Coping and Social Support in Relation to Minority Stress and Cigarette Smoking Among Lesbian, Gay, and Bisexual Veterans

Abstract

The intersection between a minoritized sexual orientation identity and a U.S. military Veteran status places lesbian, gay, and bisexual (LGB) Veterans at increased risk for cigarette smoking. Guided by the Minority Stress Model, this study assessed whether coping and three types of social support (general, Veteran-specific, and lesbian, gay, bisexual, and transgender [LGBT]-specific) moderated the association between minority stressors and past-year smoking among LGB Veterans. Participants were recruited online for a prospective cohort study. We conducted secondary data analysis of baseline surveys collected from September 2019 to December 2020. The study sample included cisgender LGB Veterans (N = 463). Adjusted multivariable logistic regression models estimated the odds of past-year smoking with interaction terms between minority stressors and coping/social support to test for moderation. Four statistically significant interaction terms were found. Higher versus lower levels (i.e., one-point score increase) of coping buffered the relationship between victimization and past-year smoking; Veteran-specific social support buffered the relationship between interpersonal LGB military stress and past-year smoking; and LGBT-specific social support buffered the relationship between intrapersonal LGB military stress and past-year smoking. However, general social support strengthened the relationship between social exclusion and past-year smoking. Findings provide some evidence for the minority stress model; however, regarding cigarette smoking, coping and social support may mitigate stress in some cases and exacerbate stress in others. LGB Veterans may benefit from learning positive coping skills and leveraging social support linked to LGB and Veteran identities to support smoking cessation.

INTRODUCTION

The high prevalence of cigarette smoking among lesbian, gay, and bisexual (LGB) individuals is well-documented.^1–5 LGB tobacco disparities are often attributed to the notion that LGB individuals are exposed to sexual orientation-related prejudice and stigma, which results in stress. Indeed, the Minority Stress Model^6–8 is often used to explain how occurrences of minority stressors influence the continued use of cigarettes among LGB individuals. Prior studies have primarily examined minority stressors as risk factors for smoking; however, the evidence in support of these associations is mixed.^2,9–15

Fewer studies have explored potential protective factors that may mitigate the relationship between minority stressors and smoking. The Minority Stress Model proposes that an individual’s characteristics may lead them to build and maintain stronger relationships with others who are alike, which can create an affirming and validating community that provides resources to cope with stress.^7,8 For example, individuals may acquire coping skills from others,⁷ and the availability of social support provides opportunities to receive emotional support, including receiving help from, counting on, and talking with others when in need.^16,17 Thus, the Minority Stress Model hypothesizes that coping and social support act as moderators that minimize the perception and internalization of stress and build resiliency, which can buffer the adverse effects of stress.^6–8 Studies on social support among LGB individuals have found that higher social support was associated with lower risk for smoking among LGB young adults (18–24 years),¹⁸ and moderates the association between sexual orientation-related discrimination and tobacco use disorder among gay and bisexual adult men.¹⁹ More broadly, studies suggest that higher social support buffers the effects of stress and is beneficial for smoking cessation outcomes, including cessation attempts, reduced relapse, and abstinence.^20,21 However, there is a lack of comprehensive evidence on coping and social support in relation to minority stress and cigarette smoking among LGB individuals.

In the United States, military Veterans also have a higher prevalence of cigarette smoking than the general population.^22–26 Evidence suggests that the intersection between a minoritized sexual orientation identity and Veteran status places LGB Veterans at increased risk for smoking.^27,28 LGB Veterans may be exposed to minority stressors in their day-to-day civilian life and when they served in the military. For example, LGB service members may have hid their sexual orientation particularly during the “Don’t Ask, Don’t Tell” policy,²⁹ where service members could be discharged for disclosing their minoritized sexual orientation. The high stigmatizing military environment may have also limited the availability of and access to social support from other LGB service members for fear of being outed.³⁰

On the other hand, the intersection between a minoritized sexual orientation identity and Veteran status may provide LGB Veterans additional resources to cope with stress. LGB Veterans may rely on their coping skills; general social support from friends, family, and/or a significant other; and social support from the LGB community and from the military/Veteran community. Findings among active-duty service members indicate that those who smoke were less likely to talk to a family member or friend for support to cope with military stress (e.g., being deployed) or family stress (e.g., balancing family responsibilities while serving) compared with those who did not smoke.³¹ Among Veterans, some findings suggest that low social support was a barrier to smoking cessation, while others found no differences in smoking cessation when increasing family social support.^32,33 For LGB Veterans, it is unclear whether they rely on all forms of social support equally, differentially, or not at all.

We tested whether coping and three types of social support (i.e., general, Veteran-specific, and lesbian, gay, bisexual, and transgender [LGBT]-specific) moderated the association between minority stressors and past-year cigarette smoking among LGB Veterans (Supplemental Figure 1; available at https://connect.springerpub.com/journals in the PDF view). Based on the Minority Stress Model,^6–8 we hypothesized that the strength of the association between stressors and past-year smoking would be smaller among those with higher versus lower levels of coping and/or social support.

METHODS

Data Source

The Health for Every Veteran Study used a prospective cohort design to understand mental health problems and health risk behaviors among heterosexual and LGBT Veterans. Details on recruitment and eligibility criteria have been previously reported.^15,34 Briefly, potential participants were directed to the study website, which provided an information statement that described the study and its risks and benefits. If they consented to participate, they completed an eligibility screener. Eligible participants were at least 18 years old; reported having ever served in active duty in the U.S. Armed Forces in the past, but not on active duty now; received an honorable or general discharge (under honorable conditions); were currently living in the United States; and had routine access to the internet. Based on self-reported responses to birth sex, current gender identity, and sexual orientation identity (straight/heterosexual, gay/lesbian, or bisexual) items from the eligibility screener, participants were assigned into subgroups (e.g., heterosexual, LGB, or transgender group) and were shown specific questions on the baseline survey based on their identities.³⁴ Participants then completed four web-based surveys every nine months for 27 months. The baseline survey took approximately 60–90 minutes to complete, and participants received $30 as compensation. This research was approved by the Institutional Review Board of University of Washington.

Analytic Sample

We conducted secondary analysis of baseline surveys collected from September 2019 to December 2020. Our analytic sample (N = 463) included cisgender (cis; current gender identity the same as sex assigned at birth) Veterans who self-identified as LGB. We excluded cis heterosexual Veterans (n = 343) because our primary independent variables focused on LGB-related stressors. Transgender (n = 189) and gender diverse (n = 66; TGD) Veterans did not receive the same measures as cis LGB participants and were excluded from the current study. We acknowledge that we may have excluded TGD Veterans who identify as LGB. One cis LGB Veteran was excluded due to missing data on primary variables.

Measures

Dependent Variable

Past-Year Cigarette Smoking Status. Participants who reported having ever smoked more than 100 cigarettes in their life and having smoked a cigarette (even one or two puffs) in the past year were categorized as those with “past-year smoking,” and all others were categorized as those with “no past-year smoking.”

Independent Variables

Discrimination.

The 10-item Everyday Discrimination Scale asked participants to report the frequency of discriminatory experiences in the past year, ranging from 1 (never) to 6 (almost everyday), and the main reason for the experience (e.g., race, age, sexual orientation)^35,36; for example, “You were treated with less courtesy than other people.” To assess sexual orientation-related discrimination, responses were recoded to 1 (at least once in the past year because of sexual orientation) and 0 (never or for some other reason). The total sum score (range 0–10) was calculated, with higher scores indicating more occurrences of sexual orientation-related discrimination in the past year (α = .83).

Harassment, Victimization, Trauma, Social Exclusion, and Family Rejection.

The Daily Heterosexist Experiences Questionnaire is a 50-item measure of sexual orientation-related experiences and events of distress.³⁷ We used a six-item subscale to assess harassment (e.g., “being verbally harassed by strangers for being LGBT”; α = .88), a four-item subscale to assess victimization (e.g., “being punched, hit, kicked, or beaten because you are LGBT”; α = .86), a six-item subscale to assess trauma (e.g., “hearing about LGBT people you know being treated unfairly”; α = .85), a four-item subscale to assess social exclusion (e.g., “having very few people you can talk to about being LGBT”; α = .82), and a six-item subscale to assess family rejection (e.g., “your family avoiding talking about your LGBT identity”; α = .87). Participants were asked about the occurrence and how much these experiences and events distressed or bothered them in the past year. Response options presented on the questionnaire ranged from 0 (did not happen/not applicable to me) to 5 (it happened and it bothered me extremely). We followed recommended scoring guidelines,³⁷ which dichotomize occurrences to 0 (did not happen/not applicable) and 1 (yes, it happened). We calculated the total sum scores, with higher scores indicating more occurrences within each subscale.

Identity Concealment.

The six-item Sexual Orientation Concealment Scale asked participants about behaviors that LGB individuals sometimes use to hide their sexual orientation³⁸; for example, “In the last year, I have concealed my sexual orientation by telling someone that I was straight or denying that I was LGB.” Participants indicated the frequency of these behaviors on a 5-point Likert scale ranging from 1 (not at all) to 5 (all the time). We calculated the total mean score (range 1–5), with higher scores indicating higher identity concealment (α = .86).

Internalized Homophobia.

The 10-item Internalized Homophobia Scale assessed participants’ thoughts and feelings about being LGB³⁹; for example, “If someone offered you the chance to be completely heterosexual, you would have accepted the offer.” Participants indicated the frequency of these thoughts and feelings in the past year on a 4-point Likert scale ranging from 1 (never) to 4 (often). The total mean score was calculated (range 1–4), with higher mean scores indicating higher internalized homophobia (α = .86).

Interpersonal LGB Military Stressors.

The eight-item Military Events and Sexual Orientation Scale was used to assess interpersonal events that may have been initiated by the U.S. military to investigate or punish individuals based on their LGB sexual orientation⁴⁰; for example, “While in the military, were you ever interrogated or investigated regarding your sexual orientation?” Participants indicated whether these events occurred: 1 (yes) or 0 (no). Scoring guidelines dichotomize responses to 1 (yes) if participants answered “yes” to any of the items or 0 (no) if they answered “no” to all items, since the events are considered relatively rare (α = .76).

Intrapersonal LGB Military Stressors.

An adapted eight-item Military Experience and Sexual Orientation Scale was used to measure intrapersonal anxiety about sexual orientation and identity concealment.⁴¹ Participants rated their agreement or disagreement with statements about challenging situations LGB individuals could have faced while in the military. Response options ranged from 1 (strongly disagree) to 5 (strongly agree). Example items include: “In the service, I was constantly trying to conceal my sexual orientation.” The total sum score was calculated (range 8–40), with higher sum scores indicating higher anxiety about concealing sexual orientation during military service (α = .78).

Moderator Variables

Coping.

Coping was assessed with the 13-item Coping Self-Efficacy Scale.⁴² Participants rated their confidence and certainty in coping when things are not going well or when they are having problems on an 11-point scale ranging from 0 (cannot do at all) to 10 (certain can do); for example, “How confident or certain are you that you can: make a plan of action and follow it when confronted with a problem.” The total mean score was calculated (range 0–13), with higher scores indicating higher levels of coping (α = .94).

General, Veteran-Specific, and LGBT-Specific Social Support.

The 12-item Multidimensional Scale of Perceived Social Support (MSPSS) was used to assess general social support from a significant other, family, or friends¹⁷; for example, “I can talk about my problems with my family” (α = .95). An adapted five-item MSPSS was used to assess Veteran-specific social support from relationships with other Veterans or friends in the military¹⁷; for example, “I can count on my military or Veteran friends when things go wrong” (α = .97). An adapted four-item MSPSS was used to assess LGBT-specific social support from relationships with other LGBT people or people with related identities¹⁷; for example, “My LGBT friends really try to help me” (α = .97). For all social support scales, participants indicated their agreement or disagreement with each statement on a 7-point Likert scale ranging from 1 (very strongly disagree) to 7 (very strongly agree). We calculated the total mean scores, with higher scores indicating greater social support within each scale.

Covariates

Sociodemographic.

Sociodemographic characteristics included age in years, gender identity (man or woman), sexual orientation (gay/lesbian or bisexual), race and ethnicity (non-Hispanic White, non-Hispanic another racial identity, or Hispanic), relationship status (in a relationship, i.e., married, domestic partnership, or dating; or single), education (some college/high school or less, or college graduate or more), income ($39,999 or less, $40,000–$79,999, or $80,000 or more), and healthcare coverage (yes or no).

Mental Health Conditions.

Posttraumatic stress disorder (PTSD) symptoms were assessed using the 20-item Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition PTSD Checklist.^43,44 Participants indicated to what extent problems, originating from stressful experiences or events that occurred at any point in their life, bothered them in the past month. Response options ranged from 0 (not at all) to 4 (extremely). The total symptom score was calculated (range 0–80), with a cutoff score of ≥33 indicating probable PTSD symptoms.⁴³ Depression was assessed using the eight-item Patient Health Questionnaire, which measures the frequency of depression symptoms in the past two weeks.^45,46 Participants indicate how often they were bothered by the symptoms, ranging from 0 (not at all) to 3 (nearly everyday). The total score was calculated (range 0–24), with a cutoff score of ≥10 indicating probable depression symptoms.

General Stress.

The 16-item Social Stress Scale was used to measure instances of social and personal life experiences and events in the past year that could have been stressors.⁴⁷ Participants were asked whether these stressors occurred: 1 (yes) or 0 (no). The total score was calculated (range 0–16), with higher scores indicating more occurrences of general stressors.

Data Analysis

We calculated descriptive statistics for all variables. Bivariate analyses were conducted to assess the associations between coping/social support with past-year smoking. Adjusted multivariable logistic regression models with 95% confidence interval (CI) estimated the odds of past-year smoking with an interaction term between stressors and coping/social support to test for moderation, with separate models for each interaction. Input variables were centered to reduce issues of multicollinearity.^48,49 Models with statistically significant interaction terms at α = .05 were probed using simple slopes.^49–51 For interactions between continuous stressors and continuous moderators, simple slopes were plotted at three selected values, the overall sample mean and ± 1 standard deviation (SD) from the mean. For interactions between categorical stressors and continuous moderators, simple slopes were plotted using dummy coding to examine differences between groups. Covariates for adjusted analyses included age, education, income, healthcare coverage, PTSD, depression, and general stressors. Covariates were selected based on previously reported findings among LGB Veterans, which found significant bivariate associations between these covariates and past-year smoking.¹⁵ In post-hoc exploratory analyses, we examined whether baseline survey completion date relative to COVID-19 had potential impacts on smoking status and our main findings. Survey completion was categorized as (1) period prior to COVID-19/international but not U.S. concern (September 1, 2019–January 21, 2020), (2) period of growing U.S. concern (January 22, 2020–March 10, 2020), and (3) period of global concern (March 11, 2020–December 31, 2020) based on key dates from the COVID-19 timeline.⁵² All analyses were conducted using Stata/MP Version 17.0 (StataCorp, College Station, TX).

RESULTS

Participant Characteristics

Table 1 presents participant characteristics for the overall sample and by past-year cigarette smoking status (N = 463). The mean age was 49.53 years (SD = 14.71), and most participants were men (54.0%, n = 250), non-Hispanic White (82.1%, n = 380), in a relationship (66.9%, n = 310), and were at least a college graduate (58.5%, n = 271). In the sample, 21.2% (n = 98) were categorized as LGB Veterans with past-year smoking.

TABLE 1.

Participant Characteristics of the Overall Sample and by Past-Year Cigarette Smoking Status

	Overall Sample (N = 463)	Past-Year Smoking (n = 98)	No Past-Year Smoking (n = 365)	P value
	Mean (Standard Deviation)
Age (years)	49.53 (14.71)	43.49 (11.93)	51.13 (14.97)	<.001*
General stressors (range 0–16)	2.73 (2.29)	3.64 (2.59)	2.48 (2.13)	<.001*
	Frequency (%)
Gender				.51
Man	250 (54.0)	50 (51.0)	200 (54.8)
Woman	213 (46.0)	48 (49.0)	165 (45.2)
Sexual orientation				.07
Gay/Lesbian	346 (74.7)	66 (67.3)	281 (77.0)
Bisexual	117 (25.3)	33 (33.7)	84 (23.0)
Race and ethnicity				.25
Non-Hispanic White	380 (82.1)	78 (79.6)	302 (82.7)
Non-Hispanic another identitya	46 (9.9)	10 (10.2)	36 (9.9)
Hispanic	36 (7.8)	9 (9.2)	27 (7.4)
Relationship status				.93
In a relationshipb	310 (66.9)	65 (67.0)	245 (67.1)
Single	150 (32.4)	32 (33.0)	118 (32.3)
Education				<.001**
Some college/high school or less	192 (41.5)	56 (57.1)	136 (37.3)
College graduate or more	271 (58.5)	42 (42.9)	229 (62.7)
Income (annual household)				<.001**
$39,999 or less	128 (27.7)	40 (40.8)	88 (24.1)
$40,000–$79,999	167 (36.1)	39 (39.8)	128 (35.1)
$80,000 or more	167 (36.1)	19 (19.4)	148 (40.6)
Healthcare coverage, yes	401 (86.6)	77 (78.6)	324 (88.8)	.02***
PTSD symptomsc	182 (39.3)	49 (50.0)	133 (36.4)	.01***
Depression symptomsd	182 (39.3)	49 (50.0)	133 (36.4)	.01***
Ever smoking, yes	374 (80.8)
Lifetime smokinge, yes	258 (69.0)

Note. Frequencies and proportions may not equal 100% due to rounding and/or missing data.

Abbreviation: PTSD = posttraumatic stress disorder.

^*Welch’s t-test significant at P < .01.

^**χ² test of independence significant at P < .01.

^***χ² test of independence significant at P < .05.

^aIncludes non-Hispanic Black, multiracial, and “other” race identities. Race and ethnic groups were combined due to small cell sizes.

^bIncludes being married, in a domestic partnership, or dating.

^cDiagnostic and Statistical Manual of Mental Disorders, Fifth Edition PTSD Checklist cutoff score of ≥33 indicating probable PTSD symptoms.

^dPatient Health Questionnaire cutoff score of ≥10 indicating probable depression symptoms.

^eSmoked 100 or more cigarettes (five packs or more) in entire life.

Differences in Coping and Social Support by Past-Year Cigarette Smoking Status

In bivariate analyses, LGB Veterans with past-year smoking reported lower coping (M = 4.97 [SD = 1.93] vs. M = 6.01 [SD = 2.05]; P < .001), lower general social support (M = 4.08 [SD = 1.56] vs. M = 4.62 [SD = 1.70]; P = .003), and lower LGBT social support (M = 3.81 [SD = 1.89] vs. M = 4.34 [SD = 1.94]; P = .016) versus those with no past-year smoking (results not shown in tables).

Moderation Analyses

The results of adjusted multivariable logistic regression models estimating the odds of past-year smoking with an interaction term between stressors and coping/social support are presented in Table 2. Four statistically significant interaction terms were found.

TABLE 2.

Adjusted Logistic Regression Models Estimating the Odds of Past-Year Smoking With Interaction Terms Between Stressors and Coping or Social Support

	Adjusted Odds Ratio (95% Confidence Interval)
	Coping	General Social Support	Veteran Social Support	LGBT Social Support
Interpersonal Minority Stressors
Discrimination	0.98 (0.93, 1.03)	0.99 (0.94, 1.04)	1.01 (0.96, 1.06)	1.00 (0.95, 1.05)
Harassment	0.95 (0.90, 1.00)	1.00 (0.95, 1.06)	0.96 (0.91, 1.00)	0.98 (0.93, 1.03)
Victimization	0.91 (0.84, 0.99)*	1.02 (0.94, 1.12)	0.95 (0.88, 1.03)	1.01 (0.93, 1.10)
Trauma	0.95 (0.88, 1.01)	1.04 (0.95, 1.13)	1.01 (0.92, 1.12)	1.00 (0.93, 1.08)
Social exclusion	0.97 (0.89, 1.06)	1.12 (1.00, 1.25)*	1.00 (0.92, 1.10)	1.05 (0.96, 1.15)
Family rejection	0.97 (0.92, 1.02)	1.00 (0.94, 1.06)	0.99 (0.93, 1.05)	0.99 (0.93, 1.05)
Intrapersonal minority stressors
Identity concealment	0.91 (0.78, 1.05)	1.03 (0.89, 1.19)	1.08 (0.93, 1.27)	1.05 (0.91, 1.21)
Internalized homophobia	0.94 (0.74, 1.21)	0.97 (0.71, 1.31)	0.99 (0.79, 1.25)	1.04 (0.81, 1.33)
LGB-specific military stressors
Interpersonal LGB military stress	1.09 (0.86, 1.39)	0.99 (0.74, 1.32)	0.76 (0.58, 0.99)*	1.06 (0.81, 1.40)
Intrapersonal LGB military stress	1.00 (0.98, 1.02)	0.99 (0.97, 1.01)	0.99 (0.97, 1.00)	0.98 (0.97, 0.99)*

Note. Models adjusted for age, education, income, healthcare coverage, general stress, PTSD symptoms, and depression symptoms.

Abbreviations: LGB = lesbian, gay, and bisexual; LGBT = lesbian, gay, bisexual, and transgender; PTSD = posttraumatic stress disorder.

^*Significant at P < .05.

First, the association between victimization and past-year smoking was different for higher versus lower levels of coping (adjusted odds ratio [aOR] = 0.91 [95% CI: 0.84, 0.99], P = .05). Figure 1 shows this association at below-average, average, and above-average levels of coping. Overall, we observed positive slopes for those with below-average or average coping, and a negligible slope for those with above-average coping. In postestimation tests, the slope for those with below-average coping was significantly different from 0 (P < .001), indicating that as occurrences of victimization increase so does the probability of past-year smoking for these individuals.

Figure 1.

Adjusted predictions of past-year smoking with 95% confidence intervals: interaction between victimization and coping.

Second, the strength of the association between social exclusion and past-year smoking increased for higher versus lower levels of general social support (aOR = 1.12 [95% CI: 1.00, 1.25], P = .05). For those with below-average general social support, we observed a negative slope, and the opposite trend for those with average or above-average general social support (see Figure 2). However, the simple slopes were not statistically different from 0.

Figure 2.

Adjusted predictions of past-year smoking with 95% confidence intervals: interaction between social exclusion and general social support.

Third, the association between interpersonal LGB military stress and past-year smoking was significantly different for higher versus lower levels of Veteran-specific social support (aOR = 0.76 [95% CI: 0.58, 0.99], P = .04). Figure 3 shows the simple slopes for those who did not experience versus those who experienced interpersonal LGB military stress. For those who experienced interpersonal LGB military stress, we observed a negative slope and the opposite trend for those who did not experience stress. The simple slopes were not statistically different compared with 0.

Figure 3.

Adjusted predictions of past-year smoking with 95% confidence interval: interaction between interpersonal LGB military stress and Veteran-specific social support.

Abbreviation: LGB = lesbian, gay, and bisexual.

Lastly, the strength of the association between intrapersonal LGB military stress and past-year smoking decreased for higher versus lower levels of LGBT-specific social support (aOR = 0.98 [95% CI: 0.97, 0.99], P = .04). We observed a positive slope for those with below-average LGBT support, a negligible slope for those with average LGBT support, and a negative slope for those with above-average LGBT support (see Figure 4). In postestimation tests, the simple slopes were not significantly different from 0.

Figure 4.

Adjusted predictions of past-year smoking with 95% confidence intervals: interaction between interpersonal LGB military stress and LGBT-specific social support.

Abbreviations: LGB = lesbian, gay, and bisexual; LGBT = lesbian, gay, bisexual, and transgender.

Sensitivity Analyses

Most participants (73.7%, n = 341) completed their baseline survey during a period of global COVID-19 concern, defined as the period between the World Health Organization’s declaration of COVID-19 as a pandemic (March 11, 2020) and the end of study enrollment (December 31, 2020).⁵² In post-hoc sensitivity analyses, we did not find that baseline survey completion date was associated with any bivariate differences in past-year smoking status (P = .210). Additionally, the findings from the multivariable logistic regression models with interaction terms and covariate adjustment for baseline survey completion date show consistent results with our main interaction findings (results not shown).

DISCUSSION

This study was guided by the Minority Stress Model,^6–8 which suggests that LGB individuals are exposed to unique sexual orientation-related stressors that accrue over time, leading to stress and adverse health behaviors. The model also suggests that coping and social support may act as protective factors that mitigate the association between stress and smoking.

Based on the model, we hypothesized that the strength of the association between minority stressors and past-year smoking would be reduced among LGB Veterans with higher versus lower levels of coping and/or social support. We found that coping moderated the association between victimization and past-year smoking, such that the probability of past-year smoking decreased by 9% for a one-point increase in mean coping scores. Our findings suggest that LGB Veterans with below-average levels of coping are at increased risk for smoking if they experience more occurrences of victimization and may benefit from learning additional skills that can help them cope with stress.

We found that LGBT-specific social support moderated the association between intrapersonal LGB military stress and past-year smoking, but the effect size was small (i.e., a 2% decrease). We also found the Veteran-specific social support moderated the association between interpersonal LGB military stress and past-year smoking, suggesting that as Veteran-specific social support mean scores increase by one point, the probability of past-year smoking decreases by 24%. These findings suggest that receiving social support from others who possess similar identities can serve as an important protective factor.⁵³ In our sample, social support from LGBT peers served as a protective factor for minority stressors (i.e., anxiety about sexual orientation and identity concealment during military service), whereas social support from Veteran peers served as a protective factor for LGB military stressors (i.e., being investigated or punished by the military because of LGB identity). These findings also extend previous qualitative studies, which found that during military service LGB members created and maintained underground communities and were most “out” to other LGB unit friends.^54,55 These communities allowed them to develop connections without feeling threatened, reduce feelings of isolation, and helped them cope with minority and military stressors.^54,56 Our measures of social support were not mutually exclusive, and it is plausible that LGB Veterans received support from other LGB Veterans, but additional studies are needed to confirm this possibility.

Contrary to our hypothesis, we found that a one-point increase in mean general social support scores strengthened the association between social exclusion and past-year smoking, such that there was a 12% increase in the probability of past-year smoking. This finding may be explained by the social influence of peers, which can inform pro-tobacco norms.⁵³ Prior studies have found that having stronger social ties and a social network with more peers who smoke is associated with an increased likelihood of current smoking.^30,53,57 This finding may also be explained by lack of social safety. Diamond and Alley argue that social safety, defined as social connection, belonging, inclusion, recognition, and protection, is often missing among LGB individuals and contributes to health disparities.⁵⁸ While we might expect higher social support to be protective, it may exacerbate smoking if received from people whom LGB Veterans cannot fully be themselves around and cannot talk to about being LGB. Additional studies are needed to inform why general social support moderates the relationship between social exclusion and smoking.

There are several limitations to this study. First, a small sample size limited the ability to conduct stratified analyses by gender identity and/or sexual orientation identity, a consistent gap in the literature. We used a single sexual identity item to categorize LGB individuals in our sample, which has been found to define sexual orientation strongly.⁵⁹ However, we acknowledge that sexual orientation is a complex construct consisting of sexual identity, behavior, and attraction/desire,⁵⁹ and may have excluded individuals who do not identify as LGB but whose sexual behavior or attraction is to the same or multiple genders. Additionally, TGD Veterans were not included in our analytic sample because they received measures that assessed minoritized gender identity-related stressors. Future research is needed to understand how exposure to minoritized gender identity-related stressors is associated with smoking among TGD Veterans. Second, it is difficult to detect moderation in cross-sectional studies with continuous variables,⁶⁰ and these detections are made more difficult with small sample sizes and low statistical power. Our findings are best considered hypothesis-generating rather than conclusive given these limitations.

Despite these limitations, this study provides new insight into whether coping and social support moderate the associations between minority stressors and past-year cigarette smoking among LGB Veterans. While the Minority Stress Model is a compelling framework for understanding tobacco use disparities among LGB individuals, research has largely taken a deficit approach seeking to identify risk factors for smoking. Few studies have focused on resiliencies and protective factors that may mitigate the relationship between stress and smoking. Our findings provide some evidence for the Minority Stress Model; however, with respect to the outcome of cigarette smoking, coping and social support may buffer stress in some cases, but may also exacerbate stress in others.

CONCLUSIONS

Our findings show that coping, LGBT-specific social support, and Veteran-specific social support buffer the relationship between stressors and past-year smoking, yet general social support may exacerbate the risk for past-year smoking among those with higher levels of social exclusion related to sexual orientation. At an individual level, teaching LGB Veterans positive coping skills can build resiliency allowing them to withstand the adverse effects of stress, which may ultimately reduce cigarette use. At a social level, social support linked to LGB and Veteran identities could be leveraged to support smoking cessation. Further study of the interaction between general social support and social exclusion is needed to understand the implications of these findings for tobacco cessation among LGB Veterans.

Supplementary Material

1

Supplementary Figure 1. Conceptual model.

Abbreviations: LGB = lesbian, gay, and bisexual; LGBT = lesbian, gay, bisexual, and transgender.

Statement of Public Health Significance:

Coping and social support are potential protective factors that may mitigate the relationship between sexual orientation-related minority stress and cigarette smoking. Findings among LGB Veterans suggest that with respect to the outcome of cigarette smoking, coping and social support may buffer stress in some cases, but may also exacerbate stress in others.