Chronic Health Conditions, Disability, and Physical and Cognitive Limitations among LGBTQ+ Cancer Survivors

Improvements in the early detection and treatment of cancer have reduced cancer-related mortality and rapidly increased the number of cancer survivors in the United States. As of 2022, there were approximately 18.1 million cancer survivors in the United States (1). Although effective at prolonging life, cancer treatments can exacerbate existing health conditions and cause new long-term health effects that impact survivors’ physical and psychological health. Chronic diseases such as diabetes, kidney disease, liver disease, chronic obstructive pulmonary disease (COPD), and heart disease may be more common among cancer survivors than among individuals without a cancer diagnosis (2).

Currently, Lesbian, Gay, Bisexual, Transgender, Queer, Plus (LGBTQ+) populations represent at least 7.1% of the US population and are likely to experience health inequities related to LGBTQ+ stigma, discrimination, and other social determinants of health (3, 4). In a recent systematic review, gay and bisexual men were found to be more likely to have asthma and chronic kidney disease than heterosexual men (5). Further, one secondary analysis of the National Health Interview Survey revealed that among women, lesbian and bisexual cancer survivors were more likely to have heart conditions, COPD, and psychological distress than non-lesbian or bisexual cancer survivors (6). These health differences may be due to experiences of stress related to LGBTQ+ stigma in health care settings and general life, as well as differences in health behaviors such as smoking (7, 8). Few national samples have been used to evaluate LGBTQ+ disparities among cancer survivors about prevalence of chronic health conditions, disability, and physical and cognitive limitations, which includes both cisgender and transgender and other gender expansive identities.

The purpose of our study was to use Behavioral Risk Factor Surveillance System (BRFSS) data to assess differences in the age-adjusted prevalence of chronic health conditions, disabilities, as well as physical and cognitive limitations reported by LGBTQ+ and non-LGBTQ+ cancer survivors. Further, we aimed to understand the differences in burden of chronic health conditions and physical and cognitive limitations by sexual minority (lesbian, gay, bisexual; LGB) and gender minority (transgender and gender nonconforming; TGNC) status. Estimating LGBTQ+ disparities in chronic health conditions, disabilities, and limitations among cancer survivors is crucial to the development of future equity-based preventive health interventions.

The BRFSS, managed by the Centers for Disease Control and Prevention (CDC), is an ongoing nationwide telephone survey system that has collected data on health-related behaviors, health status, and healthcare access since 1984 [CDC. BRFSS. (cited May 21, 2024). Available from: https://www.cdc.gov/brfss/]. Eligible BRFSS participants are noninstitutionalized adults aged 18 years or older residing in the United States and were consented during original data collection by the CDC. Our study population for secondary analysis included BRFSS participants from the 23 states that administered both the sexual orientation and gender identity module and the cancer survivorship module in 2020, 2021, or 2022. This study was deemed exempt by the University of North Carolina Institutional Review Board, which reviews in accordance to the Declaration of Helsinki.

The primary outcomes of our analysis were chronic health conditions (angina or heart disease, asthma, COPD, depressive disorder, heart disease, kidney disease, stroke, and diabetes), disabilities, and physical and cognitive limitations (hearing disability, vision disability, cognitive limitation, difficulty walking, difficulty dressing, and difficulty running errands). Chronic health condition outcomes were categorized as chronic health conditions in alignment with the CDC’s definition of chronic disease [CDC. (cited May 21, 2024). Available from: https://www.cdc.gov/chronic-disease/?CDC_AAref_Val=https://www.cdc.gov/chronicdisease/about/index.html]. Participants who answered “Yes” to the following questions were considered as having the above conditions, disabilities, and limitations: “(i) (Ever told; you had) angina or coronary heart disease?” “(ii) (Ever told; you had) asthma?” “(iii) (Ever told; you had) COPD, emphysema or chronic bronchitis?” “(iv) (Ever told; you had) a depressive disorder (including depression, major depression, dysthymia, or minor depression)?” “(v) (Ever told; you had) a had a heart attack, also called a myocardial infarction?” “(vi) Not including kidney stones, bladder infection or incontinence, were you ever told you had kidney disease?” “(vii) (Ever told; you had) a stroke?” “(viii) (Ever told; you had) diabetes?” “(ix) Are you deaf, or do you have serious difficulty hearing?” “(x) Are you blind or do you have serious difficulty seeing, even when wearing glasses” “(xi) Because of a physical, mental, or emotional condition, do you have serious difficulty concentrating, remembering, or making decisions?” “(xii) Do you have serious difficulty walking or climbing stairs?” “(xiii) Do you have difficulty dressing or bathing?” “(xiv) Because of a physical, mental, or emotional condition, do you have difficulty doing errands alone such as visiting a doctor’s office or shopping?”.

LGBTQ+ Identity, our primary exposure of interest, was determined using self-reported responses to the following questions: “(i) Sexual Orientation: Which of the following best represents how you think of yourself?: Straight, Lesbian or Gay, Bisexual, Something else” “(ii) Gender Identity: Do you consider yourself to be transgender? (If yes, ask “Do you consider yourself to be male-to-female, female-to-male, or gender nonconforming?).” People who identified themselves as “Lesbian or Gay,” “Bisexual,” or “Something else,” for Question (i) or “Yes, Transgender, male-to-female,” “Yes, Transgender, female-to-male,” or “Yes, Transgender, gender nonconforming” for Question (ii) were combined as LGBTQ+ in our study. If individuals selected “Lesbian or Gay,” “Bisexual,” or “Something else,” for Question (i) but no to “Do you consider yourself to be transgender?” were considered cisgender and sexual minority (LGB) in our study. Responding with “something else” to the sexual orientation question was interpreted as another non-heterosexual sexual orientation such as asexual, pansexual, or queer. Based on National Academies report on measuring SOGI, there is little evidence that suggests heterosexual individuals select “something else” when asked about sexual orientation (9). Furthermore, regardless of sexual orientation, if individuals selected “Yes, Transgender, male-to-female,” “Yes, Transgender, female-to-male,” or “Yes, Transgender, gender nonconforming” for Question (ii) they were considered a gender minority (TGNC) in our study. Cancer survivor status was ascertained using one question from the cancer survivorship module that asked participants “How many different types of cancer have you had?” Any response of one or more were considered cancer survivors. Other self-reported sociodemographic factors, including age at interview, race and ethnicity, household income, education level, and smoking status, were used as covariates in our analysis. Covariates were selected based on existing literature.

Sociodemographic characteristics included in this analysis were race and ethnicity, age at interview, household income, education, age at cancer diagnosis, diagnosis, and smoking status. Race and ethnicity were collapsed to non-Hispanic White, non-Hispanic person of color, and Hispanic due to the limited sample size. Age was collapsed into three age groups based on the developmental literature including 18 to 39, 40 to 64, and 65 and older. Education was collapsed to high school graduate or less, attended college or technical school, and graduated from college or technical school because of the small sample size of participants who did not complete high school education. Finally, smoking status was collapsed to current smokers, former smokers, and never smokers.

Descriptive statistics were used to characterize the survey sample by LGBTQ+ status (Table 1). All analyses were weighted using weights inherent to the survey sampling methods of BRFSS (CDC. BRFSS. [cited May 21, 2024]. Available from: https://www.cdc.gov/brfss/). The χ² test and Wilcoxon rank-sum test were conducted to describe the differences in the distribution of covariates and outcomes between different LGBTQ+ subgroups. Weighted age-adjusted prevalence of chronic health conditions and disabilities and limitations was standardized using 2021 US Census data [Table 2; United States Census Bureau. Annual Estimates for the Resident Population for Selected Age Groups by Sex for the United States: (cited May 21, 2024) Available from: https://www.census.gov/data/tables/time-series/demo/popest/2020s-national-detail.html]. Two-proportion z-test was used to test the difference by LGBTQ+ status. Statistical tests were not run to detect differences between LGB and TGNC prevalence estimates due to being underpowered. Weighted multivariate logistic regression models were used to calculate adjusted ORs (aOR) and corresponding 95% confidence intervals (CI) to assess the associations between LGBTQ+ status and each chronic health condition, disability, and limitation (Table 3), as well as associations between LGB status and each outcome (Table 4) and TGNC status and each outcome (Table 5). Logistic regression models and aORs were used due to the rarity of some outcomes and to maximize comparability to the existing literature. Four models were run for each outcome and each group (LGBTQ+, LGB, TGNC), controlling for (i) age; (ii) age, race, and ethnicity; (iii) age, race, and ethnicity, and smoking status; and (iv) age, race, and ethnicity, smoking status, education, and household income. All models included year and state fixed effects. Age was included in all multivariable as a potential confounder to reflect generational shifts in reporting of LGBTQ+ status and the association between age and the outcomes of interest. Race and ethnicity were considered in subsequent models to improve precision of outcome associations. Smoking, education, and household income potentially lie on the pathway from LGBTQ+ identity to chronic conditions and therefore were considered in separate models to evaluate the potential impact of statistical adjustment for these factors.

Two sensitivity analyses were conducted. First, we re-ran LGBTQ+ prevalence estimates and LGBTQ+ multivariable models, while our sample was restricted to the five states that collected both the SOGI and cancer survivorship BRFSS modules across all 3 years (Hawaii, New Jersey, New Mexico, Utah, and Wisconsin). This analysis was conducted to assess potential between state effects as the SOGI and cancer survivorship modules were not consistently collected by the 23 states in the main analysis. Second, we re-ran LGBTQ+ prevalence estimates and LGBTQ+ multivariable models, while excluding survivors from the sample who reported a skin cancer diagnosis. Statistical significance was set at P < 0.05; analyses do not adjust for multiple testing (10, 11). All analyses were conducted using R, version 4.2.1.

The BRFSS data are publicly available on the CDC website.

From 2020 to 2022, 4.2% of 40,990 cancer survivors in our BRFSS analysis were identified as LGBTQ+ (n = 1,715). A total of 638 (1.5%) were identified as lesbian or gay, 551 (1.6%) as bisexual, 458 (1.2%) as another nonheterosexual sexual orientation (total LGB = 1,601, 3.9%). Among the 114 (0.3%) TGNC survivors, 38 (0.1%) identified as transgender men, 43 (0.1%) identified as transgender women, and 33 (0.1%) identified as gender nonconforming. LGBTQ+ cancer survivors were significantly more racially and ethnically diverse, had lower household income, younger age at survey, and younger age at diagnosis than non-LGBTQ+ cancer survivors (Table 1). LGBTQ+ cancer survivors were also more likely to report being current smokers than non-LGBTQ+ cancer survivors. The most common diagnoses among LGBTQ+ cancer survivors were skin cancer, melanoma, other cancers, breast cancer, whereas the most common diagnoses among non-LGBTQ+ cancer survivors were skin cancer, melanoma, breast cancer, and prostate cancer. State, year, and education did not significantly differ by LGBTQ+ status.

When adjusting for age at interview in Table 2, LGBTQ+ cancer survivors had significantly higher prevalence of all chronic health conditions, disabilities, and limitations included in this analysis. The highest prevalence estimates among LGBTQ+ cancer survivors included asthma (30.7%; 95% CI, 23.8–37.5; Table 2), depressive disorders (48.5%; 95% CI, 40.3–56.8), cognitive limitations (37.9%; 95% CI, 29.2–46.5), and difficulty walking (31.4%; 95% CI, 24.7–38.1). The largest differences in age-adjusted prevalence estimates were in cognitive limitations (37.9% vs. 15.9%; P < 0.001), depressive disorders (48.5% vs. 29.5%; P < 0.001), difficulty walking (31.4% vs. 18.5%; P < 0.001), and difficulty running errands (23.7% vs. 11.6%; P < 0.001). Smaller but still significant differences were observed in COPD, hearing disabilities, heart disease, as well as other conditions and limitations (Table 2).

When age-adjusted prevalence estimates for each outcome were generated for LGB and TGNC population separately, both LGB and TGNC estimates for each outcome remained higher than non-LGBTQ+ prevalence estimates (Table 2). For most outcomes TGNC survivors had higher age-adjusted prevalence than cisgender LGB survivors including asthma (38.8% vs. 29.5%), diabetes (42.9% vs. 17.5%), vision disabilities (30.1% vs. 11.5%), difficulty dressing (33.4% vs. 11.6%), and difficulty running errands (34.3% vs. 21.1%). However, cisgender LGB cancer survivors had higher prevalence of depressive disorders (49.9% vs. 34.3%) and cognitive limitations (35.7% vs. 21.9%) than TGNC survivors.

In multivariable models controlling for age, survey year, and state (Table 3, A models); models controlling for age, survey year, state, and race and ethnicity (Table 3, B models); and models control age, survey year, state, race and ethnicity, and smoking status (Table 3, C models), LGBTQ+ cancer survivors were at higher odds of reporting nearly every outcome. The addition of race and ethnicity and smoking status in B and C models resulted in small decreases in the magnitude of association between the LGBTQ+ status and each outcome (Table 3). In final models controlling for all previous factors, as well as socioeconomic, factors such as education and household income (D models) resulted in LGBTQ+ cancer survivors being at higher odds than non-LGBTQ+ survivors of reporting asthma (aOR: 1.54; 95% CI, 1.24–1.91), depressive disorders (aOR: 1.94; 95% CI, 1.94–1.6–2.36), kidney disease (aOR: 1.54; 95% CI, 1.12–2.11), stroke (aOR: 1.71; 95% CI, 1.27–2.31), diabetes (aOR: 1.30; 95% CI, 1.02–1.64), vision disability (aOR: 1.61; 95% CI, 1.20–2.18), cognition limitation (aOR: 2.30; 95% CI, 1.81–2.91), difficulty walking (aOR: 1.65; 95% CI, 1.33–2.04), difficulty dressing (aOR: 2.02; 95% CI, 1.48–2.74), and difficulty running errands (aOR: 1.63; 95% CI, 1.25–2.13).

When multivariable models were run comparing LGB survivors to non-LGB survivors, LGB specific models resulted in nearly identical results to those for the primary LGBTQ+ models (Table 4). However, when multivariable models were run to compare TGNC survivors to non-TGNC (i.e., cisgender of any sexual orientation) survivors, estimate magnitudes increased substantially across most outcomes (Table 5). In multivariable models controlling for age, survey year, state, race and ethnicity, smoking status, education, and household income, TGNC survivors were at a higher odds than cisgender survivors of reporting heart disease (aOR: 3.67; 95% CI, 1.27–10.63), asthma (aOR: 2.34; 95% CI, 1.21–4.55), COPD (aOR: 3.07; 95% CI, 1.53–6.17), myocardial infarction (aOR: 5.70; 95% CI, 2.53–12.86), kidney disease (aOR: 6.03; 95% CI, 2.73–13.34), stroke (aOR: 4.39; 95% CI, 1.63–11.83), diabetes (aOR: 3.85; 95% CI, 1.69–8.75), hearing disabilities (aOR: 3.05; 95% CI, 1.11–8.37), vision disabilities (aOR: 3.68; 95% CI, 1.80–7.52), cognitive limitations (aOR: 4.37; 95% CI, 2.33–8.20), difficulty walking (aOR: 2.66; 95% CI, 1.49–4.72), difficulty dressing (aOR: 4.37; 95% CI, 2.08–9.21), and difficulty running errands (aOR: 3.08; 95% CI, 1.43–6.65).

In sensitivity analyses restricting analyses to states that had consistent SOGI and cancer survivorship module data collection 2020 to 2022, LGBTQ+ cancer survivors prevalence estimates for outcomes and aORs were similar (Supplementary Tables S1 and S2). Further, in sensitivity analyses excluding skin cancer survivors, LGBTQ+ cancer survivor prevalence estimates for outcomes and aORs were similar to main primary analysis estimates (Supplementary Tables S3 and S4).

In our analysis, LGBTQ+ cancer survivors had a significantly higher age-adjusted prevalence of all chronic health conditions, disabilities, and limitations than non-LGBTQ+ cancer survivors. Further, when comparing cisgender LGB and TGNC populations, transgender and gender nonconforming survivors (of any sexual orientation) had higher age-adjusted prevalence estimates for most outcomes except for depression and cognitive limitations in which cisgender lesbian, gay, and bisexual survivors had higher prevalence. Furthermore, after controlling for risk factors, such as smoking and socioeconomic status, LGBTQ+ cancer survivors remained at higher odds of reporting asthma, depressive disorders, kidney disease, stroke, diabetes, cognitive limitations, as well as difficulty walking, dressing, and running errands. These findings align with other recent studies indicating that LGBTQ+ cancer survivors, as well as the general LGBTQ+ population, are at an elevated risk of chronic health conditions (12–14).

However, our findings demonstrate higher prevalence estimates of many chronic conditions in LGBTQ+ cancer survivors in comparison to LGBTQ+ individuals without cancer and the overall general population from the literature including individuals with asthma, stroke, kidney disease, and diabetes (13, 15–21). These findings can be contextualized, for example, in the LGBTQ+ smoking cessation and stroke risk factor literature, which highlights upstream factors that drive smoking and stroke in LGBTQ+ populations, including structural forces such as identity-related stigma and minority stress (8, 22). Such structural forces have also been observed to drive higher rates of behaviors such as smoking and alcohol use—risk factors for many of the chronic conditions in this analysis, as well as many cancers (23, 24). Future inquiry would benefit from further disentangling the influence of structural forces, health behaviors, and chronic health conditions in order to tailor interventions to the LGBTQ+ cancer survivor population.

Findings from our study demonstrate a higher prevalence of difficulty with activities of daily living (ADLs, e.g., difficulty walking) and instrumental ADLs (IADLS, e.g., running errands) in LGBTQ+ cancer survivors than non-LGBTQ+ cancer survivors. Specifically, TGNC survivors had greatest prevalence of difficulty with ADLs and IADLs than LGB cancer survivors. These findings build upon recent literature that demonstrated that older LGB adults (aged > 50 years old) had a higher prevalence of difficulty with ADLs and IADLs (35.9%) than non-LGB populations (20.9%; ref. 25). Integration of prehabilitation (i.e., preparation of the patient to improve their physical functioning prior to cancer treatment) tailored to the LGBTQ+ community may be particularly important for LGBTQ+ cancer survivors (26). Barriers to health care utilization (e.g., worry or avoidance of health care settings because of discrimination or cost) may contribute to associations observed with ADLs and IADLs (27, 28).

Further our findings are consistent with previous literature that found that TGNC cancer survivors were at elevated odds of some of the conditions that we report in this study, including diabetes (29). However, our findings provide a more recent and granular look at increased prevalence of nearly all chronic health conditions, disabilities, and limitations among transgender and gender nonconforming cancer survivors in the United States. These findings can be contextualized in the literature that demonstrated that transgender individuals in the general population report more chronic conditions than their cisgender counterparts (30, 31). Transgender and gender diverse populations also have higher prevalence of risk factors for chronic health conditions such as smoking and heavy alcohol use (32). Our findings and the existing literature can all be contextualized in cissexism and transphobia that effects every level of society from anti-transgender legislation to healthcare discrimination to minority stress and ultimately poor health outcomes (33–35).

In our sample, half of LGBTQ+ cancer survivors reported a depressive disorder in their lifetime, reinforcing the need for LGBTQ+ specific cancer-related mental health services. Recent literature and the National LGBT Cancer Network reported that LGBTQ+ cancer survivors experience a high number of mentally unhealthy days (8.9 days per month), as well as 70% of LGBTQ+ survivors who had never had access to LGBTQ+ tailored mental health resources (36, 37). Depressive disorders among LGBTQ+ cancer survivors may also be intrinsically linked to cognitive limitations, as depression and other psychological factors are predictive of cognitive function in cancer survivors (38). Our findings are consistent with a higher prevalence of cognitive limitations in LGBTQ+ populations in general (39). Interventions to address risk factors for cognitive limitations such as stress and trauma, smoking, financial hardship, poor exercise and diet, and inflammation among LGBTQ+ cancer survivors may have benefit for cognitive function after cancer treatment (38).

Our findings must be considered within the context of the limitations of our data. While self-reporting is the preferred modality for ascertaining LGBTQ+ status, self-reported cancer status and chronic conditions may be subject to potential recall errors. Furthermore, the available smoking status variable lacks the specificity to assess pack–years, which may have further explained the findings. Our study is limited by a lack of cancer treatment data available, as specific treatments are associated with conditions later in life. Less than 5% of respondents indicated an LGBTQ+ identity; in national samples from the Pew Research Center, this percentage exceeded 7% and may suggest that participants did not feel comfortable disclosing this information or that LGBTQ+ individuals were less likely to respond to BRFSS (3). By using a cross-sectional study design, cancer survivors additionally needed to survive and be well enough to respond to the BRFSS survey. This requirement may have resulted in the selection of healthier cancer survivors with less severe disease or treatment in our sample. Cancer survivorship and sexual and gender identity (SOGI) modules are optional for states to include and may not reflect the experiences of cancer survivors in states that did not elect to include cancer survivorship or the SOGI module.

In our analysis, LGBTQ+ cancer survivors had a significantly higher age-adjusted prevalence of most chronic health conditions, disabilities, and physical and cognitive limitations. When controlling for various sociodemographic factors, LGBTQ+ status remained significantly associated with asthma, depressive disorders, kidney disease, stroke, diabetes, vision disabilities, cognitive limitations, as well as ADL and IADLs including difficulty walking, dressing, and running errands. Our findings highlight major disparities in the health of LGBTQ+ cancer survivors—with higher prevalence of most outcomes among transgender and gender nonconforming cancer survivors. Our findings reinforce the need for multilevel interventional approaches, from structural changes to patient-level to mitigate health disparities experienced by LGBTQ+ cancer survivors.

No disclosures were reported.

A.R. Waters: Conceptualization, data curation, investigation, visualization, methodology, writing–original draft, project administration, writing–review and editing. M. Jin: Data curation, formal analysis, methodology, writing–original draft, writing–review and editing. S.R. Jones: Conceptualization, investigation, writing–review and editing. G.D. Datta: Conceptualization, investigation, methodology, writing–review and editing. E.N. Butler: Conceptualization, investigation, methodology, writing–review and editing. E.E. Kent: Conceptualization, supervision, investigation, methodology, writing–review and editing. H.B. Nichols: Conceptualization, resources, supervision, investigation, methodology, writing–review and editing. K. Tan: Conceptualization, resources, supervision, investigation, methodology, writing–original draft, writing–review and editing.

Research reported in this publication was supported by the NCI of the NIH. A.R. Waters is supported by the NCI’s National Research Service Award sponsored by the Lineberger Comprehensive Cancer Center at the University of North Carolina (T32 CA116339). H.B. Nichols is supported by the NCI of the NIH (P01CA233432).