Differences in Substance Use Disorder Severity Among Sexual Minority Adolescents and Emerging Adults: Results from the 2023 National Survey on Drug Use and Health

Abstract

Purpose:

The purpose of this study was to describe disparities in the severity of substance use disorders (SUDs) between sexual minority (SM) and heterosexual youth.

Methods:

The current study analyzed data from the 2023 National Survey on Drug Use and Health to examine differences in SUDs between SM and heterosexual youth. Analyses were limited to participants aged 12–25 years (n = 23,333). Participants reported on their sexual identity and completed items on past-year substance use. SUDs were classified using the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5) criteria. A separate multinomial regression model was conducted for each outcome (alcohol use disorder [AUD] severity, cannabis use disorder [CUD] severity, and drug use disorder severity), stratified by age and sex. Each model included sexual identity as the primary predictor.

Results:

Odds for meeting criteria for an SUD at any severity level were generally highest for bisexual youth compared with all other youth, with few exceptions. AUD severity varied slightly among adolescents, and CUD varied slightly among males.

Conclusion:

This study is the first to identify differences in DSM-5 SUD severity between SM and heterosexual youth. The findings underscore the importance of developing clinical interventions tailored to the distinct challenges faced by SM youth, especially bisexual youth, to help reduce disparities in substance use outcomes.

Keywords

adolescent alcohol cannabis sexual minority substance use disorder severity

Introduction

Drug overdose has become the third leading cause of death among adolescents and second in emerging adults in the United States,^1–3 underscoring substance use as a major public health concern in young people. In 2023, an estimated 2.2 million adolescents (aged 12–17) and 9.2 million emerging adults (aged 18–25) met criteria for a substance use disorder (SUD).⁴ Sexual minority (SM) adolescents and emerging adults, referred to as SM youth hereafter, are at heightened risk for using substances,⁵ yet research from national data sources describing risk for developing SUDs among SM youth is limited.^6,7

SUDs are diagnostically defined according to a set of criteria outlined in the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5).⁸ Indicators of disordered use include physiological dependence, escalating tolerance, impaired control over use, and recurrent substance-related negative consequences. The severity of SUDs is determined by the number of diagnostic criteria that are met and include mild (2–3 symptoms), moderate (4–5 symptoms), and severe (6 or more symptoms). The severity of an SUD serves as an important indicator of the individual’s level of functional impairment, which can guide clinicians in determining the most appropriate level of care and ensure that treatment is tailored to meet individual needs.

Adolescence is a developmental period of increased risk-taking,⁹ which can heighten vulnerability to the onset of SUDs.^10,11 Studies with nonrepresentative samples frequently report that SM youth use substances at rates two to three times higher than their heterosexual peers, with studies indicating that substance use risk is highest among bisexual youth.⁵ Although not all youth who use substances develop SUDs, SM youth face a higher risk that may persist into adulthood. Consistent with this, a retrospective qualitative study with SM adults (aged 21–34) who met criteria for an SUD described beginning to use substances during adolescence to cope with experiences of discrimination and internalized heterosexism.¹² They further described substance use as a cultural norm for SM communities, where socialization with SM peers often occurred in substance-saturated environments and events.

Large epidemiological surveys in the United States have assessed patterns of substance use among adolescents and emerging adults for decades, but they have historically excluded data collection on sexual identity, particularly among adolescents, and lack information on DSM-5 SUDs disaggregated by severity.⁸ The National Longitudinal Study of Adolescent to Adult Health (Add Health) was among the first to add a measure of sexual identity in the third wave of data collection in 2001.¹³ A longitudinal analysis of Waves I (participants aged 12–19) to III (participants aged 18–26) of the Add Health study found that SM youth not only exhibited higher initial levels of alcohol, nicotine, and cannabis use but also experienced more rapid increases in use over time compared with their heterosexual peers.¹⁴ The Youth Risk Behavior Survey later added an item asking adolescents about their sexual identity in 2015.¹⁵ Analysis revealed that SM youth were at higher risk for lifetime use of 18 out of 19 substances, with relative risk ratios ranging from 1.15 (alcohol) to 3.58 (methamphetamine).¹⁶

The 2023 National Survey on Drug Use and Health (NSDUH) became the first national survey in the United States to include both measures of DSM-5 SUD criteria and sexual identity in youth under 18-years-old.^8,17 Yet disparities in SUDs between SM and heterosexual youth remain largely unexplored outside of the initial NSDUH summary report, which found that SM adolescents (aged 12–17) were two to three times more likely to meet criteria for an SUD in the past year compared with their heterosexual peers and that SM emerging adults (18–25) were at the highest risk for an SUD compared with all other age groups.^6,7 Although this represents an important first step in identifying disparities in SUDs among SM youth, it assumes equal risk across all SM identities and does not assess variation in SUD severity.

Existing research examining substance use disparities across sexual identities predominantly compares heterosexual individuals with an aggregated category of SM identities,^14,16 potentially obscuring important differences between SM identities. When studies do include multiple SM identities, they frequently compare each group to heterosexual individuals,^18,19 further overlooking meaningful disparities and heterogeneity in substance use patterns. Although research on between-SM substance use disparities is limited,^20,21 bisexual youth may be at the greatest risk for meeting criteria for SUD. A meta-analysis found that when SM identities were aggregated, SM youth were 190% more likely to use substances than heterosexual youth; when bisexual youth were analyzed separately from gay and lesbian youth, their risk increased to 340% compared with heterosexual peers.⁵ Consistent with this pattern, research among SM adults has shown that bisexual adults use substances at higher rates than gay and lesbian adults.^22–24

Understanding SUD severity disparities between-SM identity groups is critical for guiding evidence-based policy decisions and developing targeted prevention and intervention efforts to address the unique needs of SM youth. Using 2023 NSDUH data, the current study examined differences in SUD severity between bisexual youth and youth with other sexual identities including heterosexual, gay or lesbian, another sexual identity, and those unsure of their sexual identity. Post hoc analyses further describe differences in SUD severity between heterosexual youth and SM youth. This study hypothesized that bisexual youth would have higher odds of meeting criteria for mild, moderate, and severe SUDs compared with all other youth.

Methods

Data source

Data from the 2023 NSDUH survey were analyzed.²⁵ NSDUH is a nationally representative cross-sectional survey of U.S. residents that provides data on substance use and mental health. Current analyses were restricted to adolescents and emerging adults aged 12–25 years (n = 23,333). Details regarding sample design and data collection are available in NSDUH technical reports.²⁵ The Indiana University Institutional Review Board reviewed this study and exempted it as nonhuman subjects research.

Measures

Participant characteristics

Participants reported on their sex assigned at birth (“What sex were you assigned at birth?”), race and ethnicity (“Which of these groups describes you?”), and age. Response options for race and ethnicity are listed in Table 1. Age was collapsed into two categories: adolescents (aged 12–17) and emerging adults (aged 18–25).

Table 1.

Summary of Participant Characteristics

	Unweighted		Weighted
	n = 23,333		n = 56,165,374
	n	%	n	%
Sex assigned at birth
Male	11,413	48.9	28,413,582	50.6
Female	11,920	51.1	27,751,792	49.4
Race and ethnicity
Asian	1046	4.5	3,291,523	5.9
Black	3001	12.9	7,483,364	13.3
Hispanic	6037	25.9	14,003,565	24.9
Multiracial	1576	6.8	1,926,346	3.4
Native American	605	2.6	293,574	0.5
Native Hawaiian	109	0.5	158,774	0.3
White	10,959	47.0	29,008,228	51.6
Age
12–13	3238	13.9	7,228,734	12.9
14–15	3832	16.4	8,585,708	15.3
16–17	3649	15.6	8,375,082	14.9
18–19	4399	18.9	11,314,251	20.1
20–23	4851	20.8	12,527,325	22.3
24–25	3364	14.4	8,134,273	14.5
Adolescents (aged 12–17)	10,719	45.9	24,189,525	43.1
Emerging adults (aged 18–25)	12,614	54.1	31,975,849	56.9
Sexual identity
Heterosexual	17,550	75.2	42,837,848	76.3
Gay or lesbian	812	3.5	2,030,706	3.6
Bisexual	2922	12.5	6,464,463	11.5
Another identity	854	3.7	2,012,228	3.6
Unsure	1195	5.1	2,820,129	5.0

Sexual identity

Participants reported their sexual identity with a single question, “Which of the following do you consider yourself to be?” Response options included “heterosexual, that is, straight,” “gay or lesbian,” “bisexual,” “I use a different term,” “I am not sure about my sexual identity,” and “I do not know what this question is asking.” Participants who responded “I do not know what this question is asking” were excluded from the analyses.

Substance use disorders

Participants were classified as meeting criteria for an SUD if at least 2 of the 11 criteria were met in the past 12 months. SUDs were further classified by DSM-5 severity including mild (2–3 criteria), moderate (4–5 criteria), and severe (6 or more criteria). Current analyses utilized past-year alcohol use disorder (AUD) severity and past-year cannabis use disorder (CUD) severity.⁸ Given the low frequency of all other SUDs, a third variable named “drug use disorder (DUD) severity” was created to represent the maximum severity level across other SUDs including cocaine, methamphetamine, prescription stimulant misuse, heroin, opioid misuse, hallucinogen, inhalant, sedative, and tranquilizer.

Statistical analysis

Data analyses were completed in the Statistical Package for the Social Sciences version 29 (SPSS; IBM, Armonk, New York). All estimates were weighted to 2020 United States Census population estimates to account for selection probabilities and no-response patterns using the person-level weighting variable “ANALWT2_C” provided by the Substance Abuse and Mental Health Services Administration in the NSDUH analytic data file.¹⁷ Six multinomial regression models (one for each severity outcome—AUD, CUD, and DUD), stratified by age (adolescent 12–17 and emerging adult 18–25) and sex independently, were calculated. Multinomial regression models included sexual identity (reference group is bisexual) as the primary predictor and incorporated sex (in models stratified by age) and age (in models stratified by sex) as covariates. Post hoc analysis replicated these models with heterosexual as the sexual identity reference group (see Supplementary Data).

Results

Participant characteristics

Table 1 depicts a complete summary of participant characteristics. After applying person-level weights, most of the sample identified as heterosexual (76.3%), followed by bisexual (11.5%), unsure of their sexual identity (5.0%), another identity (3.6%), and gay or lesbian (3.6%). Approximately half of the sample identified as female (49.4%) and non-Hispanic White (51.6%). More than half of the sample were emerging adults aged 18–25 years (56.9%).

AUD severity

Table 2 depicts a complete summary of the multinomial logistic regression predicting AUD severity, stratified by age. Compared with bisexual adolescents, heterosexual adolescents and most other SM adolescents had lower odds of meeting criteria for a mild, moderate, or severe AUD (adjusted odds ratios [aORs] = 0.10–0.95). Three exceptions were observed: gay and lesbian adolescents were more likely to meet criteria for moderate (aOR = 4.78) or severe (aOR = 1.72) AUD compared with bisexual adolescents. Similarly, adolescents unsure of their sexual identity had higher odds of meeting criteria for a severe AUD (aOR = 2.21). Compared with bisexual emerging adults, heterosexual emerging adults and all other SM emerging adults had lower odds of meeting criteria for a mild, moderate, or severe AUD (aORs = 0.31–0.78). Supplementary Figure S1 and Supplementary Figure S2 illustrate the aORs of meeting criteria for a mild, moderate, or severe AUD across sexual identity groups among adolescents and emerging adults.

Table 2.

Summary of Multinomial Logistic Regression Predicting Past-Year Alcohol Use Disorder Severity, Stratified by Age

	Mild	Moderate	Severe
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Adolescents (aged 12–17)
Sexual identity (ref = bisexual)
Heterosexual	0.58 (0.57–0.58)	0.53 (0.52–0.54)	0.71 (0.69–0.72)
Gay or lesbian	0.95 (0.94–0.97)	4.78 (4.69–4.87)	1.72 (1.67–1.77)
Another identity	0.10 (0.09–0.10)	0.20 (0.19–0.22)	0.38 (0.37–0.40)
Unsure	0.20 (0.20–0.21)	0.39 (0.38–0.40)	2.21 (2.17–2.26)
Male (ref)
Female	1.69 (1.68–1.70)	1.58 (1.56–1.60)	2.43 (2.40–2.46)
Emerging adults (aged 18–25)
Sexual identity (ref = bisexual)
Heterosexual	0.71 (0.71–0.71)	0.56 (0.56–0.57)	0.49 (0.79–0.49)
Gay or lesbian	0.65 (0.65–0.66)	0.62 (0.61–0.62)	0.74 (0.73–0.75)
Another identity	0.73 (0.73–0.74)	0.64 (0.64–0.65)	0.50 (0.49–0.50)
Unsure	0.76 (0.75–0.76)	0.78 (0.78–0.79)	0.31 (0.30–0.31)
Male (ref)
Female	0.91 (0.90–0.91)	1.03 (1.03–1.04)	0.98 (0.98–0.99)

The reference group for the dependent variable, alcohol use disorder severity, is “no alcohol use disorder.”

aOR, adjusted odds ratio; CI, confidence interval; ref, reference.

Table 3 depicts a complete summary of the multinomial logistic regression predicting AUD severity, stratified by sex. Compared with bisexual males, heterosexual males and most other SM males had lower odds of meeting criteria for mild, moderate, or severe AUD (aORs = 0.04–0.89). The only exception was that males with another sexual identity had higher odds of meeting criteria for a moderate AUD compared with bisexual males (aOR = 1.29). Compared with bisexual females, heterosexual females and most other female SM adolescents had lower odds of meeting criteria for a mild, moderate, or severe AUD (aORs = 0.38–0.94). Two exceptions were observed: lesbian females had higher odds of meeting criteria for a moderate (aOR = 1.20) or severe AUD (aOR = 1.08) compared with bisexual females. Supplementary Figure S3 and Supplementary Figure S4 illustrate the aORs of meeting criteria for a mild, moderate, or severe AUD across sexual identity groups among males and females.

Table 3.

Summary of Multinomial Logistic Regression Predicting Past-Year Alcohol Use Disorder Severity, Stratified by Sex

	Mild	Moderate	Severe
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Male
Sexual identity (ref = bisexual)
Heterosexual	0.84 (0.83–0.84)	0.66 (0.65–0.66)	0.45 (0.44–0.45)
Gay or lesbian	0.82 (0.81–0.82)	0.64 (0.63–0.65)	0.53 (0.52–0.53)
Another identity	0.68 (0.67–0.69)	1.29 (1.27–1.31)	0.29 (0.28–0.30)
Unsure	0.89 (0.88–0.90)	0.04 (0.03–0.04)	0.35 (0.34–0.36)
Adolescents (aged 12–17)
Emerging adults (aged 18–25)	6.76 (6.72–6.79)	11.89 (11.77–12.01)	14.07 (13.91–14.23)
Female
Sexual identity (ref = bisexual)
Heterosexual	0.64 (0.64–0.64)	0.52 (0.52–0.52)	0.52 (0.51–0.52)
Gay or lesbian	0.69 (0.69–0.70)	1.20 (1.18–1.21)	1.08 (1.07–1.09)
Another identity	0.58 (0.58–0.59)	0.38 (0.37–0.38)	0.56 (0.55–0.56)
Unsure	0.49 (0.49–0.50)	0.94 (0.93–0.95)	0.65 (0.65–0.66)
Adolescents (aged 12–17)
Emerging adults (aged 18–25)	3.64 (3.63–3.66)	7.06 (7.00–7.11)	5.33 (5.30–5.37)

The reference group for the dependent variable, alcohol use disorder severity, is “no alcohol use disorder.”

CUD severity

Table 4 depicts a complete summary of the multinomial logistic regression predicting CUD severity, stratified by age. Compared with bisexual adolescents, heterosexual adolescents and most other SM adolescents had lower odds of meeting criteria for mild, moderate, or severe CUD (aORs = 0.25–0.95). One exception was observed: adolescents with another sexual identity had higher odds of meeting criteria for a mild CUD (aOR = 2.09) compared with bisexual adolescents. Compared with bisexual emerging adults, heterosexual emerging adults and most other SM emerging adults had lower odds of meeting criteria for mild, moderate, or severe CUD (aORs = 0.32–0.95). One exception was observed: emerging adults with another sexual identity had higher odds of meeting criteria for a moderate CUD (aOR = 1.57) compared with bisexual emerging adults.

Table 4.

Summary of Multinomial Logistic Regression Predicting Past-Year Cannabis Use Disorder Severity, Stratified by Age

	Mild	Moderate	Severe
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Adolescents (aged 12–17)
Sexual identity (ref = bisexual)
Heterosexual	0.85 (0.84–0.86)	0.25 (0.25–0.25)	0.33 (0.32–0.33)
Gay or lesbian	0.75 (0.74–0.77)	0.85 (0.84–0.87)	0.95 (0.93–0.96)
Another identity	2.09 (2.06–2.11)	0.76 (0.75–0.77)	0.52 (0.51–0.53)
Unsure	0.41 (0.40–0.41)	0.50 (0.49–0.50)	0.48 (0.48–0.49)
Male
Female	1.43 (1.42–1.43)	1.82 (1.80–1.84)	1.40 (1.39–1.41)
Emerging adults (aged 18–25)
Sexual identity (ref = bisexual)
Heterosexual	0.38 (0.38–0.38)	0.48 (0.47–0.48)	0.32 (0.31–0.32)
Gay or lesbian	0.52 (0.51–0.52)	0.74 (0.73–0.74)	0.70 (0.70–0.71)
Another identity	0.89 (0.88–0.89)	1.57 (1.56–1.59)	0.59 (0.58–0.60)
Unsure	0.44 (0.44–0.45)	0.95 (0.94–0.95)	0.60 (0.60–0.61)
Male
Female	0.83 (0.83–0.83)	0.56 (0.56–0.56)	0.48 (0.48–0.48)

The reference group for the dependent variable, cannabis use disorder severity, is “no cannabis use disorder.”

Table 5 depicts a complete summary of the multinomial logistic regression predicting CUD severity, stratified by sex. Compared with bisexual males, heterosexual males had lower odds of meeting criteria for mild, moderate, or severe CUD (aORs = 0.69–0.99). Gay males and males unsure of their sexual identity had lower odds of mild CUD (aORs = 0.51–0.56), but higher odds of moderate or severe CUD (aORs = 1.07–1.40) compared with bisexual males. Males with another sexual identity had higher odds of mild or moderate CUD (aORs = 1.65 and 4.10), but lower odds of severe CUD (aOR = 0.44) compared with bisexual males. Compared with bisexual females, heterosexual females and all other SM females had lower odds of meeting criteria for mild, moderate, or severe CUD (aORs = 0.25–0.90).

Table 5.

Summary of Multinomial Logistic Regression Predicting Past-Year Cannabis Use Disorder Severity, Stratified by Sex

	Mild	Moderate	Severe
	aOR (95% CI)	aOR (95% CI)	aOR (95% CI)
Male
Sexual identity (ref = bisexual)
Heterosexual	0.69 (0.69–0.70)	0.99 (0.98–1.00)	0.52 (0.52–0.52)
Gay or lesbian	0.56 (0.55–0.57)	1.07 (1.06–1.08)	1.11 (1.10–1.12)
Another identity	1.65 (1.63–1.66)	4.10 (4.05–4.16)	0.44 (0.43–0.45)
Unsure	0.51 (0.50–0.52)	1.40 (1.38–1.43)	1.28 (1.26–1.29)
Adolescents (aged 12–17)
Emerging adults (aged 18–25)	5.45 (5.43–5.48)	10.05 (9.97–10.12)	5.52 (5.49–5.55)
Female
Sexual identity (ref = bisexual)
Heterosexual	0.35 (0.35–0.35)	0.30 (0.30–0.30)	0.25 (0.25–0.25)
Gay or lesbian	0.68 (0.67–0.68)	0.89 (0.88–0.90)	0.67 (0.66–0.67)
Another identity	0.89 (0.88–0.90)	0.90 (0.88–0.90)	0.64 (0.64–0.65)
Unsure	0.39 (0.38–0.39)	0.72 (0.71–0.72)	0.35 (0.35–0.36)
Adolescents (aged 12–17)
Emerging adults (aged 18–25)	3.77 (3.76–3.78)	2.59 (2.57–2.60)	1.86 (1.85–1.87)

The reference group for the dependent variable, cannabis use disorder severity, is “no cannabis use disorder.”

Other DUD severity

Table 6 depicts a complete summary of the multinomial logistic regression predicting DUD severity, stratified by age. Compared with bisexual adolescents, heterosexual adolescents had lower odds of meeting criteria for mild DUD (aOR = 0.47). In contrast, all other SM males had higher odds of meeting criteria for mild DUD (aORs = 1.21–1.56). Heterosexual adolescents and all other SM adolescents had lower odds of meeting criteria for a moderate/severe DUD compared with bisexual adolescents (aORs = 0.21–0.97). Compared with bisexual emerging adults, heterosexual emerging adults and emerging adults unsure of their sexual identity had lower odds of meeting criteria for a mild DUD (aORs = 0.44–0.77). In contrast, gay and lesbian emerging adults and emerging adults with another sexual identity had higher odds of meeting criteria for a mild DUD (aORs = 1.13–1.71). Heterosexual emerging adults and all other SM emerging adults had lower odds of meeting criteria for a moderate/severe DUD compared with bisexual adolescents (aORs = 0.24–0.34).

Table 6.

Summary of Multinomial Logistic Regression Predicting Past-Year Drug Use Disorder Severity, Stratified by Age

	Mild	Moderate/Severe
	aOR (95% CI)	aOR (95% CI)
Adolescents (aged 12–17)
Sexual identity (ref = bisexual)
Heterosexual	0.47 (0.46–0.47)	0.18 (0.18–0.18)
Gay or lesbian	1.50 (1.47–1.52)	0.97 (0.96–0.99)
Another identity	1.56 (1.54–1.58)	0.47 (0.46–0.48)
Unsure	1.21 (1.20–1.23)	0.21 (0.20–0.21)
Male
Female	1.10 (1.10–1.11)	2.37 (2.35–2.40)
Emerging adults (aged 18–25)
Sexual identity (ref = bisexual)
Heterosexual	0.44 (0.43–0.44)	0.29 (0.29–0.29)
Gay or lesbian	1.13 (1.11–1.14)	0.24 (0.24–0.25)
Another identity	1.71 (1.70–1.73)	0.34 (0.33–0.34)
Unsure	0.77 (0.75–0.78)	0.31 (0.30–0.31)
Male
Female	0.87 (0.87–0.88)	0.62 (0.61–0.62)

The reference group for the dependent variable, drug use disorder severity, is “no drug use disorder.” Drug use disorder excludes cannabis use disorder.

Table 7 depicts a complete summary of the multinomial logistic regression predicting DUD severity, stratified by sex. Compared with bisexual males, heterosexual adolescents and most other SM males had lower odds of meeting criteria for mild DUD (aORs = 0.36–0.77). One exception was observed: males with another sexual identity were more likely to meet criteria for mild DUD (aOR = 1.35). Heterosexual males and all other SM males had lower odds of meeting criteria for a moderate/severe DUD compared with bisexual males (aOR = 0.21–0.40). Compared with bisexual females, heterosexual females had lower odds of meeting criteria for a mild DUD (aOR = 0.48). In contrast, all other SM females had higher odds of meeting criteria for a mild DUD (aORs = 1.13–1.82). Heterosexual females and all other SM females had lower odds of meeting criteria for a moderate/severe DUD compared with bisexual females (aORs = 0.19–0.60).

Table 7.

Summary of Multinomial Logistic Regression Predicting Past-Year Drug Use Disorder Severity, Stratified by Sex

	Mild	Moderate/Severe
	aOR (95% CI)	aOR (95% CI)
Male
Sexual identity (ref = bisexual)
Heterosexual	0.36 (0.35–0.36)	0.40 (0.39–0.40)
Gay or lesbian	0.56 (0.55–0.57)	0.35 (0.34–0.36)
Another identity	1.35 (1.33–1.37)	0.21 (0.21–0.22)
Unsure	0.77 (0.76–0.78)	0.23 (0.23–0.24)
Adolescents (aged 12–17)
Emerging adults (aged 18–25)	1.02 (1.02–1.03)	4.36 (4.31–4.40)
Female
Sexual identity (ref = bisexual)
Heterosexual	0.48 (0.48–0.48)	0.19 (0.19–0.19)
Gay or lesbian	1.82 (1.81–1.84)	0.60 (0.59–0.60)
Another identity	1.74 (1.72–1.76)	0.45 (0.45–0.46)
Unsure	1.13 (1.12–1.14)	0.26 (0.26–0.27)
Adolescents (aged 12–17)
Emerging adults (aged 18–25)	0.80 (0.78–0.79)	0.87 (0.87–0.88)

The reference group for the dependent variable, drug use disorder severity, is “no drug use disorder.” Drug use disorder excludes cannabis use disorder.

Supplementary analysis

Results using heterosexual youth as the reference group are presented in Supplementary Analyses and summarized in Supplementary Tables S1, S2, S3, S4, S5, and S6. With few exceptions, heterosexual SM youth generally had lower risk for AUD, CUD, and DUD at any severity level compared with SM youth.

Discussion

This is the first study to reveal differences in DSM-5 SUD severity between SM and heterosexual youth.⁸ Before the release of the 2023 NSDUH data, national estimates of SUD prevalence and severity among SM youth were not available. Official NSDUH reports indicated that SM youth were two to three times more likely to meet criteria for an SUD compared with their heterosexual peers.^6,7 Building on these findings, this study further examined disparities in the severity of SUD symptoms between bisexual, heterosexual, and other specific SM identities. Patterns of SUD severity resemble studies of substance use in both nonrepresentative and representative samples.^5,14,16 With few exceptions, patterns generally revealed that bisexual youth, regardless of age or sex, were at higher odds of meeting criteria for a mild, moderate, or severe SUD compared with all other youth. Supplementary analysis further demonstrated that SM youth, regardless of their specific sexual identity, generally met criteria for a mild, moderate, or severe AUD, CUD, and DUD at higher rates than their heterosexual peers.

AUD severity varied slightly among adolescents, with gay and lesbian adolescents being 1.72–4.78 times more likely to meet criteria for a moderate or severe AUD. However, these patterns reversed in emerging adulthood, where all other youth (including gay and lesbian) exhibited lower odds of meeting criteria for AUD at any severity level. Although future longitudinal research is needed, this shift may reflect developmental differences in alcohol use trajectories and changes in social environments. CUD severity also varied among males, with nearly all other SM identities being 1.07–4.10 times more likely to meet criteria for a mild or severe CUD. In contrast, among females, bisexual youth were at the highest risk for CUD at any severity level compared with all other youth. These patterns highlight sex-specific differences in CUD risk across sexual identities, suggesting that both sexual identity and sex may interact to shape the likelihood and severity of cannabis use.

These findings build on initial reports from the 2023 NSDUH data by identifying bisexual youth as a particularly vulnerable SM group. These findings are consistent with research on nonrepresentative SM youth, which has reported larger disparities in substance use between bisexual and heterosexual youth than between gay or lesbian youth compared with heterosexual youth.^5,14 The bisexual minority stress framework posits that stressors that are unique to bisexual individuals may contribute to the heightened vulnerability for SUDs, such as bisexual invisibility and negative stereotypes associated with identifying as bisexual.²² Bisexuality challenges the commonly accepted binary model of sexual identity, which assumes that individuals are attracted exclusively to either the same or another gender. The dominance of this binary model often leads to bisexual (and other plurisexual identities) being overlooked or erased in contemporary culture. Moreover, bisexual individuals often experience negative attitudes and beliefs about their sexual identity from heterosexual, gay, and lesbian individuals.^22,26 These negative attitudes can lead to bisexual individuals experiencing frequent microaggressions that express hostility and perpetuate harmful assumptions about their identity,^27,28 which may heighten the risk for substance use and SUDs.

Supplementary analyses further support and extend previous research,^5,14,19,21 showing that SM youth (regardless of specific identity) generally have higher odds of meeting criteria for SUDs at any severity level. Disparities in SUDs are likely attributed, in part, to unique identity-based stressors that are experienced by SM individuals.²⁹ Minority stress theory emphasizes the potential impact of sexual identity-based stressors on SM health.^30,31 These stressors include external experiences including experiences of discrimination, victimization, and rejection because of one’s sexual identity, and internal experiences including negative beliefs about oneself because of their sexual identity (e.g., internalize heterosexism). The added burden of these stressors activates cognitive (e.g., hopelessness), affective (e.g., maladaptive coping), and social processes (e.g., social isolation) that heighten the risk of psychopathology, including SUDs.³²

Culturally competent and identity-affirming treatment services are vital to addressing mental health and SUD disparities experienced by SM youth. Unfortunately, the availability of these services is limited in the United States, with only 12.6% of mental health and 17.6% of substance use facilities providing services specifically for LGBT+ individuals.³³ Substance use interventions that are tailored to meet the needs of SM youth are also limited.^34–36 Many existing programs focus on HIV and are designed primarily for adult SM males.^37–39 A self-paced, three-session online intervention for SM adolescents showed promise in improving coping and substance refusal skills.⁴⁰ Affirmative transdiagnostic approaches that build cognitive and behavioral skills to manage minority stress may also be effective,^41,42 although randomized clinical trials are needed.

These results emphasized the importance of analyzing SM identities as distinct identities whenever possible. Although collapsing data is necessary at times to preserve statistical power, the effect of collapsing data across sexual identity categories is not without risk. Collapsing data is a methodological procedure that assumes equal risk within categories, which may over- or underestimate disparities for specific subgroups. Large epidemiological data are uniquely situated to collect sexual identity data, as they are designed to systematically capture information across diverse and representative populations.

It should be noted that these findings were made possible by the inclusion of items assessing DSM-5 SUD criteria and sexual identities for youth under 18 years old in the 2023 NSDUH.⁸ DSM-5 criteria should be routinely incorporated into epidemiological studies involving adolescents, as these data are necessary for monitoring the prevalence and trends of SUD. Expanding data collection tools to capture both sexual identity and gender identity in nationally representative studies is an essential step toward identifying health disparities among sexual and gender minority youth more accurately. This information can inform evidence-based policy decisions, guide resource allocation for prevention and intervention efforts, and improve efforts to evaluate the impact of societal and legal changes on health and well-being.

Limitations

This study is not without limitations. Although there is growing evidence that transgender and gender expansive individuals use substances at higher rates than their cisgender peers,⁴³ the publicly accessible 2023 NSDUH dataset does not include a measure of gender identity.¹⁷ Therefore, estimates in the current study are limited to sex assigned at birth and are unable to investigate disparities among transgender and gender expansive youth.

Race and ethnicity were not included as a covariate in these analyses as a result of NSDUH suppression rules, which are in place to better ensure reliable estimates.²⁵ When multiple cross-sections of a dataset are computed simultaneously (e.g., age range × sexual identity × sex × SUD severity), the number of individuals within each cell can become very small, especially when an outcome is comparatively rare within a population. Calculation of SUD severity estimates across sexual identities stratified by race and ethnicity would require recruiting greater numbers of SM people of color in future epidemiological studies.

The disparities in SUD severity observed in this study may also be underestimates. Although the NSDUH provides a nationally representative sample of the eligible population, the lack of sexual identity in the sampling frame makes it unlikely that this sample is a representative sample of SM youth. Sampling methods also excluded active-duty military, individuals living in institutions, and unhoused individuals not living in shelters. This is particularly relevant for SM youth, who account for as much as 20%–50% of unhoused youth.^44–46

Conclusion

Secondary analysis of the 2023 NSDUH revealed notable differences in SUD severity between SM youth and their heterosexual peers. Effect sizes varied across SM identities, with bisexual youth generally showing the highest odds of having a more severe SUD compared with their heterosexual peers. Clinical interventions that are tailored to meet the unique needs of SM youth are needed to address these SUD disparities.

Data Availability

The 2023 NSDUH data are publicly available from SAMHSA at https://www.samhsa.gov/data/data-we-collect/nsduh-national-survey-drug-use-and-health/datafiles. Syntax for current analyses will be made upon request.

Disclaimer

The content is solely the responsibility of the authors and does not represent the official views of the funding agencies.

Authors’ Contributions

T.V.D.: Conceptualization (lead), formal analysis (lead), funding acquisition (lead), writing—original draft (equal), and writing—review and editing (equal). P.J.: Formal analysis (supporting), writing—original draft (equal), and writing—review and editing (equal). Z.W.A.: Conceptualization (supporting), writing—original draft (equal), writing—review and editing (equal).

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

The lead author was supported through the Loan Repayment Program for Health Disparities Research from the National Institute on Drug Abuse (1L60DA061409; principal investigator: Dellucci).

Supplemental Material

References

Goldstick

, Cunningham

, Carter

. Current causes of death in children and adolescents in the United States. N Engl J Med, 2022; 386(20):1955–1956; doi: 10.1056/nejmc2201761

Centers for Disease Control and Prevention. About Underlying Cause of Death, 2018-2023, Single Race. CDC WONDER: Atlanta, GA; 2024. Available from: http://wonder.cdc.gov/ucd-icd10-expanded.html [Last accessed: February 5, 2026].

Centers for Disease Control and Prevention. Leading Causes of Death: 2023, Unintentional Injuries, All Sexes, All Races, All Ethnicities. Atlanta, GA; 2023. Available from: https://wisqars.cdc.gov/lcd/?o=LCD&y1=2023&y2=2023&ct=10&cc=UNI&g=00&s=0&r=0&ry=2&e=0&ar=lcd1age&at=groups&ag=lcd2age&a1=0&a2=199 [Last accessed: June 30, 2025].

Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2023 National Survey on Drug Use and Health (HHS Publication No. PEP24-07-021, NSDUH Series H-59). Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. Rockville, MD; 2024. Available from: https://www.samhsa.gov/data/report/2023-nsduh-annual-national-report [Last accessed: January 8, 2026].

Marshal

, Friedman

, Stall

, et al. Sexual orientation and adolescent substance use: A meta‐analysis and methodological review. Addiction, 2008; 103(4):546–556; doi: 10.1111/j.1360-0443.2008.02149.x

Substance Abuse and Mental Health Services Administration. Behavioral Health of Adolescents Across Sexual Identities: Results from the 2023 National Survey on Drug Use and Health (SAMHSA Publication No. PEP24-07-028). Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. Rockville, MD; 2024. Available from: https://www.samhsa.gov/data/report/lgb-adolescent-behavioral-health-2023 [Last accessed: January 8, 2026].

Substance Abuse and Mental Health Services Administration. 2023 National Survey on Drug Use and Health. Among the Lesbian, Gay, Bisexual, or Some other Term (LGB+) Population Aged 12 or Older. Center for Behavioral Health Statistics and Quality. Rockville, MD; 2024. Available from: https://www.samhsa.gov/data/sites/default/files/reports/rpt53159/2023-nsduh-pop-slides-lbgplus.pdf [Last accessed: January 8, 2026].

American Psychiatric Association . Diagnostic and Statistical Manual of Mental Disorders (5th ed., text rev.). American Psychiatric Publishing: Washington, DC; 2022.

Steinberg

. A dual systems model of adolescent risk‐taking. Dev Psychobiol, 2010; 52(3):216–224; doi: 10.1002/dev.20445

10.

Substance Abuse and Mental Health Services Administration. Key Substance Use and Mental Health Indicators in the United States: Results from the 2022 National Survey on Drug Use and Health (HHS Publication No. PEP23-07-01-006, NSDUH Series H-58). Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration. Rockville, MD; 2023. Available from: https://www.samhsa.gov/data/report/2022-nsduh-annual-national-report [Last accessed: January 8, 2026].

11.

Chassin

, Ritter

. Vulnerability to substance use disorders in childhood and adolescence. In: Vulnerability to Psychopathology: Risk Across the Lifespan. ( Ingram

, Price

, eds.). The Guilford Press: New York, NY; 2001; pp. 107–134.

12.

Felner

, Wisdom

, Williams

, et al. Stress, coping, and context: Examining substance use among LGBTQ young adults with probable substance use disorders. Psychiatr Serv, 2020; 71(2):112–120; doi: 10.1176/appi.ps.201900029

13.

Harris

, Halpern

, Whitsel

, et al. Cohort profile: The National Longitudinal Study of Adolescent to Adult Health (Add Health). Int J Epidemiol, 2019; 48(5):1415–1415k; doi: 10.1093/ije/dyz115

14.

Marshal

, Friedman

, Stall

, et al. Individual trajectories of substance use in lesbian, gay and bisexual youth and heterosexual youth. Addiction, 2009; 104(6):974–981; doi: 10.1111/j.1360-0443.2009.02531.x

15.

Kann

, McManus

, Harris

, et al. Youth Risk Behavior Surveillance—United States, 2015. MMWR Surveill Summ, 2016; 65(6):1–174; doi: 10.15585/mmwr.ss6506a1

16.

Caputi

, Smith

, Strathdee

, et al. Substance use among lesbian, gay, bisexual, and questioning adolescents in the United States, 2015. Am J Public Health, 2018; 108(8):1031–1034; doi: 10.2105/ajph.2018.304446

17.

Center for Behavioral Health Statistics and Quality. 2023 National Survey on Drug Use and Health Public Use File Codebook. Substance Abuse and Mental Health Services Administration: Rockville, MD; 2024.

18.

Herrick

, Marshal

, Smith

, et al. Sex while intoxicated: A meta-analysis comparing heterosexual and sexual minority youth. J Adolesc Health, 2011; 48(3):306–309; doi: 10.1016/j.jadohealth.2010.07.008

19.

Schuler

, Stein

, Collins

. Differences in substance use disparities across age groups in a national cross-sectional survey of lesbian, gay, and bisexual adults. LGBT Health, 2019; 6(2):68–76; doi: 10.1089/lgbt.2018.0125

20.

Phillips

2nd , Turner

, Salamanca

, et al. Victimization as a mediator of alcohol use disparities between sexual minority subgroups and sexual majority youth using the 2015 National Youth Risk Behavior Survey. Drug Alcohol Depend, 2017; 178:355–362; doi: 10.1016/j.drugalcdep.2017.05.040

21.

Mereish

. Substance use and misuse among sexual and gender minority youth. Curr Opin Psychol, 2019; 30:123–127; doi: 10.1016/j.copsyc.2019.05.002

22.

Feinstein

, Dyar

. Bisexuality, minority stress, and health. Curr Sex Health Rep, 2017; 9(1):42–49; doi: 10.1007/s11930-017-0096-3

23.

Green

, Feinstein

. Substance use in lesbian, gay, and bisexual populations: An update on empirical research and implications for treatment. Psychol Addict Behav, 2012; 26(2):265–278; doi: 10.1037/a0025424

24.

Schuler

, Collins

. Sexual minority substance use disparities: Bisexual women at elevated risk relative to other sexual minority groups. Drug Alcohol Depend, 2020; 206:107755; doi: 10.1016/j.drugalcdep.2019.107755

25.

Substance Abuse and Mental Health Services Administration. 2023 National Survey on Drug Use and Health (NSDUH): Methodological Summary and Definitions. Center for Behavioral Health Statistics and Quality: Rockville, MD; 2024. Available from: https://www.samhsa.gov/data/report/2023-methodological-summary-and-definitions [Last accessed: January 8, 2026].

26.

Taylor

. Bisexual mental health: A call to action. Issues Ment Health Nurs, 2018; 39(1):83–92; doi: 10.1080/01612840.2017.1391904

27.

Nadal

, Issa

, Leon

, et al. Sexual orientation microaggressions: “Death by a thousand cuts” for lesbian, gay, and bisexual youth. J LGBT Youth, 2011; 8(3):234–259; doi: 10.1080/19361653.2011.584204

28.

Bostwick

, Hequembourg

. ‘Just a little hint’: Bisexual-specific microaggressions and their connection to epistemic injustices. Cult Health Sex, 2014; 16(5):488–503; doi: 10.1080/13691058.2014.889754

29.

Goldbach

, Tanner-Smith

, Bagwell

, et al. Minority stress and substance use in sexual minority adolescents: A meta-analysis. Prev Sci, 2014; 15(3):350–363; doi: 10.1007/s11121-013-0393-7

30.

Meyer

. Prejudice, social stress, and mental health in lesbian, gay, and bisexual populations: Conceptual issues and research evidence. Psychol Bull, 2003; 129(5):674–697; doi: 10.1037/0033-2909.129.5.674

31.

Frost

, Meyer

. Minority stress theory: Application, critique, and continued relevance. Curr Opin Psychol, 2023; 51:101579; doi: 10.1016/j.copsyc.2023.101579

32.

Hatzenbuehler

. How does sexual minority stigma “get under the skin”? A psychological mediation framework. Psychol Bull, 2009; 135(5):707–730; doi: 10.1037/a0016441

33.

Williams

, Fish

. The availability of LGBT‐specific mental health and substance abuse treatment in the United States. Health Serv Res, 2020; 55(6):932–943; doi: 10.1111/1475-6773.13559

34.

Blume

. Advances in substance abuse prevention and treatment interventions among racial, ethnic, and sexual minority populations. Alcohol Res, 2016; 38(1):47–54; doi: 10.35946/arcr.v38.1.06

35.

Mereish

. Addressing research gaps in sexual and gender minority adolescents’ substance use and misuse. J Adolesc Health, 2018; 62(6):645–646; doi: 10.1016/j.jadohealth.2018.03.011

36.

Coulter

, Egan

, Kinsky

, et al. Mental health, drug, and violence interventions for sexual/gender minorities: A systematic review. Pediatrics, 2019; 144(3):e20183367; doi: 10.1542/peds.2018-3367

37.

Starks

, Adebayo

, Kyre

, et al. Pilot randomized controlled trial of motivational interviewing with sexual minority male couples to reduce drug use and sexual risk: The Couples Health Project. AIDS Behav, 2022; 26(2):310–327; doi: 10.1007/s10461-021-03384-9

38.

Starks

, Dellucci

, Gupta

, et al. A pilot randomized trial of intervention components addressing drug use in couples HIV testing and counseling (CHTC) with male couples. AIDS Behav, 2019; 23(9):2407–2420; doi: 10.1007/s10461-019-02455-2

39.

Knight

, Karamouzian

, Carson

, et al. Interventions to address substance use and sexual risk among gay, bisexual and other men who have sex with men who use methamphetamine: A systematic review. Drug Alcohol Depend, 2019; 194:410–429; doi: 10.1016/j.drugalcdep.2018.09.023

40.

Schwinn

, Thom

, Schinke

, et al. Preventing drug use among sexual-minority youths: Findings from a tailored, web-based intervention. J Adolesc Health, 2015; 56(5):571–573; doi: 10.1016/j.jadohealth.2014.12.015

41.

Pachankis

, Harkness

, Jackson

, et al. Transdiagnostic LGBTQ-affirmative cognitive-behavioral therapy: Therapist guide. Oxford University Press: New York, NY; 2022.

42.

Burton

, Wang

, Pachankis

. Psychotherapy for the spectrum of sexual minority stress: Application and technique of the ESTEEM treatment model. Cogn Behav Pract, 2019; 26(2):285–299; doi: 10.1016/j.cbpra.2017.05.001

43.

Day

, Fish

, Perez-Brumer

, et al. Transgender youth substance use disparities: Results from a population-based sample. J Adolesc Health, 2017; 61(6):729–735; doi: 10.1016/j.jadohealth.2017.06.024

44.

Choi

, Wilson

, Shelton

, et al. Serving our youth 2015: The Needs and Experiences of Lesbian, Gay, Bisexual, Transgender, and Questioning Youth Experiencing Homelessness. The Williams Institute with True Colors Fund: Los Angeles, CA; 2015.

45.

Rosario

, Schrimshaw

, Hunter

. Risk factors for homelessness among lesbian, gay, and bisexual youths: A developmental milestone approach. Child Youth Serv Rev, 2012; 34(1):186–193; doi: 10.1016/j.childyouth.2011.09.016

46.

Ormiston

. LGBTQ youth homelessness: Why we need to protect our LGBTQ youth. LGBT Health, 2022; 9(4):217–221; doi: 10.1089/lgbt.2021.0324

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.33 MB

41.60 MB

0.33 MB

0.02 MB

0.03 MB

0.02 MB

37.84 MB

0.00 MB

0.03 MB

0.61 MB

0.03 MB

41.90 MB

0.02 MB

37.84 MB

0.40 MB

0.02 MB