There is a growing body of evidence that lesbian, gay and bisexual (LGB) people have higher rates of cigarette smoking in comparison to heterosexuals.1^–4 This excess risk for smoking puts the same communities at greater risk for the myriad of health, social and economic consequences related to smoking. In the state of Washington, USA, with a well-funded comprehensive tobacco prevention and control program since 2000, LGB populations remain a priority population in tobacco control, with a combined smoking prevalence of 32% compared to 18% in the general population.5

Suggested explanations for elevated smoking in LGB communities have included high levels of stress associated with homophobia and discrimination; social norms that accept smoking in LGB venues, particularly bars and clubs; other risk factors associated with smoking such as alcohol and drug use; and targeted marketing/promotion of smoking to the LGB community by the tobacco industry.1 It is likely that the disparate rates of smoking are the result of a combination of these influences.

The tobacco industry has demonstrated an ability to design marketing and promotion campaigns that effectively encourage smoking behaviours, despite widespread knowledge about the harm that smoking causes. For the purpose of planning public health activities to reduce smoking-related health disparities, limiting pro-tobacco marketing targeted to specific communities represents an obvious point of intervention. For example, research among adolescents has shown that there is a causal linkage between exposure to tobacco industry marketing activities, receptivity and susceptibility to tobacco use and uptake of tobacco use.6^–10 To limit the influence of marketing on youth, the Master Settlement Agreement (MSA) between US state Attorneys General and the tobacco industry banned advertising that especially appealed to young people.11 Aside from regulatory approaches, community-based campaigns can also stop the influence of targeted tobacco marketing. For example, the African–American community was able to generate outrage sufficient to stop the launch of RJ Reynolds’ “Uptown” cigarettes, a brand designed for African–American smokers.12

The tobacco industry has specifically designed marketing campaigns to reach the LGB community. The industry began advertising in the LGB press during the early 1990s.13 14 To win the loyalty of the LGB community, Philip Morris pledged significant financial support to AIDS research,15 and numerous tobacco companies have sponsored local LGB events.13

A recent analysis of tobacco-related content in LGB periodicals revealed that although the number of ads for tobacco cessation was greater than the number of pro-tobacco advertisements, the pro-tobacco ads occupied much more space than the cessation ads and were far more appealing. Many non-tobacco ads and non-advertising content also showed tobacco in a positive way so that the prevalent message in LGB periodicals was one that reinforced pro-tobacco norms.16 17 Unfortunately, these studies did no comparison to the mainstream press, to compare the frequency of exposure to pro-tobacco marketing for respective readerships.

The industry also has a history of successful advertising campaign placement in bars and clubs.18 19 LGB-oriented bars have historically been a non-discriminatory social environment, and LGB individuals have been shown to have a high prevalence of frequent bar attendance.20^–22 If targeted advertising to LGB communities occurred in gay or lesbian bars, this might contribute to higher rates of marketing exposure.

Recently, Smith et al studied LGB attitudes about targeted pro-tobacco marketing in their communities and described community attitudes that might make LGB populations more vulnerable to pro-tobacco marketing.23 Targeted advertising of tobacco products was considered by some to be positive, an acknowledgement of the community. Others noted that LGB populations might be particularly receptive to advertising designed specifically to reach them, because few campaigns have been designed to do so.

The purpose of our study was to compare receptivity and exposure to tobacco industry marketing between LGB and heterosexual people in the state of Washington, USA. Evidence of excess receptivity or exposure to tobacco industry marketing in LGB communities could demonstrate the need for regulatory or community-based actions to counteract marketing in order to improve community health.

METHODS

We used Washington State’s Behavioral Risk Factor Surveillance System (BRFSS) data from 2003–2006 combined to gather information about receptivity and exposure to tobacco industry marketing. The BRFSS is a population-based, random-digit-dialled annual cross-sectional survey. The original sample is taken from telephone numbers stratified on telephone bank (listed and unlisted numbers) and by county. Eligible respondents were English or Spanish-speaking, non-institutionalised and lived in a household with a telephone landline. Once an eligible household was reached, an adult was randomly selected to participate from among adults who lived in the household. Sampling weights were calculated based on each respondent’s probability of selection, and data were also post stratified to the age and sex distribution of the Washington State adult population.

Measures

Sexual orientation:

Respondents were asked: “Now I’m going to ask you a question about sexual orientation. Do you consider yourself to be (A) heterosexual or straight (B) homosexual, gay, or lesbian (C) bisexual, or (D) something else? Remember, your answers are confidential”. Respondents who answered “something else” or “don’t know” or “refused” were not included in further analysis.

Demographic characteristics:

Respondents provided their exact age and the highest level of education they had completed. Separate questions were asked about race (multiple responses were allowed, including an additional question about “preferred race” for those with multiple race responses) and Hispanic ethnicity. We combined race and ethnicity responses and classified respondents according to their reported single race or preferred race for non-Hispanics and as Hispanic (any race) for those who indicated Hispanic ethnicity.

Current smoking:

Respondents had smoked at least 100 cigarettes in their lifetime and currently smoke “every day” or “some days”.

Bar attendance:

Respondents went to a bar during the past year. This question was asked only in 2005.

Belief in tobacco industry rights:

Respondents gave any answer other than “strongly disagree” when asked about agreement with the statement “Tobacco companies should have the same rights to advertise their products as other companies”.

Use/wear tobacco industry merchandise:

Respondents indicated that they were open to using or wearing something with a tobacco company logo or picture on it. In 2003 responses were “yes” and “no”. In subsequent years responses were “definitely yes”, “probably yes”, “probably no” and “definitely no”. We combined “yes” from 2003 with responses other than “definitely no” from other years.

Currently own tobacco industry promotional item:

Respondents currently have a piece of clothing or other item with a tobacco brand or logo on it.

At an event with free tobacco samples:

Respondents indicated “yes” when asked “During the past year have you been somewhere, such as a concert or special event, where tobacco companies were having a promotion—for example, giving away free samples or having a special give-away?”

Received free tobacco samples:

Respondents indicated “yes” to “During the past year have you received a free sample or coupon for a free sample of cigarettes or tobacco products?”

Data analysis

We used the Pearson χ² test of independence to determine whether age, race and education distributions varied by sexual orientation. Prevalence estimates from BRFSS are weighted to adjust for sampling design and also post stratified by age. We used multiple logistic regression models to test for associations between sexual orientation and tobacco industry marketing/receptivity measures, stratified by gender. All models were adjusted for age and education. We also adjusted all models for smoking status, because smokers might receive more tobacco marketing materials and thus the excess risk for smoking in the LGB population was a potential confounder for excess marketing exposure and receptivity. We conducted exploratory analyses fitting bar attendance into multiple logistic models to see if there was any systematic effect on the odds for marketing receptivity or exposure. Because distributions were not substantially different, and because the number of minority respondents was relatively small, we did not adjust models for race/ethnicity. Finally, we combined non-smoking lesbian/bisexual women and gay/bisexual men and compared the prevalence of indicators, post stratified by age group. Analyses were completed using Stata V. 9.0 statistical software (Stata, College Station, Texas, USA) and using a significance level of p = 0.05.

RESULTS

A total of 17 851 men and 28 807 women were included in our dataset. Among men, 249 (unweighted estimate: 1.3%) self-identified as gay and 132 (0.7%) as bisexual. Among women, 308 (1.0%) self-identified as lesbian and 323 (1.1%) as bisexual. Approximately 0.2% of each gender (36 men and 59 women) reported being “other” orientation. Additionally, 0.7% of men and 1.1% of women reported “don’t know” and 1.2% of men and 1.6% of women refused to answer the question. Respondents who reported “other” orientation, “don’t know” and “refused” were excluded from remaining analyses.

Population demographics and smoking status

Table 1 presents demographic descriptive data for men and women, stratified by sexual orientation. For men and women, age was associated with sexual orientation (p<0.001); bisexuals were youngest and straight men and women oldest within each gender group.

The racial/ethnic composition of our study group reflected the distribution of Washington State’s population, which is more than 80% white non-Hispanic. Race/ethnicity was not significantly associated with sexual orientation among women or men, likely due to small numbers of minority respondents. However, we noted that there were fewer minority respondents for gay men in comparison to straight men, and more minority respondents for bisexual men and women in comparison to straight men and women.

Educational attainment was significantly different by sexual orientation (p<0.001 for men and women). For men and women, bisexuals appeared most likely to have less than a high school education and were least likely to be college graduates. Gays and lesbians were most likely to be college graduates.

Smoking status was significantly associated with sexual orientation within both gender groups (p<0.001 for male and female groups). Current smoking prevalence was highest among bisexuals in each group (43.3% among women and 37.0% among men), followed by lesbian women (33.5%) and gay men (31.5%) and lowest among heterosexuals (19.9% among men and 17.2% among women).

Having gone to a bar during the past year was not significantly associated with sexual orientation among men, although gay and bisexual men had a higher measured prevalence of bar attendance in comparison to straight men. Among women there was a significant association between sexual orientation and bar attendance (p = 0.001), with straight women less likely to go to bars in comparison to lesbians and bisexuals.

Receptivity and exposure to tobacco industry marketing: men

At least two-thirds of men in all sexual orientation groups agreed that the tobacco industry should have the same rights to advertise as other industries, and more than a third of all men would use or wear tobacco industry merchandise (see table 2). There was no significant difference in receptivity to tobacco industry marketing by sexual orientation after demographic and smoking adjustments.

Relatively few men said they currently own a tobacco industry marketing item, and although unadjusted prevalences were higher among gay and bisexual men in comparison to straight men these differences were not significant after demographic and smoking adjustments. Relatively few men said they were at an event during the past year where free samples or coupons were being distributed (8.3–10.4%); however adjusted odds for being exposed to free samples/coupons for tobacco at events were almost double for gay and bisexual men in comparison to straight men once demographic differences were taken into account. Differences sampling event attendance were non-significant for bisexual men, however the smaller sample size for bisexual men limited our power to detect statistical differences and the similar effect size to gay men is suggestive of a similar level of exposure.

About one in every seven straight men and about one in every five gay and bisexual men reported that they had received a free sample or coupon for tobacco during the past year. This higher prevalence of receiving free tobacco samples/coupons among gay and bisexual men was not significant after adjustment for demographic covariates.

Receptivity and exposure to tobacco industry marketing: women

Approximately two-thirds of lesbian and bisexual women reported believing that the tobacco industry has the same rights to advertise as other industries, which was not significantly more than among straight women (table 3). Differences did not persist after adjustment for age, education and smoking status.

Lesbian and bisexual women were more likely to report willingness to use or wear tobacco industry marketing items than straight women. These differences persisted after adjustment for covariates; odds of willingness to use tobacco industry marketing items were about double among bisexual women in comparison to straight women.

Lesbian and bisexual women were also more likely than straight women to own tobacco industry marketing items. Odds for owning tobacco industry marketing items were about doubled among lesbians and increased by 60% among bisexual women in comparison to straight women, after adjustment for covariates.

Exposure to tobacco industry sampling at events was also significantly greater among lesbian and bisexual women in comparison to straight women. The prevalence of receiving free samples was higher among bisexual women than straight women. Odds for having exposure to tobacco sampling during the past year were approached double among lesbian and bisexual women in comparison to straight women after adjustment for covariates, although odds for actually receiving free samples were not significantly different after adjustment.

Receptivity and exposure to tobacco industry marketing: non-smokers

Finally, we combined lesbian and bisexual (LB) women and gay and bisexual (GB) men, to describe the prevalence of receptivity and exposure to tobacco industry marketing among non-smokers alone in comparison to their heterosexual counterparts (fig 1). For almost all comparisons, non-smoking LB and GB groups, in comparison to non-smoking straight women or men, reported greater receptivity and exposure to tobacco industry marketing. Non-smoking LB women were significantly more likely than non-smoking straight women to report being willing to use or wear tobacco industry marketing items; to own an industry marketing item; and to have been at an event with tobacco industry marketing.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1 Prevalence of receptivity and exposure to tobacco industry marketing among non-smokers alone in comparison to their heterosexual counterparts.

DISCUSSION

While previous studies revealed tobacco industry plans to target LGB communities with advertising and other marketing strategies, our study suggests that industry has been successful in doing so. We found that tobacco industry marketing has disproportionately impacted LGB populations in comparison to straight men and women.

Lesbian/bisexual women appeared to have greater disparities in reported exposure to tobacco advertising than gay/bisexual men when compared to their straight counterparts. This complements Smith’s finding that gay periodicals had more antitobacco messages than lesbian periodicals.14

One possible explanation for differences in receptivity and exposure to marketing could be tobacco industry promotional activities in bars. To explore this possibility, we compared the prevalence of bar attendance by sexual orientation. We added bar attendance to logistic regression models and the odds for marketing exposure did not appear to be systematically decreased (data not shown). Thus, different rates of bar attendance do not explain increased odds of exposure to tobacco industry marketing for LGB populations. This suggests that higher exposure to industry marketing among LGB populations is the result of marketing activities beyond bar-based strategies.

An alternative explanation for higher levels of reported marketing exposure in the LGB community could be that these populations do not have large numbers of advertisements directed at them; thus, any advertising is more memorable. In other words, it could be that there are similar amounts of pro-tobacco advertising directed toward straight men and women as to LGB populations, but the straight population is so inundated with advertising that they are less affected by it. This theory is somewhat supported by recent findings suggesting that some LGB community members see few general LGB-targeted advertisements and may pay close attention to those they do see, even viewing tobacco marketing as a validation of the community.23 However, several sources document the intent of the tobacco industry to specifically reach LGB communities.13^–15 We do not have a way to quantitatively compare the frequency of marketing advertisements directed to the LGB community vs the general population.

We adjusted for smoking status in our analyses to avoid confounding due to differential smoking prevalences, and excess exposure to tobacco industry marketing remained after this adjustment. To provide further justification that differences were not the result of confounding, we compared the prevalence of receptivity and exposure to tobacco marketing for non-smoking straight and lesbian/bisexual or gay/bisexual groups, respectively. For almost all comparisons, non-smoking LGB populations were more likely than straight men or women to report greater receptivity and exposure to tobacco marketing; small numbers of respondents likely prevented more differences from achieving statistical significance. Thus, differences in the entire population are not the result of active smokers receiving tobacco industry marketing in the course of purchasing cigarettes.

Opportunities remain for LGB health advocates to change social norms around tobacco by influencing the LGB media. Editors and publishers of LGB media should be encouraged to refuse pro-tobacco advertising, to examine their non-tobacco advertising and to be aware of ubiquitous use of tobacco imagery in non-advertising content, to assure that their periodicals are not promoting and reinforcing norms that harm the LGB community by displaying tobacco in a positive way. LGB event sponsors and bar owners should refuse tobacco industry sponsorship and not allow marketing activities. Antitobacco marketing should be designed to reach LGB communities through targeted media and use of inclusive language and imagery. The community should be educated about industry schemes to “buy” their loyalty through charitable contributions and sponsorships. State or community-based tobacco control programs should build alliances in the LGB community and establish a presence at LGB events such as Gay Pride.

Of great importance, we note that we were able to describe tobacco-related indicators among LGB populations, generalise to a state-wide population, and include a straight comparison group only because Washington State added a sexual orientation question to the state BRFSS. The addition of sexual orientation as a demographic question in established public health surveillance systems is crucial to a better understanding of the health risks faced by the LGB community. Public health advocates should strongly encourage state and federal partners to follow the lead of Washington and other states that have begun to take this approach.

Our study provides additional information suggesting a causal pathway between targeted tobacco industry marketing activities and high smoking prevalence in specific communities. Additional studies of specific communities that have been the focus of tailored or excessive tobacco industry marketing activities, to describe the impact of tobacco marketing-related indicators, may be useful.

Limitations

The question we used to assess sexual orientation or sexual identity was based on self-identification, which may be a less sensitive measure than sexual behaviours or attraction.24 However, determination of LGB based on self-identification is satisfactory from the standpoint that public health interventions would likely be tailored to those who self-identify as LGB, and therefore are more likely to read LGB-specific periodicals, attend LGB events or venues and respond to marketing campaigns addressing LGB populations. In other words, we expect that people who have LGB behaviours but do not identify as LGB may be less likely to respond to health promotion activities designed for LGB populations. Investigation of the influence of LGB-targeted health promotions (or pro-tobacco marketing) on people with LGB behaviours who do not self-identify as LGB may be useful.

The BRFSS has a number of inherent limitations, including exclusion of individuals who live in homes without telephones or those who have only cellular telephone service, people living in institutions (such as college dorms and assisted living facilities) and those who do not speak English or Spanish. Also, the response rates (calculated using Council of American Survey Research Organizations (CASRO) method) ranged between 43–47% for each year of our study. We do not know whether potential LGB respondents are differentially likely to respond to BRFSS than the general population. However, the purpose of our study was to highlight the importance of identifying LGB individuals, within the population of individuals who are represented by the BRFSS. Generalisation of findings from our study to the true LGB population should be considered carefully, bearing in mind the overall limitations of the BRFSS.

Despite combining multiple years of data, we had relatively small numbers of LGB respondents, particularly bisexual men. This reduced our power to detect statistical differences; thus, there may be significant disparities in some of our indicators that will be revealed with inclusion of additional years of data.

CONCLUSIONS

LGB communities, particularly lesbian/bisexual women, appear to be excessively impacted by tobacco industry marketing. Strategies to limit marketing or increase resistance to marketing may be critical to reducing tobacco-related health disparities for LGB communities.