Cardiovascular disease (CVD) is the second leading cause of death among Hispanic/Latino adults [1]. This population bears a heavy burden of obesity, diabetes, poorly controlled hypertension, and other cardiovascular risk factors [2]. Recent studies have identified heterogeneity in cardiovascular risk among Hispanic/Latino adults by heritage group, gender, acculturation, and duration of US residency [3,4]. However, variability in cardiovascular risk factors has not been systematically examined by sexual orientation or gender identity.

The inclusion of sexual orientation measures in national, population-based surveys such the National Health Interview Survey, the National Health and Nutrition Examination Survey, and the Behavioral Risk Factor Surveillance Survey (BRFSS) have provided data on the elevated prevalence of cardiovascular risk factors among sexual minorities, including smoking [5-7], obesity [6-8], high blood pressure [6], and glycosylated hemoglobin [6]. An analysis of data from the National Longitudinal Study of Adolescent to Adult Health identified sexual orientation inequities in cardiovascular risk behaviors (eg, smoking), as well as clinical measures (eg, blood pressure) and biomarkers (eg, C-reactive protein), beginning in young adulthood [9].

The BRFSS is the only national population-based survey to have published data that includes a module to assess gender identity. BRFSS data indicate transgender adults experience disparities in weight [7], smoking [10], and myocardial infarction [11]. The aforementioned studies have also identified variations in cardiovascular risk among sexual and gender minorities (SGM) by sex assigned at birth (eg, male, female), gender identity (eg, transgender woman, cisgender woman), and sexual identity (eg, gay, bisexual) [5,6,8]. However, the small number of ethnic minority SGM in these samples have precluded specific analyses within Hispanic/Latino SGM, even though Hispanic/Latino SGM are the fastest growing ethnic group of SGM in the United States [12].

Psychosocial stressors play an important role in CVD risk [13]. A global investigation of the determinants of CVD outcomes, in 52 countries over 7 continents, found that psychosocial stress was a powerful predictor of myocardial infarction, comparable in impact to smoking [14,15]. Baseline data from the Hispanic Community Health Study/Study of Latinos (HCHS/SOL) found that a composite score for psychological distress, including depressive symptomatology and trait anxiety, was significantly associated with obesity for women, diabetes for men, and smoking for all participants [16].

Stigma and discrimination are common psychosocial stressors for minority groups and are associated with a range of negative health outcomes, including increased risk for CVD [17-21]. A recent nationally representative sample of 489 SGM reported prevalent harassment or threats (57%) and violence (51%) because of their sexual orientation or gender identity [22]. The HCHS/SOL Sociocultural Ancillary Study found that most Hispanics/Latinos (80%) reported exposure to ethnic discrimination during their lifetime [23]. While exposure to ethnic discrimination varied based on education, income, and acculturation, lack of sexual orientation and gender identity measures in HCHS/SOL precluded examination by SGM status. Other research suggests that Hispanic/Latino SGM experience the unique stressors of both racism and ethnocentrism within SGM communities and rejection of their sexual orientation and/or gender identity by their Hispanic/Latino families and communities [24].

Resilient coping and social support may mitigate the impact of psychosocial stressors on CVD risk. Hispanic/Latino cultural values are hypothesized to engender strong social supports that buffer health risks [25-27]. For example, lower acculturation and foreign-born nativity, often used as proxies for stronger cultural values, are associated with lower CVD prevalence [28]. However, SGM Hispanic/Latino individuals may lose access to these cultural buffers if they are rejected by their families or communities, or forced to conceal their sexual orientation or gender identity [24]. Connection to SGM community may mitigate minority stress for Hispanic/Latino SGM [29,30]. However, whether this connection reduces CVD risk for Hispanic/Latino SGM has not been studied.

This study, Stress Gender and Minority Status in the Study of Latinos (SGM SOL), aims to advance scientific knowledge of how stigma at the intersection of ethnic identity, sexual orientation, and gender identity impacts CVD outcomes for Hispanic/Latino SGM. It draws on two complementary conceptual frameworks: intersectionality theory and the minority stress model. Intersectionality addresses how the stress of holding multiple stigmatized identities, such as being a racial/ethnic minority, SGM, and/or an immigrant, may compound inequities [31,32]. Intersectionality theory guided our decision to measure multiple social identities including sexual orientation and gender identity, socially ascribed race, and immigration background. In keeping with a fundamental tenet of intersectionality theory—that social identities are mutually constituted—we plan to use adapted measures of stigma, discrimination, and social stress that are inclusive of multiple simultaneous identities without requiring participants to attribute experiences to a particular identity [33]. The minority stress model posits that stigma and discrimination create excess minority stress for SGM populations that result in health inequities, and that these inequities can be mitigated by resilience factors, such as positive coping and social support [34-36].

SGM SOL aims to assess relationships among CVD risk and stigma, discrimination, stress, and coping and social support by comparing SGM with non-SGM Hispanic/Latino adults, thereby testing relationships theorized by the minority stress model, as depicted in Figure 1.

The primary aims of this study are to (1) examine the influence of sexual orientation or gender identity on CVD risk during visit 3 of HCHS/SOL (N~9300); (2) model pathways from sexual orientation or gender identity to CVD risk through stigma, discrimination, and stress in an HCHS/SOL subsample (n~1680); and (3) examine the influence of resilience factors on sexual orientation or gender identity and CVD risk relationships in the HCHS/SOL subsample (n~1680).

SGM SOL builds on the infrastructure of HCHS/SOL, a multicenter, longitudinal, observational cohort study designed to evaluate prevalence, incidence, and risk and protective factors for cardiovascular disease and other chronic conditions among Hispanic/Latino adults in the United States. It is the largest well-characterized longitudinal cohort study of diverse self-identified US Hispanic/Latinos with 16,415 adults (aged 18-74 years) enrolled at baseline. Using a 2-stage probability sample, households and participants were randomly selected using stratification and oversampling at each stage in 4 US sites—the Bronx, NY; Chicago, IL; Miami, FL; and San Diego, CA [37]. Full details about the study design have been previously published [37,38].

SGM SOL is a longitudinal observational cohort study of HCHS/SOL adults who will participate in visit 3. HCHS/SOL collected in-person data in 2008-2011 (visit 1), and 81% of eligible participants (n=11,623) returned for visit 2 (2014-2017). We expect 80% of the visit 2 participants to return for visit 3. Thus, we project approximately 9300 visit 3 participants. Based on national data, which indicate that 6.1% of the Hispanic/Latino population in the United States identify as SGM [39], we anticipate that 560 HCHS/SOL participants will meet at least one of the SGM inclusion criteria.

As shown in Figure 2, the first aim will be addressed using the entire HCHS/SOL visit 3 cohort (N~9300) who complete the screening measures related to sexual orientation or gender identity. The second and third aims will be addressed by enrolling 560 SGM and 1120 non-SGM who are matched 2:1 to the SGM participants by site, age, and sex assigned at birth (n~1680). All study procedures have been approved by the central institutional review board (IRB) at the University of North Carolina Chapel Hill and by the local IRBs of all participating institutions.

All new SGM SOL study measures are listed in Table 1, organized by visit and topic area. We also list related measures for visit 3 or previously collected at visits 1 or 2. All HCHS/SOL visit 3 participants will complete the measures listed in Table 1 in the visit 3 column as well as the 4 SGM screening questions. The SGM SOL study participants will complete the additional measures listed in Table 1 in the new measures column. Except where noted below, all study instruments have existing, validated Spanish translations. In the few cases where translations were not available, items were translated into Spanish by culturally knowledgeable bilingual staff, then back-translated by different culturally knowledgeable bilingual staff members to ensure accuracy [40]. Below, we describe each of the main measures listed in the conceptual framework.

Each site recruits, consents, and collects data for approximately 420 participants (140 SGM and 280 non-SGM) for the SGM SOL ancillary study. Data collected at the sites and entered into the web-based data management system are identified by participant and staff ID numbers. Measures taken to ensure the security of the data include, but are not limited to requiring valid IDs and passwords to access the web-based data management system, using a firewall and user logins to shield the local area network from web users, and using secure sockets layer standard to provide encryption and user authentication. The data management system includes procedures to support efficient, high-quality data acquisition, tracking, and quality assurance. When there is active data collection, the Coordinating Center prepares monthly reports to track enrollment and data quality.

Site personnel completed central training on sexual and gender minority cultural awareness, SGM SOL research protocols, and data collection instruments. Since collection of data on sexual orientation and gender identity was new for HCHS/SOL staff, we made every effort to raise cultural awareness and ensure comfort with the questions. Site principal investigators and all site personnel were invited to an SGM cultural awareness training that included a discussion of concepts and terminology related to sexual orientation or gender identity. The interactive training included opportunities to match terms with definitions, discuss relevant case scenarios, and ask questions in a judgement-free confidential manner. More than 50 personnel completed this training. This training was recorded and made available on the study website for future refresher training or to train new staff. Subsequently, personnel engaged in data collection or supervision completed study-specific training to increase familiarity with the questions, model appropriate ways to respond to potential participant questions, and provide opportunities to practice.

The aim 1 analysis will include the entire visit 3 sample (N~9300) with approximately 560 meeting our criteria for SGM. With the 1:2 match, 560 SGM and 1120 non-SGM participants will be available for analyses of aims 2 and 3. To ensure adequate power, we have conducted power analyses for aims 2 and 3 using a smaller sample size of 1125 (375 SGM and 750 non-SGM). Previous HCHS/SOL ancillary studies have successfully recruited >70% of eligible participants and the lower bound for our sample of SGM participants (n=375) represents 67% of likely SGM participants in HCHS/SOL visit 3.

For aim 1, we assume a total sample size of 9300 with 375 SGM participants (most conservative), power=0.8, variance inflation factor=2, and α=.05. For stratified analysis, we focus on our smallest subgroup (males) who comprise 36% of the sample. Based on these conservative assumptions, the power in both overall and stratified analyses is over 0.8 to detect prevalence ratios as small as 1.2 given a CVD risk factor prevalence ranging from 35% for the MetS to 63% for the PCE and as low as 20% for some age-sex subgroups [2,3]. For stratified analyses, the minimum detectable mean is 0.35 SD among males. The power is over 0.8 to test for each outcome (PCE, MetS, CRM) separately whether the mean is different by SGM and whether the patterns of change over time are the same.

For aim 2, power analysis was performed in Mplus using Monte Carlo simulation (10,000 replicates) to test separately the mediated effect of stigma, discrimination, and stress scores between SGM and CVD risk factor z-scores, given a conservative sample size of 1125, α=.05, and different effect sizes of direct paths based on Cohen’s guidelines for R² (amount of explained variance in the outcome) [82]. We defined small, medium, and large effect sizes as 0.02, 0.13, and 0.26, given that we do not have estimates of the size of the indirect effects from the literature. We assumed the following parameters based on the literature [23,58,82]: minority stress (mean 2.08, SD 6.6 in Hispanics), discrimination (mean 1.5, SD 0.1), correlations between the SGM and CVD risk factors z-score ranging from 0.1 to 0.4, and residual variances of 0.95. The power is at least 0.9 for each effect (total, total indirect, and direct) from SGM to the CVD risk factors z-scores.

For aim 3, we performed a Monte Carlo simulation study in SAS to determine the minimum interaction term (in SDs) between high social support (ISEL-12 >17) and minority stress (continuous) on CVD risk factors z-scores assuming a conservative sample size of 1125, power=0.8, and α=.05. We generated 1000 samples and assumed the following parameters based on the literature [13,58,63]: minority stress (mean 2.08, SD 6.6 in Hispanics) and social support (range 0-36, mean ISEL-12 total 25.9, SD 6.6). The power is greater than 0.8 for an interaction term as small as 0.3 SD. Similarly, for resilient coping (range 4-20, mean 14.81, SD 2.95), the power is greater than 0.8 for an interaction term between high resilient coping and minority stress as small as 0.2 SD.

Despite the heavy impact of CVD in Hispanic/Latino communities and the data that Hispanic/Latino populations are more likely to report being SGM, remarkably little research has focused on CVD in the Hispanic/Latino SGM population [83]. Leveraging the infrastructure of the existing HCHS/SOL cohort, SGM SOL will examine relationships among stigma, discrimination, minority stress, and CVD in this population [84]. Specifically, SGM SOL will test whether there are significant associations between minority stress and CVD and whether coping and social support mitigate the negative impact of minority stress, taking into consideration acculturation. SGM SOL will contribute to knowledge about the effect of psychosocial factors on cardiovascular health and advance scientific knowledge on how intersectional stigma and discrimination become embodied as health inequities.

SGM SOL is the first CVD study of this size in the fastest growing SGM ethnic group in the United States. While most existing studies of minority stress and CVD use cross-sectional designs [85], SGM SOL captures cross-sectional and retrospective longitudinal data, strengthening the ability to make causal inferences about the nature of stress-health relationships and the impact of coping and social support on these outcomes, providing key data for future interventions. Implementation of SGM SOL provides a template for low burden measurement of SGM status in other population-based studies as well as strategies for ensuring data collection staff are comfortable collecting data with multiply marginalized populations.

Understanding the influence of stigma-induced stress as well as resilient coping on CVD risk among Hispanic/Latino SGM adults has important implications for efforts aimed at improving health in this growing minority population. These advances in our understanding are key to our ability to identify ways clinical providers can more effectively tailor their care to meet the needs of Latino/Hispanic SGM populations. Additionally, results can be used to develop family and community interventions that reduce SGM-related stigma and build on identified Hispanic/Latino cultural strengths by adapting and testing family acceptance strategies [86] that have been only been tested among youth to date.

It is possible that we will have fewer SGM participants than anticipated; therefore, we powered the study based on highly conservative estimates of the sample size. Maximizing sample size and power, aim 1 utilizes the full visit 3 HCHS/SOL cohort and will allow for exploratory analyses among specific SGM groups if the size of the SGM population is sufficiently large. Although HCHS/SOL is an excellent platform for a population-based study of Hispanic/Latino SGM health, the number of questions on sexual orientation and gender identity we could include for the entire cohort were limited by the need to reduce the burden on participants who are already completing 120 minutes of core data collection activities as part of visit 3. However, we will have more detailed information on sexual orientation and gender identity for all participants in the subsample. Hypothesized mediators (ie, stigma, discrimination, and minority stress) will be measured for the first time at visit 3, limiting the ability to assess temporality. However, CVD risk measures will be available for all 3 visits and measures of chronic stress for 2 visits. By assessing age at which SGM participants identified their current sexual orientation and gender identity, we will be able to assess temporality for relationships between SGM status, chronic stress, and CVD outcomes.