In CDC’s “Heightened National Response to the HIV Crisis in African Americans,” incarceration is proposed as one factor that elevates HIV risk among black Americans.⁷ Indeed, black men represent the largest proportion of all incarcerated men in federal or state prisons and local jails.⁸ The overall incarceration rate for black men in 2008 was 6.5 times the rate of white men.⁹ In addition, black men in state and federal prisons represent the largest proportion to report being HIV-positive among inmates who had ever been tested for HIV.¹⁰ However, the pathways by which incarceration may increase HIV risk is unclear, and it is unknown whether these mechanisms may be different for black MSM and blacks with other primary risks for HIV acquisition, such as injection drug use.

Six hundred seventy-nine (6.8%) of the men reported recent arrest. Table 1 summarizes demographic and behavioral characteristics of respondents included in the analysis. The participants reported a median age of 33.0 years. Nearly half of the men identified as non-Hispanic white, with lower numbers of non-Hispanic black, and Hispanic men. The majority of men identified as homosexual, although over one in six identified as bisexual. Most men were recruited in bars, dance clubs, raves, and circuit parties. The majority of men reported completing more than high school. Nearly half of men reported non-injection drug use, but few men reported injection drug use, in the past 12 months.

Most men reporting recent arrest had a main male partner and more than one male partner in the past 12 months (Table 2). The median number of male partners in the past 12 months was 4.0. A minority of men reported having a female partner or having an exchange sex partner in the past 12 months. Nearly half of men reported having a UAI partner in the past 12 months. When the men were asked about the last time they had sex with a main or casual male partner, 24.4% reported engaging in unprotected insertive anal sex and 18.4% engaged in unprotected receptive anal sex (data not shown in Table ​Table22).

Table 2 shows the results from the multivariable analysis of factors associated with recent arrest. Men who had identified as black or Hispanic were more likely to report recent arrest than men who had identified as white. Bisexual and heterosexual sexual identities were significantly associated with recent arrest compared to homosexual identity. Men who reported having completed less education, using non-injection and injection drugs, exchanging sex, or engaging in UAI were more likely to report recent arrest than men who did not report these characteristics. Men who had a female partner were more likely to report recent arrest than men who did not have a female partner. Having six or more male sex partners was associated with reporting recent arrest. The difference in distribution of age by recent arrest was significant (P<0.0001) and inverse.

Black MSM were more likely to report arrest history than white MSM. Our data provide additional information in this area where prior studies have reported mixed results. For example, no racial differences were found in reported incarceration history in a study of Young Men’s Survey (YMS) participants in Los Angeles,¹⁹ but a similar study using national YMS data found that young black MSM were more likely to report history of incarceration than other MSM.²⁰

MSM who were more gay-identified were less likely to report history of arrest. Similarly and independently, there was an association between reporting a recent female sex partner and higher odds of recent arrest. Approximately 17% of this sample was non-gay identifying (NGI) MSM (did not identify as homosexual) and 15% reported having a female partner; the majority of men exhibiting these characteristics were black. In comparison to other MSM, black MSM are less likely to identify as gay or to disclose their sexual identity²¹ and are more likely to be bisexually active.²² However, racial differences found in gay identity and disclosure of homosexuality among MSM were controlled for in our analysis, and thus, the association of black race and recent arrest was not the result of confounding with sexual identity. Regardless of race/ethnicity, MSM and men who have sex with men and women (MSMW) are much less likely to have ever been or recently been arrested and to have ever been charged with a criminal offense than heterosexual men.^16,23,24 Even among MSM, men who have sex with men and women are more likely to report incarceration history than men who have sex only with men.²⁵

Men who reported using non-injection and injection drugs and exchange sex activity were more likely to report arrest history. These findings are consistent with previous studies.^13–16 Although there are no racial differences in drug use and exchange sex among MSM,⁶ it is important to address these behaviors because they are associated with sexual risk behaviors in this population.^26–31 In focus group interviews conducted among NGI black MSM/MSMW, the men revealed that they use drugs, alcohol, and exchange sex as coping mechanisms to resolve conflicts between their racial and sexual identities.³² Among the interviewed men, those in treatment for substance abuse indicated that self-acceptance of their sexuality was necessary for recovery. Interventions that assist NGI black MSM/MSMW in cultivating positive self-identification may prove to be an effective strategy for reducing their HIV-risk behavior.³³ Future research is needed to better understand the sociocultural contexts in which racial and sexual identities are reconciled among black men.^32–34

The finding of our secondary analysis—that arrest history was associated with insertive, but not receptive UAI—is consistent with a previous study with a small group of black MSM in North Carolina, which also found that criminal justice involvement was associated with subsequent unprotected insertive, but not receptive, anal intercourse.³⁵ The reason for this role-specific association is not clear, but the consistency of this finding in two studies suggests it may merit further investigation. Black MSM have been more likely to report insertive anal sex than white MSM³⁶ regardless of incarceration history, but this association would not explain why insertive UAI would be associated with recent arrest in a model controlling for race. Ideally, we would have liked to control for sexual risk behaviors before arrest in our analysis of the association of UAI and arrest, but these data were not available. Although our cross-sectional design precludes our knowing when these men engaged in insertive UAI with respect to their arrest, our data underscore the need to strengthen preventive programs and harm reduction strategies in correctional facilities.

Most men with HIV in prison come into prison with HIV; we have previously reported that, in the Georgia prison system, only 11% of prevalent infections in prison were acquired in prison.³⁷ Nonetheless, prison is a high-risk setting for sex: Prevalent HIV infection rates were 4.4 times as high in prisons as in the general population in 2002,¹⁰ although prevalence rates have since declined 26%.³⁸ Furthermore, there is limited access to condoms in most correctional facilities, with less than 1% of national prisons and jails permitting condom distribution.^10,39

There are a number of limitations in this study. First, the cross-sectional design of the study makes it difficult to interpret associations because the direction of causation cannot be established. Second, the study sample was restricted to men who frequented MSM-identified venues in selected MSAs and therefore it is not representative of all MSM. Moreover, our findings may not be generalized to MSM in rural areas or smaller cities in the USA. The majority of men in our study were recruited in bars, dance clubs, raves, and circuit parties. Because attendance in these venues has been associated with drug use and sexual risk behaviors,⁴⁰ it is possible that our results may overestimate the true extent of drug use and sexual risk behavior among MSM. Also, the NHBS sampling strategy used for recruitment did not sample men via the Internet; there is growing recognition of the important role of Internet use for meeting sexual partners among MSM, and MSM who meet sex partners online may have higher levels of sexual risk behaviors.⁴¹ Also, it did not sample current prisoners, whose MSM behaviors may have been limited to while they were incarcerated.⁴² Third, the participants may not have accurately disclosed their risk behaviors during the interview. Although their responses are subject to social desirability bias, it is believed to be minimal because the interview was anonymous and confidentiality was assured. In addition, the participants may not have accurately recalled behaviors for all three time frames (recall bias): ever, the 12 months before interview, and at most recent sex. Finally, the assessment of history of arrest reflects differential involvement in crime: “In the past 12 months, have you been arrested—that is, picked up by the police?” Given that only the most serious offenses lead to imprisonment, it is possible that our proxy measure of arrest may overestimate the true extent of incarceration among MSM. Our use of arrest history as a proxy for incarceration history also assumes that racial disparities in arrest are similar to those in incarceration, and this assumption is not well validated.⁴³ Although 27% of persons arrested nationally were black in 2004, a higher percentage of jail populations, and an even higher percentage of state prison systems, was black.⁴⁴

Given that our data suggest that criminal justice involvement is more common among black than white MSM, incarceration is a candidate factor that might help explain black/white disparities in HIV prevalence and incidence among MSM. However, the mechanisms by which such an effect might operate among MSM are not clear. A greater portion of black men with HIV infection were reported to have passed through a jail or prison in the past year (estimated between 22% and 28%), compared to white men with HIV infection (estimated between 11% and 17%).⁴⁵ Incarceration has been proposed to impact HIV risk by decreasing the number of marriageable men in black communities with high rates of incarceration, thereby promoting partner concurrency,^32,46 but this consideration is less relevant for MSM. Men may have additional HIV risk behaviors while incarcerated. A study in the Georgia prison showed that risk from male–male sex was a primary driver of within-prison transmission and that most men who had male sex partners in prison did not have male sex partners before entering prison.³⁷ Further, being incarcerated may change patterns of protective behaviors for MSM: Incarcerated men and men with a history of incarceration may become habituated to unprotected sex.⁴⁷

The association between history of insertive, but not receptive, UAI, is intriguing. In the context of the discussion of how incarceration may shape men’s risk for HIV, it will be important to ascertain whether the tendency to engage in insertive UAI and a propensity to be arrested are both related to some underlying trait of certain men, or whether the experience of being incarcerated leads to an adoption of insertive UAI behavior. This will be best done by collecting more detailed data about sex behaviors before, during, and after incarceration, using both quantitative and qualitative methods.