A Comparison of the Mental Health and Well-Being of Sexual Minority and Heterosexual First-Year Medical Students

A Report From the Medical Student CHANGE Study

Abstract

Purpose 

Research is lacking on psychological distress and disorder among sexual minority medical students (students who identify as nonheterosexual). If left unaddressed, distress may result in academic and professional difficulties and undermine workforce diversity goals. The authors compared depression, anxiety, and self-rated health among sexual minority and heterosexual medical students.

Method 

This study included 4,673 first-year students who self-reported sexual orientation in the fall 2010 baseline survey of the Medical Student Cognitive Habits and Growth Evaluation Study, a national longitudinal cohort study. The authors used items from published scales to measure depression, anxiety, self-rated health, and social stressors. They conducted bivariate and multivariate analyses to estimate the association between sexual identity and depression, anxiety, and self-rated health.

Results 

Of 4,673 students, 232 (5.0%) identified as a sexual minority. Compared with heterosexual students, after adjusting for relevant covariates, sexual minority students had greater risk of depressive symptoms (adjusted relative risk [ARR] = 1.59 [95% confidence interval, 1.24–2.04]), anxiety symptoms (ARR = 1.64 [1.08–2.49]), and low self-rated health (ARR = 1.77 [1.15–2.60]). Sexual minority students were more likely to report social stressors, including harassment (22.7% versus 12.7%, P < .001) and isolation (53.7% versus 42.8%, P = .001). Exposure to social stressors attenuated but did not eliminate the observed associations between minority sexual identity and mental and self-reported health measures.

Conclusions 

First-year sexual minority students experience significantly greater risk of depression, anxiety, and low self-rated health than heterosexual students. Targeted interventions are needed to improve mental health and well-being.

According to a 2011 Institute of Medicine report, lesbian, gay, and bisexual (LGB) adults experience more mood and anxiety disorders than heterosexual individuals.¹ While there has been increasing research on the mental health and well-being of LGB patients, few studies have examined the mental health and well-being of LGB medical students and health care providers in the United States.²

Medical students, in general, experience high rates of psychological distress, including depression and anxiety.^3–5 However, compared with heterosexual students, students who identify as a sexual minority may face a higher risk of poor health due to minority stress—the additive and chronic stress resulting from the stigma and discrimination these individuals experience because of their marginalized social status.⁶ The social stressors that comprise minority stress include overt acts of discrimination, such as violence, harassment, and name-calling; institutionalized heterosexism, such as marriage inequality and employment discrimination; and more implicit microaggressions, or everyday forms of unintended discrimination, such as social exclusion, tokenizing, and heteronormativity. Together, these stressors contribute to the health disparities experienced by sexual minorities.⁷

Poor mental health among sexual minority medical students may lead to greater burnout and attrition from medical school,⁸ which in turn may diminish the diversity of the physician workforce. Both the American Medical Association (AMA) and the Association of American Medical Colleges (AAMC) have articulated the importance of diversity—including sexual orientation diversity—among physicians to the provision of accessible, quality care.^9–11

In the present study, we hypothesized that sexual minority students entering medical school would be at greater risk of depression, anxiety, and low self-rated health when compared with heterosexual medical students, and that these disparities in risk would differ by gender. We also hypothesized that sexual minority medical students would be more likely than their heterosexual counterparts to report having experienced social stressors, and that exposure to these social stressors would attenuate the association between sexual identity and health.

Method

Data source

This study used baseline data collected as part of the Medical Student Cognitive Habits and Growth Evaluation Study (the Medical Student CHANGE Study; hereafter, CHANGES), a national longitudinal study of individuals who matriculated at U.S. medical schools in fall 2010.¹² For detailed study protocol information, please refer to Supplemental Digital Appendix 1, available at https://links.lww.com/ACADMED/A258. Briefly, we sampled medical students using a stratified multistage sampling design. In the first stage, we stratified medical schools by geographic region and public/private status. From these strata, we sampled 50 medical schools from the total of 131 MD-granting U.S. schools using a proportional to (first-year class) size sampling methodology.¹³ One of the 50 schools sampled for our study was a military school that had highly unique features, including acceptance policies, curriculum structure, timing, and student characteristics. Because of concerns about the generalizability of our study findings, we excluded this school (n = 169 first-year students) from the analysis.

In the second stage, we recruited first-year students from the 49 selected schools, using three strategies to obtain their contact information because no complete list of students was available. First, we obtained e-mail addresses of students who indicated interest in participating in the study via a question included as part of the 2010 AAMC Matriculating Student Questionnaire, a voluntary annual survey sent to all students entering medical school. Second, we purchased an incomplete list of first-year medical students from an AMA-licensed vendor. Third, we used referral (i.e., snowball) sampling through recruited survey respondents. Ascertained students were invited to participate in the CHANGES survey via e-mail and/or postal mail. Those who consented to study participation completed an extensive online questionnaire and were randomized to complete various implicit association tests (e.g., race, sexuality, obesity). The University of Minnesota and Mayo Clinic internal review boards approved the study. All students who completed the survey received a $50 incentive for participation.

Study sample

Between October 2010 and January 2011, 4,732 first-year medical students completed the baseline survey, representing 81% of the 5,823 students invited to participate in the study and 55% of the 8,594 first-year students enrolled at the 49 sampled schools (for participant recruitment flowchart, see Supplemental Digital Appendix 2 at https://links.lww.com/ACADMED/A258). Our overall response rate was comparable to other published studies of medical students.^3,14

Study measures

Sexual identity.

We assessed sexual identity, the main independent variable in this study, by asking all students in our sample, “What is your sexual orientation: heterosexual, bisexual, homosexual, or other?” We combined students who selected homosexual, bisexual, or other into a larger category of sexual minority (nonheterosexual) students. We did this for two reasons. First, although sexual minority individuals are not a monolithic group, they nonetheless share a common marginalized social status relative to society’s heterosexual cultural norm.¹ Second, because of limited sample sizes, there was not sufficient power to distinguish among these three sexual minority groups.

Depression, anxiety, and self-rated health.

We assessed depression using the Patient Reported Outcomes Measurement Information System (PROMIS) Emotional Distress–Depression Short Form 8b, a validated eight-item instrument that evaluates negative mood, decrease in positive affect, negative views of self, and negative social cognition.¹⁵ As per the PROMIS scoring manual, we standardized raw scores so that a score of 50 represents the average score of the general population of the United States, with 10 as the standard deviation.¹⁶ We analyzed the depression score as a continuous dependent variable. We also sought to assess the risk of a clinically meaningful depressive symptomology by dichotomizing the depression score, whereby those students who scored one standard deviation above the general population mean (i.e., a score of 60) were categorized as exhibiting depressive symptoms. This represents a score that exceeds the minimally important difference from the mean as suggested by Yost et al.¹⁷

We assessed anxiety using the PROMIS Emotional Distress–Anxiety Short Form 7a, a validated seven-item instrument that measures self-reported fear, anxious misery, and hyperarousal.¹⁵ We standardized the raw scores so that a score of 50 represents the general population mean, with 10 as the standard deviation.¹⁶ Similar to the depression measure, we considered anxiety both as a continuous outcome and a dichotomized outcome. For the dichotomized anxiety variable, students who scored two standard deviations above the mean (i.e., a score of 70) were classified as having clinically meaningful anxiety. We chose this higher cutoff in order to increase the likelihood of capturing a meaningful difference between anxious and nonanxious medical students. (In a comparison analysis using a cutoff of one standard deviation [results not shown], we found a statistically significant, though smaller, effect size.)

We assessed students’ self-rated health by asking, “In general, would you say your health is: excellent, very good, good, fair, or poor?” We categorized students who indicated that their health was fair or poor as having low self-rated health, and students who indicated all other categories as having good self-rated health.¹⁸

Social stressors.

We evaluated the differences in social stressors reported by sexual minority and heterosexual first-year medical students using two items from the Everyday Discrimination Scale (called names or insulted at least a few times a year, and harassed or threatened at least a few times a year)¹⁹ and three items from the UCLA Loneliness Scale (whether they felt they lack companionship, felt left out, and felt isolated from others at least some of the time).²⁰

Other variables.

We used standard demographic questions to measure student age, gender, race and ethnicity, and relationship status. We asked students to provide their parents’ highest level of education to assess family socioeconomic status. Additionally, to assess social desirability response bias, we administered an abridged version of the Marlowe–Crowne Social Desirability Scale.²¹

Apart from the sexual identity question, none of the measures included in the study were specific to sexual identity. Further, some questions may have assessed exposures prior to medical school matriculation (e.g., stressors in the past year). Finally, for all multi-item scales, scores were computed only for respondents who completed at least half of the scale’s items.

Analysis

We used descriptive summary statistics of demographic characteristics to characterize sexual minority and heterosexual students. We calculated the prevalence of all measures included in the study for sexual minority and heterosexual students, and we used Pearson chi-square tests to determine whether there were significant differences between sexual minority and heterosexual students.

To estimate the association between sexual identity and depression, anxiety, and self-rated health, we fit three models to our data. In model 1, we conducted separate bivariate analyses to estimate the association between sexual identity and depression, anxiety, and self-rated health. For the two continuous main outcome measures (depression score and anxiety score), we fit simple linear regressions to estimate association coefficients; for the three dichotomous main outcome measures (depressive symptoms, anxiety symptoms, and self-rated health), we fit generalized linear models (GLMs) with a binomial distribution and a log link function to estimate relative risks. In model 2, we conducted multivariate analyses that included as covariates gender, age, race and ethnicity, relationship status, parental education, and social desirability bias, as well as the number of days the student had spent in medical school at the time of survey completion to account for differences in exposure to the medical school environment. In model 3, we examined the multivariate association between sexual identity and depression, anxiety, and self-rated health, while controlling for all the covariates included in model 2 as well as the social stressors included in this study. For one of the dichotomous outcomes (depressive symptoms), model 3 did not converge (which is known to occur with this GLM approach), and therefore we employed a GLM with a Poisson distribution and a robust variance estimator, an alternative analytic approach that has been demonstrated to reliably estimate relative risks for binary outcomes.^22,23

Acknowledging the consistently documented gender differences in mental health and self-rated health,^24–30 we also conducted a gender-stratified analysis in which we modeled the associations between sexual orientation and the outcomes considered in the study separately for men and women.

All analyses took into account the sampling probability, stratification, and clustering in the two-stage design of CHANGES. We obtained 95% confidence intervals (CIs) and P values for the model-estimated associations between each outcome and the independent variable. We set the statistical significance threshold a priori at alpha equal to .05. We conducted all analyses in Stata version 12 (StataCorp LP, College Station, Texas).

Results

The demographic characteristics of the first-year students in the CHANGES sample were similar to the demographics of all students who matriculated at U.S. medical schools in 2010, as reported by the AAMC³¹ (for a comparison, see Supplemental Digital Appendix 3 at https://links.lww.com/ACADMED/A258). Sexual orientation data were available for 4,673 (98.8%) of the 4,732 students in the CHANGES sample; of these 4,673 students, 123 (2.6%) identified as homosexual, 93 (2.0%) as bisexual, 16 (0.3%) as other, and 4,441 (95.0%) as heterosexual. Overall, 5.0% (n = 232) of our analytic sample selected a category other than heterosexual. Characteristics of the sexual minority and heterosexual students in our sample are presented in Table 1. Compared with heterosexual students, sexual minority students had a higher mean age, were more likely to be men, and were less likely to be in a relationship or to have a parent with a graduate degree.

Table 1: Characteristics of 4,673 First-Year Medical Students by Sexual Identity, Medical Student CHANGE Study Baseline Survey, 2010a

Sexual minority students were more likely than heterosexual students to report depressive symptoms, anxiety symptoms, and low self-rated health (Table 2). For depression, 20.7% (48/232) of sexual minority students reported depressive symptoms compared with 12.7% (563/4,441) of heterosexual students (P < .001). In terms of anxiety, 10.8% (25/232) of sexual minority students reported anxiety symptoms compared with 6.1% (270/4,441) of heterosexual students (P = .004). Low self-rated health was reported by 10.1% (23/228) of sexual minority students compared with 6.4% (281/4,418) of heterosexual students (P = .03).

Table 2: Prevalence of Depression, Anxiety, Low Self-Rated Health, and Social Stressors Among 4,673 First-Year Medical Students by Sexual Identity, Medical Student CHANGE Study Baseline Survey, 2010a

Sexual minority students also experienced more social stressors than their heterosexual peers. Sexual minority students were more likely than heterosexual students to report being called names or insulted at least a few times a year (34.9% [80/229] versus 26.6% [1,161/4,368], P = .01) and were more likely to report being harassed or threatened at least a few times a year (22.7% [52/229] versus 12.7% [553/4,369], P < .001). In addition, sexual minority students were more likely than heterosexual students to report feeling a lack of companionship (53.0% [122/230] versus 42.1% [1,846/4,383], P = .001), left out (50.7% [116/229] versus 42.1% [1,843/4,382], P = .01), and isolated (53.7% [123/229] versus 42.8% [1,873/4,374], P = .001).

Results of the bivariate and multivariate analyses are presented in Table 3. In both models 1 and 2 (unadjusted and adjusted analyses, respectively), we found that sexual minority medical students were at significantly greater risk than their heterosexual peers of depression and anxiety symptoms as well as low self-rated health.

Table 3: Depression, Anxiety, and Low Self-Rated Health Among 4,673 First-Year Medical Students by Sexual Identity and Gender, Medical Student CHANGE Study Baseline Survey, 2010a

After adjusting for relevant covariates (model 2), we found that sexual minority students had a mean depression score 2.45 points higher (95% CI = 1.26–3.64) and had a significantly greater risk of being classified as having depressive symptoms (adjusted relative risk [ARR] = 1.59; 95% CI = 1.24–2.04) than heterosexual students. For anxiety, sexual minority students scored 1.67 points higher (95% CI = 0.62–2.73) and had a significantly greater risk of being classified as having anxiety symptoms (ARR = 1.64; 95% CI = 1.08–2.49) compared with heterosexual students. Finally, sexual minority students had a significantly greater risk of reporting low self-rated health (ARR = 1.77; 95% CI = 1.15–2.60) than heterosexual students.

In the gender-stratified multivariate analysis (model 2), we found that sexual minority men had approximately twice the risk of reporting depressive symptoms (ARR = 2.00; 95% CI = 1.45–2.77) compared with heterosexual men, but these two groups did not differ significantly in terms of reporting anxiety symptoms or low self-rated health. Sexual minority women were twice as likely as heterosexual women to report low self-rated health (ARR = 2.04; 95% CI = 1.14–3.64), but these two groups were not significantly different in terms of reporting depressive symptoms or anxiety symptoms.

When we controlled for social stressors (model 3), we found that the associations remained significant between minority sexual identity and depression score, depressive symptoms, anxiety score, and low overall health, although the magnitude of the associations decreased. The association between minority sexual identity and anxiety symptoms was no longer significant.

Discussion

In this national cross-sectional study of individuals entering U.S. medical schools, we found that 5% of students identified as a sexual minority, a proportion slightly higher than the general population estimate of 3.5%.³² Consistent with our hypothesis, sexual minority students were at significantly greater risk of depression and anxiety, as well as low self-rated health, compared with their heterosexual peers. Our findings are consistent with an established body of research that reports higher rates of depression and anxiety among sexual minority individuals compared with heterosexual individuals in the general population.^33–35

These disparities in mental health status have been attributed primarily to the stigma and discrimination faced by LGB individuals as a marginalized minority group in the United States.^1,6 Indeed, as hypothesized, we found that sexual minority medical students were significantly more likely than their heterosexual peers to report experiencing harassment, insults, and social isolation during the previous year. When we adjusted for these social stressors in our analysis, we found that sexual minority students’ increased exposure to stressors appeared to attenuate the observed association between sexual identity and depression, anxiety, and self-rated health. Our findings contribute to a growing body of evidence about the significant detrimental impact of minority stress on the health of sexual minority individuals.

Prior studies suggest that sexual minority medical students experience discrimination and other social stressors throughout their medical training.^36–38 In a survey conducted by the AAMC, 17% of LGB medical students described the social, personal, and learning environments at their institutions as hostile toward sexual minority students.³⁹ In a survey of LGB physicians, 15% reported experiencing harassment and 22% reported social ostracizing by professional colleagues.⁴⁰ Fear of discrimination may decrease LGB students’ comfort and willingness to disclose their sexual orientation, which has been identified as an important issue for these students.⁴¹

We found that male sexual minority students were significantly more likely than their heterosexual counterparts to report depressive symptoms, but we did not observe a corresponding difference between sexual minority and heterosexual female students. The greater risk of depression among male sexual minority students may be a consequence of sexual minority men’s greater exposure to discrimination and stigma, as reported in the literature.^42,43 On the other hand, female sexual minority students had a significantly greater risk of reporting low self-rated health than their heterosexual counterparts, a finding that was not observed for male students. More research is necessary to understand the source of this health disparity among women by sexual orientation.

The additional social stress experienced by sexual minority students may exacerbate the stress experienced by medical students overall. Numerous studies have demonstrated that medical students, in general, experience higher rates of mental distress, including depression and anxiety, than the general population of young adults.^4,5,44–48 The mental health burden among medical students has implications for their academic performance and retention. Several studies have found a link between mental health distress, academic difficulties, and dropping out of medical school.^3,49 Another potential consequence of psychological distress is burnout, which is a syndrome of emotional exhaustion, depersonalization, and reduced sense of personal accomplishment.⁵⁰ Sexual minority medical students and physicians may experience more burnout than their heterosexual colleagues as a result of stigma, concealment of their sexual orientation, and discrimination.

Our findings, along with previous research, suggest that sexual minority students may face greater difficulties in medical school due to psychological distress when compared with their heterosexual counterparts. This, in turn, may lead to the underrepresentation of LGB providers, undermining AAMC and AMA calls to increase sexual orientation diversity among future physicians.^9,11 Furthermore, findings from a number of studies demonstrate that greater contact with sexual minority individuals is strongly correlated with more positive attitudes of heterosexual individuals toward sexual minority groups.^51–54 Underrepresentation of sexual minority individuals among medical students and providers may limit opportunities for contact, which may, in turn, limit heterosexual physicians’ opportunities to increase their comfort and competence in interacting with LGB patients. Taken together, the underrepresentation of LGB providers and heterosexual providers’ discomfort in caring for LGB people may exacerbate sexual minority patients’ experiences of discrimination and receipt of poor-quality health care.^55–58

Increasing attention and resources have been committed to reducing psychological distress and improving well-being among U.S. medical students.^8,59,60 Although sexual minority students are likely to benefit from these general interventions, it is doubtful that these programs will effectively address the mental health disparities we observed between heterosexual and sexual minority students. Our finding that sexual minority students are beginning their medical education facing a greater mental health burden than their heterosexual peers demonstrates a need for early, targeted interventions. Medical schools can play a key role in promoting students’ well-being and academic success by implementing evidence-based interventions that protect and improve students’ mental health^61–65 and by creating a learning environment that is inclusive of all students, including sexual and gender minorities.⁶⁶ For example, medical education programs should implement policies that promote the equal treatment of lesbian, gay, bisexual, and transgender (LGBT) medical students, such as adopting nondiscrimination policies that include sexual orientation and gender identity/expression as protected classes, offering school-sponsored health insurance to students’ same-sex and different-sex partners on an equal basis, and including coverage of gender affirmation health care services in school-sponsored health plans.^67,68 Given that bias against lesbian and gay individuals persists among heterosexual medical students despite shifts in public opinion over the past decade,⁵⁴ medical schools can also promote respectful interactions among students by implementing diversity programs and adopting a zero-tolerance policy toward discrimination and harassment. In addition, medical schools can increase the visibility of sexual and gender minority people by hiring openly LGBT faculty and staff, creating LGBT resource centers, and including instruction on the health of LGBT people as part of the general curriculum. More information on strategies to enhance medical school experiences for LGBT students is available in a GLMA: Health Professionals Advancing LGBT Equality 2013 white paper.⁶⁷

This study has several limitations. First, the cross-sectional nature of the data limits our ability to examine causal and mediational relationships between sexual identity, mental health outcomes, and social stressors. While we consider the role that social stressors may play in attenuating the association between sexual identity and health, we cannot determine empirically whether the stressors do, in fact, mediate this relationship. It is possible that students who experience more psychological distress are also more likely to feel socially isolated or stigmatized.

Second, although the mental health and well-being of sexual minority individuals are shaped by a common experience of discrimination, combining all nonheterosexual students into one larger category may have obscured differences between them. Similarly, we believe that more research on the health and wellness of transgender medical students is necessary. Although we underscore that sexual orientation and gender identity are not equivalent constructs, transgender and gender-nonconforming people share LGB people’s historically marginalized social status as a result of their departure from dominant gender norms.¹ It is therefore likely that transgender students similarly experience a greater mental health burden than their cisgender (i.e., nontransgender) peers, and more research in this population is needed.

A third limitation may stem from our inability to ascertain and invite participation of all medical students entering the sampled schools in 2010, creating potential selection bias. Despite a robust response rate (81% of the students invited to participate and 55% of all first-year medical students at the 49 schools), it is possible that our results do not generalize to the students who did not participate. Nevertheless, the demographic characteristics of the CHANGES sample are similar to those of all students matriculating in U.S. medical schools in 2010, as reported by the AAMC,³¹ suggesting that our findings are likely generalizable to medical students overall.

Conclusion

Using baseline data from a national survey of students matriculating at 49 U.S. medical schools in 2010, we found disparities in depression, anxiety, and self-rated health between sexual minority students and heterosexual students. As articulated in an AAMC report, “diversity is an essential component for promoting excellence in medical education and accessible, quality health care.”⁹ Ensuring the mental health and well-being of sexual minority medical students is crucial to attracting, retaining, and graduating a diverse population of future physicians. Our findings suggest that medical schools should implement tailored programs aimed at improving the mental health and well-being of sexual minority students.

Acknowledgments: The authors would like to acknowledge Deborah Finstad for her invaluable contributions to the Medical School Cognitive Habits and Growth Evaluation Study.