Introduction

Smoking is the leading preventable cause of cancer. About 40% of cancers diagnosed in the United States (U.S.) are linked to smoking [1]. Harmful chemicals found in tobacco and nicotine products can affect people’s immune system and DNA mutations that cause cancer [2]. In addition, smoking has been found to cause lung, liver, and breast cancer [3].

Current smoking prevalence in the U.S. adult population is 12.5% (30.8 million) [4]. While smoking prevalence has declined over time, smoking disparities persist for the lesbian, gay, bisexual, transgender, queer, and intersex plus (LGBTQI +) population. Lesbian, gay, and bisexual (LGB) adults have a higher smoking prevalence (16.1%) than straight/heterosexual adults (12.3%) [4]. Furthermore, smoking prevalence is also higher in transgender adults (35.5%) compared to straight and cisgender adults (20.7%) [5]. Among racial and ethnic groups in the LGBTQ + population, smoking is highest in Black (27%) individuals and those categorized as other (26%), which included Multiracial and Native American individuals, while it is lowest in Asian and Pacific Islander individuals (6%) [6]. It is important to note that when examining the smoking prevalence among racial and ethnic minorities who identify as straight/heterosexual, smoking is lower among Black adults (15%) relative to American Indian/Alaska Native adults (21%) and White adults (16%) [7], showing that taking into account gender and sexual orientation increased the risk of smoking. Studies on correlates of smoking in LGBTQI + populations found that younger age, lower educational status, no health insurance, alcohol use and use of other substances, and poor mental health were associated with smoking [8,9,10]. One known outcome of smoking is increased risk for cancer, yet, despite higher smoking prevalence in the LGBTQI + population, missing surveillance data limits the number of studies that have explored the relationship between smoking and cancer in this population.

Cancer incidence rates are expected to increase by 49% between 2015 and 2050 due to population growth and aging in the U.S. [11]. However, the prevalence of cancer in the LGBTQI + population is unclear because national cancer registries do not include sexual orientation in their medical records [12]. Although gender identity is included using the “transexual” category, North American Cancer Registry officials report that category has poor data quality. Recently, transgender, natal male and transgender, natal female were added in the data [12]. This highlights a critical gap that prevents an understanding of health disparities needed to design and implement cancer prevention efforts for LGBTQI + communities. Previous studies relied on estimates of cancer incidence rates among the general population using national data to understand this gap. For instance, in 2017, it was estimated that there would be about 84,000 new cases of cancer and 30,000 cancer-related deaths in the LGBTQI + population [13]. The dearth of cancer-related information in the LGBTQI + community is further compounded by discriminatory and systemic barriers that make it difficult for LGBTQI + individuals to seek cancer-related care, such as late-stage cancer screening and reduced survival rates [14,15,16].

Whereas cancer survivors may quit smoking after diagnosis or treatment, about 9 to 18% of cancer survivors continue to report smoking [17]. Furthermore, LGB cancer survivors are twice as likely to smoke than their heterosexual counterparts [18]. Continuing to smoke during treatment reduces the success of the treatment and can lead to complications as well as higher chances of cancer re-occurrence after remission [19]. Studies show that cancer survivors with less than a college degree, younger, widowed or divorced, and having nonexistent to partial healthcare access, and women with cervical cancer were more likely to smoke [20]. Cancer survivors who are unable to successfully quit from smoking may be facing additional barriers because of the motivation and confidence to quit smoking, limited familial support, and accessibility to affordable health insurance that can provide them access to smoking cessation programs and services that can help them quit [21,22,23].

The devastating effects of the COVID-19 pandemic have affected millions worldwide. While LGBTQI + communities were already vulnerable before the pandemic, the pandemic may have further exacerbated the negative experiences and conditions of LGBTQI + individuals [24]. A nationally representative sample of U.S. adult smokers reported that since learning about COVID-19 about 45% did not change the amount of cigarettes they smoked, 21% decreased their cigarette use, and 33% increased their cigarette use [25]. Data on smoking during the pandemic among LGBTQI + individuals are limited; however, the few that have been published indicated that LGBTQI + individuals were smoking to help alleviate the stress and anxiety they were feeling because of the pandemic [26]. Studies show that LGBTQI + individuals have exhibited increased depressive symptoms and increased use of substances such as alcohol [27, 28]. Transgender individuals were particularly affected concerning difficulties in accessing healthcare services related to hormone interventions and gender-affirming surgeries, homelessness, and living with family members who are discriminatory towards their gender identity and/or sexual orientation [29]. In addition, cancer patients and survivors faced additional stressors, such as accessing ongoing cancer treatment, and feasibility of adapting to telemedicine [30]. These stressors can further encourage LGBTQI + cancer survivors to engage in at-risk behaviors like smoking as a form of co**. Thus, cancer survivors in the LGBTQI + community are particularly vulnerable to smoking during the pandemic and it is important to understand the factors associated with smoking in this population during this period. The purpose of this study is to explore the correlates of cigarette use and other tobacco and nicotine products during the COVID-19 pandemic in the LGBTQI + cancer survivor population.

Methods

Setting and population

We implemented a secondary data analysis using OUT: The National Cancer Survey conducted by the National LGBT Cancer Network. The purpose of the survey is to assess the experiences of LGBTQI + cancer survivors through partnerships with numerous organizations across the country to increase and diversify the study’s sample. Participants were recruited online using social media and in collaboration with over 100 community partners. The survey was administered online from September 2020 through March 2021 and was available in both English and Spanish languages. The inclusion criteria included being previously diagnosed with cancer, 18 years old or older, identified as LGBTQI + , and resided in the U.S. Survey items included questions about their experiences during the COVID-19 pandemic (e.g., isolation, accessing healthcare services, and plans to get vaccinated), cancer diagnosis and treatment experience, cancer survivorship resources, cancer screening experiences, health and health behaviors, and demographic information. All survey questions were optional, with the exception of the screener questions. The survey was anonymous and participants did not receive any compensation. All participants provided consent to participate in the study. The survey was approved by the WCG Institutional Review Board.

Measures

Tobacco and other nicotine products

Participants were asked, “Have you smoked 100 or more cigarettes in your life?” A response of yes indicated ever use. Current use was assessed with the question, “Do you currently use any of the following tobacco products?” Because of the small sample size of participants who are currently using tobacco and other nicotine products (3.31%) separate from cigarettes (10.01%), two categories were created: no current use (I do not currently use tobacco products, don’t know/prefer not to answer) and current use (cigarettes, cigars or cigarillos, electronic cigarettes, vapes, or JUUL, hookah, chewing tobacco, snus, or snuff).

Correlates of tobacco use

Psychological distress

Psychological distress was assessed using the question, “In the past 30 days, for how many days was your mental health poor?” Participants responded with the number of days they felt their mental health was poor. We employed 14 or more days as an indicator of psychological distress as used by previous studies [31].

Binge drinking

Binge drinking was assessed using the question, “On average, how many alcoholic drinks do you drink on an average day?” Participants responded with the number of alcoholic drinks they have consumed. Although some studies have used four or more drinks for women and five or more drinks for men as thresholds for binge drinking, there is no general consensus on binge drinking by sex and/or gender identity, particularly for transgender individuals [32]. Thus, for this study, we are employing four or more drinks to indicate binge drinking.

Socio-demographic factors

Socio-demographic factors included gender identity (male, female, transgender, genderqueer/gender non-conforming, nonbinary, another gender identity), sexual orientation (lesbian, gay, bisexual, asexual spectrum, straight, pansexual, queer, another orientation), race or ethnicity (Alaska Native, American Indian, Asian/Asian American, Biracial/Multiracial, Black/African American, Middle Eastern/North African, Native Hawaiian/Pacific Islander, White/European American, other), age, educational attainment (some high school, high school, some college or vocational school, college or vocational school degree/certificate, graduate, school), health insurance, lost employment during COVID-19, disability status, and neighborhood (urban, suburban, rural, or remote). Participants were able to report multiple answers for their gender identity and sexual orientation.

Statistical analyses

We first analyzed the characteristics and frequency of distribution of the variables. We then tested the differences and associations between independent and dependent variables using chi-squares and t-tests. We analyzed unadjusted and fully adjusted models using logistic regressions to identify correlates of ever use and current use. We also implemented a sensitivity analysis using multinomial logistic regression using three categories of current use: not currently using cigarettes, current cigarette use, and current use of other tobacco and nicotine products. Missing data were handled using listwise deletion. The data were analyzed in Stata 16 SE.

Results

Of the 1629 participants in our sample, 52% reported ever using cigarettes (Table 1). About 13% were currently using cigarettes or other tobacco and nicotine products. The participants primarily identified as cisgender men (63%), gay (61%), White (89%), older (Mage = 59; SD = 10.39), obtained a graduate or professional degree (41%), had health insurance (98%), and resided in a suburban neighborhood (47%). Notably, 35% were disabled and 27% lost their employment during the pandemic. Assessing their health and well-being, 26% experienced psychological distress and close to 3% engaged in binge drinking. Some of the top types of cancers reported by the participants were prostate cancer (n = 322), breast cancer (n = 287), skin cancer (n = 143), colorectal cancer (n = 136), and non-Hodgkin’s lymphoma (n = 116). In addition, 21% (n = 341) of the sample disclosed currently living with cancer.

Table 1 Descriptive statistics for smoking and socio-demographic characteristics of LGBTQI + cancer survivors (N = 1629)
Full size table

In our fully adjusted model (Table 2), age, education, and binge drinking were identified as correlates of ever using cigarettes. Participants who were older (adjusted odds ratio [AOR] = 1.02; 95% confidence intervals [CIs]: 1.01, 1.03) and those who engaged in binge drinking (AOR = 2.47; 95% CI: 1.17, 5.20) were associated with higher odds of ever using cigarettes. Compared to participants with a high school degree or less, those with a graduate or professional degree were associated with lower odds of ever using cigarettes (AOR = 0.40; 95% CI: 0.23, 0.71).

Table 2 Logistic regression of correlates of smoking in the LGBTQI + cancer survivor population (N = 1629)
Full size table

Table 2 highlights the logistic regression analysis for current use of cigarettes and other tobacco and nicotine products, with not currently using as the reference. Gender, race, age, education, binge drinking use, health insurance, and disability were identified as correlates of current use of either cigarette or tobacco and nicotine products. Compared to cisgender men, cisgender women were associated with lower odds of current use (AOR = 0.30; 95% CI: 0.12, 0.77). Latine participants were associated with higher odds of current use than White participants (AOR = 1.89; 95% CI: 1.07, 3.36). For age, younger participants were associated with lower odds of current use than older participants (AOR = 0.98; 95% CI: 0.96, 0.99). For educational attainment, participants with a graduate or professional degree were associated with lower odds of current use than participants with a high school degree or less (AOR = 0.33; 95% CI: 0.15, 0.70). Participants who engaged in binge drinking were associated with higher odds of current use (AOR = 3.18; 95% CI: 1.56, 6.48). In examining health insurance status, compared to those with a health insurance, those without health insurance were associated with higher odds of current use (AOR = 2.37; 95% CI: 1.10, 5.10). Disabled participants were associated with lower odds of current use than non-disabled participants (AOR = 1.64; 95% CI: 1.19, 2.26). The findings from our sensitivity analyses showed similar results, although the 95% confidence intervals were wider compared to the model where we combined all cigarette, other tobacco, and nicotine products into one category (Supplementary Table 1).

Discussion

Over half of our sample reported ever using cigarettes in their lifetime. This finding is consistent with the existing literature of LGBTQI + individuals, including cancer survivors, reporting a higher proportion of smoking than their cisgender and heterosexual counterparts [5, 7, 18]. Notably, while ever use was high (53%), current use was dramatically lower (13%). Cancer survivors may be less likely to currently smoke than people who never had cancer because of increased risk for cancer re-occurrence [19] and being immunocompromised that can be further complicated by contracting COVID-19 [30].

Our findings found several correlates of smoking in LGBTQI + cancer survivors during the COVID-19 pandemic. Compared to cisgender men, cisgender women were less likely to report current use. While cisgender women were less likely to smoke than cisgender men, the prevalence of smoking may change when we consider the intersections of gender and sexual orientation. Lesbian and bisexual women ages 25 years old or older have higher proportions of smoking than gay and bisexual men in the same age range [33]. During the pandemic, data did show that among smokers in the U.S., cisgender women reported higher current cigarette use than cisgender men [25]. Compared to cisgender men, gender diverse individuals had higher odds of current use. Our study has separate categories for transgender and gender diverse individuals, although individuals can identify with multiple gender identities. The sample of gender diverse individuals in most studies is already limited, the few that have been published combine transgender individuals with other gender diverse individuals [34]. Future studies need to consider oversampling gender diverse individuals to further understand differences health behaviors and outcomes in understudied populations. Moreover, there are a limited number of smoking interventions for cancer survivors and virtually none for LGBTQI + populations [19] which underscores a critical need for more smoking intervention studies that centers the need of cancer survivors especially those who identify as LGBTQI + .

Educational attainment, specifically obtaining a graduate degree, was associated with lower odds of ever use and current use. Prior studies have shown the protective effect of having higher educational attainment from likelihood of smoking, though these protective effects have a reduced effect for racial and ethnic minorities [35, 36] and LGBTQI + individuals [37]. Binge drinking was associated with higher odds of ever use and current use. Studies have established the concomitant relationship of smoking and drinking in the LGBTQI + population [38,39,40]. Given the added stressors brought on by the pandemic, smoking [41] and drinking [42] can be a potential co** mechanism in the LGBTQI + population. Having no health insurance and being disabled were also associated with higher odds of current use. The COVID-19 pandemic may have further disadvantaged LGBTQI + cancer survivors with no health insurance and disabled individuals since LGBTQI + are less likely to have health insurance [24] and healthcare providers are less informed how to provide gender-affirming care to LGBTQI + cancer survivors [43] and individuals with a disability [44], adding additional stress to this population.

We also found correlates that are different from findings from previous studies. Age was inversely associated with ever use and current use. Older participants were associated with higher odds of ever use while younger participants were associated with lower odds of current use. However, existing literature suggest that younger LGBTQI + are more likely to currently smoke than older individuals [45], with even a higher proportion who use e-cigarettes [46]. The results from our sensitivity analysis showed that younger participants have lower odds of currently using nicotine and other tobacco products. However, we want to note that this result should be interpreted with caution because we measured age as a continuous variable; therefore, what is considered younger participants refer to middle-aged adults given that the mean age of our sample is 59 years old. Compared to White participants, Latine participants were associated with higher odds of current use which differed from the literature where White LGBTQI + have a higher smoking prevalence than Latine LGBTQI + [6]. Latine LGBTQI + cancer survivors may experience more discriminatory events than their White counterparts due to the intersection of these identities which may lead to higher levels of stress.

Limitations

Our results should be considered with several limitations. While our sample is unique and an understudied population, the diversity in our sample has some limitations. It was comprised predominantly of cisgender men, gay, white, older, and were highly educated, close to 80% have a college degree or higher, suggesting that they are also more likely to have high socioeconomic status (SES). Thus, our findings cannot be generalized to the larger LGBTQI + cancer survivor population. Follow-up studies should consider oversampling underrepresented individuals within the LGBTQI + cancer survivor with a particular attention to people of color, transgender and gender diverse populations, individuals with other sexual orientation (e.g., pansexual and asexual), and low-income individuals who are at higher risk for smoking because of systemic forms of discrimination. The results for current use combined cigarettes and other tobacco and nicotine products may make it difficult to distinguish how the correlates we included in the model were associated with the use of different products. However, we were able to address this issue by implementing a sensitivity analysis that separated these products. Finally, because the data was collected during the COVID-19 pandemic, it is imperative to further assess the impact of the pandemic in this population, specifically the effect of stress and its relationship with smoking, given that stress is a known predictor of smoking. While we included psychological distress in the models, there may be other types of measures that can better capture the concept of stress. We encourage future studies to examine the influence of stress longitudinally during the pandemic and its associations with smoking in this population.

Conclusion

This study extends the literature in our understanding of the correlates of ever use and current use of cigarettes and other tobacco and nicotine products in the LGBTQI + cancer survivor population during the pandemic. The correlates we identified showed how access to certain institutions such as higher education and having health insurance can be protective factors from smoking. However, certain marginalized identities like being a person of color and being disabled make LGBTQI + cancer survivors more vulnerable to smoking because of discriminatory institutions that they have to navigate that were further heightened by the pandemic. Although the pandemic disrupted access to healthcare programs and services in general, we encourage public health researchers and practitioners to implement proactive responses to smoking cessation programs and services that can be more accessible, affordable, de-stigmatized, and culturally responsive to LGBTQI + cancer survivors that are effective with or without a pandemic. Potential solutions may include the use of telehealth and mobile health (mHealth) technology and mandated on-going training among healthcare providers to provide gender-affirming care to close these gaps.