20.7) 5 (16.1) 29 (24.0)
 ≥ $100 000 1510 (40.5) 109 (42.6) 18 (58.1) 51 (42.2)
Number of survey waves completed from 2003 to 2010
 1 841 (18.5) 26 (8.6) < 0.0001 6 (17.7) 0.4472 16 (11.0) <0.0001
 2 844 (18.6) 36 (12.0) 3 (8.8) 11 (7.5)
 3 984 (21.7) 78 (25.9) 8 (23.5) 26 (17.8)
 4 1877 (41.3) 161 (53.5) 17 (50.0) 93 (63.7)
Open in a new tab

Using completely heterosexuals as the referent, P-values were derived using Rao–Scott χ2 tests for categorical variables and univariable models with generalized estimating equations for continuous variables, both of which were adjusted for sibling clusters. Missing data for college attendance, pregnancy and mother’s household income were excluded from this table, including the Rao–Scott χ2 tests.

Measures

Alcohol use

Three indicators assessed alcohol use in 2003, 2005, 2007 and 2010. These indicators were distributed non-normally, hence we modeled them as ordinal variables. We assessed past-year average frequency: ‘On average, in the past year, how often did you drink beer, wine or liquor? don’t drink; less than once a month; less than once a week; 1–2 days per week; 3–5 days per week; almost every day; or daily’. We combined daily and almost every day because of small cell sizes, and coded this variable from 0 (‘don’t drink’) to 6 (‘daily or almost every day’). We assessed past-year average quantity: ‘When you drink alcohol, how much do you usually drink at one time? don’t drink; less than 1; 1; 2; 3; 4; 5; or 6 or more drinks’. We coded this variable from 0 (‘don’t drink’) to 7 (‘6 drinks’). We assessed past-year binge drinking: ‘Over the past year, how many times did you drink four (for women) or five (for men) or more alcohol drinks over a few hours? none; 1; 2; 3–5; 6–8; 9–11; or 12 or more times’. We coded this from 0 (‘none’) to 6 (‘12 or more times’).

Alcohol use disorders

We measured criteria for probable AUD during the past 12 months for the first time in 2010, with items assessing symptoms based on the DSM-IV [50], adapted from the National Survey on Drug Use and Health [51]. The DSM-IV described two distinct disorders, alcohol abuse and alcohol dependence. If participants met criteria for alcohol abuse or dependence, we coded them as having a probable AUD, creating a single binary variable. We made this analytical decision based on prior research [30,40,52] and because the DSM-5 [53] integrates abuse and dependence into a single disorder. We conducted sensitivity analyses with the original DSM-IV categorizations, which yielded similar results (not shown) to the single binary AUD variable.

Sexual orientation

Sexual orientation was assessed at each survey wave included in this analysis. We classified participants’ sexual orientation based on their last report. An adapted item [54] assessed sexual orientation: ‘Which one of the following best describes your feelings?’. Completely heterosexual (attracted to people of the opposite sex); mostly heterosexual; bisexual (equally attracted to men and women); mostly homosexual; completely homosexual (attracted to persons of the same sex); or unsure’. Like prior research [24], we combined mostly and completely homosexual into a single group (henceforth gay/lesbian) to increase statistical power.

Covariates

Covariates included race/ethnicity (white versus non-white; measured at baseline), region of residence (West versus midwest, southwest and northeast; measured in 2010), mothers’ household income (reported in the 2001 NHS2 survey), number of completed surveys from 2003 to 2010 and age (we used age in 2010 because this is when AUD was assessed). We assessed college attendance in 2010 using the adapted item [55]: ‘What is the highest grade of school you have completed or the highest degree you have received?’. We dichotomized the response options into any college attendance versus none (i.e. high/trade/vocational school graduate or less). We prospectively assessed life-time pregnancy (yes/no) from 1999 to 2010 for women. We used the missing indicator method [56] for college attendance, pregnancy and mothers’ household income, which creates an additional ‘missing’ category for each variable, allowing us to analyze all available data and preserve statistical power.

Analyses

We used a classify-analyze approach to characterize longitudinal AUT classes and estimate predictors and outcomes associated with these classes [7].

Alcohol use trajectory classes

We estimated AUTs using longitudinal latent class analyses in Mplus version 7.2 (Muthén & Muthén, Los Angeles, CA, USA), allowing for the estimation of subgroup populations who differ across multiple indicators of alcohol use over time. We estimated the unconditional model (i.e. no covariates) with the three ordinal alcohol use variables at each participant’s age using a cohort-sequential design [57]. We used full information maximum likelihood estimation (to use all available observations [5,58]), the robust maximum likelihood estimator (to account for non-normality [59,60]) and the complex survey command (to account for non-independence of sibling clusters [58]).

We estimated the trajectory classes separately by gender because men drink higher quantities of alcohol on average than women [61,62]. We estimated 1–9-class solution models for women, and 1–7-class solution models for men. To determine the best-fitting number of classes, we examined several fit statistics, including the Bayesian information criterion (BIC), the Bayes factor (BF) and the correct model probability (cmP) [7,58,63,64]. We considered the best-fitting model to have the lowest BIC, BF > 10, the greatest cmP and highest interpretative validity [7,60,63]. We examined entropy and other classification qualities, i.e. average posterior probabilities (AvePP) and odds of correct classification (OCC) [7,65]. We considered good latent class separation and classification precision as classes that had AvePP > 0.7 and OCC > 5 [65]. We assigned participants to the class for which they had the highest posterior probability of membership.

Disparities in alcohol use trajectory class membership

Using SAS version 9.4 (SAS Institute, Cary, NC, USA), we used Rao–Scott χ2 tests adjusting for sibling clusters to examine the bivariate associations of trajectory groups with sexual orientation and covariates. We fitted multinomial logistic regression models to test for sexual-orientation differences in trajectory class membership (polytomous variables), controlling for covariates, using generalized estimating equations (GEE) to account for sibling clusters. Reference groups were non-drinkers (versus other trajectory classes) and completely heterosexuals (versus sexual-minority subgroups).

Alcohol use disorders

Because AUD prevalence was greater than 10%, we fitted modified Poisson regression models using GEE [66]. First, we estimated sexual-orientation differences in AUDs, controlling for covariates. Secondly, we estimated the effects of sexual orientation and alcohol trajectory classes on AUD, controlling for covariates. Subsequently, we tested whether the AUT classes mediated sexual-orientation differences in AUDs using the publicly available %MEDIATE macro, which computes the mediation proportion (i.e. indirect effect) of exposure effects explained by the effect of the exposure on proposed mediator variable (i.e. AUTs) [67]. Using the difference method, the mediation proportion is calculated as 1−(βα), where β is the coefficient of the exposure variable on the outcome variable when controlling for the proposed mediator, and α is the coefficient for the exposure variable on the outcome variable when not controlling for the proposed mediator [68].

RESULTS

Overall, 81.1% of women identified as completely heterosexual [CH], 15.0% as mostly heterosexual [MH], 2.3% as bisexual and 1.5% as lesbian (Table 1). Among men, 90.4% identified as CH, 6.0% as MH, 0.7% as bisexual and 2.9% as gay.

We selected the six-class model for women and the five-class model for men because these models met our selection criteria (i.e. thresholds for certain values), had agreement throughout the most fit statistics (Table 2) and had theoretical and meaningful interpretations. These classes had good separation and adequate assignment (all AvePP > 0.77 and OCC > 11).

Table 2.

Class enumeration fit indices and qualities for longitudinal latent class analyses, stratified by gender: Growing Up Today Study, 2003–2010.

Classes Free parameters BIC BF cmP AWE Entropy
Women
1 144 200031 <1 0.00 201 895 NA
2 289 177553 <1 0.00 181 296 0.875
3 434 170 637 <1 0.00 176 256 0.848
4 579 168055 <1 0.00 175 553 0.828
5 724 167337 <1 0.00 176 712 0.817
6 869 167 004 > 10 1.00 178 257 0.798
7 1014 167091 > 10 0.00 180 222 0.779
8 1159 167369 > 10 0.00 182 376 0.777
9a 1304 167 720 NA 0.00 184 605 0.778
Men
1 144 111 869 <1 0.00 113 791 NA
2 289 99 323 <1 0.00 103 180 0.860
3 434 95437 <1 0.00 101 228 0.838
4 579 94 504 <1 0.00 102 229 0.818
5 724 94 380 >10 1.00 104 040 0.779
6 869 94518 >10 0.00 106 113 0.762
7 1014 94953 NA 0.00 108 484 0.754
Open in a new tab

Models were estimated with samples of 7470 women and 5033 men. Bold type indicates the model we selected for each gender (i.e., the 6 class model for women, and the the 5 class model for men). BIC = Bayesian information criterion; BF = Bayes factor; cmP = correct model probability; AWE = approximate weight of evidence criterion; NA = not applicable. The likelihood ratio χ2 goodness-of-fit tests could not be computed in Mplus because the latent class indicator model was too large. Bootstrap likelihood ratio test could not be estimated because we employed the COMPLEX command in Mplus to adjust for non-independence within sibling clusters.

a

Model was not well identified.

Supporting information, Fig. S1 depicts the latent classes for women. Non-drinkers (7.0%) mostly abstained from drinking. Heavy drinkers (23.5%) used 1–2 days/week, consumed four to five drinks/occasion, and 55–80% of them engaged in monthly binge drinking. Moderate drinkers (31.8%) increased their frequency slightly from ages 18–21 to 1–2 days/week, and then plateaued until age 25. Their average quantity increased slightly from ages 18 to 19 at three drinks, and decreased to two drinks by age 25. Their binge drinking increased slightly from 18–21 to 6–8 times/year, and then plateaued until age 25. Legal drinkers (11.1%) abstained from alcohol use from ages 18–20, and nearly all of them drank alcohol from ages 21–25. Escalation to moderately- heavy drinkers (9.7%) abstained from drinking at age 18, and escalated to moderately heavy use by age 21, plateauing thereafter. Light drinkers (17.0%) consumed two drinks/occasion less than once a month and engaged in one binge drinking episode/year, on average, from ages 18 to 25.

Supporting information, Fig. S2 depicts the latent classes for men. Non-drinkers (9.1%) mostly abstained from drinking. Heavy drinkers (36.9%) consumed alcohol 1–2 days/week at age 18, which increased slightly to age 21, plateauing thereafter. Most heavy drinkers consumed six drinks/occasion at age 18, decreasing slightly from ages 19 to 25. Fifty per cent of heavy drinkers engaged in monthly binge drinking at age 18, and 75–95% engaged in monthly binge drinking from ages 19–25. Moderate drinkers (26.4%) increased their frequency slightly from ages 18–21 to 1–2 days/week and then plateaued until age 25. Their quantity was three to four drinks/occasion from ages 18 to 22, decreasing slightly thereafter. Their binge drinking remained consistent from ages 19 to 25 at approximately three to five binge drinking episodes/year. Legal drinkers (15.7%) largely abstained from alcohol use from ages 18 to 20, and nearly all drank from ages 21 to 25. Escalation to moderately heavy drinkers (12.0%) abstained from drinking at age 18, and escalated to moderately heavy use by age 20, plateauing thereafter.

Sexual orientation and covariates were associated with AUT classes (Table 3). For women, MH and bisexual participants had significantly higher odds [odds ratios (ORs) range = 2.02–3.42; P-values range < 0.0001–0.0351; Table 4] than CH participants of being heavy, moderate, escalation to moderately heavy and light drinkers versus non-drinkers. For men, MH participants had significantly higher odds (ORs range = 2.24–3.34; P-values range < 0.0001–0.0084) than CH participants of being heavy, moderate and legal drinkers versus non-drinkers. Gay men had 3.38 times the odds (P = 0.0027) of CH men of being moderate drinkers versus non-drinkers.

Table 3.

Bivariate sociodemographic characteristics of longitudinal alcohol use trajectories, stratified by gender: Growing Up Today Study, 2003–2010.

Alcohol use trajectory groups
Class 1. Nondrinkers Class 2. Heavy drinkers Class 3. Moderate drinkers Class 4. Legal drinkers Class 5. Escalation to moderately heavy drinkers Class 6. Light drinkers
n (%) n (%) n (%) n (%) n (%) n (%) P-value
Women
Total 521 (7.0) 1754 (23.5) 2373 (31.8) 830 (11.1) 723 (9.7) 1264 (17.0)
Sexual orientation
 Completely heterosexual 467 (7.7) 1337 (22.1) 1903 (31.4) 738 (12.2) 580 (9.6) 1032 (17.0) <0.0001
 Mostly heterosexual 43 (3.8) 339 (30.2) 379 (33.8) 74 (6.6) 109 (9.7) 177 (15.8)
 Bisexual 7 (4.0) 50 (28.7) 55 (31.6) 12 (6.9) 19 (10.9) 31 (17.8)
 Lesbian 4 (3.5) 28 (24.8) 36 (31.9) 6 (5.3) 15 (13.3) 24 (21.2)
Race/ethnicity
 White 480 (6.9) 1670 (23.9) 2228 (31.9) 763 (10.9) 675 (9.7) 1162 (16.7) 0.0038
 Non-white 41 (8.4) 84 (17.3) 145 (29.8) 67 (13.8) 48 (9.9) 102 (20.9)
Region
 West 102 (8.5) 258 (21.6) 378 (31.6) 157 (13.1) 100 (8.4) 201 (16.8) <0.0001
 Midwest 183 (7.6) 536 (22.3) 757(31.4) 305 (12.7) 231 (9.6) 396 (16.5)
 South 123 (9.3) 255 (19.3) 412 (31.1) 145 (11.0) 135 (10.2) 254 (19.2)
 Northeast 113 (4.5) 705 (27.8) 826 (32.6) 223 (8.8) 257 (10.1) 413 (16.3)
College attendance
 Never attended 36 (21.1) 36 (21.1) 34 (19.9) 19 (11.1) 16 (9.4) 30 (17.5) <0.0001
 Attended 383 (7.2) 1227 (22.9) 1704 (31.8) 606 (11.3) 545 (10.2) 891 (16.6)
 Missing 102 (5.3) 491 (25.3) 635 (32.8) 205 (10.6) 162 (8.4) 343 (17.7)
Life-time pregnancy
 No 280 (6.4) 1030 (23.6) 1424 (32.6) 509 (11.7) 450 (10.3) 676 (15.5) <0.0001
 Yes 156 (12.1) 250 (19.4) 349 (27.1) 131 (10.2) 125 (9.7) 278 (21.6)
 Missing 85 (4.7) 474 (26.2) 600 (33.2) 190 (10.5) 148 (8.2) 310 (17.2)
Mother’s household income in 2001
 < $50000 75 (9.8) 157 (20.4) 234 (30.4) 84 (10.9) 76 (9.9) 143 (18.6) <0.0001
 $50000–74999 128 (8.9) 291 (20.2) 445 (30.9) 185 (12.8) 142 (9.9) 251 (17.4)
 $75 000–99 999 81 (6.1) 300 (22.6) 437 (33.0) 148 (11.2) 137 (10.3) 223 (16.8)
 ≥ $100 000 131 (5.3) 681 (27.7) 816 (33.1) 221 (9.0) 229 (9.3) 385 (15.6)
Missing 106 (7.2) 325 (22.2) 441 (30.1) 192 (13.1) 139 (9.5) 262 (17.9)
Number of survey waves completed from 2003 to 2010
 1 29 (4.3) 195 (29.0) 227 (33.8) 73 (10.9) 40 (6.0) 108 (16.1) <0.0001
 2 63 (6.8) 228 (24.6) 292 (31.5) 106 (11.5) 68 (7.3) 169 (18.3)
 3 80 (5.9) 350 (25.7) 457 (33.5) 121 (8.9) 125 (9.2) 230 (16.9)
 4 349 (7.8) 981 (21.8) 1397 (31.0) 530 (11.8) 490 (10.9) 757(16.8)
Men
 Total 458 (9.1) 1853 (36.9) 1326 (26.4) 788 (15.7) 602 (12.0)
Sexual orientation
 Completely heterosexual 437 (9.6) 1696 (37.3) 1148 (25.3) 716 (15.8) 549 (12.1) <0.0001
 Mostly heterosexual 13 (4.3) 105 (34.9) 104 (34.6) 49 (16.3) 30 (10.0)
 Bisexual 1 (2.9) 12 (35.3) 15 (44.1) 3 (8.8) 3 (8.8)
 Gay 7 (4.8) 40 (27.4) 59 (40.4) 20 (13.7) 20(13.7)
Race/ethnicity
 White 422 (9.0) 1738 (37.2) 1248 (26.7) 715 (15.3) 555 (11.9) 0.0290
 Non-white 36 (10.3) 115 (33.0) 78 (22.4) 73 (20.9) 47 (13.5)
Region
 West 105 (12.4) 290 (34.2) 232 (27.3) 132 (15.6) 90 (10.6) <0.0001
 Midwest 145 (8.7) 607 (36.2) 430 (25.6) 285 (17.0) 210 (12.5)
 South 89 (10.6) 251 (30.0) 233 (27.8) 158 (18.9) 106 (12.7)
 Northeast 119 (7.2) 705 (42.4) 431 (25.9) 213 (12.8) 196 (11.8)
College attendance
 Never attended 26 (19.9) 35 (26.7) 39 (29.8) 23 (17.6) 8 (6.1) <0.0001
 Attended 245 (8.7) 1009 (35.9) 706 (25.1) 471 (16.8) 378 (13.5)
 Missing 187 (9.0) 809 (38.8) 581 (27.8) 294 (14.1) 216 (10.4)
Mother’s household income in 2001
 < $50000 65 (12.5) 165 (31.7) 130 (25.0) 93 (17.9) 67 (12.9) <0.0001
 $50000–74999 107 (11.0) 322 (33.2) 260 (26.8) 171 (17.6) 111 (11.4)
 $75 000–99 999 76 (8.0) 361 (37.8) 256 (26.8) 151 (15.8) 111 (11.6)
 ≥ $100 000 119 (7.1) 702 (41.6) 450 (26.7) 216 (12.8) 201 (11.9)
 Missing 91 (10.2) 303 (33.9) 230 (25.8) 157 (17.6) 112 (12.5)
Number of survey waves completed from 2003 to 2010
 1 56 (6.3) 379 (42.6) 264 (29.7) 143 (16.1) 47 (5.3) <0.0001
 2 91 (10.2) 344 (38.5) 256 (28.6) 105 (11.7) 98 (11.0)
 3 100 (9.1) 389 (35.5) 288 (26.3) 166 (15.2) 153 (14.0)
 4 211 (9.8) 741 (34.5) 518 (24.1) 374 (17.4) 304 (14.2)
Open in a new tab

P-values were derived using Rao–Scott χ2 tests for categorical variables and adjusted for sibling clusters. The light drinking class only emerged for women.

Table 4.

Results of multinomial logistic regression models predicting longitudinal alcohol use trajectory class memberships, stratified by gender: Growing Up Today Study, 2003–2010.

Heavy drinkers versus non-drinkers Moderate drinkers versus non-drinkers Legal drinkers versus non-drinkers Escalation to moderately heavy drinkers versus non-drinkers Light drinkers versus non-drinkers
OR (95% CI) P-value OR (95% CI) P-value OR (95% CI) P-value OR (95% CI) P-value OR (95% CI) P-value
Women
Sexual orientation
 Completely heterosexual 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent)
 Mostly heterosexual 3.07 (2.15,4.40) <0.0001 2.40 (1.68, 3.41) <0.0001 1.19 (0.79, 1.79) 0.4154 2.24 (1.51, 3.31) < 0.0001 2.02 (1.40, 2.93) 0.0002
 Bisexual 3.42 (1.49, 7.83) 0.0037 2.61 (1.15, 5.90) 0.0215 1.41 (0.54, 3.65) 0.4843 2.80 (1.14, 6.86) 0.0248 2.47 (1.07, 5.72) 0.0351
 Lesbian 2.24 (0.75, 6.66) 0.1470 2.08 (0.71’ 6.08) 0.1823 0.93 (0–26, 3.42) 0.9182 2.91 (0.95, 8.95) 0.0622 2.70 (0.92, 7.97) 0.0719
Men
Sexual orientation
 Completely heterosexual 1.00 (referent) 1.00 (referent) 1.00 (referent) 1.00 (referent)
 Mostly heterosexual 2.24 (1.23,4.09) 0.0084 3.34 (1.83, 6.08) <0.0001 2.29 (1.22,4.29) 0.0101 1.76 (0–90, 3.44) 0.0982
 Bisexual 2.62 (0.35, 19.71) 0.3491 5.27(0.71, 39.20) 0.1046 1.79 (0–18, 17.89) 0.6202 2.12 (0–21, 21.11) 0.5227
 Gay 1.48 (0.66, 3.35) 0.3444 3.38 (1.53, 7.48) 0.0027 1.64 (0–68,3.95) 0.2682 2.06 (0.86,4.93) 0.1053
Open in a new tab

Bold type indicates P < 0.05. Models were estimated with samples of 7465 women and 5027 men. OR = odds ratio; CI = confidence interval. The light drinking class only emerged for women. Models were adjusted for race/ethnicity, age, region, lifetime college attendance, number of surveys completed from 2003 to 2010 and mother’s household income. Models for women were also adjusted for life-time pregnancy.

Among women, the prevalence of probable AUD was 17.2% (8.7% with abuse only and 8.5% with dependence). Sexual-minority subgroups were 2.00–2.17 times more likely (P-values < 0.0001) than CHs to meet criteria for probable AUD in 2010 (Table 5; multivariable model 1). Heavy, moderate and escalation to moderately heavy drinkers were more likely than non-drinkers to evidence AUD [risk ratios (RRs) range = 11.79–29.82; P-values < 0.0001; multivariable model 2]. AUT classes mediated differences in AUD for MH and lesbian women (mediated proportions range = 30.8–31.1%; P-values range < 0.0001–0.0180). After controlling for AUT classes, MH, bisexual, and lesbian women remained more likely than CH women to evidence AUD (RRs range: 1.71–1.85; P-values range < 0.0001–0.0013).

Table 5.

Mediational effects of longitudinal alcohol use trajectory classes on sexual-orientation differences in alcohol use disorders, stratified by gender: Growing Up Today Study, 2003–2010.

Alcohol use disorder
Univariable Multivariable model 1 Multivariable model 2 Proportion mediated
% P-value RR (95% CI) P-value RR (95% CI) P-value % (P-value)
Women
Sexual orientation
 Completely heterosexual 14.0 <0.0001 1.00 (referent) 1.00 (referent)
 Mostly heterosexual 30.8 2.17(1.91, 2.47) <0.0001 1.78 (1.57, 2.01) <0.0001 30.8 (<0.0001)
 Bisexual 29.5 2.12 (1.59, 2.83) <0.0001 1.85 (1.43, 2.38) <0.0001 16.7 (0.0609)
 Lesbian 29.3 2.00 (1.41, 2.83) <0.0001 1.71 (1.23, 2.38) 0.0013 31.1 (0.0180)
Alcohol use trajectory classes
 Non-drinkers 1.3 <.0001 1.00 (referent)
Heavy drinkers 40.3 29.82 (12.34, 72.08) <0.0001
 Moderate drinkers 17.4 13.09 (5.40, 31.71) <0.0001
 Legal drinkers 3.4 2.61 (0.98, 6.94) 0.0538
 Escalation- to moderately heavy drinkers 16.3 11.79 (4.80, 28.94) <0.0001
 Light drinkers 3.4 2.56 (0.99, 6.61) 0.0519
Men
Sexual orientation
 Completely heterosexual 26.6 0.0015 1.00 (referent) 1.00 (referent)
 Mostly heterosexual 36.7 1.34 (1.10, 1.64) 0.0044 1.40 (1.18, 1.67) 0.0002 not mediated
 Bisexual 38.9 1.54(0.84, 2.80) 0.1604 1.67 (0.93, 3.01) 0.0859 not mediated
 Gay 41.4 1.59 (1.25, 2.02) 0.0001 1.72 (1.41, 2.10) <0.0001 not mediated
Alcohol use trajectory classes
 Non-drinkers 1.3 <0.0001 1.00 (referent)
 Heavy drinkers 50.8 38.69 (12.69, 117.94) <0.0001
 Moderate drinkers 19.2 13.72 (4.46, 42.21) <0.0001
 Legal drinkers 7.6 5.60 (1.75, 17.89) 0.0037
 Escalation to moderately heavy drinkers 29.6 22.39 (7.28, 68.90) <0.0001
Open in a new tab

Bold type indicates P < 0.05. Models were estimated with samples of 5122 women and 2510 men. Univariable P-values were derived using Rao–Scott χ2 tests for categorical variables and adjusted for sibling clusters. Multivariable models were adjusted for race/ethnicity, age, region, life-time college attendance (yes/no), number of surveys completed from 2003 to 2010 and mother’s household income. Models for women were also adjusted for life-time pregnancy (yes/no). RR = rate ratio; CI = confidence interval.

Among men, the prevalence of probable AUD was 28.0% (13.1% with abuse only and 14.9% with dependence). Compared with CH men, MH men were 1.34 times more likely (P = 0.0044), and gay men were 1.59 times more likely (P = 0.0001) to evidence AUD (Table 5; multivariable model 1). Heavy, moderate, legal and escalation to moderately heavy drinkers were more likely than non-drinkers to meet criteria for AUD (RRs range = 5.60–38.69; P-values range < 0.0001–0.0037; multivariable model 2). AUT classes did not mediate sexual-orientation differences in probable AUD for men. After controlling for AUTs, MH (RR = 1.40; P = 0.0002) and gay men (RR = 1.72; P < 0.0001) were more likely than CH men to meet criteria for AUD.

DISCUSSION

We found distinct AUT classes for women and men: this includes heavy, moderate, escalation to moderately heavy, legal, light (for women only) and non-drinkers. While our trajectories were similar to previous research [8], our study extended previous literature by simultaneously modeling three indicators of alcohol use for each year of the emerging adulthood period, providing rich longitudinal descriptions of alcohol use. Moderate and heavy drinkers comprised the largest classes for both men and women (55–63%), suggesting moderate to heavy alcohol use during emerging adulthood is normative. Chronic heavy and moderate alcohol use have myriad short- and long-term consequences, such as AUDs, reduced neurocognitive functioning, alcohol-impaired driving, high-risk sexual behaviors, cardiovascular diseases and cancers [20,69–81]. Non-drinkers were the smallest class. We also found a small class of legal drinkers which, to our knowledge, only emerged in a study conducted in Sweden [82]; this small AUT class may be related to minimum legal drinking age laws in the United States, which prohibit drinking before age 21.

We also found that several sexual-minority subgroups had greater odds of being in higher AUT classes than CHs, and these disparities were larger for women than men. Our longitudinal study extends prior research [15–26] by showing that certain sexual-minority subgroups have greater odds of having higher AUTs across emerging adulthood. This probably places sexual-minority emerging adults at greater risk than CHs for numerous alcohol-related problems [20,69–81]. Additionally, long-term problems may be heightened among sexual-minority populations because they begin drinking alcohol at earlier ages than CHs [24,43].

Several sexual-minority subgroups were at greater risk than CHs for probable AUDs. While AUDs are usually higher among sexual-minority adults [30–32], we corroborated the little research [33] quantifying sexual-orientation differences in AUDs during the emerging adulthood period. Furthermore, we added novel contributions by examining whether sexual-orientation differences in AUDs were mediated by longitudinal AUTs. We found longitudinal AUTs explained 30.8–31.1% of sexual-orientation disparities in AUDs for MH and lesbian women, but did not explain these associations for men. Therefore, decreasing sexual-minority populations’ alcohol use may reduce some of the burden of AUDs in MH and lesbian women, but not among sexual-minority men.

Social ecological factors may be determinants of sexual-orientation disparities in AUTs and AUDs. Prior research shows that minority stress—internalized, interpersonal or structural stigma—may influence sexual-orientation disparities in alcohol use and disorders [30,37–48]. For example, sexual-minority populations face chronic and acute stressors because of their minority sexual orientation, which can lead to drinking as a coping mechanism [37]. Furthermore, sexual-minority stressors and cultural norms (e.g. gay bar attendance) may interact with alcohol use to place sexual-minority populations at greater risk for AUDs, as posited by differential vulnerability models of health [83,84]. This may be particularly relevant for the paradoxical finding among gay men, wherein they had a higher risk of probable AUD than CH men but in most circumstances did not have higher odds of heavier AUTs. Alternatively, sexual-minority populations may be more likely than CHs to report AUD symptoms because they are more likely to be in mental health treatment [85], thereby increasing their awareness about how alcohol negatively impacts their wellbeing.

Our study has limitations. GUTS participants were sampled non-probabilistically from the United States, were predominantly college-attending non-Hispanic white and were children of mothers who were NHS2 participants; our results may not generalize to more globally, racially, ethnically or socio-economically diverse populations. For example, AUD prevalence in our sample was slightly higher in our study than national US estimates in 2010 [86]. Additional research on this topic is needed with more diverse and representative samples. Attrition bias may be present if non-response was related differentially to sexual orientation, AUT classes or AUD; however, the extent of this bias is unknown. Despite using prospective cohort data, we cannot confirm causality or temporality. Additionally, sexual orientation was measured using each participant’s last report of sexual identity/attraction; thus, our findings may not be generalizable to other operationalizations of sexual orientation (e.g. sexual behavior, sexual orientation trajectories). Some sexual-orientation subgroups were small, limiting statistical power. We also removed the few people who reported being ‘unsure’ of their sexual orientation (limiting our knowledge of this subgroup), and had few bisexual men in our study (limiting statistical power). GUTS measured past-year AUD for the first time in 2010 (precluding us from solely examining incident AUDs) using self-reported items (not clinical assessments) based on the DSM-IV criteria and not DSM-5, which slightly revised the definition of AUD [50,53]. There may be measurement error attributable to the classification of longitudinal latent classes. However, a sensitivity analysis comparing the results from classification-error corrected analyses (not shown) to our reported findings did not differ substantively. As such, we expect that classification error-corrected results of AUD analyses would be similar to our reported findings, even in the presence of minimal non-differential misclassification. We may have residual confounding, despite controlling for several covariates.

Our study extended the literature by investigating sexual-orientation differences in AUTs and AUDs in emerging adulthood. Our findings can inform future intervention and epidemiological studies. Given the sexual-orientation disparities in our study, sexual-minority emerging adults should be a priority target population for interventions aimed at reducing alcohol use and AUDs. However, scant intervention research aims to reduce alcohol use and disorders among sexual-minority populations [29]. Sexual-orientation disparities in AUDs were explained partially by AUTs for MH and lesbian women but not for men, highlighting the need for research testing additional causal mechanisms that can serve as modifiable targets for interventions. Epidemiological and intervention research can help to understand and eliminate sexual-orientation disparities in AUTs and AUDs, thereby fostering health equity for sexual-minority populations.

Supplementary Material

Appendix

Figure S1 Longitudinal latent class analysis profile plots for past-year alcohol frequency, quantity and binge drinking from ages 18 to 25 among women: Growing Up Today Study, 2003–10.

Figure S2 Longitudinal latent class analysis profile plots for past-year alcohol frequency, quantity and binge drinking from ages 18 to 25 among men: Growing Up Today Study, 2003–10.

Acknowledgements

The National Institute on Drug Abuse (awards F31DA037647 to R.W.S.C., K01DA023610 and R01DA033974 to H.L.C., K01DA034753 to J.P.C.), the Eunice Kennedy Shriver National Institute of Child Health and Human Development (award F32HD084000 to B.M.C.) and the National Center for Advancing Translational Sciences (TL1TR001858 to R.W.S.C.) of the National Institutes of Health supported this research paper. We would like to thank the Growing Up Today Study participants for the information they shared. The funding agencies had no involvement in the study design, analysis or interpretation of data, the writing of the report, or the decision to submit for publication. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

Footnotes

Declaration of interests

None.

Supporting Information

Additional Supporting Information may be found online in the supporting information tab for this article.

References

  • 1.National Institute on Alcohol Abuse and Alcoholism (NIAAA). NIAAA council approves definition of binge drinking. NIAAA Newsletter 2004;3:3. [Google Scholar]
  • 2.Arnett JJ Emerging adulthood: what is it, and what is it good for? Child Development Perspectives 2007; 1: 68–73. [Google Scholar]
  • 3.Substance Abuse and Mental Health Services Administration. 2010–2011 National Survey on Drug Use and Health. 2011. Available at: http://www.samhsa.gov/data/NSDUH/2k11State/NSDUHsae2011/NSDUHsaeUS2011.pdf (accessed 25 April 2013) (Archived at http://www.webcitation.org/6zMSB76GA on 12 May 2018). [PubMed]
  • 4.Centers for Disease Control and Prevention. Behavioral Risk Factor Surveillance System Survey Data. Atlanta, Georgia: US Department of Health and Human Services, Centers for Disease Control and Prevention; 2010. [Google Scholar]
  • 5.Muthén B Latent variable analysis: growth mixture modeling and related techniques for longitudinal data In: Kaplan D, editor. Handbook of Quantitative Methodology for the Social Sciences. Newbury Park, CA: Sage Publications; 2004, pp. 345–68. [Google Scholar]
  • 6.Maggs JL, Schulenberg JE Trajectories of alcohol use during the transition to adulthood. Alcohol Res Health 2004; 28: 195. [Google Scholar]
  • 7.Masyn KE Latent class analysis and finite mixture modeling In: Little TD, editor. The Oxford Handbook of Quantitative Methods in Psychology, Vol. 2 New York, NY: Oxford University Press; 2013, pp. 551–611. [Google Scholar]
  • 8.Nelson SE, Van Ryzin MJ, Dishion TJ Alcohol, marijuana, and tobacco use trajectories from age 12 to 24 years: demographic correlates and young adult substance use problems. Dev Psychopathol 2015; 27: 253–77. [DOI] [PubMed] [Google Scholar]
  • 9.Chassin L, Pitts SC, Prost J Binge drinking trajectories from adolescence to emerging adulthood in a high-risk sample: predictors and substance abuse outcomes. J Consult Clin Psychol 2002; 70: 67–78. [PubMed] [Google Scholar]
  • 10.Tucker JS, Ellickson PL, Orlando M, Martino SC, Klein DJ Substance use trajectories from early adolescence to emerging adulthood: a comparison of smoking, binge drinking, and marijuana use. J Drug Issues 2005; 35: 307–32. [Google Scholar]
  • 11.Schulenberg J, O’Malley PM, Bachman JG, Wadsworth KN, Johnston LD Getting drunk and growing up: trajectories of frequent binge drinking during the transition to young adulthood. J Stud Alcohol 1996; 57: 289–304. [DOI] [PubMed] [Google Scholar]
  • 12.Tucker JS, Orlando M, Ellickson PL Patterns and correlates of binge drinking trajectories from early adolescence to young adulthood. Health Psychol 2003; 22: 79–87. [DOI] [PubMed] [Google Scholar]
  • 13.Roerecke M, Rehm J Alcohol use disorders and mortality: a systematic review and meta-analysis. Addiction 2013; 108: 1562–78. [DOI] [PubMed] [Google Scholar]
  • 14.Sacks JJ, Gonzales KR, Bouchery EE, Tomedi LE, Brewer RD 2010 national and state costs of excessive alcohol consumption. Am J Prev Med 2015; 49: e73. [DOI] [PubMed] [Google Scholar]
  • 15.Coulter RWS, Marzell M, Saltz R, Stall R, Mair C Sexual-orientation differences in drinking patterns and drinking contexts among college students. Drug Alcohol Depend 2016; 160: 197–204. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16.Schauer GL, Berg CJ, Bryant LO Sex differences in psychosocial correlates of concurrent substance use among heterosexual, homosexual and bisexual college students. Am J Drug Alcohol Abuse 2013; 39: 252–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17.Ridner SL, Frost K, LaJoie AS Health information and risk behaviors among lesbian, gay, and bisexual college students. J Am Acad Nurse Pract 2006; 18: 374–8. [DOI] [PubMed] [Google Scholar]
  • 18.McCabe SE, Hughes TL, Bostwick W, Boyd CJ Assessment of difference in dimensions of sexual orientation: implications for substance use research in a college-age population. J Stud Alcohol Drugs 2005; 66: 620–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19.Eisenberg M, Wechsler H Substance use behaviors among college students with same-sex and opposite-sex experience: results from a national study. Addict Behav 2003; 28: 899–913. [DOI] [PubMed] [Google Scholar]
  • 20.Kerr DL, Ding K, Chaya J Substance use of lesbian, gay, bi-sexual and heterosexual college students. Am J Health Behav 2014; 38: 951–62. [DOI] [PubMed] [Google Scholar]
  • 21.Talley AE, Sher KJ, Steinley D, Wood PK, Littlefield AK Patterns of alcohol use and consequences among empirically derived sexual minority subgroups. J Stud Alcohol Drugs 2012; 73: 290–302. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22.Talley AE, Sher KJ, Littlefield AK Sexual orientation and substance use trajectories in emerging adulthood. Addiction 2010; 105: 1235–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Dermody SS, Marshal MP, Cheong J, Burton C, Hughes T, Aranda F et al. Longitudinal disparities of hazardous drinking between sexual minority and heterosexual individuals from adolescence to young adulthood. J Youth Adolesc 2014; 43: 30–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24.Corliss HL, Rosario M, Wypij D, Fisher LB, Austin SB Sexual orientation disparities in longitudinal alcohol use patterns among adolescents: findings from the Growing Up Today Study. Arch Pediatr Adolesc Med 2008; 162: 1071–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25.Marshal MP, Friedman MS, Stall R, Thompson AL Individual trajectories of substance use in lesbian, gay and bisexual youth and heterosexual youth. Addiction 2009; 104: 974–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26.Hatzenbuehler ML, Corbin WR, Fromme K Trajectories and determinants of alcohol use among LGB young adults and their heterosexual peers: results from a prospective study. Dev Psychol 2008; 44: 81–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27.Corliss HL, Rosario M, Wypij D, Wylie SA, Frazier AL, Austin SB Sexual orientation and drug use in a longitudinal cohort study of US adolescents. Addict Behav 2010; 35: 517–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 28.Talley AE, Gilbert PA, Mitchell J, Goldbach J, Marshall BD, Kaysen D Addressing gaps on risk and resilience factors for alcohol use outcomes in sexual and gender minority populations. Drug Alcohol Rev 2016; 35: 484–93. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Institute of Medicine. The Health of Lesbian, Gay, Bisexual, and Transgender People: Building a Foundation for Better Understanding. Washington, DC: National Academies Press; 2011. [PubMed] [Google Scholar]
  • 30.Hatzenbuehler ML, McLaughlin KA, Keyes KM, Hasin DS The impact of institutional discrimination on psychiatric disorders in lesbian, gay, and bisexual populations: a prospective study. Am J Public Health 2010; 100: 452–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31.McCabe SE, Hughes TL, Bostwick WB, West BT, Boyd CJ Sexual orientation, substance use behaviors and substance dependence in the United States. Addiction 2009; 104: 1333–45. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32.King M, Semlyen J, Tai SS, Killaspy H, Osborn D, Popelyuk D et al. A systematic review of mental disorder, suicide, and deliberate self harm in lesbian, gay and bisexual people. BMC Psychiatry 2008; 8: 1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33.Goldberg S, Strutz KL, Herring AA, Halpern CT Risk of substance abuse and dependence among young adult sexual minority groups using a multidimensional measure of sexual orientation. Public Health Rep 2013; 128: 144–52. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34.Center for Behavioral Health Statistics and Quality. 2015 National Survey on Drug Use and Health: Detailed tables. Rockville, MD: Substance Abuse and Mental Health Services Administration; 2016. [Google Scholar]
  • 35.Sher KJ, Gotham HJ, Watson AL Trajectories of dynamic predictors of disorder: their meanings and implications. Dev Psychopathol 2004; 16: 825–56. [DOI] [PubMed] [Google Scholar]
  • 36.Chassin L, Flora DB, King KM Trajectories of alcohol and drug use and dependence from adolescence to adulthood: the effects of familial alcoholism and personality. J Abnorm Psychol 2004; 113: 483–98. [DOI] [PubMed] [Google Scholar]
  • 37.Prejudice Meyer I. H., social stress, and mental health in lesbian, gay, and bisexual populations: conceptual issues and research evidence. Psychol Bull 2003; 129: 674–97. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38.Coulter RWS, Birkett M, Corliss HL, Hatzenbuehler ML, Mustanski B, Stall RD Associations between LGBTQ-affirmative school climate and adolescent drinking behaviors. Drug Alcohol Depend 2016; 161: 340–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39.Coulter RWS, Kinsky SM, Herrick AL, Stall RD, Bauermeister JA Evidence of syndemics and sexuality-related discrimination among young sexual-minority women. LGBT Health 2015; 2: 250–7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40.Hatzenbuehler ML, Keyes KM, Hasin DS State-level policies and psychiatric morbidity in lesbian, gay, and bisexual populations. Am J Public Health 2009; 99: 2275–81. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41.Rosario M, Schrimshaw EW, Hunter J Disclosure of sexual orientation and subsequent substance use and abuse among lesbian, gay, and bisexual youths: critical role of disclosure reactions. Psychol Addict Behav 2009; 23: 175–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 42.Bontempo DE, d’Augelli AR Effects of at-school victimization and sexual orientation on lesbian, gay, or bisexual youths’ health risk behavior. J Adolesc Health 2002; 30: 364–74. [DOI] [PubMed] [Google Scholar]
  • 43.Talley AE, Hughes TL, Aranda F, Birkett M, Marshal MP Exploring alcohol-use behaviors among heterosexual and sexual minority adolescents: intersections with sex, age, and race/ethnicity. Am J Public Health 2014; 104: 295–303. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44.Espelage DL, Aragon SR, Birkett M, Koenig BW Homophobic teasing, psychological outcomes, and sexual orientation among high school students: what influence do parents and schools have? School Psych Rev 2008; 37: 202. [Google Scholar]
  • 45.Amadio DM, Chung YB Internalized homophobia and substance use among lesbian, gay, and bisexual persons. J Gay Lesbian Soc Serv 2004; 17: 83–101. [Google Scholar]
  • 46.Amadio DM Internalized heterosexism, alcohol use, and alcohol-related problems among lesbians and gay men. Addict Behav 2006; 31: 1153–62. [DOI] [PubMed] [Google Scholar]
  • 47.Lehavot K, Simoni JM The impact of minority stress on mental health and substance use among sexual minority women. J Consult Clin Psychol 2011; 79: 159–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48.Hatzenbuehler ML, Pachankis JE, Wolff J Religious climate and health risk behaviors in sexual minority youths: a population-based study. Am J Public Health 2012; 102: 657–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49.Field AE, Camargo CA, Taylor CB, Berkey CS, Frazier AL, Gillman MW et al. Overweight, weight concerns, and bulimic behaviors among girls and boys. J Am Acad Child Adolesc Psychiatry 1999; 38: 754–60. [DOI] [PubMed] [Google Scholar]
  • 50.American Psychiatric Association. Diagnostic and Statistical Manual of Mental Disorders, 4th edn. Washington, DC: American Psychiatric Association; 1994. [Google Scholar]
  • 51.Research Triangle Institute. 2010 National Survey on Drug Use and Health. 2009. Available at: https://www.samhsa.gov/data/sites/default/files/NSDUH2010MRB/NSDUH2010MRB/2k10Q.pdf (accessed 9 July 2013) (Archived at http://www.webcitation.org/6zMT65Vhr on 12 May 2018). [Google Scholar]
  • 52.Cochran SD, Mays VM, Alegria M, Ortega AN, Takeuchi D Mental health and substance use disorders among Latino and Asian American lesbian, gay, and bisexual adults. J Consult Clin Psychol 2007; 75: 785–94. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 53.American Psychiatric Association Diagnostic and Statistical Manual of Mental Disorders, 5th edn. Arlington, VA: American Psychiatric Publishing; 2013. [Google Scholar]
  • 54.Remafedi G, Resnick M, Blum R, Harris L Demography of sexual orientation in adolescents. Pediatrics 1992; 89: 714–21. [PubMed] [Google Scholar]
  • 55.US Department of Health and Human Services. National Survey of Family Growth. Hyattsville, MD: US Department of Health and Human Services; 2003. Available at: http://www.cdc.gov/nchs/about/major/nsfg/nsfgback.htm (accessed 11 November 2003) (Archived at http://www.webcitation.org/6zMSTZJuF on 12 May 2018). [Google Scholar]
  • 56.Horton NJ, Kleinman KP Much ado about nothing: a comparison of missing data methods and software to fit incomplete data regression models. Am Stat 2007; 61: 79–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57.Prinzie P, Onghena P Cohort sequential design In: Everitt B, Howell D, editors. Encyclopedia of Statistics in Behavioral Science, Vol. 1 New York, NY: John Wiley and Sons; 2005, pp. 319–22. [Google Scholar]
  • 58.Muthén L, Muthén B Mplus User’s Guide, 7th edn. Los Angeles, CA: Muthén & Muthén; 1998–2012. [Google Scholar]
  • 59.Yuan KH, Bentler PM Three likelihood-based methods for mean and covariance structure analysis with nonnormal missing data. Sociol Methodol 2000; 30: 165–200. [Google Scholar]
  • 60.Muthén B, Shedden K Finite mixture modeling with mixture outcomes using the EM algorithm. Biometrics 1999; 55: 463–9. [DOI] [PubMed] [Google Scholar]
  • 61.Pedersen DE Gender differences in college binge drinking: examining the role of depression and school stress. Soc Sci J 2013; 50: 521–9. [Google Scholar]
  • 62.Ham LS, Hope DA College students and problematic drinking: a review of the literature. Clin Psychol Rev 2003; 23: 719–59. [DOI] [PubMed] [Google Scholar]
  • 63.Nylund KL, Asparouhov T, Muthén BO Deciding on the number of classes in latent class analysis and growth mixture modeling: a Monte Carlo simulation study. Struct Equ Model 2007; 14: 535–69. [Google Scholar]
  • 64.Kass RE, Raftery AE Bayes factors. J Am Stat Assoc 1995; 90: 773–95. [Google Scholar]
  • 65.Nagin D Group-based Modeling of Development. Cambridge, MA: Harvard University Press; 2005. [Google Scholar]
  • 66.Zou G A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol 2004; 159: 702–6. [DOI] [PubMed] [Google Scholar]
  • 67.Hertzmark E, Pazaris M, Spiegelman D The SAS MEDIATE macro. Boston: Brigham and Women’s Hospital, Channing Laboratory; 2012. [Google Scholar]
  • 68.Nevo D, Liao X, Spiegelman D Estimation and inference for the mediation proportion. Int J Biostat 2017; 10.1515/ijb-2017-0006. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 69.Squeglia LM, Jacobus J, Tapert SF The influence of substance use on adolescent brain development. Clin EEG Neurosci 2009; 40: 31–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 70.Casey B, Jones RM Neurobiology of the adolescent brain and behavior: implications for substance use disorders. J Am Acad Child Adolesc Psychiatry 2010; 49: 1189–201. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 71.US Department of Health and Human Services. The Health Consequences of Smoking—50 years of Progress: A Report of the Surgeon General. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health; 2014, p. 17. [Google Scholar]
  • 72.Naimi TS, Lipscomb LE, Brewer RD, Gilbert BC Binge drinking in the preconception period and the risk of unintended pregnancy: implications for women and their children. Pediatrics 2003; 111: 1136–41. [PubMed] [Google Scholar]
  • 73.Baliunas D, Rehm J, Irving H, Shuper P Alcohol consumption and risk of incident human immunodeficiency virus infection: a meta-analysis. Int J Public Health 2010; 55: 159–66. [DOI] [PubMed] [Google Scholar]
  • 74.Taylor B, Irving H, Kanteres F, Room R, Borges G, Cherpitel C et al. The more you drink, the harder you fall: a systematic review and meta-analysis of how acute alcohol consumption and injury or collision risk increase together. Drug Alcohol Depend 2010; 110: 108–16. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 75.Macdonald S, Greer A, Brubacher J, Cherpitel C, Stockwell T, Zeisser C Alcohol consumption and injury In: Boyle P, Boffetta P, Lowenfels AB, Burns H, Brawley O, Zatonski W et al. , editors. Alcohol: Science, Policy and Public Health. Oxford: Oxford University Press; 2013, pp. 171–8. [Google Scholar]
  • 76.Cherpitel CJ Focus on: the burden of alcohol use—trauma and emergency outcomes. Alcohol Res 2014; 35: 150. [PMC free article] [PubMed] [Google Scholar]
  • 77.Farmer GW, Bucholz KK, Flick LH, Burroughs TE, Bowen DJ CVD risk among men participating in the National Health and Nutrition Examination Survey (NHANES) from 2001 to 2010: differences by sexual minority status. J Epidemiol Community Health 2013; 67: 772–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 78.Farmer GW, Jabson JM, Bucholz KK, Bowen DJ A population-based study of cardiovascular disease risk in sexual-minority women. Am J Public Health 2013; 103: 1845–50. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 79.Cochran SD, Mays VM Risk of breast cancer mortality among women cohabiting with same sex partners: findings from the national health interview survey, 1997–2003. J Womens Health 2012; 21: 528–33. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 80.Austin SB, Pazaris MJ, Rosner B, Bowen D, Rich-Edwards J, Spiegelman D Application of the Rosner–Colditz risk prediction model to estimate sexual orientation group disparities in breast cancer risk in a US cohort of premenopausal women. Cancer Epidemiol Biomark Prev 2012; 21: 2201–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 81.Boehmer U, Miao X, Ozonoff A Cancer survivorship and sexual orientation. Cancer 2011; 117: 3796–804. [DOI] [PubMed] [Google Scholar]
  • 82.Virtanen P, Nummi T, Lintonen T, Westerlund H, Hägglöf B, Hammarström A Mental health in adolescence as determinant of alcohol consumption trajectories in the Northern Swedish Cohort. Int J Public Health 2015; 60: 335–42. [DOI] [PubMed] [Google Scholar]
  • 83.Hamman C, Henry R, Daley S Depression and sensitization to stressors among young women as a function of childhood adversity. J Consult Clin Psychol 2000; 68: 782–7. [PubMed] [Google Scholar]
  • 84.Nolen-Hoeksema S, Girgus JS The emergence of gender differences in depression during adolescence. Psychol Bull 1994; 115: 424–43. [DOI] [PubMed] [Google Scholar]
  • 85.Cochran SD, Sullivan JG, Mays VM Prevalence of mental disorders, psychological distress, and mental health services use among lesbian, gay, and bisexual adults in the United States. J Consult Clin Psychol 2003; 71: 53–61. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 86.Substance Abuse and Mental Health Services Administration (SAMHSA). Substance Abuse and Mental Health Services Administration (SAMHSA)’s public online data analysis system (PDAS): National Survey on Drug Use and Health, 2010. 2018. Available at: https://pdas.samhsa.gov/ (accessed 1 April 2018) (Archived at http://www.webcitation.org/6zMSe8VMD on 12 May 2018).

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Appendix

Figure S1 Longitudinal latent class analysis profile plots for past-year alcohol frequency, quantity and binge drinking from ages 18 to 25 among women: Growing Up Today Study, 2003–10.

Figure S2 Longitudinal latent class analysis profile plots for past-year alcohol frequency, quantity and binge drinking from ages 18 to 25 among men: Growing Up Today Study, 2003–10.

NIHMS1038334-supplement-Appendix.doc (622KB, doc)

RESOURCES