Prevalence and Sociodemographic and Behavioral Correlates of Psychological Distress among Lesbian, Gay, and Bisexual Adults in the United States, 2013-2018

Background : The lesbian, gay, bisexual


Introduction
Mental health problems exact a substantial toll on the overall health and well-being of adolescents, youth, and adults and are leading causes of morbidity and mortality in the United States. 1,2ealthcare and social costs associated with mental health problems are also considerable. 3,415][16] The broader LGBT population is a sizable community, comprising approximately 5.6% of the total US population. 17In this study, we focus on LGBs instead of the broader LGBT group because most national databases, including the National Health Interview Survey on which this study is based, do not include information on gender identity, and hence transgender adults could not be included along with LGBs in the analysis. 18,19 addition to the limited information on prevalence, social determinants of mental health outcomes among lesbian, gay, bisexual (LGB) adults and the underlying psychosocial and behavioral mechanisms are not well studied.It is not known whether the sociodemographic and behavioral correlates of psychological distress among LGB adults are similar to those observed for the general population.][22] To address the research gaps in the literature, we use recent data from the National Health Interview Survey 23 to examine variations in psychological distress in the US according to sexual orientation and other social, demographic, and behavioral determinants and to identify specific groups of LGB adults who may be at increased risk of psychological distress and who may therefore require appropriate social and mental healthcare services.Specifically, we: (1) examine prevalence and levels of psychological distress among LGB adults in the US and compare these estimates with those for the straight/ heterosexual population using large, nationally representative samples of US adults and (2) examine a wide range of socioeconomic, demographic, and behavioral predictors of psychological distress among LGB adults and the general US population.

Data
Pooled cross-sectional data on mental health and selected socioeconomic, demographic, and behavioral characteristics for LGB and straight/heterosexual populations were derived from the 2013-2018 National Health Interview Surveys (NHIS). 23,24The NHIS is a national sample household survey in which data on socioeconomic, demographic, behavioral, morbidity, health, and healthcare characteristics are collected via personal household interviews. 7,23,24ll information collected in the survey is based on self-reports.The NHIS uses a complex, multistage probability design and is representative of the civilian non-institutionalized population of the United States.The NHIS, one of the longest-running federal health surveys, has been conducted annually since 1957 by the National Center for Health Statistics. 24esponse rate for an annual NHIS generally exceeds 87%.Data are obtained via in-home person interviews. 23,24Substantive and methodological details of the NHIS are described elsewhere. 7,23,24

Measurement of Psychological Distress (Dependent Variable)
We pooled 6 years of NHIS data from 2013-2018 to ensure sufficient sample sizes for analyzing mental health patterns by LGB status and other sociodemographic characteristics.The latest NHIS data were available for 2019 and 2020, but they did not include information on psychological distress and several of the covariates, such as occupation, alcohol use, and physical activity, used in our study. 23ifferentials in mental health outcomes were analyzed for 183,020 adults aged ≥18 in 2013-2018 for whom information on psychological distress and sexual orientation was available.Psychological distress was based on 6 questions that asked respondents how often during the past 30 days they felt (1) so sad that nothing could cheer them up, (2) nervous, (3) restless or fidgety, (4) hopeless, (5) that everything was an effort, or (6) worthless. 7,23ach question had 5 response categories: all of the time (coded 4), most of the time (coded 3), some of the time (coded 2), a little of the time (coded 1), or none of the time (coded 0).The response values to these six items were summed to create a scale (K6), ranging in value from 0 to 24, with a score of 13 or more used to define serious psychological distress (SPD). 4,5,8,25,26 addition to the dichotomous measure, we defined psychological distress as a continuous, composite index.The psychological distress index was constructed using principal components analysis of the above six items for all 183,020 adults aged ≥18.The factor loadings for the index items were as follows: sadness (0.79), nervousness (0.75), restlessness (0.73), hopelessness (0.83), everything an effort (0.77), and worthlessness (0.78).The index had a high-reliability coefficient (Cronbach's alpha=0.87).The psychological distress index scores ranged from a low of 87.18 to a high of 210.27 (mean=100; SD=20).Higher scores on the index indicate higher levels of psychological distress.

Covariate of Interest (Independent Variable)
Starting with the 2013 NHIS, respondents were, for the first time, asked questions about their sexual orientation, the primary covariate of interest in the study.Male respondents were asked: "Which of the following represents how you think of yourself: (1) gay; (2) straight, that is not gay; (3) bisexual; (4) something else; (5) I don't know the answer?"Female respondents were asked: "Which of the following represents how you think of yourself: (1) lesbian or gay; (2) straight, that is not lesbian or gay; (3) bisexual; (4) something else; (5) I don't know the answer?"For this study, we defined sexual orientation as a dichotomous variable by combining "gay or lesbian and bisexual" as one single category of LGB men and women and the second category consisting of straight/heterosexual individuals.Those with missing data on sexual orientation and with responses of "something else" and "I don't know the answer" were excluded from the analysis, 23 resulting in an effective/final sample size of 183,020 for analysis.

Sociodemographic and Behavioral Covariates
][8][9][10]12 Race/ethnicity was classified into 5 major categories: non-Hispanic Whites, non-Hispanic Blacks/African Americans, American Indians/Alaska Natives (AIANs), Asian/Pacific Islanders (APIs), and Hispanics.US-born were those born in one of the 50 US states or Washington, DC.Immigrants or foreignborn referred to those born outside these territories.
Educational attainment was measured by four categories: <12, 12, 13-15, ≥16 years of completed schooling.Income/poverty level, measured as the ratio of annual family income to the federal poverty threshold, was defined by 5 categories, ranging from <100% to ≥500% of the poverty level.Occupational class was defined in terms of 5 broad categories: professional and managerial occupations, sales/ clerical and technical support occupations, service, craft and repair, and laborers.These occupational groups, derived from the major occupational groups defined by the US Census Bureau, are consistent with previously defined social class positions of upper white-collar, lower white-collar, upper blue-collar, and lower blue-collar jobs. 27All other covariates were measured as shown in Tables 1 and 2. Less than 2.7% of observations for all variables except income/ poverty level were missing.For income/poverty level, the proportion missing values was 7.2%.We included in the analysis covariate categories for all missing values in order to avoid losing a substantial number of observations due to listwise deletion.

Statistical Methods
Multivariate logistic regression was used to examine the association between the binary outcome of SPD and sexual orientation before and after controlling for selected socioeconomic and demographic factors.Since the composite psychological distress index was a continuous variable, least squares regression was used to model mean psychological distress index scores.Interactions effects on psychological distress of sexual orientation with race/ethnicity, education, and poverty levels were examined, leading to race/ethnicity-, education-, and poverty-levelspecific stratified models of psychological distress.Additionally, sociodemographic and behavioral predictors of SPD and mean psychological distress scores were examined by limiting the sample to LGB adults only.Adjusted prevalence or mean scores were derived from the fitted logistic and least-squares models respectively.To account for the complex sample design of the NHIS, SUDAAN software was used to conduct all statistical analyses. 28The psychological distress index was created using the SAS Factor procedure. 29The Chi-square statistic was used to test the overall association between each covariate and psychological distress, while the two-sample t test was used to test the difference in prevalence or mean scores between any two groups.

Sociodemographic Characteristics of LGB and Straight Populations
LGB and straight populations differed significantly in their sociodemographic and behavioral characteristics.Compared with the straight population, LGB adults were more likely to be younger, non-Hispanic White, US-born, single, and had higher education, unemployment, and poverty levels (Table 1).Compared with the straight population, LGB adults were more likely to smoke and drink alcohol but less likely to be physically inactive than straight adults.LGB females had higher obesity but gay and bisexual males had lower obesity rates than their straight counterparts (Table 1).

Disparities in SPD Prevalence and Mean Psychological Distress Scores
The prevalence of SPD was 8.0% for the LGB population aged ≥18, 7.0% for gay and bisexual males, and 8.9% for LGB females, compared with 3.4% for the total straight/heterosexual population, 2.8% for straight males, and 3.9% for straight females (Table 2).In the general population, SPD prevalence varied from 1.6% for API males to 11.4% for AIAN males.After controlling for sociodemographic characteristics including gender, SES, health insurance, and functional limitation, LGB adults had 101% higher odds of SPD than their straight counterparts (Table 3).Adjusting for additional risk factors such as smoking, drinking, physical inactivity, and obesity/overweight reduced differentials only slightly; LGB adults still had 89% higher odds of SPD than straight adults (Table 3).After adjusting for sociodemographic and behavioral characteristics, compared to their straight counterparts, gay or bisexual males had 151% higher odds of SPD (OR=2.51;95% CI=1.86-3.40)and LGB females had 60% higher odds of SPD (OR=1.60;95% CI=1.31-1.96)[data not shown].
Mean psychological distress index scores were highest among LGB females, followed by gay and bisexual males, straight females and males (Table 4).When stratified by race/ethnicity, mean psychological distress index scores varied from 96.8 for straight APIs to 118.5 for gay and bisexual AIANs.After adjusting for sociodemographic and behavioral characteristics, gay or bisexual males had significantly higher psychological distress levels than straight males (mean psychological distress index score=104.3vs. 97.7)and LGB females had significantly higher psychological distress levels than straight females (mean index score=106.4vs. 100.7).
Other sociodemographic characteristics were associated with psychological distress in an expected manner (Tables 2-4).Those aged ≥65 years had the    (Table 3, Model 3).Immigrants with <15 years of residence in the US had 32% lower adjusted odds of SPD than US-born adults.Education, income, and homeownership were inversely related to SPD and psychological distress index scores, while smoking, alcohol consumption, physical inactivity, and obesity were associated with increased risks of SPD and with higher psychological distress index scores.

Predictors of SPD and Psychological Distress Level in LGB Adults
Younger age, unmarried status and marital disruption, lower-income, divorce/separation, lack of health insurance, functional limitation, smoking, physical inactivity, and obesity were significant predictors of SPD and higher psychological distress levels in LGB adults (Table 5).
LGB adults under the age of 45 had more than 5.6 times higher adjusted odds of SPD than LGB adults aged ≥65.Mean psychological distress scores were substantially higher among LGB adults aged under 35 than among LGB adults aged ≥65 (110.8 vs. 96.5).
LGB adults who experienced divorce or separation had 2.2 higher adjusted odds of SPD than their married counterparts.Poverty was strongly linked to SPD and higher psychological distress levels in LGB adults.SPD prevalence and psychological distress scores were 18.3% and 118.6, respectively, for LGB adults below the poverty level, compared with 1.8% and 99.1 for LGB adults with income ≥500% of the poverty level (Table 5).After adjusting for covariates, LGB adults below the poverty level and with income at 100-199% of the poverty level had 6.9 and 4.9 times higher odds of SPD respectively than LGB adults with income at ≥500% of the poverty level.
LGB adults who did not have health insurance had 2.0 times higher odds of SPD than LGB adults who did.Psychological distress levels were nearly 9 points higher among LGB adults lacking health insurance.
LGB adults with a functional limitation or disability had 3.2 times higher adjusted odds of SPD and a 15-points higher distress score than LGB adults with a limitation.LGB smokers had 86% adjusted higher odds of SPD and a 6-points higher distress score than LGB non-smokers.Physically inactivity    was associated with an 85% adjusted higher odds of SPD in LGB adults, while obesity was associated with a significant (4-point) increase in psychological distress level among LGB adults (Table 5).

Discussion
Our study has shown substantially higher risks of SPD and higher psychological distress levels among LGB adults in the US compared to their straight/heterosexual counterparts.Estimates of SPD prevalence and psychological distress levels for LGB adults had not been previously available at the national level, and a comparison of their mental health outcomes with the straight population had not been made by controlling for differences in sociodemographic and behavioral characteristics.
Our study confirmed findings from previous studies that show significant socioeconomic and health disparities in mental health among LGB adults.Our study is one of the few studies that have examined mental health disparities in the LGB population by using nationally representative samples and by controlling a number of risk factors that are known to influence mental health outcomes.Documenting national estimates of psychological distress among LGB adults from various gender and racial/ethnic groups is new to the literature.Analysis of mental health disparities using a composite, continuous factor-based psychologic distress index is another novel feature of the study.
Our finding that LGB adults have approximately two times higher risks of SPD and psychological distress levels, compared with heterosexual adults, is consistent with previous studies.One paper using meta-analyses of 25 studies found that lesbian, gay, and bisexual people experienced at least 1.5 times higher risk for depression and anxiety disorders over a 12-month period or lifetime. 30Recent studies using the Behavioral Risk Factor Surveillance System (BRFSS) also found that LGB adults were more likely to have 1.3 to 2.7 times more days with poor mental health, compared with heterosexual adults. 31,32tudies using the National Survey on Drug Use and Health (NSDUH) found that LGB adults had a higher prevalence of drug use and any or severe mental illness including any mental, behavioral, or emotional disorder in the past year, compared with heterosexual adults. 15,16r findings on differences in mental health status by intersectionality between sexual orientation and gender are consistent with previous findings that gay or bisexual males have higher odds of SPD than female LGBs.According to one study using the NHIS, gay and bisexual men and women were more likely to report severe psychological distress than heterosexual OR=odds ratio; CI=confidence interval; BMI=body mass index. 1 SPD is measured by a 6-item scale (K6) that asks respondents how often they experienced each of 6 symptoms of psychological distress in the past 30 days (feelings of sadness, nervousness, restlessness, hopelessness, everything an effort, worthlessness).The K6 varies from 0 to 24, with a score of 13 or more used to define SPD. 2 Unadjusted for other covariates. 3This logistic regression model includes survey year, age, sexual orientation, gender, race/ethnicity, length of immigration, marital status, region of residence, education, poverty status, occupation housing tenure, health insurance, and activity limitation as covariates. 4This logistic regression model includes all covariates of Model 2 plus smoking, drinking, physical activity and BMI.individuals, while lesbians were more likely to report moderate psychological distress. 14The NSDUH study also found that gays and both male and female bisexual adults aged 50 and older were more likely to have mental illness compared with their heterosexual counterparts, but not lesbians. 16Another study, focusing on LGBT cancer survivors, found that gay, bisexual, and transgender males had a higher prevalence of depressive symptoms than heterosexual males, whereas lesbian, bisexual, and transgender females were not significantly different from heterosexual females. 33The BRFSS studies also showed higher odds of SPD for gays than lesbians. 31,32Further studies are needed to estimate the intersectionality in SPD between sexual orientation and gender.
A higher prevalence of psychological distress among LGB might be explained by the minority stress model, under which excess in social stressors related to stigma and prejudice toward LGB populations is postulated to cause mental disorders among them. 22exual orientation discrimination, stressful life events, and adverse childhood experiences among LGB individuals are associated with co-occurrence of alcohol or tobacco use disorder with anxiety, mood disorders, and posttraumatic stress disorder. 34onsidering minority stress during lifetime, various policy and clinical interventions might be helpful for mental health improvement among LGBs.For example, providing cognitive behavioral therapy or clinical care to youth with gender dysphoria, or a state-level antibullying law that enumerated sexual orientation as a protected class might reduce mental health problems among sexual and gender minority youth. 35

Limitations
This study has some limitations.The K6 items in NHIS used to define psychological distress levels and SPD prevalence are based on self-reports, which may underestimate the actual prevalence of psychological distress among various sociodemographic groups, including LGB adults. 1 Second, the cross-sectional nature of the NHIS limits the estimation of the mental health impacts of socioeconomic variables and health-risk behaviors.However, the measures of psychological distress refer to the experiences during the 30 days preceding the survey, whereas some of the socioeconomic variables in the survey precede the psychological distress in their temporality.For example, family income/poverty level relates to the average income earned during the year preceding the survey.Similarly, education for most adults aged 30 and older is attained long before the time of the survey interview.
Third, self-reported data on the marital status by LGB adults in 2013 and 2014 NHIS may have been affected prior to the legalization of same-sex marriages in the US on June 26, 2015. 36Fourth, information on stressors (such as social stigma, discrimination, personal stress, financial stress, relationship problems, job strain, and family health problems) underlying psychological distress among LGB adults is lacking in the NHIS, and data on social and familial support is also limited.Fifth, since mental health, health-risk factors, and social determinants are likely to vary for specific LGBT groups, studies of psychological distress need to consider disaggregated data for lesbians, gays, bisexuals, and  transgender people. 14,16,20Finally, since NHIS did not include data on gender identity, we were unable to include transgender adults who have been shown to be significantly more disadvantaged in their social and health-risk profile, compared with lesbians, gays, and bisexuals. 18,19

Conclusion and Implications for Translation
Significant disparities in mental health exist, with LGB adults at a substantially increased risk of psychological distress and likely in greater need of appropriate social and mental health services.Health policies aimed at improving the material conditions and social environments may lead to improved mental health outcomes among LGB adults and the general population.Evidence-based social and public health interventions to reduce psychological distress among LGB adults are lacking.Further research is urgently needed to address the mechanisms (such as social stigmatization, discrimination, and stress) through which social determinants influence mental health outcomes and psychological well-being among LBG adults across different racial/ethnic and socioeconomic groups. 18,22Continued monitoring of social conditions and mental health disparities among sexual minorities is essential in tracking progress towards achieving the national goal of eliminating health inequities. 2,18Increasingly, a number of national, state, and community health surveys have started to include variables on sexual orientation and gender identity, which should enable monitoring of health disparities and a better understanding of health, healthcare, and social needs of the LGB population. 18,37mpliance with Ethical Standards

Table 1 : Descriptive socioeconomic, demographic, and behavioral characteristics of LGB (Lesbian, Gay, or Bisexual) and Straight/Heterosexual populations aged ≥18 years in the United States: The National Health Interview Survey, 2013-2018
0.32 41.93 1.41 31.310.37 34.38 1.29 31.85 0.34 (Contd....)SE=standard error.All Chi-square tests for differences in characteristics by sexual orientation (except for health insurance status and activity limitation for men and region for women) were statistically significant at P<0.05.

Table 1 : (Continued)
lowest odds of SPD and levels of psychological distress than adults in younger age groups.AIANs and Hispanics had significantly higher SPD prevalence (10.5% and 4.2% respectively) than non-Hispanic Whites, while APIs had a significantly lower prevalence of SPD (1.8%).Adjusting for

Table 2 : (Continued)
socio-behavioral characteristics reduced racial/ ethnic differentials in SPD prevalence, with only non-Hispanic Blacks experiencing 30% lower adjusted odds of SPD than non-Hispanic Whites

Table 3 : (Continued)Table 4 : Observed and adjusted mean Psychological Distress Index scores by sexual orientation and its interactive effects with gender and race/ethnicity, US population aged 18+ years: The National Health Interview Survey, 2013-2018 (N = 183,265) Interaction of LGB status with demographic covariates Observed Psychological Distress Index Adjusted 1 Psychological Distress Index Score
The percent variance explained (R 2 ) for the adjusted model was 15.25 (total), 14.06 (male) and 15.73 (female).All Chi-square tests indicate statistically significant associations between each covariate and psychological distress levels at p < 0.05 in both observed and adjusted models, except for insurance status (p=0.06) in the adjusted model for men.
1E=standard error.1Adjustedby least squares regression model for survey year, age, sexual orientation, gender, race/ethnicity, length of immigration, region of residence, education, marital status, poverty status, occupation, housing tenure, health insurance, activity limitation, smoking, drinking, physical activity, and BMI.

Table 5 : (Continued)
(Contd....) 2E=standard error.OR=odds ratio; CI=confidence interval; BMI=body mass index.1Unadjustedforothercovariates.2Adjusted by logistic or least squares regression model for survey year, age, gender, race/ethnicity, length of immigration, region of residence, marital status, education, poverty status, occupation, housing tenure, health insurance, activity limitation, smoking, drinking, physical activity, and BMI.The percent variance explained (R 2 ) in psychological distress scores for the adjusted least squares model was 20.36.

Table 5 : (Continued)
No IRB approval was required for this study, which is based on the secondary analysis of a public-use federal database.Acknowledgments: None.Disclaimer: The views expressed are the authors' and not necessarily those of their institutions.
Conflicts of Interest:The authors declare that they have no conflict of interest.Financial Disclosure: None to report.Funding/Support: None.Ethical approval: