Assessing self-reported discrimination among men who have sex with men (MSM)

Abstract

Objectives:

To determine the prevalence of four forms of sexual identity discrimination among MSM in 23 US metropolitan statistical areas, examine racial/ethnic and socioeconomic disparities in each form of discrimination.

Methods:

We examined interview data collected during 2017 for National HIV Behavioral Surveillance (n = 10 029 respondents) and used generalized linear models to assess the association between the prevalence of reported discrimination during the previous 12 months and selected sociodemographic groups.

Results:

Overall, 34% of participants reported experiencing verbal discrimination; 16%, discrimination in a workplace, school, or a healthcare setting; and 8%, physical assault. MSM who had reported experiencing discrimination were most likely to be young, had achieved lower education, and had lower incomes. High prevalence of reported discrimination was associated with young age, lower education, lower household income, sexual identity disclosure, and lower perceived community tolerance of gay or bisexual persons.

Conclusion:

MSM discrimination affects different groups and occurs in multiple settings. Addressing discrimination should be an integral aspect of multifaceted efforts to improve MSM health.

Introduction

In investigating HIV infection risk factors, a research shift toward understanding the role of broader social and societal pressures on the experiences of individual MSM has occurred in recent years. Stigma, which occurs when a person is socially devalued on the basis of perceived negative attributes of their identity or behavior [1], is a central component of such understanding and has been theorized as a driver of HIV infection risk [2]. Stigma leads to discrimination, which occurs when persons are treated unfairly as they are perceived to belong to a particular group [3].

In the United States, despite indications that acceptance toward homosexuality has increased over time [4], discrimination toward sexual minorities is still common. A previous study of MSM discrimination across 20 metropolitan regions that used 2011 National HIV Behavioral Surveillance (NHBS) data reported that, during the previous year, 32% of participants had experienced verbal harassment; 23%, discrimination in businesses, work or school, or healthcare; and 9%, physical assault [5]. Other studies conducted during 2014 [6] and 2015 [7] reported similar levels of MSM discrimination in the United States. An analysis of HIV-positive persons determined that 32.1% of patients experiencing discrimination attributed it to their sexual identity [8].

Our analysis serves as an update to both the 2011 NHBS study as well as to the growing body of literature by identifying sociodemographic associations with MSM discrimination in the 23 metropolitan statistical areas (MSAs) participating in the 2017 MSM cycle of NHBS. Our objectives were to determine the prevalence of four specific forms of discrimination attributable to sexual identity among MSM in 23 MSAs and examine racial/ethnic and sociodemographic disparities in each discrimination form.

Methods

National HIV behavioral surveillance

Data for this study were collected during 2017 as part of NHBS, which conducts rotating cycles of biobehavioral surveys among MSM, persons who inject drugs, and heterosexually active adults at increased risk for HIV infection. MSM were recruited by using venue-based, time-space sampling (VBS) in 23 MSAs with high HIV prevalence: Atlanta, Georgia; Baltimore, Maryland; Boston, Massachusetts; Chicago, Illinois; Dallas, Texas; Denver, Colorado; Detroit, Michigan; Houston, Texas; Los Angeles, California; Memphis, Tennessee; Miami, Florida; Nassau-Suffolk Counties, New York; New Orleans, Louisiana; New York City, New York; Newark, New Jersey; Philadelphia, Pennsylvania; Portland, Oregon; San Diego, California; San Francisco, California; San Juan, Puerto Rico; Seattle, Washington; Virginia Beach, Virginia; and Washington, DC. Each MSA attained its own institutional review board approval.

Eligibility was limited to participants aged at least 18 years who resided in a participating MSA, had not already participated in NHBS during 2017, were able to complete the survey in English or Spanish, and were able to provide informed consent. Further details regarding the NHBS methodology have been published previously [9].

Measures

Participants were asked if, during the 12 months before interview, they had experienced any of five types of discrimination as someone knew or assumed that they were attracted to men. Out of these five questions, we developed four variables: verbal harassment, defined as name-calling or insults; service discrimination, defined as poor service received in restaurants, stores, or other businesses; workplace, school, or healthcare discrimination, defined as unfair treatment in the workplace, school, or a healthcare setting; and physical attack, defined as a physical attack or injury. Variables 1, 2, and 4 were standalone questions in the interview, whereas variable three combined the standalone workplace or school discrimination and healthcare settings discrimination questions. These three outcomes were combined as an interaction in a workplace, school, or healthcare setting might be more likely to involve some familiarity between persons rather than a service setting interaction that might be more likely to occur between strangers.

In addition to the four discrimination outcomes, we included seven sociodemographic variables as independent variables: age, race/ethnicity, education, household income, sexual identity (or whether the respondent identified as being heterosexual or bisexual versus gay), sexual identity disclosure to anyone, and self-perceived community tolerance of gay and bisexual persons. Age, race/ethnicity, education, and household income all have been previously associated with different types of discrimination [10–13] and are core components of persons’ social and cultural status. Sexual identity and disclosure are key aspects of personal self-esteem and social dimensions within the overall MSM population and have been linked to exposure to potential discrimination [14]. Finally, self-perceived community tolerance to gay and bisexual men was not originally examined in the NHBS study that used 2011 data. This measure can provide context to structural support or lack thereof for MSM discrimination [15,16].

Analyses

We assessed sociodemographic differences in each of the four types of MSM discrimination outcomes. Adjusted prevalence ratios (aPRs) and corresponding 95% confidence intervals (CIs) were calculated from log-linked Poisson regression models with generalized estimating equations clustered on VBS recruitment event and MSA and adjusting for age, race/ethnicity, education, and income. All analyses were conducted during October 2019.

Results

Our sample consisted of 10 029 respondents, after exclusion of 75 records of eligible participants because of missing data for at least one of the five discrimination items. The sample included a diversity of ages, races/ethnicities, education levels, and household incomes (Table 1). Referring to the 12 months before participation in the survey, 34% of participants reported experiencing verbal harassment; 15%, service discrimination; 16%, discrimination in the workplace, school, or a healthcare setting; and 8%, physical assault (Table 2). In the adjusted models, age was associated with each type of discrimination, with the prevalence of discrimination decreasing with increasing age. For instance, MSM aged 18–24 years were approximately twice as likely to report workplace or healthcare discrimination as MSM aged at least 50 years (aPR = 1.85; 95% CI 1.58–2.17). Additionally, black participants had the lowest prevalence of verbal harassment (aPR = 0.77; 95% CI = 0.72–0.84), followed by Hispanic/Latino participants (aPR = 0.80; 95% CI = 0.74–0.86). The percentage of persons who reported verbal discrimination was approximately identical between black and Latino participants at 31.9 and 31.5%, respectively. Participants with less than a high school education had higher prevalence of service discrimination (aPR = 1.42; 95% CI = 1.11–1.82), workplace, school or healthcare setting discrimination (aPR = 1.27; 95% CI = 1.01–1.60), and physical assault (aPR = 2.20; 95% CI = 1.66–2.90) than those with a bachelor’s degree or higher. Participants whose reported annual household income was less than $20 000 had higher prevalence of verbal harassment (aPR = 1.22; 95% CI = 1.13–1.32), workplace, school, or healthcare discrimination (aPR = 1.33; 95% CI = 1.17–1.50), and physical assault (aPR = 1.96; 95% CI = 1.62–2.36) than those whose annual household income was greater than $50 000. MSM living in communities self-perceived as intolerant of gay and bisexual persons were more likely to experience discrimination (aPR = 1.49; 95% CI = 1.37–1.61), compared with MSM living in communities self-perceived as tolerant.

Table 1.

Characteristics of MSM in 23 metropolitan statistical areas in the United States (n = 10 029 respondents).

Characteristic	No.	%
Age (years)
18–24	1569	15.6
25–29	2355	23.5
30–39	2980	29.7
40–49	1534	15.3
≥50	1591	15.9
Race/ethnicity
Black, non-Hispanic	3002	29.9
Hispanic/Latinoa	2606	26.0
Other/multiracialb	849	8.5
White, non-Hispanic	3519	35.1
Education level
Less than high school	292	2.9
High school	1997	19.9
Some college	3333	33.2
College degree or more	4404	43.9
Household income
<$20 000	2491	24.8
$20 000–$49 999	3298	32.9
≥$50 000	4161	41.5
Sexual identity
Heterosexual	111	1.1
Gay	8063	80.4
Bisexual	1818	18.1
Self-perceived community tolerance of gay and bisexual persons
Strongly agree/agree the majority of people are tolerant	7402	73.8
Neither agree nor disagree	1619	16.1
Strongly disagree/disagree the majority of people are tolerant	1000	10.0

MSA, metropolitan statistical area.

^aMen of Hispanic/Latino ethnicity might be of any race.

^bIncludes American Indian, Alaskan Native, Asian, Native Hawaiian, Pacific Islander, or multiple races.

Table 2.

Associations between characteristics of MSM and experiences of sexual identity discrimination (n=10 029 respondents).

	Verbal harassment				Service discrimination				Workplace, school, or healthcare discrimination				Physical assault
Characteristic	No.	Row %	aPRa,b	95% CI	No.	Row %	aPRa,b	95% CI	No.	Row %	aPRa,b	95% CI	No.	Row %	aPRa,b	95% CI
Age (years)
18–24	657	41.8	1.75	1.59–1.92	300	19.1	1.54	1.27–1.87	394	25.1	1.85	1.58–2.17	177	11.3	1.58	1.24–2.02
25–29	904	38.3	1.64	1.46–1.84	411	17.5	1.48	1.25–1.76	406	17.2	1.46	1.28–1.67	204	8.7	1.53	1.22–1.91
30–39	1015	34.0	1.45	1.31–1.60	452	15.2	1.30	1.10–1.53	426	14.3	1.30	1.18–1.43	235	7.9	1.40	1.14–1.72
40–49	425	27.6	1.18	1.07–1.31	190	12.4	1.07	0.86–1.33	162	10.6	0.98	0.86–1.11	114	7.4	1.34	1.10–1.64
≥50	383	23.9	–—c	–—	182	11.4	–—	–—	173	10.9	–—	–—	85	5.3	–—	–—
Race/ethnicity
Black, non-Hispanic	956	31.9	0.77	0.72–0.84	494	16.5	1.09	0.96–1.24	576	19.2	1.11	0.99–1.26	281	9.4	0.97	0.81–1.16
Hispanic/Latino	822	31.5	0.80	0.74–0.86	420	16.1	1.11	0.97–1.26	456	17.5	1.09	0.97–1.24	204	7.8	0.97	0.81–1.17
Other/multiracial	309	36.5	0.92	0.83–1.02	135	15.9	1.08	0.90–1.30	151	17.8	1.14	0.96–1.34	88	10.4	1.34	1.06–1.69
White, non-Hispanic	1268	36.1	–—	–—	478	13.6	–—	–—	493	14.1	–—	–—	234	6.7	–—	–—
Education
Less than high school	116	39.7	1.17	1.00–1.36	60	20.5	1.42	1.11–1.82	63	21.6	1.27	1.01–1.60	60	20.5	2.20	1.66–2.90
High school	686	34.4	1.02	0.94–1.11	308	15.4	1.08	0.94–1.23	367	18.4	1.09	0.96–1.24	210	10.5	1.33	1.09–1.61
Some college	1186	35.6	1.07	1.00–1.15	585	17.6	1.25	1.11–1.40	595	17.9	1.18	1.06–1.32	290	8.7	1.27	1.07–1.50
College deg or higher (Ref.)	1385	31.4	–—	–—	582	13.2	–—	–—	536	12.2	–—	–—	255	5.8	–—	–—
Household income
<$20 000	949	38.1	1.22	1.13–1.32	419	16.8	1.01	0.89–1.15	512	21.9	1.33	1.17–1.50	344	13.8	1.96	1.62–2.36
$20 000–$49 999	1117	33.9	1.08	1.00–1.15	516	15.6	0.98	0.87–1.09	521	16.9	1.08	0.97–1.21	224	6.8	1.09	0.90–1.32
≥$50 000	1279	30.7	–—	–—	590	14.2	–—	–—	509	13.7	–—	–—	236	5.7	–—	–—
Sexual identity
Heterosexual or bisexual	615	31.9	0.91	0.85–0.98	282	14.6	0.90	0.79–1.01	340	17.6	0.96	0.86–1.07	185	9.6	1.00	0.85–1.18
Gay	2750	34.1	–—	–—	1246	15.5	–—	–—	1342	16.6	–—	–—	628	7.8	–—	–—
Disclosed sexual identity
No	99	21.7	0.66	0.55–0.78	44	9.6	0.60	0.45–0.81	62	13.6	0.75	0.59–0.96	27	5.9	0.55	0.36–0.82
Yes	3275	34.2	–—	–—	1491	15.6	–—	–—	1626	17.0	–—	–—	788	8.2	–—	–—
Self-perceived community tolerance of gay and bisexual persons
Strongly agree/agree the majority of people are tolerant	447	44.7	1.49	1.37–1.61	266	26.6	2.06	1.81–2.34	288	28.8	1.85	1.62–2.11	161	16.1	2.07	1.72–2.49
Neither agree nor disagree	596	36.8	1.19	1.11–1.27	302	18.7	1.41	1.25–1.58	328	20.3	1.29	1.15–1.45	156	9.6	1.32	1.11–1.58
Strongly disagree/disagree the majority of people are tolerant	2330	31.4	–—	–—	966	13.1	–—	–—	1072	14.5	–—	–	497	6.7	–—	–—
All	3375	33.7	–—	–—	1535	15.3	–—	–—	1561	15.6	–—	–—	815	8.1	–—	–—

aPR, adjusted prevalence ratio; CI, confidence interval.

^aAdjusted for age, race/ethnicity, education, and income. Clustered on metropolitan statistical area and recruitment event.

^bClustered on metropolitan statistical area and recruitment event.

^cNot applicable.

Discussion

Overall, discrimination against MSM persists and was similar between the 2017 and 2011 NHBS studies. Our analysis illustrates that 34% of participants reported experiencing verbal harassment during the previous 12 months, compared with 32% in the 2011 NHBS study. Lastly, 8% of MSM reported experiencing a physical attack during the 12 months before the survey in both studies. The lack of major change between 2011 and 2017 is striking, given improving attitudes toward homosexuality over time [4] and the Obergefell v. Hodges US Supreme Court decision in 2015 [17].

Both this study and the 2011 NHBS study revealed that the likelihood of being young was highest among those who had experienced discrimination versus those who had not; similar patterns existed regarding lower education and income. The age effect is of particular interest and might be partially explained by older men having more effective social support networks than younger men, which might insulate them against discrimination [18]. Both analyses also indicate that participants who had reported discrimination were more likely to be white than black or Hispanic/Latino. This result is a particularly interesting similarity between our study and the 2011 NHBS study. The intersectionality between racism and sexual identity discrimination among MSM is often difficult to untangle, and why white MSM might experience higher relative discrimination than black or Hispanic/Latino MSM is unclear. Previous efforts limited to specific cities have reported that discrimination among black MSM is higher than among white MSM [19,20], which contradicts both NHBS studies. One possible explanation for our results is that groups who regularly experience discrimination might minimize it to preserve self-esteem or experience discrimination so commonly they do not notice it [21]. However, this hypothesis requires further testing and, overall, we support further investigation into this topic, especially given the disproportionate burden of HIV faced by racial/ethnic minority MSM.

In all four types of discrimination, participants who had reported experiencing discrimination were less likely to have a positive perception of their community’s tolerance of gay and bisexual persons, whereas participants who had not experienced discrimination were more likely to report a positive perception. This association might provide quantitative evidence that community characteristics are important structural context for discrimination, although it might also be a result of discriminated persons perceiving their community to be less tolerant.

Our study is subject to at least three limitations. Overall, as this study was designed to identify cross-sectional associations between MSM discrimination and selected respondent attributes, we cannot claim to have identified causal linkages between predictor and outcome. Additionally, because of the sampling methodology, we were restricted to recruiting urban-dwelling men who visited a gay-related event or venue; thus, our results cannot be extrapolated to the general population. Lastly, participants were solely questioned about enacted MSM discrimination. The prevalence of other forms of discrimination (e.g. internalized racism, internalized homophobia, or HIV-related discrimination) should be addressed in subsequent studies. We look forward to the examination of novel variables, methods, and samples so that the role of discrimination in MSM health can be further explored and understood.

Overall, despite improvements in social attitudes toward MSM in the United States, discrimination persists. The prevalence, type, and severity of MSM discrimination might change as societal norms continue to evolve, yet the minimal change is cause for concern. Thus, regular assessments or follow-up investigations can be useful in tracking temporal trends. As we look toward ending the HIV epidemic [22], addressing structural and social barriers like stigma and discrimination might help protect and empower MSM, improving mental and physical wellbeing.