Perceived substance use risks among never-users: Sexual identity differences in a sample of U.S. young adults

Abstract

Introduction:

Lower perceived risk is a well-established risk factor for initiating substance use behaviors and an integral component to many health behavior theories. An established literature has shown that many substance use behaviors are more prevalent among individuals who identify as lesbian, gay, or bisexual (LGB) compared to those who identify as heterosexual. However, potential differences in perceived risk by sexual identity among individuals with no lifetime use have not been well-characterized to date.

Methods:

Data on 111,785 adults aged 18–34 (including 11,377 LGB adults) were from the 2015–2019 National Survey on Drug Use and Health. Perceived risks (classified as “great risk” vs. “less than great risk”) were assessed with 11 NSDUH survey items regarding 6 different substances (alcohol, cigarettes, marijuana, cocaine, LSD, heroin). Survey-weighted and sex-stratified logistic regression models were used to estimate sexual identity differences regarding perceived “great risk” among those reporting no lifetime use. Analyses were conducted in 2021–2022.

Results:

Gay men, bisexual men, lesbian/gay women, and bisexual women were all significantly less likely than heterosexual peers to perceive “great risk” associated with specific marijuana, cocaine, LSD, and heroin use behaviors. Bisexual men and women were also significantly less likely than heterosexual peers to perceive “great risk” associated with binge drinking behaviors and smoking 1+ packs of cigarettes daily.

Conclusions:

This novel investigation among never-users provides evidence that LGB adults perceive significantly lower risks associated with multiple substance use behaviors than heterosexual adults, which may indicate important sexual identity differences in susceptibility to substance use initiation.

INTRODUCTION

An established literature has identified substance use disparities by sexual identity, finding that young adults who identify as lesbian, gay, or bisexual (LGB) are more likely than young adults who identify as heterosexual to use tobacco,^{1, 2} alcohol,^2–4 marijuana,^{2, 5, 6} and other drugs.^{2, 7, 8} Notably, substance use risk is heterogeneous among LGB individuals – disparities are generally larger among LGB women compared to gay and bisexual men^{2, 9, 10} and also varies between bisexual and lesbian/gay individuals (especially women).^11–13 For those who have not initiated substance use, understanding and addressing key precursors that increase risk for initiation are central to prevention efforts.

Multiple health behavior theories – including the Health Belief Model and Theory of Planned Behavior – emphasize the important role of risk perceptions as a key precursor to engaging in potentially harmful behaviors including substance use.^14–17 These theories posit that perceived risk is one of several factors (including perceived benefits, perceived barriers, and perceived norms) that shape behavioral motivation, which in turn impacts health-related behaviors. Theoretically, greater perceived risk leads to a lower likelihood of engaging in a given behavior. Risk perception is an individual’s subjective appraisal of the nature and severity of harm associated with a given behavior (e.g., smoking a pack of cigarettes a day).¹⁸ Risk perceptions may be dynamic across time and impacted by factors such as information acquisition (e.g., as a result of a public health campaign), protective action (e.g., use a designated driver), and social factors (e.g., changing social acceptability of marijuana use).^19–21 Health behavior-related interventions often seek to change risk perceptions, which has been shown to be linked with changes in health behaviors across a variety of domains.²²

The hypothesized link between lower perceived risk and greater likelihood of substance use has been borne out in the empirical literature. National survey data among high school seniors has shown that lower perceived risk of marijuana use is robustly associated with actual marijuana use across several cohorts.^{21, 23} Other studies have demonstrated an association between lower perceived risk and use of alcohol, tobacco, and illicit drugs among young adults.^24–27 Multiple studies have found that lower risk perceptions of alcohol, tobacco, and other drug harms among never users are strongly associated with subsequent initiation of those substances, highlighting the importance of understanding and addressing risk perceptions among those at risk for initiation.^{21, 28–30}

Despite elevated rates of substance use among LGB individuals, few studies to date have examined the potential role of perceived risks among LGB individuals. Among a national sample, bisexual and lesbian/gay women reported significantly lower perceived risk of trying heroin compared to heterosexual women, whereas no differences were observed among gay, bisexual, and heterosexual men.³¹ Additionally, in a study of adolescents and young adults ages 15–34, lesbian/gay and bisexual individuals reported significantly lower perceived risk of prescription opioid misuse compared to heterosexual peers.³² However, prior studies among LGB individuals have primarily focused on perceived risk without regard to lifetime use status – if perceived risk is not assessed prior to initiation, it cannot inform whether lower perceived risk contributes to initiation risk among LGB individuals.

This study examined sexual identity differences in risk perceptions related to use of alcohol, cigarettes, and other drugs (marijuana, cocaine, LSD, and heroin) among a national sample of young adults (ages 18–34) reporting no lifetime use. From a prevention perspective, understanding perceived risk among individuals with no history of use is of particular interest. All results are stratified by sex and compare young adults who identify as gay, lesbian, or bisexual to young adults who identify as heterosexual. This study is novel both in its focus on perceived risks among never users as well as its inclusion of perceived risks of multiple different substances. Consideration of multiple substances is critical given variation in the timing, context, and consequences of initiation across substances. Study findings can serve to inform efforts to prevent substance use initiation.

METHODS

Study Population

Data were from the 2015–2019 NSDUH, an annual nationally-representative cross-sectional survey on drug use among the civilian, non-institutionalized US population ages 12 and older. The study sample was restricted to individuals ages 18–34 year old who identified as heterosexual, lesbian/gay, or bisexual (n=111,785, including 11,377 LGB adults). Respondents ages 12–17 were excluded as the NSDUH does not ask minors about sexual identity, as were respondents who did not respond or answered “don’t know” to the sexual identity question. This study was deemed exempt from review by RAND’s institutional review board, as it involved de-identified survey data.

Measures

Sexual identity was assessed with the following item: “Which one of the following do you consider yourself to be?” with response choices of “Heterosexual, that is, straight,” “Lesbian or gay,” “Bisexual,” and “Don’t know.” For each substance examined (alcohol, cigarettes, marijuana, cocaine, LSD, and heroin), respondents were classified as “never users” based on the corresponding NSDUH item assessing any lifetime use. Perceived risk was assessed using NSDUH’s 11 survey items that ask respondents to assess risk (as “no risk;” “slight risk;” “moderate risk;” and “great risk”) of specific substance use behaviors (e.g., “How much do people risk harming themselves physically and in other ways when [e.g., consuming 4–5 alcoholic drinks/day]”). Analyses used the NSDUH recoded variables that dichotomized responses into 1=“great risk” and 0=“less than great risk” as “great risk” is of key interest from a prevention perspective, as this highest level of perceived risk is theorized to have the strongest deterrent effect on substance use initiation. Demographic covariates included: age (categorized in NSUDH as: 18, 19, 20, 21, 22–23, 24–25, 26–29, 30–34), sex (male or female), race/ethnicity (non-Hispanic White; non-Hispanic Black; Hispanic; non-Hispanic Asian; non-Hispanic Native American/Alaskan Native; non-Hispanic Native Hawaiian/Other Pacific Islander; non-Hispanic other race/multiracial), employment (full-time; part-time; student; unemployed; other), household income (less than $20,000; $20,000-$49,999; $50,000-$74,999; $75,000+), and urbanicity (large metro area, small metro area, non-metro area).

Statistical analysis

Survey-weighted prevalences of lifetime never use were calculated by sex and sexual identity for each substance. Among never-users, survey-weighted prevalences of perceived “great risk” were calculated by sex and sexual identity for each perceived risk item. Primary analyses compared risk perceptions of LGB and heterosexual individuals who reported no lifetime use. For each NSDUH risk item, logistic regression model (comparing “great risk” to “less than great risk”) that included sexual identity, sex, and their interaction (as well as demographic factors and year fixed effects) was estimated to obtain sexual identity- and sex-specific adjusted odds ratios (ORs). Adjusted ORs are presented for lesbian/gay vs. heterosexual and bisexual vs. heterosexual comparisons; significant bisexual vs. lesbian/gay contrasts are reported. Unadjusted OR estimates are presented in the Appendix. Secondary analyses that stratified by age (18–25 and 26–34) were also conducted. All analyses were conducted during 2021–2022 in Stata version 17.0 using the svy suite to account for NSDUH survey design.

RESULTS

Descriptive characteristics of study population

The study sample included 1,499 lesbian/gay women, 6,661 bisexual women, 1,482 gay men and 1,735 bisexual men as well as 50,559 heterosexual women and 49,849 heterosexual men (Table 1). In weighted analyses, 3.0% of men identified as gay and 3.1% identified as bisexual; among women, 2.5% identified as lesbian/gay and 10.4% identified as bisexual. Bisexual men and women were younger, on average, than their heterosexual peers. Racial/ethnic composition was generally similar across sexual identity groups, although gay men and lesbian/gay women were less likely to be non-Hispanic white compared to other groups. Bisexual men and women had lower prevalence of full time employment and lower household incomes, compared to other sexual identity groups – this may, in part, reflect their younger age.

Table 1.

Demographic characteristics of 2015–2019 NSDUH respondents ages 18–34 (n=111,785)

Demographic characteristics	Heterosexual Men	Gay Men	Bisexual Men	Heterosexual Women	Lesbian/Gay Women	Bisexual Women
	n=49,849	n=1,482	n=1,735	n=50,559	n=1,499	n=6,661
Age
18–25	46.6%	45.6%	56.3%	45.0%	49.5%	58.4%
26–34	53.4%	54.4%	43.7%	55.0%	50.5%	41.6%
Employment
Full time	62.0%	60.3%	49.4%	49.2%	55.3%	42.0%
Part time	14.5%	17.5%	20.8%	21.3%	17.3%	23.2%
Student	9.5%	7.4%	11.8%	17.9%	13.3%	18.9%
Unemployed	8.4%	9.1%	10.9%	6.0%	9.5%	10.9%
Other	5.7%	5.7%	7.1%	5.6%	4.6%	5.1%
Race/ethnicity
NH White	55.7%	51.2%	59.9%	55.2%	49.8%	59.4%
NH Black	13.0%	14.1%	8.9%	13.9%	22.7%	14.3%
Hispanic/Latinx	21.6%	24.3%	21.9%	20.9%	19.8%	17.4%
NH Asian	6.6%	6.1%	5.6%	7.0%	3.1%	3.8%
NH multiracial	2.1%	3.0%	2.6%	2.0%	3.8%	4.1%
NH Native American / Alaskan Native	0.6%	0.7%	0.5%	0.7%	0.6%	0.7%
NH Native Hawaiian / Pacific Islander	0.4%	0.5%	0.7%	0.4%	0.2%	0.3%
Household income
Under $20,000	18.2%	18.0%	25.0%	21.8%	28.9%	29.2%
$20,000 – $49,999	31.8%	32.4%	34.4%	31.9%	31.8%	36.1%
$50,000 – $74,999	16.7%	15.1%	15.4%	16.0%	12.1%	13.3%
$75,000 or More	33.2%	34.5%	25.3%	30.3%	27.1%	21.4%
Urbanicity
Large metro	57.6%	70.4%	59.6%	57.8%	61.8%	56.3%
Small metro	30.2%	22.3%	30.4%	29.9%	28.3%	31.7%
Non-metro	12.2%	7.3%	10.0%	12.3%	10.0%	12.0%

Alcohol use

Across sexual identity groups, 9–16% of adults ages 18–34 reported no lifetime use of alcohol (Table 2; Figure 1). Among those with no lifetime use, perception of “great risk” associated with consuming 5+ alcoholic drinks 1–2 times per week was significantly lower for bisexual men (aOR=0.64 [0.44–0.93]) and bisexual women (aOR=0.73 [0.58–0.90]) compared to heterosexual adults of the same sex (henceforth “heterosexual peers”). Bisexual women were also significantly less likely to perceive “great risk” associated with consuming 4–5 alcoholic drinks per day compared to heterosexual women (aOR=0.72 [0.57–0.92]).

Table 2.

Perceived risk associated with specific substance use behaviors by sexual identity and sex

	Hetero-sexual Men	Gay Men (vs. Heterosexual Men)		Bisexual Men (vs. Heterosexual. Men)		Hetero-sexual Women	Lesbian/Gay Women (vs. Heterosexual Women)		Bisexual Women (vs. Heterosexual Women)
N (%) / Perceived risk items	%	%	aOR [95% CI]	%	aOR [95% CI]	%	%	aOR [95% CI]	%	aOR [95% CI]
N (%) of lifetime never-users of alcohol	7,626 (13.8%)	164 (9.8%)		280 (16.2%)		7,776 (14.5%)	156 (8.9%)		648 (9.2%)
Perceived “great risk” of:
5+ alc drinks 1–2x/wk	51%	47%	0.80 [0.53, 1.22]	40%	0.64 [0.44, 0.93]	62%	55%	0.76 [0.50, 1.16]	53%	0.73 [0.58, 0.90]
4–5 alc drinks/day	69%	69%	0.95 [0.62, 1.45]	67%	0.91 [0.62, 1.34]	78%	70%	0.69 [0.44, 1.09]	71%	0.72 [0.57, 0.92]
N (%) of lifetime never-smokers	19,808 (38.3%)	584 (38.3%)		664 (38.0%)		25,744 (50.5%)	494 (32.3%)		2,135 (31.8%)
Perceived “great risk” of:
1+ packs cigs/day	70%	71%	1.01 [0.79, 1.29]	64%	0.77 [0.62, 0.97]	77%	78%	1.07 [0.81, 1.40]*	71%	0.78 [0.68, 0.89]*
N (%) of lifetime never-users of marijuana	21,624 (42.7%)	498 (31.4%)		668 (39.7%)		25,416 (50.3%)	463 (30.1%)		1,842 (27.8%)
Perceived “great risk” of:
Marijuana use 1x/month	24%	14%	0.51 [0.36, 0.72]	17%	0.69 [0.50, 0.95]	31%	17%	0.47 [0.34, 0.65]	15%	0.43 [0.35, 0.51]
Marijuana use 1–2x/wk	30%	21%	0.64 [0.47, 0.88]	24%	0.80 [0.60, 1.05]	39%	19%	0.41 [0.30, 0.55]	18%	0.39 [0.33, 0.47]
N (%) of lifetime never-users of cocaine	41,769 (82%)	1,177 (75%)		1,395 (78%)		45,318 (89%)	1,252 (81%)		5,296 (78%)
Perceived “great risk” of:
Cocaine use 1x/month	67%	59%	0.72 [0.61, 0.85]	56%	0.67 [0.58, 0.78]	76%	69%	0.71 [0.59, 0.85]	69%	0.74 [0.68, 0.81]
Cocaine use 1–2x/wk	85%	82%	0.81 [0.66, 1.01]	82%	0.89 [0.73, 1.09]	91%	87%	0.69 [0.54, 0.88]	88%	0.77 [0.68, 0.87]
N (%) of lifetime never-users of LSD	44,033 (88%)	1,302 (87%)		1,372 (78%)		47,798 (95%)	1,344 (89%)		5,628 (84%)
Perceived “great risk” of:
Try LSD 1–2 times	56%	44%	0.60 [0.52, 0.71]	41%	0.58 [0.50, 0.68]	66%	56%	0.63 [0.54, 0.74]	50%	0.54 [0.50, 0.58]
LSD use 1–2x/wk	74%	66%	0.69 [0.58, 0.81]	65%	0.73 [0.62, 0.86]	83%	75%	0.62 [0.52, 0.74]	71%	0.55 [0.51, 0.60]
N (%) of lifetime never-users of heroin	48,557 (97%)	1,435 (97%)		1,664 (95%)		49,775 (99%)	1,447 (97%)		6,357 (95%)
Perceived “great risk” of:
Try heroin 1–2 times	83%	76%	0.65 [0.54, 0.78]	75%	0.66 [0.56, 0.78]	86%	82%	0.68 [0.56, 0.82]	84%	0.84 [0.76, 0.92]
Heroin use 1–2x/wk	94%	93%	0.89 [0.67, 1.19]	91%	0.68 [0.52, 0.87]	95%	94%	0.83 [0.61, 1.12]	95%	0.99 [0.84, 1.16]

Note: Adjusted odds ratios (aOR) are adjusted for age, race/ethnicity, household income, employment status, urbanicity, and survey year. Reference category is heterosexual adults of the same sex. Bold denotes estimates that are significant at the 0.05 level.

^*denotes significant differences between lesbian/gay and bisexual individuals of the same sex (bisexual vs. lesbian/gay contrast: OR = 0.73 [0.54, 0.98]

Figure 1.

Perceived risk associated with alcohol, cigarette, and marijuana behaviors by sexual identity and sex among never-using NSDUH respondents ages 18–34

Cigarette smoking

Half of heterosexual women; 38% of heterosexual, gay, and bisexual men; and 32% of LGB women reported no lifetime cigarette use (Table 2; Figure 1). Among those with no lifetime smoking, perception of “great risk” associated with smoking 1+ packs per day was significantly lower among bisexual men (aOR=0.77 [0.62–0.97]) and bisexual women (aOR=0.78 [0.68–0.89]) compared to heterosexual peers. Bisexual women were significantly less likely to perceive “great risk” associated with smoking 1+ packs per day relative to lesbian/gay women.

Marijuana use

Half of heterosexual women, approximately 40% of heterosexual and bisexual men, and approximately 30% of gay men and LGB women reported no lifetime marijuana use (Table 2; Figure 1). Among those with no lifetime use, perception of “great risk” associated with marijuana use once per month was significantly lower among all LGB groups relative to heterosexual peers – aORs ranged from 0.43 [0.35–0.51] for bisexual women to 0.69 [0.50–0.95] for bisexual men. Additionally, perception of “great risk” associated with marijuana use 1–2 times a week was significantly lower among gay men (aOR=0.64 [0.47–0.88]), lesbian/gay women (aOR=0.41, [0.30–0.55]), and bisexual women (aOR=0.39 [0.33–0.47]) compared to heterosexual peers.

Cocaine use

The proportion of individuals reporting no lifetime cocaine use ranged from 75% of gay men to 89% of heterosexual women (Table 2; Figure 2). Among those with no lifetime use, perception of “great risk” associated with cocaine use once per month was significantly lower among all LGB groups relative to heterosexual peers (aORs ranged from 0.67 [0.58–0.78] for bisexual men to 0.74 [0.68–0.81] for bisexual women). Perception of “great risk” associated with cocaine use 1–2 times a week was significantly lower among lesbian/gay women (aOR=0.69 [0.54–0.88]) and bisexual women (aOR=0.77 [0.68–0.87]) compared to heterosexual women.

Figure 2.

Perceived risk associated with cocaine, LSD, and heroin behaviors by sexual identity and sex among never-using NSDUH respondents ages 18–34

LSD use

The percentage of individuals reporting no lifetime LSD use ranged from 78% of bisexual men to 94% of heterosexual women (Table 2; Figure 2). Among those with no lifetime use, perception of “great risk” associated with trying LSD 1–2 times was significantly lower among all LGB groups compared to heterosexual peers; aORs ranged from 0.54 [0.50–0.58] for bisexual women to 0.63 [0.54–0.74] among lesbian/gay women. Perception of “great risk” associated with LSD use 1–2 times per week was also significantly lower among all LGB groups relative to heterosexual peers (aORs ranged from 0.55 [0.51–0.60] for bisexual women to 0.73 [0.62–0.86] for bisexual men).

Heroin use

The vast majority of individuals reported no lifetime heroin use (95–98% across sexual identity groups) (Table 2; Figure 2). Among those with no lifetime use, perception of “great risk” associated with trying heroin 1–2 times was significantly lower among gay men (aOR=0.65 [0.54–0.78]), bisexual men (aOR=0.66 [0.56–0.78]), lesbian/gay women (aOR=0.68 [0.56–0.82]), and bisexual women (aORs=0.84 [0.76–0.92]) relative to heterosexual peers. Perception of “great risk” associated with heroin use 1–2 times per week was significantly lower among bisexual men compared to heterosexual men (aOR=0.68 [0.52–0.87]).

Secondary analysis

As reported in the Appendix, for many substance use behaviors, LGB subgroups reported significantly lower perceived risks compared to their heterosexual peers for both age categories (18–25 and 26–34). However, several notable age differences emerged. For gay men and lesbian/gay women, lower perceived risks associated with cocaine behaviors were only observed for those aged 26–34. For bisexual men and bisexual women, lower perceived risks associated with binge drinking and smoking behaviors were only observed among those aged 18–25. For bisexual men, lower perceived risks associated with marijuana behaviors were only observed for those aged 18–25.

DISCUSSION

Using a national sample, this study examined perceived risks of 11 specific substance use behaviors among lesbian/gay, bisexual, and heterosexual adults aged 18–34. This study is novel in examining potential differences in perceived risks among LGB and heterosexual young adults who have no history of use and, thus, are still at risk for initiation. Perceived risk is recognized as an important factor influencing behavioral motivation and health-related behaviors^14–17 and lower perceived risk has empirically been linked to a higher likelihoods of initiating and engaging in substance use behaviors.^{21, 24, 26} Few studies to date have examined whether potential differences in perceived risks by sexual identity may be a contributing factor to substance use disparities among LGB individuals. Findings show that LGB young adults report significantly lower perceived risks associated with multiple substance use behaviors than their heterosexual peers and highlight variation in perceived risks across LGB subgroups. These findings may shed light on mechanisms behind greater substance use risk among LGB populations.

Multiple factors may contribute to LGB young adults perceiving lower substance use-related risks than heterosexual young adults. First, substance use prevalence was likely higher among the social networks of LGB individuals – even those not personally using substances – due to greater substance use among LGB individuals and social network homophily (i.e., tendency for individuals to associate with others who share similarities (such as LGB identity) with them).^33–35 Greater exposure to substance use among peers may impact risk perceptions, in part, via influence on social norms.^36–38 Additionally, perceived risks and social norms are also shaped by perceived use among peers (in addition to observed use) – studies have found that LGB individuals generally overestimate their peers’ substance use.^{39, 40} Taken together, the social context regarding substance use among LGB young adults may differ significantly from that of heterosexual young adults, contributing to lower perceived risks.

Another contributory factor may be substance use-related marketing and media depiction, given evidence that alcohol and tobacco companies have strategically targeted advertisement campaigns towards LGBTQ+ individuals.^41–43 Broadly, marketing can reduce perceived risks among adolescents and young adults, particularly for alcohol, tobacco, and marijuana.^{44, 45} While adolescents and young adults in general have high exposure and engagement with alcohol and tobacco marketing,^{45, 46} LGB young adults may have even greater exposure due to targeted campaigns (e.g., digital ads, venue ads, event sponsorship). Understanding the influence of media and marketing is particularly salient, given the rapid growth both in terms of product types (e.g., proliferating cannabis products) and advertising platform (e.g., social media, online games).

Additionally, stressors related to LGB identity may also contribute to differential risk perceptions. Many LGB individuals report using substances to cope with minority stress -- namely the stigma, prejudice, and discrimination uniquely experienced by socially marginalized individuals.^{38, 47} Using substances as a coping mechanism may directly impact perceived risk – i.e., LGB individuals may systematically discount the potential harms of substance use, relative to the severity of the phenomena driving the coping response. Experiences of discrimination and social exclusion may also negatively influence LGB adult’s sense of self-worth and, in turn, their risk calculations of potentially harmful behaviors.⁴⁸ Furthermore, an established literature has highlighted that bisexual individuals experience unique forms of minority stress compared to gay and lesbian individuals that can manifest as questioning the legitimacy of bisexuality or negative stereotypes.^{13, 49} Prior work has shown that bisexual individuals experience sexual identity-related discrimination from heterosexual individuals as well as gay and lesbian individuals and that bisexual individuals report more negative consequences of disclosing their sexual identity compared to gay and lesbian individuals.^{49, 50} These experiences of bisexual-specific stressors may contribute to findings that bisexual men and bisexual women frequently reported lower perceived risks than lesbian/gay peers.

While level of perceived risk is theorized to correspond to likelihood of use, study results indicated that perceived risks (among non-users) were not strictly correlated with overall rates of lifetime use. Notably, while marijuana ranked third behind alcohol and cigarettes in terms of lifetime use, perceived risk was lowest for marijuana use behaviors – approximately 20–40% of young adults perceived “great risk” associated with marijuana use 1–2 times a week. This may, in part, be explained by non-comparability of items – notably, the alcohol and cigarette items specify both quantity and frequency of use, yet the illicit drug items only specify frequency (having no established “safe” dosage). Perceived risk of marijuana may be more accurately assessed in future work by specifying quantity in addition to frequency. Furthermore, the salience of perceived risk may vary across time – although marijuana use among young adults has historically tracked perceived risk, emerging evidence suggests that perceived risk has increasingly diverged from likelihood of marijuana use (i.e., “decoupling of risk”)⁵¹ amid increasing social acceptability of marijuana.^52–54

Implications for Prevention Efforts

Study findings that LGB young adults reported significantly lower perceived risks across multiple substance use behaviors suggest that they may be more susceptible to initiation compared to heterosexual peers. Prevention efforts that target perceived risks and norms regarding substance use – which prior work suggests may be responsive to interventions^{55, 56} – may help to address excess substance use among LGB individuals. Tailoring these efforts to resonate with LGB young adults – e.g., utilizing peer influencers – may be particularly effective. Age-stratified findings underscore that differential perceived risks – and possibly heightened risk of initiation – persist throughout young adulthood for LGB adults. Thus, prevention efforts, which commonly focus on adolescents and college-age youth, should be inclusive of young adult populations up to age 34.

Effective efforts to change risk perceptions – and in turn substance use disparities – likely requires intervention across multiple levels, as risk perceptions are shaped by factors on individual, interpersonal, and societal levels.⁵⁷ For example, campaigns to limit or counteract targeted alcohol and tobacco marketing may represent one potential avenue of intervention. Additionally, a broad range of interventions aiming at reducing sexual identity-based stigma and discrimination (e.g., policies to reduce school-based bullying, workplace anti-discrimination laws, improving access to affirming behavioral health services) may positively impact LGB young adults’ sense of self-worth and acceptance, which may, in turn, impact their risk calculations regarding substance use.

Limitations and Directions for Future Research

As measures of sexual identity and substance use were self-reported, measurement error may be present. NSDUH data is cross-sectional, precluding examination of the longitudinal relationship between level of perceived risk and subsequent initiation. Analyses could only examine differences across sexual identity groups as assessed in the NSDUH and thus could not characterize perceived risk among individuals with sexual identities not assessed in NSDUH (e.g., pansexual). As NSDUH does not comprehensively assess gender identity, it was not possible to examine disparities by gender identity, including among transgender and non-binary individuals. Finally, sample size constraints precluded examination of potential additional heterogeneity among LGB individuals with regard to other intersecting identities, such as race/ethnicity or socioeconomic status.

CONCLUSION

Study findings indicate that, among adults aged 18–34 reporting no lifetime use of specific substances, LGB adults report lower perceived risks associated with multiple substance use behaviors compared to heterosexual adults. These findings may indicate important sexual identity differences in susceptibility to substance use initiation among young adults.

Appendix Table 1: Unadjusted odds ratio estimates for perceived risk of various substance use behaviors among never-using NSDUH respondents ages 18–34

Perceived “great risk” of:	Gay Men (vs. Heterosexual Men)	Bisexual Men (vs. Heterosexual Men)	Lesbian/Gay Women (vs. Heterosexual Women)	Bisexual Women (vs. Heterosexual Women)
	OR [95% CI]	OR [95% CI]	OR [95% CI]	OR [95% CI]
5+ alc drinks 1–2x/wk	0.85 [0.56, 1.29]	0.64 [0.45, 0.91]	0.74 [0.50, 1.11]	0.68 [0.55, 0.85]
4–5 alc drinks/day	1.00 [0.65, 1.55]	0.92 [0.63, 1.34]	0.65 [0.42, 1.01]	0.67 [0.53, 0.85]
1+ packs cigs/day	1.04 [0.81, 1.32]	0.75 [0.60, 0.93]	1.04 [0.79, 1.37]	0.73 [0.64, 0.83]
Marijuana use 1x/month	0.53 [0.38, 0.73]	0.67 [0.49, 0.93]	0.44 [0.33, 0.60]	0.37 [0.31, 0.45]
Marijuana use 1–2x/wk	0.65 [0.48, 0.89]	0.76 [0.57, 1.00]	0.37 [0.28, 0.50]	0.34 [0.29, 0.41]
Cocaine use 1x/month	0.71 [0.60, 0.84]	0.63 [0.54, 0.73]	0.73 [0.61, 0.87]	0.71 [0.65, 0.77]
Cocaine use 1–2x/wk	0.81 [0.65, 0.99]	0.84 [0.69, 1.03]	0.67 [0.52, 0.85]	0.72 [0.64, 0.81]
Try LSD 1–2 times	0.60 [0.52, 0.71]	0.54 [0.46, 0.63]	0.65 [0.56, 0.75]	0.51 [0.48, 0.55]
LSD use 1–2x/wk	0.68 [0.58, 0.80]	0.67 [0.57, 0.78]	0.61 [0.51, 0.72]	0.51 [0.47, 0.55]
Try heroin 1–2 times	0.66 [0.55, 0.79]	0.63 [0.53, 0.74]	0.72 [0.59, 0.87]	0.82 [0.74, 0.90]
Heroin use 1–2x/wk	0.92 [0.69, 1.23]	0.65 [0.50, 0.84]	0.82 [0.61, 1.11]	0.93 [0.80, 1.09]

Note: Reference category is heterosexual adults of the same sex. Bold denotes estimates that are significant at the 0.05 level.

Appendix Table 2: Adjusted odds ratio estimates for perceived “great risk” stratified by age: never-using NSDUH respondents ages 18–25 and ages 26–34

	Gay Men (vs. Heterosexual Men)	Bisexual Men (vs. Heterosexual Men)	Lesbian/Gay Women (vs. Heterosexual Women)	Bisexual Women (vs. Heterosexual Women)
Age 18–25 results:	aOR [95% CI]	aOR [95% CI]	aOR [95% CI]	aOR [95% CI]
Perceived “great risk” of:
5+ alc drinks 1–2x/wk	0.92 [0.60, 1.41]	0.63 [0.44, 0.92]	0.79 [0.50, 1.25]	0.73 [0.58, 0.92]
4–5 alc drinks/day	0.99 [0.64, 1.53]	0.99 [0.69, 1.42]	0.85 [0.51, 1.42]	0.75 [0.58, 0.97]
1+ packs cigs/day	0.94 [0.71, 1.24]	0.72 [0.57, 0.91]	0.92 [0.68, 1.26]	0.77 [0.67, 0.89]
Marijuana use 1x/month	0.54 [0.37, 0.80]	0.44 [0.29, 0.67]	0.49 [0.34, 0.71]	0.36 [0.29, 0.44]
Marijuana use 1–2x/wk	0.74 [0.51, 1.07]	0.63 [0.45, 0.87]	0.44 [0.31, 0.63]	0.34 [0.28, 0.42]
Cocaine use 1x/month	0.84 [0.69, 1.01]	0.68 [0.57, 0.80]	0.90 [0.73, 1.10]	0.80 [0.73, 0.88]
Cocaine use 1–2x/wk	0.99 [0.77, 1.26]	0.85 [0.68, 1.05]	0.92 [0.70, 1.22]	0.89 [0.77, 1.01]
Try LSD 1–2 times	0.69 [0.57, 0.83]	0.62 [0.51, 0.74]	0.64 [0.54, 0.77]	0.55 [0.51, 0.61]
LSD use 1–2x/wk	0.77 [0.64, 0.94]	0.75 [0.63, 0.90]	0.59 [0.48, 0.71]	0.60 [0.54, 0.66]
Try heroin 1–2 times	0.72 [0.59, 0.89]	0.70 [0.58, 0.84]	0.76 [0.62, 0.95]	0.92 [0.82, 1.02]
Heroin use 1–2x/wk	0.87 [0.64, 1.19]	0.71 [0.54, 0.93]	0.99 [0.70, 1.41]	1.14 [0.94, 1.37]
Age 26–34 results:
Perceived “great risk” of:
5+ alc drinks 1–2x/wk	0.58 [0.20, 1.63]	0.64 [0.27, 1.54]	0.72 [0.27, 1.90]	0.74 [0.44, 1.26]
4–5 alc drinks/day	0.89 [0.28, 2.79]	0.75 [0.29, 1.95]	0.40 [0.15, 1.08]	0.68 [0.39, 1.21]
1+ packs cigs/day	1.11 [0.73, 1.70]	0.90 [0.55, 1.45]	1.39 [0.81, 2.39]	0.78 [0.58, 1.05]
Marijuana use 1x/month	0.49 [0.28, 0.84]	1.04 [0.65, 1.67]	0.45 [0.27, 0.75]	0.52 [0.38, 0.70]
Marijuana use 1–2x/wk	0.56 [0.34, 0.94]	1.05 [0.66, 1.65]	0.37 [0.23, 0.61]	0.46 [0.35, 0.62]
Cocaine use 1x/month	0.62 [0.47, 0.81]	0.65 [0.49, 0.87]	0.55 [0.41, 0.74]	0.66 [0.57, 0.78]
Cocaine use 1–2x/wk	0.68 [0.49, 0.95]	0.94 [0.64, 1.40]	0.51 [0.35, 0.75]	0.63 [0.51, 0.79]
Try LSD 1–2 times	0.55 [0.43, 0.69]	0.54 [0.41, 0.71]	0.62 [0.48, 0.79]	0.52 [0.46, 0.60]
LSD use 1–2x/wk	0.62 [0.48, 0.80]	0.70 [0.52, 0.92]	0.67 [0.49, 0.93]	0.51 [0.43, 0.59]
Try heroin 1–2 times	0.60 [0.45, 0.80]	0.60 [0.45, 0.80]	0.60 [0.44, 0.83]	0.76 [0.64, 0.91]
Heroin use 1–2x/wk	0.91 [0.55, 1.50]	0.63 [0.40, 0.99]	0.67 [0.40, 1.11]	0.83 [0.62, 1.09]

Note: Adjusted odds ratios (aOR) are adjusted for race/ethnicity, household income, employment status, urbanicity, and survey year. Reference category is heterosexual adults of the same sex. Bold denotes estimates that are significant at the 0.05 level.