In the United States (U.S.), men who have sex with men (MSM), including persons who identify as gay and bisexual, account for 67% of all HIV diagnoses (Department of Health and Human Services 2015). Among MSM, black/African American (hereafter referred to as “black”) and/or Hispanic/Latino (hereafter referred to as Latino) men who have sex with men (BLMSM) make up 64.1% (37.6% and 26.5%, respectively) of HIV diagnoses, compared with non-Hispanic white MSM who comprise 30.3% of HIV diagnoses (Department of Health and Human Services 2015). In New York City (NYC), MSM account for approximately 74% of all new HIV diagnoses (New York City Department of Health and Mental Hygiene 2015). HIV-related racial/ethnic disparities also persist; black and Latino MSM represent 72.5% (38.6% and 33.9%, respectively) of new HIV diagnoses in NYC compared with 21.3% of whites (New York City Department of Health and Mental Hygiene 2015). Increased HIV risk associated with condomless anal sex contributes to the disproportionate HIV diagnosis rate among MSM in general, including MSM of color (Centers for Disease Control and Prevention 2016).

Gay identity development can be an important element of psychological well-being for MSM (Garnets 2002). This process can significantly shape social interactions for MSM with other individuals from various communities (Howarth 2001). However, many communities do not accept or embrace same-sex relationships. As a result, gay-related stigma may occur in these communities leaving many MSM feeling socially ostracized (Herek 1984). Consequently, these negative experiences can contribute to psychological distress making gay identity formation challenging for many gay men (Brooks-Gunn and Garber 1999).

Gay identity formation can be particularly difficult for BLMSM because of the potentially challenging process of reconciling their sexual and ethnic identities (Garnets 2002; Vu et al. 2011). Similar to gay identity, ethnic identity is important for self-identification and sense of belonging for BLMSM (Parham and Helms 1985). Ethnic identity is salient particularly for blacks and Latinos living in the United States to help cope with negative social and structural circumstances related racial and ethnic discrimination (Gayles et al. 2016; Hatzenbuehler and Pachankis 2016). However, BLMSM also experience gay-related stigma and related sequelae (e.g., violence, micro aggression) from both within and outside their racial and ethnic group; such circumstances can potentially create personal distress for BLMSM during stages of identity formation (Glock 2001; Han 2008; Jeffries et al. 2013; Mizuno et al. 2015; Nadal et al. 2011; Quinn et al. 2015; Rosario et al. 2006). Therefore, gay identity formation for BLMSM in the United States can be significantly more challenging compared with their white MSM counterparts.

Engagement with established gay communities can also influence gay identity formation for BLMSM. However, the literature shows that gay community involvement, particularly with communities comprised of mostly white gay men, may be less prevalent for BLMSM (Lewis et al. 2015). For example, studies show that black MSM are more likely to select sex partners within their own racial group compared to MSMs from other minority racial groups and ethnic groups (Lewis and Wilson 2017; Marks et al. 2009; Wei et al. 2011). In addition, Latino MSM, particularly foreign-born Latino MSM, may not engage with a developed gay scene even though engaging in condomless anal sex is prevalent (Lewis and Wilson 2017; Rhodes et al. 2012). Also, migration-related HIV risk syndemic factors play a role in community engagement, but these have been understudied among MSM of color (Lewis and Wilson 2017). In addition, little is known about the relationship of gay identity formation and gay community involvement among BLMSM.

For BLMSM, gay identity formation may have significant implications for HIV risk; but limited existing data are inconclusive. For example, gay-identified MSM were more likely to report higher rates of condom use and taking recent HIV tests compared to non-gay identified MSM (Pathela et al. 2006). However, Hampton et al. (2013) reported that non-gay identified black MSM engaged in more HIV risk behaviors compared with gay identified black MSM in NYC (Hampton et al. 2013). For Latino MSM, gay identity may serve as a protective factor for HIV by acting as a psychological buffer against gay-related social stressors, and thereby result in more condom use, less substance use and less general anxiety (Rosario et al. 2001, 2006; Waldo et al. 2000). In a study comprised of majority white MSM, gay-identified MSM were more likely to have had an HIV test compared to non-gay identified MSM (Rietmeijer et al. 1998).

Due to conflicting data regarding gay identity and HIV risk for BLMSM, we sought to fill a gap by examining gay identity, gay community involvement, HIV risk, and selected social determinants of health among BLMSM living in NYC. Our goal was to inform HIV prevention strategies that may reduce HIV racial and ethnic disparities.

Data were collected from spring 2011 through summer 2012 as part of a larger study (Project BROTHA) that explored the relationship between interpersonal communication about HIV prevention and HIV testing among MSM who identify themselves as black/African American and/or Latino/Hispanic; and their social networks (ages 18–64 years) in the NYC area (Aholou et al. 2017; Nanin et al. 2009). The study was led by the Center for HIV Educational Studies and Training (CHEST) in partnership with Gay Men of African Descent (GMAD), a trusted local, community-based organization that offered services for gay-identified/same-gender loving men of color in NYC.

A purposive sampling strategy was used to recruit BLMSM. The research team used structured outreach recruitment that combined active and passive strategies online and offline (e.g., banner ads on websites, such as Adam 4 Adam, Manhunt.com; distribution of palm cards with study contact information at community events; and large gay-focused events such as NYC Gay Pride Parade). In addition, the team used respondent-driven sampling (RDS), a network-based recruitment method (Heckathorn 1997). Eligible participants were men who were: born male, ages 18–64 years, reported sex with a man at least once within the previous 3 months, had no HIV antibody test within the past year, were able to communicate in English, and were able to identify two other MSM from their social network. After signed consent, participants were enrolled and completed an Audio Computer-Assisted Self-Interview (ACASI) survey which obtained demographic information (e.g., age, sexual identity, educational background) and assessed psychosocial and behavioral issues.

We used the Gay Identity Questionnaire (GIQ) (Brady and Busse 1994) as part of the ACASI survey. The GIQ was derived from the Homosexual Identity Formation Model (HIF), which originally identified a six-stage process that an individual experiences during gay identity development (Cass 1979). Later, other studies reported data showing that the HIF six-stage process can be collapsed from six to two stages instead (Stages I and II) (Brady and Busse 1994). Therefore, for this study, data were collected from the men using six-stage Likert responses that were collapsed into two stages [Stage I (not fully accepting) and Stage II (fully accepting)] to describe gay identity for the purposes of data analysis. We examined the association between gay identity with several demographic (age, sexual orientation), social determinant factors (gay community involvement, employment, poverty, insurance, homelessness, incarceration), and HIV risk-related behavioral measures (condomless sex, injection drug use, sex exchange for food, drugs, shelter or money). We used logistic regression (adjusting for age) to calculate adjusted odds ratios (aOR) and 95% confidence intervals (CI) using SAS, version 9.3.

A total of 111 BLMSM participated in the study; 58.6% self-identified as homosexual, gay or queer; 33.3% self-identified as bisexual or other; 46.0% were unemployed; 19.8% had graduated from high-school; 40.5% had some college; 27.9% had a bachelor’s degree or higher. Eighty-two percent reported living above the poverty line, 12.6% reported being homeless in the past 3 months, and 72.1% report having health insurance coverage (Table 1). Additionally, the racial and ethnicity composition included 75.7% (n = 84) black/African American, 12.6% (n = 14) Hispanic/Latino, 9.9% (n = 11) identifying with both groups and 1.8% (n = 2) not reporting race or ethnicity (not shown in tables).

GIQ assessment indicated that 76.6% of BLMSM sample were “fully accepting” of their gay identity and 19.8% were “not fully accepting.” About 67.6% of participants reported high involvement with the gay community; 46.8% reported having condomless anal insertive sex in the past 3 months; 34.2% had condomless anal receptive sex in the past 3 months; 88.3% reported being negative on most recent HIV test; 7.2% reported exchanging sex for drugs, money, food or shelter, and 2.7% reported injection drug use in the previous 3 months (Table 1).

The regression model is presented in Table 2. We excluded injection drug use and recent HIV test results from the analyses due to low cell sizes. Our bivariate analyses indicated that age, sexual orientation and homelessness during past 3 months were associated with Stage II gay identity, but statistically insignificant in the multivariate model. The multivariate logistic regression models showed that Stage II gay identity was more likely seen among BLMSM with high involvement in the gay community (aOR 3.2; CI 1.00, 10.26) and less likely among BLMSM who exchanged sex for food or shelter (aOR 0.15; CI 0.02, 0.98).

Most BLMSM in our study (80%) were fully accepting of their gay identity. Our analysis revealed a strong link between fully accepting of one’s gay identity (Stage II) and high involvement in gay community. This finding is consistent with other studies indicating similar links between affirmative gay identity and having a strong social support system among black MSM (Crawford et al. 2002; Lapinski et al. 2010). Also, our findings are congruent with a study indicating high correlation between gay identity and community involvement among Latino MSM living in New York (Ramirez-Valles and Diaz 2005). Other studies identified social activism and socializing as factors related to gay community involvement among a sample of predominantly white MSM (LeBeau and Jellison 2009). Although social activism information was not part of data collection for this study, it may serve as a mediator between gay identity and gay community involvement and should be examined among BLMSM in future studies.

Our findings regarding gay identity and gay community involvement may be particularly salient for BLMSM due to racial/ethnic psychosocial challenges. Other studies indicate that disclosing sexual identity and engaging in HIV risk reduction behaviors are often inhibited by stigmatizing environments within some black and Latino communities, particularly in the context of homophobia and religiosity (Bogart et al. 2014; Jeffries et al. 2014). These anti-gay intra-group dynamics are further compounded by race- and ethnic-based discrimination that many MSM report encountering outside of their communities (Solorzano et al. 2000). Given the large and diverse black and Latino communities in New York City (United States Census Bureau 2016), our BLMSM participants likely engage with other BLMSM within their and other racially and ethnically diverse communities in NYC (Tieu et al. 2016). This notion is further supported by studies reporting discomfort of BLMSM toward involvement with predominantly white gay communities due to factors such as such sexual objectification (Lewis and Wilson 2017; Teunis 2007). Nevertheless, further exploration of how the dynamics of race and ethnicity influences the nature of BLMSM involvement in various gay and non-gay communities, particularly in the context of gay identity formation.

We also found that fully accepting gay identity was less likely among BLMSM if they exchanged sex for drugs, food, money or shelter. These findings are consistent with other data; an exploratory study revealed that Puerto Rican MSM involved with male sex work refused to be identified as gay or bisexual (Finlinson et al. 2006). In addition, a review of surveillance reports from Latin American countries indicated that MSM who engaged in sex exchange refused to identify themselves as gay or bisexual (Caceres 2002). Caceres (2002) reported that these men rationalized the paradox between their sexual identity and behavior by sexually engaging only in hegemonic male roles (e.g., practicing insertive oral or anal sex). Both of these studies included MSM who reported living in homophobic and stigmatizing environments, which contributed to ostracized gay subcultures and secretive sexual relationships and networks (Caceres 2002; Finlinson et al. 2006). Our study suggests that some BLMSM living in NYC are perhaps living in more gay-affirming communities and are more likely to be fully accepting of their gay identity.

This study has some limitations. First, participants were recruited in New York City, therefore our findings cannot be generalized to other geographic areas (i.e., rural settings). Second, possible recruitment biases may exist since some of the outreach efforts were focused in locations and events where BLMSM may be more accepting of their gay identity (e.g., Gay Pride events). Third, the measures used in our analyses had not previously been normed with BLMSM; additional psychometric studies are warranted to ensure optimally accurate measures for BLMSM. Lastly, this was a small, convenience sample; larger, adequately powered samples of BLMSM will allow for more rigorous analyses. It is possible that the small sample size did not permit sufficient statistical power to detect associations between gay identity and other risk factors and measures we studied.

This study suggests that fully accepting gay identity is associated with less risky behaviors and higher involvement in gay communities for BLMSM. Programs that strengthen and support gay identity should be evaluated for their effectiveness as an additional HIV prevention tool. Such programs could potentially help serve as a protective factor against social stigma and other factors that contribute to increased HIV risk behaviors and could be particularly useful for MSM of color due to compounded sexual and racial minority stigma (Nadal et al. 2011; Solorzano et al. 2000; Sue et al. 2007). Additional studies with larger sample of BLMSM are needed in both urban and rural areas to inform HIV prevention strategies for this disproportionately affected population. Also, more program development is essential to reduce potential HIV risks related to transactional sex for BLMSM. Researchers and service providers should consider that gay identity and related experiences can be very different for BLMSM compared with white MSM and other MSM. Violence, stigma, discrimination, and other negative experiences throughout one’s life can negatively affect sexual identity development and experiences, particularly for BLMSM. Creating avenues to help BLMSM develop better psychosocial and behavioral tools to live and love in healthier contexts is essential for HIV prevention.