Sexual Identity Disparities in Smoking and Secondhand Smoke Exposure in California: 2003–2013

Abstract

Objectives. To determine smoking prevalence, smoking behavior, and secondhand smoke (SHS) exposure of lesbian, gay, and bisexual (LGB)–identified Californians; compare these with that of heterosexuals; and analyze changes over time.

Methods. We analyzed self-reported variables from 111 965 heterosexual, 1667 lesbian, and 1706 bisexual women, and 79 881 heterosexual, 2505 gay, and 911 bisexual men, aged 18 to 70 years, in the 2003–2013 California Health Interview Surveys.

Results. Sexual minority women had higher smoking prevalence, and female bisexual smokers were less likely to be light smokers, than heterosexuals. Smoking prevalence was higher among sexual minority men, and gay smokers were more likely to be daily smokers than were heterosexuals; and male bisexual smokers were more likely to be light smokers than were gay or heterosexual smokers. Sexual minority adults were more likely to have SHS exposure at home than were heterosexuals. Current smoking prevalence decreased annually 4% and 7% for lesbian and bisexual women, and 5% and 6% for gay and bisexual men, respectively. Exposure to SHS fell an average of 11% annually for sexual minority men and women.

Conclusions. Sexual identity disparities in smoking and SHS exposure exist in California, with bisexuals particularly at risk.

The lesbian, gay, and bisexual (LGB)–identified population in the United States has higher smoking prevalence than the rest of the population.^1–7 California LGB-identified persons also have higher smoking prevalence than their heterosexual counterparts.^8–11 Previous studies have reported that gay and bisexual men smoke at 1.4 to 1.6 times the rate of heterosexual men.^8–10 The differentials were even greater for women, with rates ranging from 1.7 times as high for lesbian women in a study that used the 2001 California Health Interview Survey (CHIS),⁸ to 2.4 to 2.5 times as high for lesbian and bisexual women according to the 2002 California Tobacco Survey,¹⁰ to a study that used the 2004 California LGB Tobacco Use Survey of 2004 that found a differential of almost 3 times for lesbian and bisexual women compared with heterosexual women.⁹ Each of these studies used a different survey, and the surveys differ in the time period covered as well as the questions used to identify sexual identity. Nonetheless, the findings consistently show that sexual minority adults have higher smoking prevalence than heterosexual adults in California. Although adult smoking prevalence has continued to decline in California, it is unknown whether this trend also occurs in the LGB population.

Other smoking behaviors, including smoking intensity and daily versus nondaily smoking, have not been examined in the California LGB population. Among the California general population of smokers, the average number of cigarettes smoked per day fell from 19 in 1992 to 14 in 2008,¹² and nondaily smokers increased from 14.8% of smokers in 1992 to 28.1% in 2008.¹² It is not known whether the declining smoking intensity and the shift from daily to nondaily smoking have also occurred in the LGB population. With the high rates of smoking prevalence reported among LGB-identified people, it is likely that secondhand smoke (SHS) exposure is also high, but this has not been documented previously.

The purpose of this study was to determine the smoking prevalence, smoking behavior patterns, and SHS exposure of LGB-identified Californians; compare these with those of heterosexuals; and analyze changes in these behaviors over the past decade.

METHODS

The CHIS is a telephone survey of California households that has been conducted since 2001. It contains data on smoking behavior, other risk behaviors, socioeconomic characteristics, and health care utilization. This survey employs a multistage geographically stratified random-digit-dial design, yielding a sample representative of California’s noninstitutionalized population. Adult sample sizes average roughly 47 000 per cycle. Since 2007, the CHIS has included a cell phone sample, which in 2011–2012 accounted for 22% of the adult interviews. For the 2001–2002 through 2009–2010 surveys, data were collected over a period of roughly 9 months within each 2-year cycle and were weighted based on the 2000 Census population for California. Beginning with the 2011–2012 CHIS, data are collected continuously across the 2-year cycle, and are weighted based on the 2010 Census. The 2013 data were released as a single year before the availability of the 2014 data. To obtain sufficient sample size for sexual minority subpopulations, we pooled data from the 2003–2004, 2005–2006, 2007–2008, 2009–2010, 2011–2012, and 2013 surveys. Our analyses were restricted to respondents aged 18 to 70 years who identified as heterosexual, lesbian, gay, or bisexual, resulting in a study sample size of 83 299 men and 115 338 women. Because the sexual identity question is only available in the confidential CHIS data set, analyses were conducted through the CHIS Data Access Center.

Measures

Sexual identity.

We determined LGB status on the basis of a sexual identity question asked of all CHIS respondents aged between 18 and 70 years: “Do you think of yourself as straight or heterosexual, as gay/lesbian or homosexual, or bisexual?” Fewer than 1% of the survey respondents responded “other”; “not sexual, celibate, or none”; or were skipped because of proxy interview. We excluded these respondents from all analyses in this study. We classified respondents as heterosexual, gay, lesbian, or bisexual.

Smoking behavior.

We defined current smokers as those who reported having smoked at least 100 cigarettes in their lifetime and now smoke every day or some days. We further classified current smokers into daily smokers and occasional smokers depending on whether they now smoke cigarettes every day or some days. We also classified current smokers into light smokers (occasional smokers and daily smokers who smoke fewer than 10 cigarettes per day) and non–light smokers.

Secondhand smoke exposure.

We analyzed SHS exposure only among nonsmokers, defined as respondents who are not current smokers. Although smokers may also suffer the consequences of SHS exposure, it is difficult to separate the impact of active and passive smoking for them, and thus we focused on nonsmokers. We defined nonsmokers who reported that someone smoked inside the home at least 1 day per week as being exposed to SHS at home. Because the SHS exposure was not assessed in the 2013 CHIS, analyses related to SHS exposure were limited to 2003 to 2012.

Covariates.

We included the following sociodemographic covariates on the basis of our review of the literature: age, race/ethnicity, education, and poverty level.^3,5,6,13,14 We measured poverty level by using the ratio of family income to the family’s poverty threshold as defined by the US Census Bureau.¹⁵ We included age and poverty level as continuous variables in all regression analyses. For descriptive purposes in Table A (available as a supplement to the online version of this article at http://www.ajph.org), we also show age in 4 categories (18–25, 26–34, 35–49, and 50–70 years) and poverty level in 4 categories (0.00–0.99, 1.00–1.99, 2.00–2.99, and ≥ 3.00). We also included age-squared in our regression analyses to control for nonlinear relationships between age and outcome measures, resulting in better model fit than inclusion of either linear or quadratic age alone. We classified race/ethnicity as Hispanic, non-Hispanic White, non-Hispanic African American, non-Hispanic Asian, and non-Hispanic other or multiple. We grouped education into 4 categories based on the highest level attained: less than high-school diploma, high-school diploma, some college, and college degree or higher.

Statistical Analyses

We conducted analyses for men and women separately. We used cross-tabulations to describe the frequency distribution of each categorical sociodemographic variable (race/ethnicity and education) and survey year, whereas we used both means and cross-tabulations to summarize continuous demographic variables (age and poverty level) and their categorical recodes by sexual identity. We performed separate χ² tests to compare gay or lesbian and bisexual subgroups with the heterosexual reference group, as well as the bisexual subgroup with the gay or lesbian reference group. We used the t test to compare continuous sociodemographic variables by sexual identity.

We analyzed the association between sexual identity and smoking behaviors as well as SHS exposure by using multivariate logistic regression models, with control for age, age squared, poverty level, year, race/ethnicity, and education. We used these models to estimate adjusted odds ratios (AORs) and their 95% confidence intervals (CIs) for each sexual identity subgroup, with heterosexual as the reference group. Wherever (3-level) sexual identity was treated as a predictor, we specified contrast estimation in the models to obtain coefficients and P values indicating any significant differences in outcome likelihood between gay or lesbian and bisexual subgroups. For both men and women, we combined gay or lesbian and bisexual subgroups for the analysis of SHS exposure at home because of small sample sizes.

We conducted time-trend analyses for each smoking-related outcome measure. For each gender∗sexual identity subgroup, we estimated multivariate logistic regression models of each outcome on year, with control for age, age squared, race/ethnicity, education, and poverty level. We used these models to estimate AORs and their 95% CIs, reflecting the annual incremental impact of year on smoking outcome likelihood between 2003 and 2013 (2003 to 2012 for SHS exposure). We conducted an additional analysis stratified by gender only to examine differences in time trends between sexual identity subgroups by using an interaction term of sexual identity∗year, as recommended by Lee et al.¹⁶

We plotted trends in annual prevalence rates of current smoking and SHS exposure by using estimated adjusted predicted prevalence from the logistic models described previously. The resulting figures reflect the effect of year on each smoking measure, when age, age squared, race/ethnicity, education, and poverty level are held at mean values.¹⁷

We used survey data analysis procedures in this study because of the complex multistage sample design of the CHIS. We conducted all analyses by incorporating the appropriate sampling weights to account for selection probabilities and to adjust for survey nonresponse. We conducted analyses with SAS version 9.3 (SAS Institute Inc, Cary, NC) with PROC SURVEYLOGISTIC for logistic regression and PROC SURVEYFREQ, PROC SURVEYREG, and PROC SURVEYMEANS for statistical calculation. We used the jackknife repeated replication method to obtain accurate standard errors of all estimates. We considered statistical significance as a 2-tailed P value less than .05.

RESULTS

The final sample analyzed included 111 965 heterosexual, 1667 lesbian, and 1706 bisexual women, and 79 881 heterosexual, 2505 gay, and 911 bisexual men. Between 2003 and 2013 in California, 3.0% of women identified as lesbian (1.3%) or bisexual (1.7%). Among men, 3.7% identified as gay (2.6%) or bisexual (1.1%).

Socioeconomic Characteristics

There were statistically significant differences in the distribution of sociodemographic characteristics between heterosexual and sexual minority women, as detailed in Table A. The racial/ethnic distribution of lesbian and of bisexual women differed from that of heterosexual women. Bisexual women were younger (mean = 33.1 years) than both heterosexual (mean = 41.7 years) and lesbian women (mean = 41.7 years). Heterosexual, lesbian, and bisexual women had different educational profiles, with lesbians having the greatest percentage of women with a college degree or higher. Lesbians were most prevalent in the highest-income group (65.1%) and bisexual women were the least prevalent (44.1%).

Gay and bisexual men differed by race/ethnicity from heterosexual men (Table A), with more gay and bisexual men being non-Hispanic White (54.9% and 48.3%, respectively) than heterosexual men (44.6%). Similar to women, bisexual men were younger (mean = 37.0 years) than both heterosexual men (mean = 40.9. years) and gay men (mean = 40.6 years). Gay men differed from both heterosexual and bisexual men in terms of education. Gay men were most prevalent in the highest-income group (71.2%) and bisexual men were the least prevalent (48.6%).

Smoking Behavior

There were significant differences in smoking behavior by sexual identity after we controlled for socioeconomic characteristics (Table 1). Among women, lesbian and bisexual women smoked at significantly higher rates than did heterosexual women (20.9%, 25.2%, and 11.6%, respectively), and were more than twice as likely to be current smokers (AOR = 2.16; 95% CI = 1.73, 2.70 for lesbian women, and AOR = 2.53; 95% CI = 2.08, 3.09 for bisexual women). Although there were no significant differences in daily smoking rates, bisexual smokers were less likely (AOR = 0.68; 95% CI = 0.48, 0.95) to be light smokers than were heterosexual smokers. Sexual minority women were almost twice as likely to be exposed to SHS at home (AOR = 1.72; 95% CI = 1.19, 2.49) as heterosexual women.

TABLE 1—

Smoking Behavior and Secondhand Smoke Exposure Among California Adults Aged 18–70 Years by Gender and Sexual Identity: California Health Interview Survey, 2003–2013

Variable	Raw No.	Weighted Prevalence, % (SE)	AORa (95% CI)
Women
Current smoker = yes
Heterosexual	14 007	11.6 (0.2)	1 (Ref)
Lesbian	316	20.9 (1.8)	2.16 (1.73, 2.70)b
Bisexual	422	25.2 (1.8)	2.53 (2.08, 3.09)b
Daily smoker = yesc
Heterosexual	9 931	67.5 (0.7)	1 (Ref)
Lesbian	233	62.1 (4.7)	0.95 (0.62, 1.45)
Bisexual	280	65.6 (3.9)	1.31 (0.93, 1.86)
Light smoker = yesc
Heterosexual	6 911	57.6 (0.7)	1 (Ref)
Lesbian	143	54.1 (4.4)	0.67 (0.44, 1.02)
Bisexual	237	60.2 (4.0)	0.68 (0.48, 0.95)b
Household SHS exposure = yesd
Heterosexual	1 833	2.5 (0.1)	1 (Ref)
Lesbian or bisexual	63	3.9 (0.7)	1.72 (1.19, 2.49)b
Men
Current smoker = yes
Heterosexual	13 831	18.9 (0.2)	1 (Ref)
Gay	523	21.1 (1.4)	1.47 (1.22, 1.75)b
Bisexual	242	29.0 (2.6)	1.77 (1.36, 2.31)b
Daily smoker = yesc
Heterosexual	9 497	62.7 (0.7)	1 (Ref)
Gay	369	70.1 (3.2)	1.52 (1.09, 2.11)b
Bisexual	169	60.0 (5.4)	0.92 (0.59, 1.44)
Light smoker = yesc
Heterosexual	6 678	57.5 (0.7)	1 (Ref)
Gay	234	50.4 (3.8)	0.79 (0.56, 1.10)
Bisexual	123	64.9 (4.5)	1.56 (1.03, 2.37)b,e
Household SHS exposure = yesd
Heterosexual	1 671	3.0 (0.1)	1 (Ref)
Gay or bisexual	110	5.6 (0.8)	2.35 (1.72, 3.21)b

Note. AOR = adjusted odds ratio; CI = confidence interval; SHS = secondhand smoke.

^aAdjusted by age, age squared, race/ethnicity, educational attainment, poverty level, and year.

^bSignificant difference at P < .05 compared with heterosexual subgroup.

^cAmong current smokers.

^d2003–2012 California Health Interview Survey only (household SHS exposure was unavailable in 2013 survey).

^eSignificant difference at P < .05 compared with gay subgroup.

Men showed similar differences in smoking behavior by sexual identity. Smoking prevalence among gay, bisexual, and heterosexual men was 21.1%, 29.0%, and 18.9%, respectively. Gay and bisexual men were more likely to be current smokers than were heterosexual men (AOR = 1.47; 95% CI = 1.22, 1.75, and AOR = 1.77; 95% CI = 1.36, 2.31, respectively), and gay smokers were more likely to be daily smokers than were heterosexuals (AOR = 1.52; 95% CI = 1.09, 2.11). Bisexual male smokers were more likely to be light smokers than were both heterosexual male smokers (AOR = 1.56; 95% CI = 1.03, 2.37) and gay smokers (AOR = 1.88; 95% CI = 1.12, 3.20). Sexual minority men were more than twice as likely to be exposed to SHS at home as were heterosexual men (AOR = 2.35; 95% CI = 1.72, 3.21).

Time Trends in Smoking Behavior

The likelihood of current smoking among bisexual women decreased significantly over the time period from 2003 to 2013, with adjusted odds falling an average of 7% each year (Table 2). These results are also shown in Figure 1, which shows the adjusted predicted current smoking prevalence over time. Prevalence fell for all sexual identity groups, but remained higher for sexual minority women and men. Heterosexual women smokers’ light smoking odds increased over time, by an average of 6% per year. Increasingly fewer heterosexual women were exposed to SHS at home, with adjusted odds falling an average of 10% every year from 2003 to 2012. Figure 2 shows that SHS exposure showed a similar pattern of decline for heterosexual and lesbian or bisexual women, but was lower each year for heterosexual women.

TABLE 2—

Analysis of Annual Time Trends in Adult Smoking Behavior and Secondhand Smoke Exposure Among California Adults Aged 18–70 Years by Gender and Sexual Identity: California Health Interview Survey, 2003–2013

Variable	AORa (95% CI)
Women
Current smoker = yes
Heterosexual	0.99 (0.98, 1.00)
Lesbian	0.96 (0.89, 1.03)
Bisexual	0.93 (0.88, 0.98)
Daily smoker = yesb
Heterosexual	0.99 (0.97, 1.01)
Lesbian	0.92 (0.79, 1.07)
Bisexual	1.00 (0.89, 1.11)
Light smoker = yesb
Heterosexual	1.06 (1.04, 1.09)
Lesbian	1.09 (0.95, 1.25)
Bisexual	1.01 (0.91, 1.12)
Household SHS exposure = yesc
Heterosexual	0.90 (0.87, 0.93)
Lesbian or bisexual	0.89 (0.79, 1.00)
Men
Current smoker = yes
Heterosexual	0.98 (0.97, 0.99)
Gay	0.95 (0.90, 0.99)
Bisexual	0.94 (0.88, 1.01)
Daily smoker = yesb
Heterosexual	0.98 (0.96, 0.99)
Gay	0.89 (0.80, 0.99)
Bisexual	0.90 (0.77, 1.05)
Light smoker = yesb
Heterosexual	1.06 (1.04, 1.08)
Gay	1.16 (1.04, 1.29)
Bisexual	1.26 (1.08, 1.46)
Household SHS exposure = yesc
Heterosexual	0.94 (0.91, 0.97)
Gay or bisexual	0.89 (0.81, 0.99)

Note. AOR = adjusted odds ratio; CI = confidence interval; SHS = secondhand smoke.

^aAdjusted by age, age squared, race/ethnicity, educational attainment, and poverty level.

^bAmong current smokers.

^c2003–2012 California Health Interview Survey only (household SHS exposure was unavailable in 2013 survey).

FIGURE 1—

Adjusted Predicted Current Smoking Prevalence Among California Adults by Sexual Identity for (a) Women and (b) Men: California Health Interview Survey, 2003–2013

Note. All models are adjusted by age, age squared, race/ethnicity, educational attainment, and poverty level.

FIGURE 2—

Adjusted Predicted Secondhand Smoke Exposure Among California Adults by Sexual Identity for (a) Women and (b) Men: California Health Interview Survey, 2003–2013

Note. All models are adjusted by age, age squared, race/ethnicity, educational attainment, and poverty level.

Heterosexual and gay men showed declining odds of current smoking, with rates of decline of 5% for gay men and 2% for heterosexual men yearly. Figure 1 shows that the adjusted predicted current smoking prevalence declined over time for all groups, but remains lowest for heterosexual men. The odds of daily smoking also declined by 2% annually among heterosexual male smokers, and by 11% annually among gay male smokers. All 3 sexual identity groups experienced increases in light smoking odds among smokers of 6% for heterosexual men, 16% for gay men, and 26% for bisexual men each year. All men showed a pattern of reduced SHS exposure at home over the 2003 to 2012 time period, with reductions averaging 6% per year for heterosexual men and 11% per year for sexual minority men. Exposure to SHS remained lowest for heterosexual men (Figure 2).

Our comparison of time trends between sexual identity groups (results not shown) revealed that current smoking is declining more rapidly for bisexual women and men compared with heterosexual women and men, and that light smoking is increasing more rapidly among bisexual male smokers compared with heterosexual male smokers.

DISCUSSION

Our findings are consistent with those of other studies that have reported higher rates of smoking in sexual minority groups than among heterosexuals.^1–11 We found that, in California, not only do LGB-identified men and women smoke at greater rates than do heterosexuals, but there were also differences by sexual identity in daily smoking rates (for men) and light smoking rates among smokers (for men and women). Bisexual women smokers were less likely to be light smokers than were heterosexuals, and gay male smokers were more likely to be daily smokers and bisexual male smokers were more likely to be light smokers than were heterosexual smokers. We also found that LGB men and women had greater rates of SHS exposure at home than did heterosexuals.

A recent report issued by the California Department of Public Health (CDPH)¹⁸ found that, during 2005 to 2010, lesbian women’s smoking prevalence was 2.5 times higher than that of heterosexual women (24.4% vs 9.8%); we found the differential to be similar at 1.8 times (20.9% vs 11.6%). The CDPH reported that gay men’s smoking prevalence was more than 1.5 times higher than for heterosexual men (25.8% vs 16.0%); we found it to be 1.1 times higher for gay men (21.1% vs 18.9%). The differences are likely attributable to our analysis covering a more extended time period, including more recent years during which we found falling prevalence, and the use of a different survey by the CDPH—the California Adult Tobacco Survey.

We found that bisexual men and women had the greatest odds of being current smokers, consistent with other studies.¹⁹ This group had lower income but similar educational attainment (for men) compared with their heterosexual counterparts, and this finding persisted after we controlled for sociodemographic covariates. Bisexual men and women may be particularly vulnerable to stressors, such as violence, substance abuse, and mental health issues, even more so than lesbians and gay men.²⁰ Bisexuals may experience lack of acceptance from both the heterosexual and the lesbian and gay community,²¹ resulting in less willingness to be open, and thus less support. Our results add to the evidence that bisexuals may be at risk for worse health outcomes than other sexual orientation groups because of higher smoking rates.

Higher smoking and SHS-exposure rates are particularly troubling, given that the LGB population has a number of other risk factors for negative health outcomes that may be compounded by tobacco exposure. For example, gay men have greater rates of HIV than do heterosexual men, and smoking has been shown to lead to poorer clinical outcomes among those with HIV, including greater likelihood of detectable viral load.²² HIV-positive smokers have been found to be at greater risk than HIV-positive nonsmokers for cardiovascular disease, human papillomavirus–associated cancers, lung cancer, tuberculosis, chronic obstructive pulmonary disease, pneumonia, bone fractures, and other serious threats, both HIV-related and not.²³

There is a paucity of information about the patterns of daily versus occasional smoking and intensity of smoking among LGB smokers. We found that bisexual women were less likely to be light smokers than were heterosexual women, bisexual men were more likely to be light smokers than were both gay and heterosexual men, and gay men were more likely to be daily smokers than were heterosexual men. The differences in smoking behaviors among sexual minority men and women highlight the importance of of developing smoking cessation interventions that consider the differences within the LGB community, not just the differences between the LGB and heterosexual communities.²⁴

In light of the higher rates of smoking in the LGB community compared with the heterosexual population, it is not surprising that SHS exposure was also higher. The good news is that exposure is falling among all groups, although sexual minority nonsmokers continue to be exposed at greater rates.

We found that current smoking prevalence has fallen during the past decade in the LGB population, particularly among bisexual men and women. Analyzing why these trends have occurred is beyond the scope of this article. However, several factors may be contributing. Perhaps most significant is the overall decline in smoking prevalence in California to second-lowest in the nation as a result of an aggressive statewide tobacco-control campaign. All Californians have been exposed to effective messaging. State-level tobacco environments are known to have an impact on LGB smoking.²⁵ The LGB men and women living in states with more restrictive tobacco environments (such as California) are less likely to have ever smoked, and in those states there are smaller sexual identity disparities in smoking. Thus, the California LGB community has likely benefitted from tobacco-control policies that do not specifically target LGB smokers.

In addition, a growing network of lesbian, gay, bisexual, and transgender tobacco-control advocates has emerged to address the degree to which tobacco harms the community and the role that the tobacco industry plays in keeping LGB people smoking.²⁶ As a result, there is increasingly favorable coverage of tobacco-control issues in the LGB media, which may be encouraging a new norm in which young people no longer link smoking with being gay.²⁷ Smoking cessation clinics²⁸ focused on the LGB population and campaigns to urge LGB elected officials not to accept money from the tobacco industry²⁹ may also promote a smoke-free norm. With a range of interest groups now available in the LGB community, many have become less dependent for socializing on a bar-focused subculture that encourages smoking. Also, as more people comfortably identify as part of the lesbian, gay, bisexual, and transgender community, they may experience less stress and feel less of a need to smoke. “Coming out” may help align smoking prevalence rates across sexual identity groups.

California ranks as relatively progressive with respect to the legal and cultural equality of sexual minorities, including, for example, same-sex marriage. Sexual identity smoking disparities may diminish over time as structural stigma, systemic discrimination, and internalized prejudice is further reduced.

Limitations

We acknowledge some limitations of our study. First, from 2005 through 2013, the CHIS question about sexual identity was only asked of respondents aged 18 to 70 years. Thus, our results cannot be generalized to those outside this age group. Second, sexual orientation is defined in the CHIS on the basis of self-identification. It is possible that there were changes over time in the willingness of respondents to identify as sexual minorities. However, we are not able to test this and our analyses all rely on the measure of self-identification that is available in the CHIS.

Third, despite the combination of multiple years of data, sample size for some of the analyses remained small. For example, we were forced to combine lesbian and bisexual women, and gay and bisexual men for analyses related to SHS exposure. Fourth, CHIS does not include questions about SHS exposure in the workplace or other nonhome settings. As a result, our analyses underestimate SHS exposure. Bars are part of the culture of the LGB community, and despite smoke-free laws, bars remain a common place of exposure, but we were not able to assess this in our study. Finally, the CHIS collects information on cigarette smoking only and we were not able to analyze the use of other tobacco products, such as cigars, smokeless tobacco, or e-cigarettes.

Conclusions

Smoking and SHS exposure are falling among all groups in California, but disparities by sexual identity are evident. The LGB population is particularly vulnerable to the negative impact of tobacco as a result of other health risk factors. Although California has a strong tobacco-control program, there remains a need to expand upon the few interventions and policies that have targeted the LGB community. Bisexual men and women are particularly at risk for smoking, and interventions should consider the unique needs of this population compared with other sexual minorities. We found that even within the sexual minority community, there are differences in the association between socioeconomic status and smoking behaviors.

Smoking prevalence and SHS exposure have fallen over the past decade among California adults, and the LGB community has benefitted from this reduction. Nonetheless, the LGB community continues to smoke and be exposed to SHS at higher rates than their heterosexual counterparts. The development of targeted tobacco-control efforts for the LGB community is merited.