Lead/Presenter: Meredith Duncan, VA Tennessee Valley Healthcare System, University of Kentucky
All Authors: Duncan MS (VA Tennessee Valley Healthcare System, University of Kentucky), Streed CG (Boston University School of Medicine, Center for Transgender Medicine and Surgery) Heier KR (VA Tennessee Valley Healthcare System, University of Kentucky) O'Leary JR (VA Connecticut Healthcare System, Yale University School of Medicine,) Skanderson M (VA Connecticut Healthcare System) Goulet JL (VA Connecticut Healthcare System, Yale School of Medicine)
Objectives:
Research has highlighted that sexual minority adults experience inequities in cardiovascular disease (CVD) risk and outcomes compared to heterosexual peers. With an increasing sexual minority population, upwards of 7 million over the age of 50 by 2030, assessing and addressing their cardiovascular health remains a critical priority for clinicians, researchers, and public health officials. However, few electronic health record (EHR) cohorts have been used to study this association. As the largest integrated health care system in the US, with over 9 million veterans enrolled in the Veteran Affairs (VA) health program, the Veteran Health Administration is a unique data source to understand the health of marginalized populations and assess relatively rare health outcomes. To this end, we used the Veterans Healthcare Administration Lesbian, Gay, Bisexual, Transgender EHR cohort to examine the distribution of risk factors and CVD prevalence in sexual minority veterans compared to non-sexual minority veterans.
Methods:
In this analysis, we excluded transgender individuals to avoid conflating sexual orientation with gender identity. The baseline date was the first primary care appointment on or after October 1, 2009. Using a natural language processing tool with 88.2% sensitivity, 91.5% specificity, and 85.9% positive predictive value, sexual orientation of 1,108,990 veterans was identified, 185,788 of whom were classified as lesbian, gay, or bisexual (LGB). First, we calculated sample statistics stratified by sex and sexual minority status; differences were assessed within sex. We then used logistic regression analyses to assess the association of sexual minority status with prevalent CVD (history of AMI, ischemic stroke, or revascularization in the year preceding the baseline date). Missing values were imputed using multiple imputation by chained equation techniques to produce ten complete datasets; results were combined over imputations according to Rubin’s rules.
Results:
Many traditional CVD risk factors, including age, body mass index, prevalence of hypertension, and lipid levels were similar among LGB veterans compared to non-sexual minority veterans of the same sex. However, LGB veterans possessed greater rates of substance use, diabetes, homelessness, military sexual trauma, suicide attempts, depression, and anxiety. Prevalent CVD was present among 25,144 veterans (4,434 LGB). Adjusting for age, sex, race, and Hispanic ethnicity, LGB veterans had 1.24 [1.20, 1.28] times the odds of prevalent CVD compared to non-sexual minority veterans. Upon additional adjustment for BMI, diabetes, hypertension, lipids, substance use, anxiety, and depression, this association remained significant (OR [95% CI]: 1.15 [1.11, 1.19]).
Implications:
This is one of the first studies to examine cardiovascular health within a veteran population stratified by sexual orientation. Given that LGB veterans have greater prevalence of overall CVD and non-traditional CVD risk factors, addressing the cardiovascular health of marginalized populations requires further understanding of the mechanisms by which stress results in the inequitable distribution of both CVD risk and outcomes.
Impacts:
The establishment and description of this cohort demonstrates the utility of large EHR cohorts to study the cardiovascular health of sexual minority adults. Furthermore, this investigation highlights the potential to further elucidate the pathogenesis of CVD in sexual minority individuals and other marginalized populations.