Cigarette Smoking Motives and Stages of Change in Smoking Cessation Among Veterans: Differences by Gender and Sexual Orientation

Abstract

Purpose:

To examine differences across gender and sexual orientation in cigarette smoking motives and stages of change in smoking cessation among Veterans.

Design:

Secondary data analysis of cross-sectional baseline surveys from a prospective cohort study.

Setting:

United States, self-administered online survey.

Participants:

Cisgender Veterans who reported past-year smoking (N = 146); 66.4% identified as lesbian, gay, or bisexual and 52.1% were men.

Measures:

Smoking motives (i.e., social, self-confidence, boredom relief, and affect regulation), with higher scores indicating stronger motivation to smoke. Stages of change categories included precontemplation, contemplation/preparation, and action/maintenance.

Analysis:

Analyses were stratified by gender and sexual orientation. Age-adjusted linear regression models estimated differences in smoking motives scores and multinomial logistic regression models estimated differences in stages of change categories relative to the precontemplation stage (reference category).

Results:

In this Veteran sample, gay men reported higher social smoking motives vs heterosexual men (β = 1.50 (95% CI: .04, 2.97), P-value = .045) and higher boredom relief smoking motives vs bisexual men (β = 1.53 (95% CI: .06, 2.29), P-value = .041) in age-adjusted models. Lesbian women were more likely to be in the action/maintenance stage relative to the precontemplation stage when compared to both heterosexual women (aRRR = 4.88 (95% CI: 1.00, 23.79) P-value = .050) and bisexual women (aRRR = 16.46 (95% CI: 2.12, 127.57), P-value = .007) after adjusting for age.

Conclusion:

Smoking cessation interventions may benefit from enhancing peer support for gay men. Given bisexual and heterosexual women were in less advance stages of change, there may be a greater need for motivational interventions to encourage quitting and additional support to assist with cessation efforts. Overall, findings highlight the diversity of cigarette use within LGB communities.

Purpose

United States (US) military Veterans have a higher prevalence of cigarette smoking compared to the general population, with recent estimates among Veterans at 21.6% compared to 12.5% in US adults.^1-5 Nearly 1 million Veterans identify as lesbian, gay or bisexual (LGB).⁶ In the general population, LGB individuals are more likely to report current smoking compared to heterosexual individuals.^4,7,8 The intersection of Veteran status and a minoritized sexual orientation identity suggest LGB Veterans may be at increased risk for smoking^9-11 and represent a population in need of tobacco prevention and control efforts.

Evidence-based interventions for smoking cessation reduce cigarette use, and several tailored interventions have been developed for Veterans¹² and LGB individuals.^13,14 For example, SmokefreeVET is a text messaging cessation program that provides Veterans with daily encouragement and advice, and strategies for coping with Veteran identified smoking triggers.^15,16 Vet Flexiquit is a web-based and text messaging program that uses an avatar who represents a supportive peer that shares their personal quitting journey.¹⁷ A similar program called Empowered, Queer, Quitting, and Living (EQQUAL) was tailored for sexual and gender minority young adults.¹⁸ Lastly, the Bitch to Quit program tailored materials for LGB individuals by using inclusive language and pictures of same sex couples, and addressed risk factors for smoking, evidence on the impact of smoking, and barriers to smoking cessation among LGBT communities.¹⁹ However, these tailored interventions have had mixed success in treatment uptake, treatment completion and abstinence rates at follow up. This suggests the need for improved knowledge about factors contributing to smoking in these populations. In addition, there is a lack of research on the intersection of Veteran status and LGB identity and its role in cigarette smoking and smoking cessation. Understanding what motivates LGB Veterans to continue smoking cigarettes and understanding where LGB Veterans are in their plans with quitting can inform the development of future interventions.

Smoking Motives

Smoking motives draw on the Theory of Planned Behavior^20,21 and Motivational Models of Alcohol Use^22,23 to describe how smoking is influenced by internal motives (e.g., an individual’s attitudes about smoking) and external motives (e.g., social pressure and norms about smoking), which reinforce the behavior over time. For example, individuals may smoke to obtain positive outcomes at the individual-level (e.g., improve self-confidence) or social-level (e.g., social group enhancement).²³ Smoking may also be negatively reinforced by helping individuals avoid unwanted outcomes at the individual-level (e.g., regulate negative affect) or social-level (e.g., protect against social rejection).²³

Little is known about smoking motives among LGB Veterans. Studies on smoking motives among Veterans have focused on motives that originate during military service, such as smoking to combat stress and boredom, build camaraderie, and because of peer pressure.^1,24,25 However, research suggests these motives influence smoking behavior even after separation from the military as Veterans build a dependency to smoke in these situations (e.g., when stressed or bored).¹ Findings among LGB individuals are limited, but one study, which only made comparisons between bisexual and heterosexual individuals, found that bisexual individuals reported higher affect regulation motives and higher social motives.²⁶ For LGB Veterans, the confluence of individual-level motives and social-level motives from both the military community^2,27 and the LGB community,^28,29 may act as motivators enabling the continued use of cigarettes.

Stages of Change (SOC) in Smoking Cessation

The Transtheoretical Model of Health Behavior (also referred to as the Stages of Change model)³⁰ posits that smoking-related behavior change progresses through 6 stages including: precontemplation (e.g., no plans to quit ever or within 6 months), contemplation (e.g., plans to quit within 6 months, but not within 1 month), preparation (e.g., plans to quit within 1 month), action (e.g., quit in the last 6 months), maintenance (e.g., quit more than 6 months ago), and termination (e.g., no plans to smoke again). The progression and time between stages varies by individual and because setbacks are common, individuals may move between stages non-linearly.³¹

Documenting patterns of plans to quit smoking can clarify whether LGB Veterans have lower, greater, or equal plans to quit compared to their heterosexual Veteran peers. With this understanding, future interventions can better support LGB Veterans in their cessation process. Further, it would allow for alignment with US Clinical Practice Guidelines which recommend smoking cessation interventions be tailored based on an individual’s plans to quit. For example, if an individual is in a more advanced stage (i.e., with plans to quit within 1 month), they should be offered cessation treatments and assistance with a quit attempt.³² For all others, brief motivational interventions based on Motivational Interviewing (MI) principles are suggested to encourage progression to quitting.³²

This exploratory study examined differences in smoking motives and SOC among Veterans with past-year smoking, which to our knowledge has not been previously studied. Based on prior research among Veterans indicating gender differences between men and women on smoking status outcomes,^27,33 we also conducted stratified analyses by gender and then compared LGB Veterans to heterosexual Veterans.

Methods

Design

This study used data from the Health for Every Veteran Study, a prospective cohort study to understand mental health problems and health risk behaviors among heterosexual and lesbian, gay, bisexual, and transgender (LGBT) Veterans. We conducted secondary data analysis of cross-sectional baseline survey data collected from September 2019 to December 2020. Participants were recruited through online social networking sites and targeted online advertisements to Veterans’ groups, LGBT Veteran groups, and the larger LGBT and Veteran communities. To be eligible, participants had to be at least 18 years old, self-report having served in the US Armed Forces, be currently living in the US, have routine access to the Internet, have a valid e-mail address, postal address, and phone number, and be willing to answer questions about key demographics (e.g., age, gender, race, ethnicity, and sexual orientation). Participants completed four web-based surveys every nine months for 27 months. Participants who completed the baseline survey received $30, which took approximately 60 to 90 minutes to complete.

Sample

Our analytic sample included cisgender Veterans with past-year smoking (N = 146). In separate items, participants were asked, “What sex was listed on your original birth certificate? (Male or Female)” and “What is your current gender identity?” Response options included: man, woman, transgender man/female to male (FTM), transgender woman/male to female (MTF), genderqueer, gender fluid, non-binary, or another gender identity. Participants who indicated a current gender identity that corresponded with their sex assigned at birth were categorized as cisgender. We excluded transgender and gender diverse (TGD) Veterans (n = 189 and n = 66, respectively) because of the exploratory nature of the study and small sample sizes among Veterans who smoke.

We assessed past-year cigarette smoking status using three separate items. Participants were asked, “Have you ever smoked a cigarette, even one or two puffs?” Response options were (1) Yes or (0) No. Participants who indicated “Yes,” were then asked, “How many cigarettes have you smoked in your entire life? A pack usually has 20 cigarettes in it.” and “In the past year, have you smoked a cigarette, even one or two puffs?” Participants who reported having smoked ≥100 cigarettes in their life and having smoked a cigarette in the past year were categorized as Veterans with past-year smoking. We excluded Veterans who reported never smoking (n = 174), smoking <100 cigarettes in their life (n = 222), or not smoking in the past year (n = 265).

Measures

Dependent Variables

Smoking Motives.

The 15-item Smoking Motives Scale^34-36 asked participants to consider the past year and rate their agreement with statements others have said about smoking cigarettes using a 5-point Likert scale ranging from (1) not at all true to (5) very true. The scale measures four dimensions of smoking motives, including 4 items on social motives (e.g., “Smoking helps you fit in with other people;” α = .86), 4 items on self-confidence motives (e.g., “Smoking makes you feel more sure of yourself;” α = .77), 2 items on boredom relief motives (e.g., “Smoking is something to do when you’re bored;” α = .93) and 5 items on affect regulation motives (e.g., “Smoking helps you forget about your worries;” α = .89). Total mean scores were calculated by summing and averaging responses within each dimension, higher scores indicated stronger motivation to smoke within that dimension.

Stage of Change.

Stages of change was assessed using a single item which asked, “What best describes your intentions regarding quitting smoking?” Response options included (1) never expect to quit, (2) may quit in the future but not in the next 6 months, (3) will quit in the next 6 months, (4) will quit in the next month, or (5) I have already quit and no longer smoke. Responses were recoded to (1) precontemplation (response options 1 and 2), to indicate no plans to quit ever or within 6 months, (2) contemplation/preparation (response options 3 and 4), to indicate plans to quit within 1-6 months, and (3) action/maintenance (response option 5), to indicate having already quit. Due to small cell sizes, we combined the contemplation and preparation stages. Additionally, we did not assess when participants quit smoking (i.e., within the last 6 months or more than 6 months ago) and therefore combined the action and maintenance stages.

Independent Variable

Sexual Orientation Identity.

Sexual orientation identity was assessed using a single item that asked, “If you had to choose one, which of these categories most closely represents you?” Response options included: lesbian or gay, straight or heterosexual, or bisexual. Sexual orientation was crossed with gender (i.e., gay man, bisexual man, heterosexual man, lesbian woman, bisexual woman, and heterosexual woman) for analyses of group differences in smoking motives and SOC.

Covariates

Sociodemographic.

Sociodemographic characteristics included age in years, race and ethnicity (non-Hispanic White, non-Hispanic other racial identity, or Hispanic), relationship status (in a relationship–i.e., married, domestic partnership, or dating; or single), educational attainment (some college/high school or less, or college graduate or more), annual household income ($39,999 or less, $40,000 to $79,999, or $80,000 or more), and any healthcare coverage (yes or no).

Severity of Nicotine Dependence.

Severity of nicotine dependence was assessed using the Heaviness of Smoking Index (HSI),^37,38 which is a combined measure of average cigarettes per day (CPD) and time to first cigarette (TTFC).^39,40 Participants were asked, “On the days that you smoke, how many cigarettes do you smoke on average?” CPD responses were recoded as (0) 1-10, (1) 11-20, (2) 21-30, or (3) ≥31. Participants were also asked, “How soon after you wake up do you smoke your first cigarette?” TTFC responses options included (3) within 5 minutes, (2) 6-30 minutes, (1) 31-60 minutes or (0) after 60 minutes after waking. We calculated the HSI by summing CPD and TTFC (range 0-6), with higher scores indicating higher nicotine dependence.

Analysis

We calculated descriptive statistics for participant characteristics, smoking motives, and SOC. Bivariate analyses compared mean smoking motives scores using Welch’s t-test and SOC using Fisher’s exact test. Age-adjusted linear regression models with 95% Confidence Interval (CI) estimated mean differences in smoking motives. Age-adjusted multinomial logistic regression models with 95% CI estimated differences in SOC categories relative to the precontemplation stage (reference category). To compare groups, we conducted stratified analyses across gender and sexual orientation. First, we compared gay, bisexual, and heterosexual men to each other. Separately, we compared lesbian, bisexual, and heterosexual women to each other. Given the exploratory nature of this study and small sample sizes, models were only adjusted for age to not over-adjust variations between groups, which can bias estimates towards the null.⁴¹ In bivariate analysis, of all participant characteristics, the one variable with the most consistent differences in pairwise comparisons across groups was age. All analyses were conducted using Stata/MP Version 17.0 (StataCorp, College Station, TX). This research was approved by the Human Subjects Division of the University of Washington.

Results

Participant Characteristics

Table 1 presents participant characteristics by gender and sexual orientation (N = 146). In the total sample, 76 (52.1%) were men; 33 (43.4%) self-identified as gay, 17 (22.4%) self-identified as bisexual, and 26 (34.2%) self-identified as heterosexual. Of the 70 women (47.9% of total sample), 31 (42.3%) self-identified as lesbian, 16 (22.9%) self-identified as bisexual and 23 (32.8%) self-identified as heterosexual.

Table 1.

Participant Characteristics by Gender and Sexual Orientation, N = 146.

	Men, n = 76			Women, n = 70
	Gay, n = 33	Bisexual, n = 17	Heterosexual, n = 26	Lesbian, n = 31	Bisexual, n = 16	Heterosexual, n = 23
	Mean (SD)
Age (years)a,b,e,f	45.15 (11.95)	43.35 (13.62)	56.88 (13.69)	44.65 (10.58)	36.88 (10.67)	49.26 (11.62)
HSI (Range 0-6)	2.03 (1.59)	2.53 (2.13)	1.90 (1.41)	2.37 (1.46)	1.57 (1.55)	1.84 (1.57)
	Frequency (%)
Race and ethnicityc
Non-hispanic white	29 (87.9)	14 (82.4)	19 (73.1)	24 (77.4)	10 (62.5)	18 (78.3)
1Non-hispanic other identity	3 (9.1)	0 (.0)	3 (11.5)	3 (9.7)	4 (25.0)	3 (13.0)
Hispanic	0 (.0)	3 (17.6)	3 (11.5)	4 (12.9)	2 (12.5)	2 (8.7)
Relationship status
2In a relationship	20 (60.6)	14 (82.4)	20 (76.9)	21 (67.7)	9 (56.3)	12 (52.2)
Single	12 (36.4)	3 (17.6)	6 (23.1)	10 (32.3)	7 (43.8)	11 (47.8)
Educational attainment
Some college/High school or less	18 (54.5)	10 (58.8)	14 (53.8)	18 (58.1)	10 (62.5)	11 (47.8)
College graduate or more	15 (45.5)	7 (41.2)	12 (46.2)	13 (41.9)	6 (37.5)	12 (52.2)
Income (Annual household)b
$39,999 or less	14 (42.4)	7 (41.2)	6 (23.1)	13 (41.9)	5 (31.3)	6 (26.1)
$40,000 to $79,999	9 (27.3)	8 (47.1)	7 (26.9)	13 (41.9)	9 (56.3)	10 (43.5)
$80,000 or more	10 (30.3)	2 (11.8)	13 (50.0)	5 (16.1)	2 (12.5)	6 (26.1)
Healthcare coveragee
Yes	29 (87.9)	14 (82.4)	25 (96.2)	24 (77.4)	9 (56.3)	20 (87.0)
No	4 (12.1)	3 (17.6)	1 (3.8)	6 (19.4)	7 (43.8)	3 (13.0)

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

HSI: Heaviness of Smoking Index; SD: Standard Deviation.

¹Includes non-Hispanic Black, multiracial, and “other” racial identities. Race and ethnic groups were combined due to small cell sizes.

²Includes being married, in a domestic partnership or dating.

^a-fSuperscripts indicate statistically significant differences at P-value <.05 between the following groups:

^agay men and heterosexual men.

^bbisexual men and heterosexual men.

^cgay men and bisexual men.

^dlesbian women and heterosexual women.

^ebisexual women and heterosexual women.

^flesbian women and bisexual women.

Compared to heterosexual men, gay men were younger (P-value = .001). Bisexual men were also younger (P-value = .003) and had lower annual household income (P-value = .037) compared to heterosexual men. Bisexual men were also more likely to be Hispanic compared to gay men (P-value = .048). Bisexual women were younger than both lesbian (P-value = .024) and heterosexual women (P-value = .002) and reported less healthcare coverage than heterosexual women (P-value = .031) (see Table 1).

Differences in Smoking Motives among Men

Table 2 presents descriptive statistics (means and standard deviations, [SD]) for smoking motives among men by sexual orientation. In bivariate analyses, gay men reported higher social smoking motives (M = 6.76 [SD] = 3.49 vs M = 4.77 [SD] = 1.31; P-value = .004) and higher affect regulation smoking motives (M = 13.88 [SD] = 5.92 vs M = 11.15 [SD] = 4.40; P-value = .047) compared to heterosexual men. Mean affect regulation motives were also higher for bisexual men compared to heterosexual men (M = 14.59 [SD] = 5.33 vs M = 11.15 [SD] = 4.40; P-value = .035).

Table 2.

Descriptive Statistics of Smoking Motives and Stages of Change by Sexual Orientation Among Men, N = 76.

	Gay, n = 33	Bisexual, n = 17	Heterosexual, n = 26	Gay vs Heterosexual	Bisexual vs Heterosexual	Gay vs Bisexual
	Mean (SD)			P-value
Smoking motives (Range)
Social (4-20)	6.76 (3.49)	5.76 (2.39)	4.77 (1.31)	.004 **	.129	.243
Self-confidence (4-20)	6.24 (3.33)	6.71 (1.96)	5.65 (2.31)	.427	.118	.539
Boredom relief (2-10)	6.67 (2.78)	5.29 (2.44)	5.38 (2.53)	.069	.908	.081
Affect regulation (5-25)	13.88 (5.92)	14.59 (5.33)	11.15 (4.40)	.047 *	.035 *	.670
	Frequency (%)			P-value
Stages of change				.878	.240	.274
Precontemplation	15 (45.5)	8 (47.1)	12 (46.2)	--	--	--
Contemplation/Preparation	11 (33.3)	8 (47.1)	8 (30.8)	--	--	--
Action/Maintenance	7 (21.2)	1 (5.9)	6 (23.1)	--	--	--

^*Welch’s t-test comparing mean smoking motives scores found to be significant at P-value <.05.

^**Welch’s t-test comparing mean smoking motives scores found to be significant at P-value <.01.

Results of age-adjusted linear regression models estimating the mean difference in smoking motives among men are presented in Table 3. Differences in social smoking motives for gay men compared to heterosexual men remained significant after adjusting for age (β = 1.50 (95% CI: .04, 2.97), P-value = .045). Of note, in bivariate analyses, we found no differences between gay men and bisexual men, but there was a difference after adjusting for age. The estimated age-adjusted mean difference in boredom relief smoking motives was 1.53 points higher among gay men than bisexual men (β = 1.53 (95% CI: .06, 2.29), P-value = .041).

Table 3.

Age-Adjusted Linear Regression Models Estimating Mean Differences in Smoking Motives and Multinomial Logistic Regression Models Estimating Relative Risk Ratios of Stages of Change by Sexual Orientation Among Men, N = 76.

	Gay vs Heterosexual		Bisexual vs Heterosexual		Gay vs Bisexual
	β (95%CI)	P-value	β (95%CI)	P-value	β (95%CI)	P-value
Smoking motives
Social	1.50 (.04, 2.97)	.045 *	.45 (−.76, 1.65)	.463	1.12 (−.45, 2.69)	.157
Self-confidence	.13 (−1.44, 1.69)	.874	.56 (−.82, 1.93)	.424	−.39 (−1.86, 1.08)	.601
Boredom relief	.90 (−.66, 2.46)	.253	−.55 (−2.13, 1.04)	.495	1.53 (.06, 2.29)	.041 *
Affect regulation	1.57 (−1.55, 4.70)	.319	2.27 (−.96, 5.51)	.165	−.51 (−3.48, 2.46)	.730
	aRRR (95%CI)	P-value	aRRR (95%CI)	P-value	aRRR (95%CI)	P-value
Stages of change
Precontemplation	Ref		Ref		Ref
Contemplation/Preparation	.80 (.23, 2.83)	.733	1.18 (.28, 4.92)	.823	.68 (.19, 2.43)	.556
Action/Maintenance	.62 (.12, 3.20)	.569	.16 (.16, 1.53)	.110	3.99 (.45, 35.53)	.215

CI = Confidence Interval; aRRR = Age-adjusted Relative Risk Ratio (RRR).

^*Mean smoking motives scores found to be significant at P-value <.05.

^**Mean smoking motives scores found to be significant at P-value <.01.

Differences in Smoking Motives among Women

Table 4 presents descriptive statistics (means and SD) for smoking motives among women by sexual orientation. In bivariate analyses and age-adjusted linear regression models (see Table 5), there were no significant differences in smoking motives among women.

Table 4.

Descriptive Statistics of Smoking Motives and Stages of Change by Sexual Orientation Among Women, N = 70.

	Lesbian, n = 31	Bisexual, n = 16	Heterosexual, n = 23	Lesbian vs Heterosexual	Bisexual vs Heterosexual	Lesbian vs Bisexual
	Mean (SD)			P-value
Smoking motives (Range)
Social (4-20)	5.58 (2.85)	7.00 (4.08)	5.43 (2.92)	.855	.199	.227
Self-confidence (4-20)	5.81 (2.68)	6.81 (3.31)	6.17 (2.52)	.609	.520	.304
Boredom relief (2-10)	5.19 (2.60)	5.69 (2.65)	5.43 (2.41)	.727	.763	.547
Affect regulation (5-25)	12.94 (5.60)	14.88 (5.12)	13.91 (5.87)	.540	.591	.242
	Frequency (%)			P-value
Stages of change				.029 ♦	.069	.017 ♦
Precontemplation	6 (19.4)	9 (56.3)	12 (52.2)	--	--	--
Contemplation/Preparation	13 (41.9)	5 (31.3)	7 (30.4)	--	--	--
Action/Maintenance	12 (38.7)	2 (12.5)	4 (17.4)	--	--	--

^♦χ² test of independence found to be significant at P-value <.05.

Table 5.

Age-Adjusted Linear Regression Models Estimating Mean Differences in Smoking Motives and Multinomial Logistic Regression Models Estimating Relative Risk Ratios of Stages of Change by Sexual Orientation Among Women, N = 70.

	Lesbian vs Heterosexual		Bisexual vs Heterosexual		Lesbian vs Bisexual
	β (95%CI)	P-value	β (95%CI)	P-value	β (95%CI)	P-value
Smoking motives
Social	−.19 (−1.69, 1.32)	.802	.66 (−1.68, 3.00)	.574	−.85 (−3.12, 1.41)	.454
Self-confidence	−.55 (−1.91, .81)	.423	.15 (−1.63, 1.93)	.865	−.70 (−2.57, 1.12)	.457
Boredom relief	−.52 (−1.87, .83)	.448	−.49 (−2.21, 1.24)	.574	−.03 (−1.65, 1.58)	.972
Affect regulation	−1.70 (−4.79, 1.39)	.275	−.99 (−4.72, 2.75)	.600	−.72 (−3.98, 2.55)	.662
	aRRR (95%CI)	P-value	aRRR (95%CI)	P-value	aRRR (95%CI)	P-value
Stages of change
Precontemplation	Ref		Ref		Ref
Contemplation/Preparation	4.16 (.98, 17.72)	.054	1.37 (.22, 8.58)	.739	3.04 (.64, 14.58)	.164
Action/Maintenance	4.88 (1.00, 23.79)	.050 *	.30 (.44, 2.02)	.215	16.46 (2.12, 127.57)	.007 **

CI = Confidence interval; aRRR = Age-adjusted Relative Risk Ratio (RRR).

^*Found to be significant at P-value <.05, adjusting for age in years.

^**Found to be significant at P-value <.01, adjusting for age in years.

Differences in Stages of Change among Men

Table 2 also presents frequencies and proportions (%) for SOC among men by sexual orientation. In bivariate analyses and age-adjusted multinomial logistic regression models (see Table 3), there were no significant differences in SOC among men.

Differences in Stages of Change among Women

Table 4 presents frequencies and proportions (%) for SOC among women by sexual orientation. Most bisexual women (56.3%) and heterosexual women (52.2%) were in the precontemplation stage, indicating no plans to quit ever or within the next 6 months. Among lesbian women, 41.9% were in the contemplation/preparation stage (i.e., plans to quit within 1-6 months) and 38.7% were in the action/ maintenance stage (i.e., having already quit). In bivariate analyses, the observed distribution of SOC categories was different between lesbian women and heterosexual women (P-value = .029) and between lesbian women and bisexual women (P-value = .017).

Table 5 presents the results of age-adjusted multinomial logistic regression models estimating the relative risk ratios (RRR) of SOC categories compared to the precontemplation stage (reference category). After adjusting for age, lesbian women were more likely to be in the action/maintenance stage relative to the precontemplation stage compared to heterosexual women (aRRR = 4.88 (95% CI: 1.00, 23.79), P-value = .050) and bisexual women (aRRR = 16.46 (95% CI: 2.12, 127.57), P-value = .007).

Discussion

The purpose of this cross-sectional study was to provide preliminary data on smoking motives and SOC among Veterans across gender and sexual orientation. This study was guided by the Theory of Planned Behavior^20,21 and Motivational Models of Alcohol Use,^22,23 which describe cigarette smoking motives across four dimensions including social, self-confidence, boredom relief, and affect regulation motives. The Transtheoretical Model of Health Behavior³⁰ was used to describe SOC. Our study addresses gaps in the literature by focusing on LGB Veterans, an understudied population with tobacco-use disparities.

Overall, we observed several differences between LGB and heterosexual Veterans by gender. Gay men reported higher social motives compared to heterosexual men and higher boredom relief motives compared to bisexual men in age-adjusted models. Our findings on social motives align with a qualitative study of LGBT individuals, which found desire for social acceptance, omnipresence of smoking among peers and bar culture influenced smoking behavior and were barriers to smoking cessation.⁴² Bars and night clubs have been considered common social settings for gay men which provide safe spaces to meet and build community. However, they also foster an environment where smoking is an accepted social norm used to build affinity and reduce stress.⁴³

Our findings for gay men also highlight the significant influence boredom relief motives have on their smoking behavior. The design of future smoking cessation interventions should consider developing tailored messaging around these motives as smoking triggers, which can help gay men identify and address boredom as a type of affective trigger than can be addressed with specific coping strategies. This finding also highlights that reasons for smoking are diverse and vary between LGB subgroups. Additional studies are needed to corroborate our findings and provide a comprehensive understanding of the contextual factors that contribute to boredom relief as a more salient smoking motive among gay men compared to bisexual men.

In our sample, we did not find differences in smoking motives among women Veterans by sexual orientation. A qualitative study of reasons for smoking between lesbian and heterosexual women found that the two groups reported several common reasons for smoking, including for affect regulation, to manage stress and to enhance social relationships. However, lesbian women reported additional, unique stressors caused by sexual orientation-related stigma (e.g., homophobic comments from family members, internalized homophobia, harassment, and discrimination).⁴⁴ These stressors, commonly referred to as minority stress,^45-47 provide a compelling framework for understanding the high prevalence of cigarette smoking among LGB individuals, albeit not the focus of this study. More broadly, studies of adults in the general population have found that women were more likely to report stress relief motives than men.^36,48 Future research should consider examining stress and other motives from the literature (e.g., for enjoyment, for weight management, to aid concentration or to combat pain) among women Veterans as a function of sexual orientation.

In terms of SOC, bisexual and heterosexual women were more likely to be in the earlier stages of quitting compared to lesbians, who reported being in more advanced stages or had already quit. Prior studies have found that lesbian Veterans are more likely to report current smoking compared to their heterosexual counterparts.^9,49 The present study did not compare rates of smoking, but our findings among Veterans with past-year smoking suggest lesbian women were more likely to have quit smoking. This finding may highlight an important positive outcome among lesbian women, and future research should explore what facilitators assisted them in quitting. Our findings also suggest that bisexual and heterosexual women who are in less advanced stages of quitting may benefit from additional support in making a quit attempt. Current guidelines recommend MI interventions for those in less advanced stages,³² but findings are mixed on the efficacy of these interventions.⁵⁰

We found no differences in SOC among Veteran men. Although plans to quit were similar among subgroups of Veteran men in our sample, prior research suggests that gay men have the highest lifetime rates of tobacco cessation treatment-seeking compared to heterosexual men⁸ and are also often overrepresented in smoking cessation studies for LGBT individuals.¹⁹ Despite an increased awareness and utilization of smoking cessation treatments, gay men continued to have a higher prevalence of smoking compared to heterosexual men. This might suggest that gay men have more difficulty making a successful quit attempt and/or experience more setbacks in their quitting process.

This study has several limitations. First, due to the exploratory nature of the study and the sample size, we conducted minimally adjusted models which are subject to residual confounding. Still, our sample sizes for bisexual men and bisexual women were small (i.e., 17 and 16, respectively) and we may have over-adjusted group differences between bisexual groups and other groups. Future studies with larger samples could provide further insight into associations between gender and sexual orientation with smoking motives and SOC. Second, eligibility criteria for the study included having routine access to the Internet, having a valid email address, postal address, and phone number, which may have excluded Veterans who might be most vulnerable and likely to smoke. Despite these limitations, this study added important preliminary information about smoking motives and SOC among Veterans, a current gap in the literature.

There are important additional directions for future research. Studies should consider additional sociodemographic and smoking characteristics and their relation to smoking motives and SOC among Veterans. Cigarette use is higher among Veterans of younger age, lower income and education, minoritized race and ethnicity, and those with mental health conditions.^1,2,27 Our analytic sample was predominantly non-Hispanic White and their cigarette use and smoking behavior may be different than that of Veterans with minoritized racial and ethnic identities. More research is also needed among TGD individuals. Prior research has found that tobacco use disparities among TGD adults are stronger for e-cigarette use and cigar use when compared to cisgender adults.⁵¹ Specific tobacco product use is an important consideration for future research among TGD individuals. Similarly, an estimated 7.0% of Veterans use multiple types of tobacco products, and research suggests that dual use of combustible cigarettes and e-cigarettes is high among certain Veterans.^2,52 Future research is needed to comprehensively understand sociodemographic characteristics, tobacco product use, and multi-tobacco use among Veterans, which has implications for smoking motives, plans to quit, and cessation intervention development. Lastly, to fully understand the range of reasons for smoking among LGB Veterans, a mixed-methods study may be valuable.

Conclusion

Understanding smoking motives and plans to quit smoking among LGB Veterans can inform the development of tailored smoking cessation interventions to increase successful quitting. Our results show that among men Veterans, gay men reported higher social and boredom relief motives. Smoking cessation interventions may need to consider how to leverage peer support for gay men to better assist with quitting. Findings among women Veterans indicate that lesbian women were more likely to be in more advanced stages of quitting or had already quit in the past year, highlighting a positive outcome. At the same time, findings suggest that bisexual and heterosexual women may benefit from MI interventions to increase readiness to quit and additional support to aid with quit attempts.

So What?

What is already known on this topic?

Evidence-based interventions for smoking cessation reduce cigarette use. Understanding smoking motives for the continued use of cigarettes and stages of change in smoking cessation can inform the development of future interventions for LGB Veterans, an understudied population at increased risk for smoking.

What does this article add?

Gay men report higher social and boredom relief smoking motives compared to heterosexual men and bisexual men, respectively. Among women Veterans, bisexual women and heterosexual women were more likely to report no plans to quit ever or within six months compared to lesbian women who were in more advanced stages of quitting or had already quit.

What are implications for health promotion practice and research?

Findings highlight the diversity of smoking motives and change readiness within LGB communities. Future research is needed to understand whether smoking cessation interventions tailored to address these smoking motives are more effective in treatment completion and abstinence rates for LGB Veterans. More research identifying barriers and facilitators for smoking cessation is also needed, particularly for those in less advanced stages of quitting.

Appendix

Abbreviations

CI

Confidence interval

CPD

Cigarettes per day

LGB

Lesbian, gay, and bisexual

LGBT

Lesbian, gay, bisexual, and transgender

MI

Motivational interviewing

RRR

Relative risk ratio

SD

Standard deviation

SOC

Stages of change

TGD

Transgender and gender diverse

TTFC

Time to first cigarette