Exploring Sexual Orientation Disparities Regarding the Interplay of Childhood Sexual Abuse, Self-Reported Diabetes Status, and Depression Among Adults in the United States

ABSTRACT Previous research has revealed a strong link between the experience of childhood sexual abuse (CSA) and diabetes in adulthood. Moreover, research has shown that sexual minorities (SM) are exposed to adverse childhood experiences (ACEs) (i.e. CSA) and experience depression at higher rates than their heterosexual counterparts. Thus, it is imperative to further investigate the role of depression and the differential associations of exposure to ACEs with diabetes prevalence by sexual orientation. We explored sexual orientation disparities regarding the relationship between CSA and diabetes and examined the moderating role of depression. A total of 29,903 participants from the 2021 Behavioral Risk Factor Surveillance System (BRFSS) were included in this study. Secondary data analysis was conducted using the survey data, and weighted logistic regression and moderation analysis were performed. Heterosexuals who experienced CSA (AOR = 1.25; p < .05) and SM who experienced CSA (AOR = 2.13; p < .05) reported higher odds of having diabetes. Among heterosexuals, depression (AOR = 1.38; p < .001) was significantly associated with having diabetes. Additionally, depression was a significant moderator among heterosexuals with and without CSA. Further understanding of the impact of ACEs on diabetes among specific subgroups of SM should be assessed in future studies.

the United States (Centers for Disease Control and Prevention, 2020c).With such life-threatening consequences, it is crucial to understand the factors associated with the development of diabetes to address this growing public health issue effectively.
Research has uncovered numerous factors related to the development of diabetes, including adverse childhood experiences (ACEs) (Jones et al., 2019;Zhang et al., 2022).ACEs are traumatic events experienced during childhood, such as enduring neglect, living in a dysfunctional household, or experiencing physical, emotional, or sexual abuse (Centers for Disease Control and Prevention, 2020b).It is estimated that 1 in every 6 adults in the U.S. has experienced at least four types of ACEs (Centers for Disease Control and Prevention, 2020b;Jones et al., 2019).These experiences can have detrimental long-term impacts on the well-being of individuals who experience them and have been associated with num erous negative life outcomes, such as poor mental health, an increase in risky behaviors, and chronic diseases such as cancer, obesity, or diabetes (National Center for Injury Prevention and Control, 2021).
Previous research has revealed a strong link between the experience of childhood sexual abuse (CSA) and diabetes in adulthood (Campbell et al., 2018a(Campbell et al., , 2016(Campbell et al., , 2018b;;Huang et al., 2015;Shields et al., 2016).One study found that having experienced CSA, categorized as experiencing unwanted sexual touching, one occurrence of forceful sexual activity, and frequent forced sexual activity, was associated with a 16%, 34%, and 69% higher risk of diabetes, respectively (Shields et al., 2016).Another study found that CSA was the most significant predictor of diabetes compared to other ACEs (Campbell et al., 2018b).Further, individuals who reported that they experienced childhood sexual assault had a 57% higher likelihood of developing diabetes than those who did not report experiencing sexual abuse as children (Campbell et al., 2018b).While there is evidence of this link, further research is needed to better understand the relationship between CSA and the development of diabetes in adulthood.
Clear disparities in the experience of ACEs have also been found.Sexual minority (SM) (i.e., non-heterosexual orientation) individuals are exposed to ACEs at higher rates relative to their heterosexual counterparts (Friedman et al., 2011;McCabe et al., 2020).Research has also indicated that compared to heterosexuals, SM individuals are at greater risk of experiencing ACEs and related negative consequences, such as chronic illnesses and poor mental health outcomes (Andersen et al., 2013;Campbell et al., 2018b).It has been found that among the different types of ACEs, the largest disparity between SM and heterosexual individuals involves the experience of CSA (Friedman et al., 2011).For example, a meta-analysis assessing ACEs concluded that SM individuals were 3.8 times more likely, on average, to experience CSA (Friedman et al., 2011).Furthermore, it is separately known that exposure to CSA is correlated with an increased risk of diabetes (Campbell et al., 2018a(Campbell et al., , 2018b;;Huang et al., 2015), and SM individuals may have an increased risk of CSA exposure (Friedman et al., 2011).To our knowledge, however, there have been no studies that have aimed to assess whether there are differential associations of exposure to adverse childhood experiences (such as CSA) with diabetes prevalence by sexual orientation.
Research has also found that depression may affect the relationship between CSA and diabetes risk, although the exact role remains unclear (Campbell et al., 2018b;Mukherjee & Chaturvedi, 2019).Despite a higher prevalence of depression and depressive symptoms among SM individuals compared to heterosexuals, there is a lack of research assessing the relationship between these factors while considering potential disparities in sexual orientation (Lucassen et al., 2017).Using the Behavioral Risk Factor Surveillance System (BRFSS), this study aims to explore sexual orientation disparities regarding the relationship between CSA and diabetes and assess how depression moderates this relationship.

Study design & data source
We used the 2021 publicly available dataset from the CDC BRFSS.The BRFSS uses a complex multi-stage cluster sampling design and sampling weights to account for the non-response bias and establish a representative populationbased sample of the adults interviewed for the survey.The 2021 BRFSS dataset includes a total number of 438,693 observations.However, our study only consists of participants with complete data on self-reported diabetes status, CSA, sexual orientation, and depressive disorder variables, which resulted in a final sample size of (N = 29,903).

Dependent variable
The primary outcome variable in this study is self-reported diabetes status, which was measured by the item, "(Ever told you had) diabetes?"with response options including "yes," "yes, but female told only during pregnancy," "no," "no, pre-diabetes or borderline diabetes," or "don't know/not sure."This variable was dichotomized to indicate diabetes status as "yes" or "no."

Independent variables
CSA (CSA) that occurred before the age of 18 was measured by three items under the ACE module, "How often did anyone at least 5 years older than you or an adult ever touch you sexually?""How often did anyone at least 5 years older than you or an adult try to make you touch them sexually?"and "How often did anyone at least 5 years older than you or an adult, force you to have sex?"These three items were collapsed into one CSA item, and the corresponding responses were dichotomized into "yes" or "no" (Brown et al., 2009;Dong et al., 2004;Felitti et al., 1998).
Sexual orientation was assessed by the following item for both males and females, "Which of the following best represents how you think of yourself?"The response options were collapsed into "straight," "gay/lesbian," "bisexual," or "other."For the descriptive/bivariate, multivariable logistic regression, and moderation analysis, we recoded the sexual orientation variable into "heterosexuals" and "SM."We pooled all the SM groups (gay/lesbian, bisexual, and other) into one "SM" group to match the categorization used in our multivariable logistic regression.
The following item measured self-reported depressive disorder, "(Ever told you had) a depressive disorder (including depression, major depression, dysthymia, or minor depression)?"The response options were collapsed into "yes" or "no" options.

Statistical analysis
All statistical analyses were performed using STATA BE software version 17.0.We calculated the weighted prevalence of diabetes status among individuals by their sexual orientation.We conducted a descriptive analysis of all variables, including unweighted frequencies and their weighted percentages by their self-reported diabetes status stratified by sexual orientation (i.e., heterosexuals and SM).We performed bivariate analysis using the chi-squared test to examine the differences in diabetes status by CSA, mental health, and sociodemographic factors stratified by sexual orientation.We then conducted multivariable logistic regression with adjusted odds ratios (AOR) with a 95% confidence interval to examine the relationship between CSA stratified by sexual orientation while controlling for all the covariates.Finally, we performed a moderation analysis to assess the role of depression on the association between CSA and self-reported diabetes status while adjusting for the sociodemographic factors based on sexual orientation.

Prevalence of self-reported diabetes status by sexual orientation
Figure 1 below portrays the overall prevalence of self-reported diabetes by sexual orientation.The figure shows that the prevalence of non-diabetic adults was considerably higher than those with diabetes.Furthermore, individuals who perceived themselves to be "other" in terms of sexual orientation had the highest prevalence of being diabetic (14.54%) compared to individuals who identified as being heterosexual (straight) (12.73%), gay/lesbian (11.27%), or bisexual (6.26%).

Multivariable regression analysis
The adjusted weighted logistic regression model results of the association between self-reported diabetes status, CSA, depression, and sociodemographic factors by sexual orientation are presented in  § § = The estimate should be interpreted with caution due to the wide 95% confidence interval.
Regarding employment status among heterosexuals, those who were unemployed reported significantly greater odds of diabetes than their employed counterparts (AOR = 1.37;CI: 1.16, 1.63; p < .001).In the context of BMI, compared to being underweight, all other BMI levels among SM individuals were significantly associated with higher odds of having diabetes.On the other hand, for heterosexuals, only obesity (AOR 4.77; CI: 2.27, 10.00; p < .001)increased the odds of having diabetes.Finally, only heterosexuals (AOR = 0.67; CI: 0.58, 0.78; p < .001)who reported being physically active indicated significantly lower odds of being diabetic.

Moderation analysis
The margins plot of the moderation analysis is displayed in Figure 2.Even though we found depression to be a significant predictor of diabetes only among heterosexuals, our findings did not find depression to be a significant moderator in the overall relationship between CSA exposure and diabetes status.However, our study did find that depressed heterosexuals without CSA (AOR = 1.34;CI: 1.13.1.58; p < .01),non-depressed heterosexuals with CSA (AOR = 1.34;CI: 1.03, 1.74; p < .05),depressed heterosexuals with CSA (AOR = 1.54;CI: 1.19, 1.99; p < .01),and non-depressed SM individuals with CSA (AOR = 2.01; CI: 1.02, 3.95; p < .05),exhibited significantly increased odds of having diabetes.

Discussion
To our knowledge, this study is the first to assess if differential associations between depression and CSA history on the odds of diabetes exist by sexual orientation.The odds of diabetes were similar among our study population, with factors such as belonging to a racial/ethnic minority group, falling within the age range of 35 to 65 years or older, and obesity being associated with an increased likelihood of diabetes in both heterosexual and SM individuals.However, a couple of divergent trends are noticed within heterosexual versus SM.Although increasing income > $15,000 was associated with decreased odds of diabetes among heterosexual individuals compared to those making <$15,000 annually, this protective factor was not seen among SM individuals, as SM individuals earning between $50,000 to <$100,000 showed higher odds of having diabetes.This suggests that the independent effect of income level on diabetes odds is not as salient of a protective factor among SM individuals compared to heterosexuals.Consistent with existing literature that highlights the heightened risk of diabetes associated with CSA exposure in adulthood (Campbell et al., 2018a(Campbell et al., , 2016(Campbell et al., , 2018b;;Huang et al., 2015;Shields et al., 2016), our study found that history of sexual abuse during childhood increased the odds of diabetes among both heterosexuals and SM, with SM individuals exhibiting higher odds of diabetes compared to heterosexuals during adulthood.

Diabetes and CSA in SM
The weighted prevalence of CSA exposure in our study was comparable between heterosexuals and SM.CSA exposure has been shown to increase the risks of diabetes due to toxic stress and psychological trauma (Huang et al., 2015;Huffhines et al., 2016), particularly in females who experienced genitalcontact CSA (Romans et al., 2002).Studies have found that female SM individuals have a higher risk of CSA and lifetime sexual victimization (Conron et al., 2010;Roberts et al., 2012).Therefore, it would be plausible to assume SM faces a higher diabetes risk due to increased CSA exposure.In our study, CSA had a significant impact on diabetes odds among both heterosexual and SM individuals.Previous studies have mainly focused on the association between sexual assault history and chronic conditions such as diabetes, mainly among SM women, with limited focus given to the broader SM population (Caceres et al., 2021;Lehavot & Simoni, 2011).Our study findings demonstrate that lifetime sexual abuse trauma can have significant consequences regarding chronic physical health conditions among both male and female SM populations.Given the small SM sample size (<8% of the sample), specific subanalysis within the SM subgroups was not performed.Future studies should evaluate the impact of CSA and depression on diabetes and related adverse outcomes, as some SM subgroups may be at more risk.

Depression in SM
Even though studies have found that the national prevalence of depression is at least double in SM than in heterosexual individuals (Kelley et al., 2018;Plöderl & Tremblay, 2015) due to factors such as societal stigma and marginalization (Diplacido, 1998, Flentje et al., 2022;Kelley et al., 2018;Mason & Lewis, 2015;Newcomb & Mustanski, 2010), our study did not find a significant correlation between diabetes and depression due to the small sample size among SM, which may not reflect the broader population of SM.However, it is worth noting that our study did find that depressed heterosexuals had significantly increased odds of having diabetes.
Internalized homophobia, which is internalized negative stigma about one's sexual orientation stemming from a heteronormative society, has been linked to adverse emotional and physical health outcomes in SM adults (Newcomb & Mustanski, 2010;Szymanski et al., 2008).Despite improving societal attitudes toward SM, internalized homophobia persists (Newcomb & Mustanski, 2010;Savin-Williams, 2008).Additionally, SM adults exhibit higher levels of the stress hormone cortisol, which is a risk factor for diabetes (Lick et al., 2013).Our use of the 2021 BRFSS survey allowed us to account for the physical and emotional health of SM individuals amid the pandemic.Notably, the pandemic has exacerbated depression symptoms among SM groups, particularly those without preexisting mental health symptoms (Flentje et al., 2020).The increasing severity of depression symptoms is associated with an increased risk for diabetes due to a decreased desire to incorporate healthy lifestyle behaviors (Flentje et al., 2020;Yu et al., 2015), resulting in poor general physical and emotional health (Lett et al., 2020).

Strengths & limitations
A strength of the study is that we used a population-based national survey to assess various covariates potentially associated with increased diabetes risk, including depression status and CSA history.The BRFSS survey allowed us to gauge the diabetes risk factors for a national, generalizable sample of SM, advancing the literature in this subpopulation.In this study, our secondary data analysis of the 2021 BRFSS survey presents a few limitations.As a cross-sectional study, the temporal relation and causation between sexual orientation, depression, or CSA history on diabetes cannot be established.The BRFSS survey also relies on participant self-reporting and is, therefore, subject to recall, selection, and social desirability bias (Beach et al., 2018).This may be the case for self-identification of sexual orientation status, as only individuals who have fully embraced and are willing to disclose their alternative sexual identity as either gay/lesbian, bisexual, or other would be included in our SM groups.The simplistic BRFSS classification of sexual identity used in this study may aggregate varied SM identities and mask diabetes inequities within distinct, uncategorized sexual identities.
In our study, approximately 92% of the BRFSS sample consisted of heterosexual individuals, while roughly 7% represented SM individuals.The small sample size of SM compared to heterosexuals resulted in wider confidence intervals and larger standard errors for a few covariates, including age and BMI.Furthermore, we are disproportionately missing homeless or institutionalized individuals since BRFSS relies on landline and cellphone users (Beach et al., 2018).Additionally, BRFSS does not employ objective laboratory measures to ascertain diabetes status and only relies on self-report (Beach et al., 2018).This is an important limitation as some individuals with diabetes may not be aware of their diabetes status and thus would be misclassified as normal in our analysis (Lin, 2006 Specifically, SM individuals who have fewer encounters with the formal healthcare system (Powers et al., 2015) may be tested less for diabetes and, therefore, unaware of actually having diabetes.Accordingly, our prevalence data for diabetes is likely underestimated.

Conclusion
The literature on SM disparities in chronic physical and mental health disease risk has not kept pace with their population growth.Overall, associated factors for self-reported diabetes remained consistent in heterosexual individuals compared to SM.Further understanding of diabetes prevalence among specific subgroups of SM, using a more diverse classification of sexual orientation than is used in the BRFSS survey, should be assessed in future studies.Other covariates, such as discrimination and psychological stress measures, should be considered in future studies to provide more insight into factors contributing to depression and diabetes in SM.Future studies should also assess how to address the institutional barriers toward accessing healthcare among SM individuals for them to obtain equity in both physical and mental health outcomes.

Figure 2 .
Figure 2. Moderation analysis of the effect of depression on the association between diabetes status and CSA based on sexual orientation.

Table 2
be diabetic than heterosexual individuals in the remaining age groups.Similarly, SM individuals aged 65 years or older (AOR = 24.93;CI: 6.93, 89.63; p < .001)had higher odds of being diabetic compared to SM

Table 1 .
Descriptive statistics and Bivariate Analyses of Sample Sociodemographic Characteristics, diabetes status, CSA, and depression stratified by sexual orientation.

Table 1 .
(Continued).The table above represents the unweighted frequencies for the various categories stratified by sexual orientation.The number in parenthesis represents the weighted proportion of each subcategory within each column.

Table 2 .
Weighted multivariable logistic regression analysis of self-reported diabetes status and its association with CSA, depression, and sociodemographic characteristics by sexual orientation.