Intersectional experienced stigma and psychosocial syndemic conditions in a sample of Black men who have sex with men engaged in sex work (BMSM-SW) from six US cities

Abstract

Black men who have sex with men (BMSM) in the United States experience a disproportionate burden of violence, substance use, physical and mental health conditions relative to other racial groups. BMSM who engage in sex work (BMSM-SW) experience a high burden of psychosocial conditions, sexually transmitted infections including HIV, and intersectional stigma. This analysis characterizes remuneration and client typologies for BMSM-SW, documents intersectional stigma experienced by BMSM-SW relative to other BMSM, and explores the impact of experienced intersectional stigma on the relationship between sex work engagement and psychosocial syndemic conditions (violence, polydrug use, and depression symptoms). Results show that a majority of BMSM-SW in the sample had female clients and that sex workers were more likely than other BMSM to hire another sex worker. BMSM-SW were more likely than other BMSM to report stigma attributed to race; sexuality; HIV status; socioeconomic status; and “other” attributes, and were more likely to report experiencing stigma across all settings assessed (schools; healthcare; employment; housing; police/courts; and in public/community). Intersectional stigma mediated the relationship between sex work engagement and psychosocial syndemic conditions, accounting for 49% (95% CI: 47.6–50.0%) of the relationship. Interventions for BMSM-SW should include resilience-building components to counteract the effects of intersectional stigma.

Introduction

Researchers have been increasingly studying men’s engagement in sex work, broadly defined as the exchange of sex for money, drugs or other goods, also called “exchange sex” or “transactional sex” (Baral et al., 2015; Bobashev et al., 2009; Nerlander et al., 2017). There is increasing recognition of the multiple forms and contexts of sex work including street-based sex work, web-based escorting (i.e., Rentboy.com and similar services), and digital sex work, which has grown with the use of the internet and mobile geolocation applications (Meunier et al., 2021; Minichiello et al., 2013, 2015). Findings from previous research suggest that male sex workers experience a constellation of health disparities related to their mental health, substance use, and biopsychosocial conditions at the intersections of racism, violence, stigma, and sexually transmitted infections (STI), including HIV, when compared to men not engaged in sex work (Baral et al., 2015; Bauermeister et al., 2017; Mgbako et al., 2019; Raine, 2021). However, research specific to Black men who have sex with men who also engage in sex work (BMSM-SW) remains limited.

One factor understudied in the health outcomes of men engaged in sex work is the impact of multiple, intersecting forms of stigma, which includes stigmas related to race, sexuality, social class and HIV (Brookfield et al., 2020). The relationship of racism, sexuality and stigma has already been documented among BMSM (Arscott et al., 2020). Researchers have begun developing tools to assess the multiple layers of stigma experienced by individuals engaged in sex work, although many of these studies have focused on female sex workers (Kerrigan et al., 2021; Logie et al., 2011; Oga et al., 2020). Qualitative research with male sex workers notes that multiple types of perceived and enacted stigmas based on sex work engagement, substance use, and economic distress, among other characteristics, may limit engagement in health services including HIV prevention (Underhill et al., 2015; Underhill et al., 2014). Further, the criminalization of sex work in the US provides an additional opportunity for structural stigma and exposure to entities such as law enforcement, potentially adding to the already stigmatized nature of male-male sex (Crofts, 2014). The required clandestine nature of sex work may also expose male sex workers to violence which male sex workers are reticent to report because they may anticipate discrimination, inaction, or imprisonment by law enforcement for engaging in illegal activities (Decker et al., 2015). There exists a need to understand experienced intersectional stigma within the context of sex work engagement for BMSM-SW in the US, particularly in social and structural settings (e.g., receiving healthcare, interactions with law enforcement) identified in previous research, which compels an intersectional approach to data collection and analysis.

Intersectionality, or the framework to explain interlocking experiences of privilege or discrimination based on multiple identities (e.g., race, social class, sexual orientation, gender) (Bowleg, 2012; Crenshaw, 1989), is an important lens to view experiences of stigma among BMSM-SW, as these experiences may differ from BMSM with no recent history of sex work (Underhill et al., 2015). An intersectional approach is useful to delineate health disparities unique to BMSM-SW and other BMSM. Intersectional stigma, an extension of intersectionality theory, has been proposed as a method to describe the complexity of overlapping social stigmas (e.g., based on race and socioeconomic status) on health outcomes like psychosocial conditions (Alvidrez et al., 2021). Ameliorating the intersectional stigma experienced by sexual and gender minorities requires an understanding of individual and structural factors, including locations where stigma is encountered and the concomitant identity-based attributions for stigmatization.

The factors affecting health outcomes may best be reviewed using syndemic theory. Syndemic theory posits that co-occurring epidemics of biological and ecosocial conditions converge to worsen health outcomes (Singer et al., 2017; Singer et al., 2006). The theory is most often used to explore psychosocial conditions such as issues in mental health, violence, substance use and risk behavior (Tsai & Burns, 2015). Syndemic theory has been applied to studies of health outcomes among MSM, including male sex workers noting that syndemic conditions have been associated with poor health outcomes (Chandler et al., 2020; Dyer et al., 2012; Dyer et al., 2020; Walters et al., 2020) and lack of engagement in the HIV continuum of care for BMSM (Quinn et al., 2018). A recent systematic review identified “a complex syndemic of social disadvantage and exclusion acting to produce and reinforce health disparities related to sexual health” where experiences of stigma among male sex workers were a predominant barrier to receiving effective sexual healthcare (Brookfield et al., 2020, p. 682). Emerging literature is applying intersectionality and syndemic theory to the study of MSM health (Kline, 2020; Quinn, 2019) including the association of intersectional stigma with the production of syndemics (Viswasam et al., 2020).

Previous work with BMSM-SW

A previous analysis of data from Promoting Our Worth, Equality and Resilience (POWER), a cross-sectional study of BMSM and Black transgender women, focused on understanding the relationship of sex work engagement and past-year sexually transmitted infection (STI) burden among a sample of BMSM. Compared to BMSM with no recent sex work engagement, BMSM-SW had significantly higher odds of reporting past-year physical assault, past-year intimate partner violence, past three-month polysubstance use, and past-week depression symptomatology, comprising a psychosocial syndemic (Chandler et al., 2020). The analysis found that psychosocial syndemic conditions partially mediated the relationship between sex work engagement and STI burden, suggesting the need to address psychosocial syndemic conditions as a part of interventions to reduce STI burden.

The current analysis

Literature is still emerging regarding men in sex work, spurring a need for further research characterizing how male sex work is performed, remunerated, and stigmatized, especially among BMSM-SW in the US. To provide a more complete understanding of the sex work ecology of BMSM-SW, more specificity in remuneration and client gender is necessary (Baral et al., 2015). This analysis describes the social ecology of sex work for a sample of BMSM-SW in the US by detailing: (1) types of remuneration for sex work; (2) gender of sex work clients; and (3) a comparison of the odds of hiring sex workers among BMSM-SW and other BMSM. As studies of experienced stigma among BMSM-SW are lacking but may provide context to the social and structural factors impacting their health, this analysis also includes: (4) correlates and associations of experienced stigma by sex work engagement; (5) settings of experienced stigma; and (6) an assessment of the contribution of experienced intersectional stigma to the relationship between sex work engagement and psychosocial syndemic conditions.

Methods

Participants and sampling

Data were collected from BMSM attending annually occurring Black Pride events in six cities across the United States from 2014–2017 as part of the POWER Study (Bukowski et al., 2018; Matthews et al., 2016). Cities/district of data collection were Philadelphia, PA, Houston, TX, Detroit, MI, Washington, DC, Memphis, TN, and Atlanta, GA. Data collection identified intercept zones and employed time-location sampling (TLS) to randomly sample BMSM attending Black Pride events in 2-hour time blocks (Magnani et al., 2005). Potential participants were approached, and those interested were informed of the study. Participants were screened using audio computer-assisted self-interview (ACASI) enabled tablets. Eligible participants were consented to the anonymous study using the ACASI tablets. Once consented, participants completed a 20-minute behavioral survey. Participants were also invited to complete an HIV screening test. Participants were compensated $10 for a completed survey and $10 for a completed HIV screening. All consent documents and procedures were approved by the institutional review board at the University of Pittsburgh.

Among the 5,858 MSM and transgender participants with submitted surveys, this analysis includes participants who: (1) were aged 18 and over; (2) reported having sex with a man in the previous 12 months; (3) were assigned male sex at birth; (4) identified as male at the time of data collection; and (5) identified as Black or African American, including participants who were multiracial. Using easily recalled information, participant data was screened for duplicate entries by the creation of a unique identifier code so that only one record was retained for each code as described in previous literature (Dilley et al., 2000; Hammer et al., 2003). This analysis removed 301 duplicate responses for an analytical sample of N=4430. More detailed methods for POWER data collection can be found elsewhere (Bukowski et al., 2018; Chandler et al., 2020; Matthews et al., 2016).

Measures

Demographics.

All participants were categorized by their city and year of data collection. In addition, all participants were asked demographic information. In previous research with this sample comparing BMSM-SW (n=254) and other BMSM (n=4176), BMSM-SW were more likely to be older, identify as Hispanic/Latinx, have an income below the federal poverty threshold, report sex with men and women, have bisexual identity and report living with HIV. More information on demographic comparisons between BMSM-SW and other BMSM can be found elsewhere (Chandler et al., 2020).

Past-year sex work engagement and client gender.

All participants were asked if they had engaged in transactional sex in the previous year using two entry questions: “In the past 12 months, did you ever give or take money, drugs or other goods for sex with a female partner?” and “In the past 12 months, did you ever give or take money, drugs or other goods for sex with a male partner?” with response options being “Yes”, “No”, “Don’t know” and “Refuse to answer.” Participants were then asked if they had received money, drugs or other goods from the corresponding gender client. Participants who responded “Yes” were coded to have engaged in past-year sex work. Partner gender was assessed by using the entry questions and follow-up questions specific to receiving money, drugs, or other goods from a female or male partner.

Remuneration type.

In the inaugural year of the study (2014), participants who answered that they had engaged in sex work were asked to specify if they had exchanged sex for money only, drugs only, some other good(s) only, or combinations of these options. In later years, a single question assessed “money, drugs, or other goods.” Remuneration analyses assess 2014 data only (n=82 BMSM-SW).

Hired sex worker.

Participants who responded that they had given or received a form of payment for sex in the past year were asked a series of follow-up questions to clarify the direction and gender of the partner which included: “Yes, I gave money, drugs or other goods for sex” with male or female partners. Participants who responded that they had given money, drugs, or other goods were recoded to have hired a sex worker.

Forms of experienced stigma.

All participants were asked if they had experienced discrimination (as a form of experienced stigma) in the previous year due to any of five identity-related attributions adapted from the Experiences of Discrimination (EOD) scale (Krieger, 1990; Krieger et al., 2005). For each attribution, participants were asked if they had experienced enacted stigma based on the identity-based attribution with response options being “Yes”, “No”, “Don’t know” and “Refuse to answer.” For example, “In the past year, have you experienced discrimination, been prevented from doing something, or been hassled or made to feel inferior because of your race?” was used to assess experienced stigma attributed to race, and the stem was replaced for each form of stigma attribution. Attributions included stigma experienced based on race, sexuality “you are gay or have sex with men?”; HIV status “your HIV status?”; social class “your income or social class?”; or some other stigma “because of some other reason?”. When participants indicated that they had experienced stigma, they were also asked to select the setting of the stigma from a generated list adapted from the EOD scale.

Settings of stigma.

Participants who identified that they had experienced stigma were asked if the stigma had taken place in any of the following settings: (1) at school or in educational settings; (2) receiving medical care; (3) at work/place of employment; (4) interacting with the police or the courts; (5) getting housing; or (6) on the street or in public. For analyses, employment-related questions were aggregated to reflect any employment-related discrimination. In the first survey year (2014), the setting where stigmatization occurred prefaced identity-based attributions for stigmatization.

Psychosocial syndemic conditions.

Psychosocial syndemic variables were included in three broad categories: violence, substance use and mental health. Violence was comprised of two questions. Physical assault was assessed with the question: “In the past year have you been physically assaulted (hit, kicked, beat up or in any other way physically harmed)?” Intimate partner violence (IPV) was assessed with the question: “In the past year, have you been in a relationship with a partner who has ever hit, kicked, slapped, beaten or in any other way physically assaulted you?” with response options being “Yes”, “No”, “Don’t know” and “Refuse to answer” for both questions. Polysubstance use was assessed by participant reports of past three-month use of two or more of the following substances: marijuana, MDMA/ecstasy, GHB, inhalant “poppers”, amphetamines, powder cocaine, crack cocaine, heroin, or opioids not prescribed to the participant. Depression symptomology was assessed with the likelihood of moderate to severe depression symptoms in the past week using the abbreviated Center for Epidemiological Studies Depression 10-item scale (CESD-10) (Andresen et al., 1994). Moderate to severe depression likelihood was indicated by a score of 10 or greater in a range of 0–30 as described in previous literature (Andresen et al., 1994). All condition assessments were dichotomized (e.g., polysubstance use or no polysubstance use).

Statistical analysis

All analyses were completed using Stata version 16SE (StataCorp, 2019) and SAS 9.4 (SAS Institute, 2013). Conventional frequencies were used to describe the type of remuneration for sex work in the first year of analysis (2014) and the gender of the clients for BMSM-SW. Chi-square analyses compared the full sample of participants (BMSM-SW vs. BMSM with no recent history of sex work) based on those who had hired sex workers in the previous year, experiences of stigma, and settings where stigma was experienced. Bivariate analyses of the summed attributions for experienced stigma and settings of stigma were compared by sex work engagement using chi-square analyses. Multivariable logistic regressions adjusted for city, year of sampling, Hispanic/Latinx ethnicity, age, income, reported sex with men and women, and HIV status were explored to understand differences in experienced stigma and locations of stigma by sex work engagement.

A mediation analysis using a structural equation model (SEM) was conducted to assess if latent experienced intersectional stigma mediated the relationship between sex work involvement (predictor) and psychosocial syndemic (latent outcome, comprised of past-year assault, past-year intimate partner violence, past three-month polysubstance use, and past-week depression symptomology). Experienced intersectional stigma (variable comprised of identity-related attributions for stigma including race; sexuality; HIV status; class; and other stigma) was also operationalized as a latent construct. These latent constructs allowed for inclusion of all mediator- and outcome-related response patterns while maximizing parsimony. The SEM was conducted using maximum likelihood estimation with an observed estimation matrix (OEM), adjusted for potential confounders (city, sampling year, Hispanic/Latinx ethnicity, age, income, reported sex with men and women, and HIV status) on mediator and outcome variables. Model fit for SEM was assessed via standardized root mean residuals (SRMR), assuming acceptable model fit at SRMR < .08; the model’s χ2, root mean squared error of approximation (RMSEA) and comparative fit index (CFI) were also assessed.

Lastly, a sensitivity analysis using established procedures for causal mediation with four-way decomposition was conducted (Jackson & VanderWeele, 2019; Valeri & VanderWeele, 2013; VanderWeele & Vansteelandt, 2009; VanderWeele & Vansteelandt, 2010). Sex work engagement was the main predictor, and a summed variable of psychosocial health conditions (0–4 of physical assault, intimate partner violence, polydrug use, and depression symptoms) was operationalized as the outcome in this model. We created a variable summing identity-related attributions for experienced stigma (total: 0–5 of race-, sexuality-, HIV-, SES-, and other-based stigma), and used this as the mediator in a cross-sectional Poisson model. This analysis offered an opportunity to assess decompositions including a controlled direct effect (the expected inequality in outcome due to sex work engagement, controlling for effects of both mediation and interaction between mediator and outcome variables) and a pure indirect effect (the mediating effect of intersectional stigma on psychosocial burden, controlling for direct and interaction effects). We also reported the portion eliminated, constituting the proportion of the effect of sex work on the outcome (summed psychosocial syndemic conditions) that would be eliminated if intersectional stigma levels among BMSM-SW were reduced to levels reported by BMSM not engaged in sex work. Sensitivity analyses were adjusted for covariates and were conducted using SAS 9.4.

Results

Table 1 describes the ecology of sex work in the sample. The sample included a total of N=4430 BMSM, of whom 5.7% (n=254) reported engaging in sex work. During the 2014 data collection year, BMSM-SW (n=82) were asked about the specific method of remuneration exchanged, with the majority (n=64) identifying that they received “money only” for sex work transactions, followed by “drugs only” (n=7); all other combinations of money, drugs and/or other goods were reported by four or less participants.

Table 1.

Ecology of sex work engagement: Remuneration characteristics, partner gender, correlations between sex work engagement and enacted stigma attributions and settings among sexually active, BMSM in 6 U.S. cities, 2014–2017 (N=4430).

	BMSM-SW (n=254)	%	Non-SW BMSM (n=4176)	%	Significance
Variable	n	%	n	%	Chi-square value; p-value
Type of remuneration for sex work (2014 BSMW-SW only: n=82)			-	-	-
Money only	64	78.0%
Drugs only	7	8.5%
Other goods only	1	1.2%
Money and drugs only	4	4.9%
Money and other goods only	1	1.2%
Drugs and other goods only	1	1.2%
Money, drugs, and other goods	4	4.9%
Gender of sex work clients (BMSM-SW only)			-	-	-
Sold sex to males only	109	42.9%
Sold sex to females only	49	19.3%
Sold sex to both males and females	97	38.2%
Hired sex
Hired any sex (overall)	24	9.4%	88	2.1%	52.2; p<.001
Hired sex from males only	13	5.1%	56	1.3%	-
Hired sex from females only	9	3.5%	12	0.3%	-
Hired sex from males and females	2	0.8%	20	0.5%	-
Reporting enacted stigma related to an identity attribute
Race	114	44.9%	817	19.6%	199.0; p<.001
Sexuality	128	50.4%	813	19.5%	92.0; p<.001
HIV status	102	40.3%	435	10.5%	136.0; p<.001
Socioeconomic status	97	38.3%	397	9.5%	199.1; p<.001
Other attribution (unspecified)	70	27.7%	405	9.7%	79.9; p<.001
Reporting enacted stigma from a specific setting
School/education	57	22.4%	393	9.4%	44.3; p<.001
Medical/healthcare	50	19.7%	170	4.1%	123.3; p<.001
Employment	107	42.1%	639	15.3%	122.5; p<.001
Law enforcement/police	74	29.1%	311	7.5%	141.3; p<.001
Housing	48	18.9%	179	4.3%	104.7; p<.001
Public/community	112	44.1%	680	16.3%	125.6; p<.001

Notes: SW = sex work; remuneration type only assessed in year 1 of the study (2014, n=82)

Gender of sex work partners and proportion of sex work hiring is included in Table 1. Among BMSM-SW, 42.9% (n=109) reported only male clients, 19.3% (n=49) reported only female clients, and a substantial portion, 38.2% (n=97), reported having both male and female clients. In a full-sample comparison of sex work hiring, BMSM-SW were more likely to report hiring other sex workers than BMSM not engaged in sex work (9.4% versus 2.1% respectively, χ2 = 52.2, p<.001).

Table 1 also shows that, relative to BMSM not engaged in sex work, BMSM-SW participants had significantly higher likelihood of reporting experienced stigma by each possible identity-based attribution, including race (44.9% versus 19.6%, χ2 = 199.0, p<.001), sexuality (50.4% versus 19.5%, χ2 = 92.0, p<.001), HIV status (40.3% versus 10.5%, χ2 = 136.0, p<.001), socioeconomic status (38.3% versus 9.5%, χ2 = 199.1, p<.001) or some other reason (27.7% versus 9.7%, χ2 = 79.9, p<.001). Similarly, relative to BMSM not engaged in sex work, BMSM-SW had higher rates of experiencing stigma in each possible setting, including in school/education (22.4% versus 9.4%, χ2 = 44.3, p<.001), in healthcare settings (19.7% versus 4.1%, χ2 = 123.3, p<.001), in employment (42.2% versus 15.3%, χ2 = 122.5, p<.001), in interactions with law enforcement/police (29.1% versus 7.5%, χ2 = 141.3, p<.001), in housing (18.9% versus 4.3%, χ2 = 104.7, p<.001) and in the community-at-large/public spaces (44.1% versus 16.3%, χ2 = 125.6, p<.001).

Experiences of Stigma

Table 2 denotes the adjusted odds ratios (aOR) and associations of experienced stigma by sex work engagement with 95% confidence intervals (CI). Adjusting models for city and year sampled, and demographic considerations (e.g., age, Hispanic ethnicity, HIV status), BMSM-SW had higher odds than BMSM not engaged in sex work of reporting experienced stigma related to race (aOR=3.07, 95% CI: 2.29, 4.12); sexuality (aOR=3.66, 95%CI: 2.74, 4.88), HIV status (aOR=5.01, 95% CI: 3.67, 6.85), socioeconomic status (aOR=4.77, 95% CI: 3.46, 6.56), as well as “other” (unspecified) stigmatized attributes (aOR=2.52, 95% CI: 1.79, 3.54).

Table 2.

Multiple logistic regressions showing adjusted odds, by sex work engagement, of reporting enacted stigma related to specific attributes and specific settings: sexually active BMSM in 6 U.S. cities, 2014–2017 (N=4430).

	aOR	s.e.	Z	95% CI (lower)	95% CI (upper)	p-value
Reporting enacted stigma related to an identity attribute
Race	3.07	0.46	7.53	2.29	4.12	p<0.001
Sexuality	3.66	0.54	8.83	2.74	4.88	p<0.001
HIV status	5.01	0.80	10.12	3.67	6.85	p<0.001
Socioeconomic status	4.77	0.78	9.58	3.46	6.56	p<0.001
Other attribution (unspecified)	2.52	0.44	5.31	1.79	3.54	p<0.001
Reporting enacted stigma from a specific setting
School/education	1.99	0.36	3.80	1.39	2.84	p<0.001
Medical/healthcare	3.76	0.76	6.58	2.54	5.59	p<0.001
Employment	3.47	0.52	8.29	2.58	4.65	p<0.001
Law enforcement/police	4.44	0.76	8.75	3.18	6.20	p<0.001
Housing	3.13	0.66	5.44	2.08	4.73	p<0.001
Public/community	3.71	0.55	8.83	2.78	4.97	p<0.001

Notes: aOR = adjusted odds ratio; all models adjusted for city and year sampled, Hispanic/Latinx ethnicity, age≥40, low-income status, bisexual behavior, and HIV status.

Settings of Stigma

Table 2 also shows that, compared to BMSM with no recent sex work engagement, BMSM-SW had higher odds of reporting stigmatization in all settings. Compared to BMSM with no recent sex work, BMSM-SW were more likely to report stigmatization in education/school settings (aOR=1.99, 95% CI: 1.39, 2.84), healthcare settings (aOR=3.76, 95% CI: 2.54, 5.59), employment settings (aOR=3.47, 95% CI: 2.58, 4.65), law enforcement/police settings (aOR=4.44, 95% CI: 3.18, 6.20), housing (aOR=3.13, 95% CI:2.08, 4.63) and in public/community settings (aOR=3.71, 95% CI: 2.78, 4.97). A post-hoc analysis of frequencies of experienced stigma by attribution and setting among BMSM-SW shows that the most common attribution-by-setting intersections were stigmatization attributed to minority sexuality occurring in public settings (n=64; 25.2%); stigmatization attributed to minority sexuality in employment settings (n=46; 18.1%); stigmatization attributed to race in public settings (n=43; 16.9%); stigmatization attributed to income or social class occurring in public settings (n=39; 15.4%); and stigmatization attributed to race in employment settings (n=37; 14.6%) (data not shown).

Mediation Analysis

Figure 1 illustrates results of the structural equation model (SEM) of the relationship between sex work engagement (predictor) and psychosocial syndemic (latent outcome) mediated by experienced intersectional stigma (latent mediator). Model fit was considered good: Standardized Root Mean Square Residual (SRMR)=0.043; Root Mean Square Error of Approximation (RMSEA)=0.068; Comparative Fit Index (CFI)=0.874.

Figure 1.

Structural equation model of total and indirect effects between sex work involvement (predictor), intersectional stigma attributions (latent mediator), and psychosocial syndemics (latent outcomes) in a mixed-serostatus population of BMSM in 6 U.S. cities (n=4430), 2014—2017.

Note: Model adjusted for city, year, HIV status, ethnicity, income, bisexual behavior, and age.

Among BMSM, there was a significant total effect between sex work engagement and the psychosocial syndemic (B = 0.15 ± 0.01; p < .001). There were also significant total effects between sex work engagement and reports of experienced intersectional stigma [stigma attribution] (B = 0.25 ± 0.02, p<.001) and between experienced intersectional stigma and the psychosocial syndemic (B = 0.29 ± 0.02, p<.001). There was a significant indirect effect of sex work engagement on the psychosocial syndemic, through experienced intersectional stigma (B = 0.07 ± 0.01, p<.001).

Table 3 displays the extent of the mediation by experienced intersectional stigma of the relationship between sex work engagement and the psychosocial syndemic among BMSM-SW. Experienced intersectional stigma was estimated to account for 49.0% (95% CI: 47.6%, 50.0%) of this relationship. Appendix 1 shows all total, direct, and indirect effects for this SEM.

Table 3.

Total and indirect path coefficients and standard errors in pathways between sex work engagement, intersectional stigma attributions (latent mediator), and psychosocial health syndemic among BMSM in the POWER study, 2014—2017 (n=4430). Beta coefficients and respective standard errors reported.

Contribution of intersectional to the relationship between past-year sex work engagement and psychosocial health syndemic (latent variable)
Predictors and covariates	Total effects		Indirect effects in model adjusting for mediator (intersectional stigma attributions)
	B (SE B)	p	B (SE B)	p	Proportion of effect mediated by intersectional stigma (% and 95% CI)
Past-year sex work engagement	0.15 (0.01)	<0.001	0.07 (0.00)	<0.001	49.0% (47.6%, 50.0%)

Notes: Model adjusted for year, city, income, Hispanic/Latinx ethnicity, HIV positive status, bisexual behavior, and age ≥40. SRMR=0.043; RMSEA=0.068; CFI=0.874; Chi-square=<0.001.

Appendix 2 shows results from a sensitivity analysis describing causal mediation by experienced intersectional stigma (summed) of the relationship between sex work engagement and psychosocial syndemic conditions (summed). Results from a 4-way decomposition show that there was a significant effect of sex work on summed psychosocial syndemic conditions (controlled direct effect): estimate=1.68; s.e.=0.28; p<0.001. There was a significant pure indirect effect (estimate=0.86; s.e.=0.12; p<0.001), signifying that intersectional stigma mediated the relationship between sex work engagement and overall summed psychosocial syndemic conditions. The estimated portion eliminated was 37.9% (95% CI: 29.9%, 45.8%; p<0.001), indicating that over one-third of the effect of sex work engagement on psychosocial syndemic conditions would be eliminated if intersectional stigma levels among BMSM-SW were reduced to mean intersectional stigma levels reported by BMSM not engaged in sex work.

Discussion

Results from this study add valuable context to the lived experiences of Black men who have sex with men engaged in sex work. Our findings provide additional data on sex work ecology, including remuneration typologies, client gender, the reciprocal nature of buying and selling sex, and experienced intersectional stigma. Prior research from this sample has shown that, compared to their counterparts not engaged in sex work, BMSM-SW experience profoundly higher rates of negative psychosocial health conditions. This study contextualizes these findings by demonstrating that psychosocial health disparities among BMSM-SW are substantially explained by their higher rates of experienced intersectional stigma when compared to other BMSM. This mediation analysis likely reflects exposure to intersecting stigmas that shape the context in which sex work takes place. This confirms the central role that experienced intersectional stigma plays in BMSM-SW’s psychosocial outcomes. This mediation analysis supports data that has suggested that men engaged in sex work are more likely to experience stigma and violence than other MSM (Brookfield et al., 2020; Raine, 2021; Walters et al., 2021) all of which will require additional study in BMSM-SW. Results from our sensitivity analysis show that 37.9% of the effect of sex work engagement on psychosocial outcomes would be eliminated if intersectional stigma levels among BMSM-SW were reduced to levels experienced by BMSM not engaged in sex work. This suggests that designing and implementing intersectional stigma reduction interventions tailored to BMSM-SW has high potential to impact their severe psychosocial health disparities.

BMSM-SW in this study appeared to show a reciprocal quality of the sexual economy, as BMSM-SW were more likely to have hired sex from other sex workers compared with BMSM with no recent history of sex work; only recently has this complex relationship been explored (Bond et al., 2019). Further, remuneration data identified that sex work typically centered around monetary transactions; however, a surprising finding is that most participants engaged in sex work reported having female clients. This finding is unique among the extant literature regarding men engaged in sex work, which have most often identified the clients of male sex workers as male, and provides confirmation that querying client gender is important to consider when conducting research with male sex workers (Baral et al., 2015).

In the framework of combination HIV prevention, these results add to the evidence behind the need to offer prevention and care strategies tailored to most-at-risk subpopulations of BMSM. BMSM-SW in multivariable analyses were significantly more likely to report stigma in healthcare settings than other BMSM (aOR=3.76, p<.001), which is consistent with other analyses (Brookfield et al., 2020). Noting the reciprocal nature of sex work in this sample, and the unique findings regarding the gender of clients, it is important to continue to characterize the risk environments wherein BMSM engage in both selling and buying sex, including frequency, types of venue, characteristics of transactional partners (including other sex workers), and holistic risk management strategies that encompass stigma, legal, psychosocial, and HIV risk domains. Given that nearly 40% of the sample of BMSM-SW was living with HIV (Chandler et al., 2020), it may be important to explore interventions for sex workers living with HIV as well as those who may be at increased risk for HIV acquisition that address the multicomponent considerations of sex work. For example, male sex workers who are living with HIV may require additional assistance in order to help them maximize safer sex practices with partners, including clients; and to help them minimize their risk of legal peril, as sex workers living with HIV may be more likely to be prosecuted under HIV transmission laws. Considering the changing climate faced by sex workers in the US, it is essential to assess the ways in which criminalizing sex work (and instituting felony charges on sex workers who are living with HIV under several states’ HIV transmission laws) serves as a policy-level barrier limiting the focus and development of innovative and necessary interventions for this population. Future research must also examine the importance of integrating strategies for the mitigation of intersectional stigma among male sex workers in order to increase HIV prevention and limit transmission, given the importance of HIV-related stigma (Beckwith, 2020; Greenwood et al., 2021).Nevertheless, there exist opportunities for universal intervention components that may address other forms of stigma and health conditions, such as violence or mental health challenges.

This analysis further differentiates the context that places BMSM-SW at increased risk of negative psychosocial outcomes when compared to other BMSM. The fact that not only were BMSM-SW much more likely to report all identity-based typologies of experienced stigma, but also were more likely to report experienced stigma in all settings surveyed, supports the importance of sex work-specific analyses and intervention design dedicated for sex workers. For example, while BMSM-SW and other BMSM both identified stigma in public most frequently (44.1% and 16.3% respectively), followed by employment settings (42.1% and 15.3% respectively), BMSM-SW cited stigmatization by law enforcement as the third-most stigmatized setting (29.1% and 7.5% respectively). In multivariable analyses of sex work engagement on settings for stigmatization, the largest adjusted odds ratio was for stigmatization from law enforcement (aOR=4.44, p<.001), which is consistent with previous literature citing concerns of reporting sexual and physical assault to authorities (Crofts, 2014). These findings demonstrate that BMSM-SW are engaged in a milieu where they are consistently stigmatized at higher rates than other populations.

We found that BMSM-SW had higher odds than other BMSM of reporting stigma due to unspecified attributions (“other” stigma). These results may indicate the presence of stigma related specifically to sex work. Recent research has begun to identify this phenomenon, but mostly among female sex workers (Kerrigan et al., 2021). Other stigma may have also been associated with substance use, as polysubstance use among BMSM-SW was higher than other BMSM in this sample. It has been established that stigma can be associated with substance use (Kulesza et al., 2013), and a recent study of working women, including sex workers, noted associations between substance use, sex work and perceived stigma (Benoit et al., 2015). Other factors which may require additional study include stigma related to desired sexual position, body composition, and gender presentation (Bhambhani et al., 2019; Brennan et al., 2015; Lichtenstein et al., 2018; Ramos et al., 2021).

Limitations

Despite the large sample size and important contribution, this analysis must be viewed with an understanding of specific limitations. Our secondary analysis was focused on experienced stigma, and was constrained by the available variables, which did not assess intersectional internalized or anticipated stigma domains. In addition, we allowed in our analyses for the possibility that certain potential attributions for stigmatization (i.e., HIV status and social class) could be selected by respondents not belonging to respective minoritized classes (i.e., HIV-negative participants; participants reporting high income). Though unlikely, it is possible for people who are HIV-negative to perceive stigmatization based on HIV status (for example, when receiving HIV prevention services in an environment geared toward PWH), just as it is possible for high-income individuals to report stigmatization based on class; conceptually, we chose not to presume who or why someone might report experiencing stigma. This study employed an inter-categorical approach to intersectional stigma, including experiences that could be discrete, without querying all 2-item, 3-item, 4-item and 5-item combinations, but rather a composite of the available experiences. This study did not include questions on the full context in which sex work took place (e.g., the specific settings of sex work such as online or street escorting); this limits our ability to assess how different conditions and drivers of sex work involvement are related to the outcomes in this analysis. The use of BMSM as a category limits a more nuanced analysis that takes into consideration heterogenous sexual and gender identities of partners. For example, among men who reported male and female clients, our survey limited the choices to “male” and “female” and did not include the full range of gender identities that exist. Because we did not conduct qualitative research with this sample, we are unable to provide more substantive context for the reciprocity in buying and selling sex that we found in this sample. Future mixed-methods research should explore more fully the characteristics of reciprocity in the sexual economy among BMSM-SW. The list of stigma typologies did not include a sex work-specific query, such as Oga et al. (2020) and Kerrigan et al. (2021), which may have been represented in the “other stigma” category. Additionally, there was not specific query of about sex worker identity (i.e., if the participant identified as a sex worker). In the identified syndemic, the general physical assault measure and the intimate partner violence measure, while separately queried, may contain respondent overlap in situations where participants were physically assaulted by intimate partners. Data were based on self-report, which is subject to recall bias and participant fatigue. Psychosocial syndemic conditions had different recall periods as suggested in previous literature, but the representativeness of these conditions could be improved with a longitudinal analysis. The participants in this study were recruited at Black Pride events in mostly eastern and southern US cities, which may influence the representativeness of the sample relative to the larger BMSM-SW population. Finally, the cross-sectional study design prevents conferring causal associations in this sample, but provides us a guide to related behaviors and outcomes for more in-depth and longitudinal analysis. In cross-sectional mediation analyses, it is not possible to determine the directionality of associations; it is possible, for example, that psychosocial health conditions may also act as a mediator of the relationship between sex work engagement and intersectional stigma, or that these relationships propel each other continually. Longitudinal samples of BMSM-SW are best suited to answer these and other questions.

Conclusion

Research and interventions that include the multiple stigmas faced by men engaged in forms of sex work must be improved. Overall, research on male sex work is limited, and research that focuses on BMSM-SW more limited still. The compounded inequities that are known to affect BMSM in general may be enhanced in BMSM-SW, and current literature does not contain data that characterize the extent and particular nature of these inequities for those involved in sex work. For example, in studying HIV, the use of BMSM as a category limits our understanding of the heterogenous identities of individuals and the implications these have for the risks and vulnerabilities they experience. This critique of the homogenizing effect of BMSM as a category has been explored by other authors (Baral et al., 2015) and may be counterbalanced by additional qualitative research focused on BMSM-SW identities and typologies; such work will be useful in exploring the “some other stigma” experienced by BMSM-SW. It will be essential to know, for instance, if there are common misconceptions about the men engaged in sex work, or if some male sex workers are already identifying resilience strategies.

Explorations among sex workers may need to include additional forms of stigma, such as internalized, and anticipated stigma in addition to experienced or enacted stigma as defined in the HIV stigma framework (Earnshaw et al., 2013). The current Ending the HIV Epidemic: A Plan for America (EHE) campaign highlights the importance of eliminating HIV-related stigma, however, to address the impact of HIV, STIs and other health disparities among BMSM-SW, an important strategy will be the alleviation of the compounded stigma of sex work though policy and public health practice complete with sex-positive and specific programming and resources. EHE presents an important opportunity to engage with communities at the intersection of HIV and sex work decriminalization, using relevant data for decision making and policy development. New research specifically focused on male sex workers (e.g., stigma attributed to sex work, sex worker identity) will allow for the identification of not only the myriad of stigmas faced, but what can be used to buffer the effects of such stigma.

Such research will be critical in designing multicomponent, combination prevention activities to address stigma and the associated psychosocial syndemic conditions noted in this analysis. Overall, the results of this analysis suggest opportunities for intervention design span multiple levels of the social ecology. Certainly, there are downstream examples in HIV programming, such as including multi-spectrum pre- and post-exposure prophylaxis, and comprehensive HIV/STI testing and care services, but also essential are resources for housing, substance use prevention, mental health treatment, employment opportunities, and education assistance concomitant with safer sex work education. Interventions designed at the individual-level for sex workers must include components to address the intersections of stigma faced by this population, by fostering resilience to these safety threats. There must also be a new understanding of sex work, in this era of online content creation where sex workers may offer direct services but also be creators of online content. This study also suggests that interventions that reduce public stigma toward sex workers would be useful, as it was the third-most reported stigma. We suggest sex-work specific research alongside the evolving field of sex work that will allow for a more dynamic approach to developing interventions alongside sex workers as new health challenges arise.