MEN WHO HAVE SEX WITH MEN (MSM) continue to be disproportionately affected by the human immunodeficiency virus (HIV) epidemic in the United States.1 MSM comprised the largest number of cumulative HIV cases among men in 2005, represented 53% of all persons diagnosed with HIV or AIDS1 and are the only risk group in the United States for whom HIV diagnoses are increasing rather than decreasing.2 Published studies often regard MSM as homogenous, but MSM include men who have sex only with men (MSM/O) and men who have sex with men and women (MSM/W), and encompass men of various racial and ethnic backgrounds. However, few studies compare HIV risk behaviors between MSM/W and MSM/O and fewer studies examine racial or ethnic differences in such comparisons.
MSM/W and MSM/O differ in HIV infection rates and HIV risk behavior. Centers for Disease Control and Prevention (CDC) surveillance data indicate that HIV prevalence is lower among MSM/W compared with MSM/O.3,4 Studies have also found that MSM/W are more likely than MSM/O to engage in commercial sex work,5,6 injection drug use,5–7 other illegal substance use,8,9 and sex with multiple partners.10 Conversely, studies also report that MSM/W engage in greater rates of protected anal intercourse with male sex partners than MSM/O.9,11
Few studies have examined whether psychosocial factors associated with HIV risk among MSM generally apply to both MSM/W and MSM/O, or whether such associations differ by racial/ethnic group. These factors include treatment optimism, behavioral intervention exposure, condom peer norms, and social support. Treatment optimism, the belief that current HIV treatments make it less likely to be infected with HIV or reduce the severity of HIV infection, has been associated with an increased willingness to engage in unprotected anal sex,12,13 whereas exposure to behavioral interventions that target sexual risk,14,15 positive peer norms regarding condom use,13,16,17 and greater social support18,19have each been associated with lower rates of unprotected anal sex among MSM.
To address these gaps in prior research, we examined whether UAI with male partners, psychosocial variables, and other HIV risk-related variables differed between MSM/W and MSM/O. We also examined whether any of these differences varied across racial/ethnic group. Finally, we examined whether any associations between HIV risk-related variables and UAI differed between MSM/W and MSM/O.
Background and Procedure
This report is based on responses to a questionnaire collected by the Community Intervention Trial for Youth. Details of the study are described elsewhere.20 Funded by the CDC, the Community Intervention Trial for Youth project recruited approximately 250 young MSM each in 13 different communities during the summers of 1999 to 2002. As part of the study design, 9 of the 13 sites collected data only from young MSM of a particular racial or ethnic group. Data collection in Atlanta, Birmingham, and Chicago included only African Americans; San Diego, and Seattle included only Asian and Pacific Islanders, Jackson Heights-Corona, and Washington Heights-South Bronx (NY), and Orange County, and San Gabriel Valley (CA), included only Hispanics or Latinos. The remaining research sites (Detroit, Milwaukee, Minneapolis, and West Hollywood) recruited young MSM regardless of ethnicity.
The sampling protocol was designed to obtain a representative sample in each community of young MSM who attended venues where the target population gathered. The project employed a method of venue-based sampling (time-space sampling) designed to approximate a probability-based sample.20,21 Eligibility requirements for participants included (1) being between 15 and 25 years of age, (2) sexual contact with a man reported within the past year, and (3) satisfying any racial-ethnic requirements of a specific research site. The interviewer-administered questionnaire required approximately 30 minutes to complete. IRB approval was obtained from the CDC and each of the 9 collaborating research institutions. Informed consent was obtained from all study participants.
Of the full sample, 8801 participants reported sex with men only in the past 3 months, 1494 reported sex with both men and women, 181 reported sex with women only, and 905 participants reported no sexual activity. An additional 108 responses to the sexual risk questions were missing due to technical problems with the translation of the questionnaire. The current report analyzes responses from the 10,295 respondents who reported sex with men and women (MSM/W) or sex with men only (MSM/O) during the past 3 months.
The survey instrument included questions about demographics, including the participants age, ethnicity, education level, sexual identity (i.e., gay, bisexual, heterosexual), HIV testing history, and self-reported HIV status. Identifying as gay, having ever been tested for HIV, and being HIV-positive were each scored as dichotomous variables. Men were asked how many times during the past 3 months they engaged in insertive or receptive anal sex with a man without using a condom. We dichotomized this variable to represent those who had reported any UAI and those who reported no UAI. Men were also asked how many times during the past three months they had engaged in unprotected vaginal sex. For the purpose of this study, we limited our analysis to a comparison of UAI with male partners and excluded data about unprotected vaginal sex.
Awareness of Combination Therapy and HIV Treatment Optimism
Participants were asked whether they had heard about combination therapy to treat HIV or AIDS. For those participants who had heard of the therapy, we asked whether they believed HIV was a less serious threat because of the new treatments and whether they practiced safer sex less often than before the new treatments were available. For consistency with previously published reports of this variable,13 treatment optimism was scored 1 if the respondent agreed strongly or somewhat strongly to one or both of these two questions and 0 otherwise.
Exposure to AIDS Prevention Interventions
Men were asked whether they had been exposed to any of six types of AIDS prevention interventions during the previous six months (attended workshops, events, or educational meetings about AIDS prevention, seen flyers or pamphlets about AIDS prevention or AIDS prevention events, or other exposure to AIDS prevention interventions). A score from 0 to 6 was computed by summing the total number of “yes” responses.
Peer Condom Norms
Four questions asked participants about condom use in their friendship network (e.g., most of my friends think you should always use condoms when having anal sex, most of my friends are using condoms these days when they have anal sex). Each item was rated 1 to 5, “strongly agree” to “strongly disagree.” A single index was created by subtracting the mean of the four items of six so that higher numbers indicated stronger condom-use norms. Cronbach’s internal consistency alpha for this sample was 0.78. These items were adapted from a longer scale22 that had a Cronbach’s α of 0.75.
Four questions assessed social support (e.g., there is someone to share concern about AIDS, or to talk to about safer sex). Each item was rated 1 to 5, “never or none of the time” to “always or all of the time.” These items were combined into a single index and a mean score was calculated. A higher score on the index indicated more social support. Cronbach’s internal consistency alpha for this measure was 0.83.
Hierarchical multiple and logistic regression were used for analyses. To examine associations by sexual partner group (i.e., MSM/W vs. MSM/O), we conducted bivariate analyses regressing UAI and HIV risk factors onto sexual partner group. To examine whether associations varied by racial/ethnic group, dummy-coded variables representing racial/ethnic group (using African American as the reference group) and appropriate product terms were added to the regression model. These methods to detect moderator effects were repeated using sexual partner group as the potential moderator. Tests of moderator effects used standard procedures (e.g., centering continuous variables; Cohen et al.23). Odds ratios (ORs) less than 1 for sexual partner group indicated a lower odds of a given outcome for MSM/W than MSM/O, and an OR greater than 1 indicated a higher odds of a given outcome for MSM/W than MSM/O.
The mean age of the sample was 21.3 (SD = 2.4, range = 15–25). Twenty-eight percent of the sample was African American, 10% Asian/Pacific Islander, 37% Latino, 22% white, and 2.1% unknown or other (for the latter, 40% gave no ethnic or racial designation, 30% were American Indian, and the remainder reported some combination). MSM/W were less likely than MSM/O to report engaging in UAI (22% vs. 31%; Table 1). Only 18% of MSM/W considered themselves gay (69% bisexual, 13% “other”) whereas 82% of MSM/O considered themselves gay (13% bisexual, 5% other). MSM/W were less likely than MSM/O to report ever being tested for HIV, to be HIV-positive, to be aware of combination therapy, to be exposed to HIV prevention interventions, or to have social support (Tables 1, 2). MSM/W and MSM/O reported comparable treatment optimism and peer condom norms.
We explored whether the prevalence of MSM/W in the sample and the prevalence of UAI varied by racial/ethnic group (Table 3). For African Americans, the prevalence of MSM/W (19%) was significantly higher, and the prevalence of UAI (24%) was significantly lower, than the corresponding percentages for other racial/ethnic groups.
Next, we considered whether any of the associations described in Tables 1 and 2 were moderated by racial/ethnic group. The difference in UAI between MSM/W and MSM/O did not vary significantly by race/ethnicity (interaction P = 0.60); for all racial ethnic groups MSM/W were less likely to report UAI with male partners than MSM/O. The relationship between sexual partner group and four of the remaining variables did interact with race/ethnicity at a statistically significant level. Race/ethnicity moderated the association between sexual partner group and gay identity, awareness of combination therapy, and social support (P = 0.036, 0.023, and 0.044, respectively). These results indicated that the association between sexual partner group and the variable of interest changed by race or ethnicity in the strength of association but overall followed the same pattern described in Tables 1 and 2. In contrast, the association between sexual partner group and HIV testing history (P < 0.001) did show a different pattern across racial/ethnic groups. Follow-up analyses showed that among African Americans, MSM/W were more likely than MSM/O to be tested for HIV (86% vs. 82%; OR = 1.37, 95% CI = 1.05–1.77), whereas among other racial/ethnic groups, MSM/W were less likely than MSM/O to be tested (Asian/Pacific Islander, OR = 0.63, 95% CI = 0.41–0.96; Latino, OR = 0.70, 95% CI = 0.56–0.87; white, OR = 0.61, 95% CI = 0.45–0.83).
Given the moderating role of race/ethnicity in the association between ever being tested and sexual partner group, we wanted to know if the association between ever being tested and UAI also varied by racial/ethnic group. Results indicated a statistically significant interaction [χ2(4, N = 10,295) = 46.0, P < 0.001]. Follow-up analyses demonstrated that associations between ever being tested for HIV and UAI varied by racial/ethnic group (Table 4). African Americans who had been tested for HIV were less likely to report UAI than those not tested. In contrast, Asian/Pacific Islanders and whites who had been tested were more likely to report UAI than those not tested, and Latinos showed little difference; these patterns characterized both MSM/W and MSM/O.
Finally, we examined the association between UAI and demographics or psychosocial variables, and whether sexual partner group moderated any of these associations. Table 5 describes the associations between UAI and these variables: UAI was associated with identifying as gay, having ever been tested for HIV, being aware of combination therapy, greater treatment optimism, and lower peer condom use norms. A statistically significant interaction with sexual partner group was detected for the association between ever being tested and UAI [χ2 (1, N = 10,295) = 5.13, P = 0.024]. This interaction with sexual partner group was marginally significant for awareness of combination therapy and UAI [χ2 (1, N = 10,295) = 3.79, P = 0.052]. All other tested interactions were not significant (P > 0.10). We conducted follow-up analyses to further explore these interactions. The association between ever being tested and UAI by sexual partner group demonstrated that this association was statistically significant for MSM/O (OR = 1.22, 95% CI = 1.09–1.38) and not for MSM/W (OR = 0.84, 95% CI = 0.63–1.13). This analysis indicated that among MSM/O, a history of ever being tested for HIV was associated with an increased likelihood of reporting UAI and that the opposite was true for MSM/W. Additionally, the statistically significant association between awareness of combination therapy and UAI remained only for MSM/W (OR = 1.35, 95% CI = 1.05–1.37) and not for MSM/O (OR = 1.04, 95% CI = 0.95–1.13).
Our study describes important findings that characterize similarities and differences between MSM/W and MSM/O. MSM/W, in comparison with MSM/O, were less likely to be tested for HIV (if they were not African American), to be aware of antiretroviral treatments, to have been exposed to AIDS prevention interventions, and to report social support. Although each of these factors have been associated with risky sexual behavior in previous studies of MSM,12,14,15,18,19,24,25 our data revealed that the associations between risk factors and UAI did not differ substantially between MSM/W and MSM/O. We found one difference between MSM/O and MSM/W that has not been previously reported in the literature: Having a prior HIV test was associated with greater UAI among MSM/O but not MSM/W. Thus, there may be a need for more in-depth risk reduction counseling for MSM/O during HIV pre- and post-test counseling than for MSM/W. Consistent with previous research, we found that MSM/W were less likely to ever be tested for HIV than MSM/O but also were less likely to engage in UAI with male partners9,11or to test HIV-positive.3–4
Another important finding from our study was differences in HIV risk across racial/ethnic groups of MSM. Sexual risk differed for African American men in our dataset from MSM of other races and ethnicities. Compared with members of other racial/ethnic groups, African American MSM were more likely to report recent sex with both men and women, as opposed to men only, and less likely to report UAI with their male partners. We also found differences in testing patterns by race. African American MSM/W were most likely to have been tested for HIV, followed by MSM/O of all racial/ethnic groups, and followed by MSM/W of racial/ethnic groups other than African American. African American MSM, both MSM/W and MSM/O, who had been tested for HIV were less likely to engage in UAI than African American MSM who had not been tested. Conversely, Asian/Pacific Islander and white MSM who had been tested for HIV were considerably more likely to report UAI than those not tested. The diametrically opposed associations with HIV testing for Asian/Pacific Islanders and whites in comparison with African Americans may be explained by the greater HIV prevalence and incidence in African American MSM populations than in white and Asian/Pacific Islander MSM populations, especially younger MSM populations.26,27 Because African American MSM are more likely to receive seropositive test results when tested for HIV than white or Asian/Pacific Islander MSM, it is possible that African American MSM engage in comparatively less unprotected intercourse with their sexual partners as a consequence.28 Consequently, the influence of HIV testing in curbing sexual risk may be greater among African American MSM than MSM of other races or ethnicities. Increased testing campaigns and other strategies that promote HIV testing may be critical to reducing HIV incidence among African American MSM.
This study has some limitations. The data reported in this study are cross-sectional and thus limit the ability to make causal inferences. Additionally, data were collected from young men who were willing to take the time to complete a questionnaire and who met study eligibility requirements. This may limit the generalizeability of the findings to similar young men. Data were also collected using face-to-face self-report methods and although some steps were taken to ensure the quality of these data, they are still subject to potential biases linked with this data collection method. Another important consideration is that sampling of men from different races/ethnicities varied to some degree by city. Therefore, some city differences may affect the race/ethnicity differences observed. Lastly, some of the specific relationships described in this paper may have been observed partly because of the time that data were collected. If such an historical effect is operating, some of the associations reported here may be different if data were collected today.
In summary, in one of the largest, most racially/ethnically diverse, multisite studies in the country, we found differences in sexual risk between young MSM/W and young MSM/O. We also found racial/ethnic differences in sexual risk and HIV testing among MSM. Each of these sets of findings must be carefully considered in future research and the design of effective future interventions for these populations.
1. Centers for Disease Control and Prevention. Trends in HIV/AIDS diagnoses—33 states, 2001 to 2004. MMWR Morb Mortal Wkly Rep 2005; 54:1149–1153.
3. Chu SY, Peterman TA, Doll LS, et al. AIDS in bisexual men in the United States: epidemiology and transmission to women. Am J Public Health 1992; 82:220–224.
4. Montgomery JP, Mokotoff ED, Gentry AC, et al. The extent of bisexual behavior in HIV-infected men and implications for transmission to their female sex partners. AIDS Care 2003; 15:829–837.
5. Diaz T, Chu SY, Frederick M, et al. Sociodemographics and HIV risk behaviors of bisexual men with AIDS: Results from a multistate interview project. AIDS 1993; 7:1227–1232.
6. Goldbaum G, Perdue T, Wolitski R, et al. Differences in risk behavior and sources of AIDS information among gay, bisexual and straight-identified men who have sex with men. AIDS Behav 1998; 2:13–21.
7. Freeman AC, Krepcho M, Hedrick A, et al. Gay, bisexual, and straight men who have sex with men: HIV risk and disclosure. Paper presented at: The 8th International Conference on AIDS, 1992; Amsterdam, Netherlands.
8. Kalichman SC, Roffman RA, Picciano JF, et al. Risk for HIV infection among bisexual men seeking HIV-prevention services and risks posed to their female partners. Health Psychol 1998; 17:320–327.
9. Stokes JP, Vanable P, McKirnan DJ. Comparing gay and bisexual men at sexual behavior, condom use, and psychosocial variables related to HIV/AIDS. Arch Sex Behav 1997; 26:383–397.
10. Lehner T, Chiasson MA. Seroprevalence of human immunodeficiency virus type-1 and sexual behaviors in bisexual African-American and Hispanic men visiting a sexually transmitted disease clinic in New York. Am J Epidemiol 1998; 147:269–272.
11. Valleroy LA, MacKellar D, Behel S, et al. The bridge for HIV transmission to women from 15- to 29-year-old men who have sex with men in 7 US cities [abstract ThOrC1367]. Paper presented at: The 15th International Conference on AIDS; July 11–16, 2004; Bangkok.
12. Crepaz N, Hart T, Marks G. Highly active antiretroviral therapy and sexual risk behavior. JAMA 2004; 292:224–236.
13. Peterson JL, Bakeman R. Impact of Beliefs about HIV treatment and peer condom norms on risky sexual behavior among gay and bisexual men. J Comm Psychol 2006; 34:37–46.
14. Johnson WD, Semaan S, Hedges LV, et al. A protocol for the analytical aspects of a systematic review of HIV prevention research. J AIDS 2002; 30(suppl 1):S62–S72.
15. Herbst JH, Sherba RT, Crepaz N, et al. A meta-analytic review of HIV behavioral interventions reducing sexual risk behavioral interventions reducing sexual risk behavior of men who have sex with men. J AIDS 2005; 39:228–241.
16. Hart T, Peterson J. Predictors of risky sexual behavior among young African American men who have sex with men. Am J Public Health 2004; 94:1122–1123.
17. Kelly JA, St Lawrence J, Stevenson LY, et al. Community AIDS/ HIV risk reduction: The effect of endorsements by popular people in the three cities. Am J Public Health 1992; 82:1483–1489.
18. Folkman S, Chesney MA, Pollack L, et al. Stress, coping and high-risk sexual behaviors. Health Psychol 1992; 11:218–222.
19. Peterson JL, Coates TJ, Catania JA, et al. High-risk sexual behavior and condom use among gay and bisexual African-American men. Am J Public Health 1992;82:1490–1494.
20. Muhib FB, Lin LS, Stueve A, et al. A venue-based method for sampling hard-to-reach populations. Public Health Rep 2001; 116(suppl 1):216–222.
21. Stueve A, O’Donnell LN, Duran R, et al. Time-space sampling in minority communities: Results with young Latino men who have sex with men. Am J Public Health 2001; 91:922–926.
22. Choi KJ, Coates TJ, Catania JA, et al. High HIV risk among gay Asian and Pacific Islander men in San Francisco. AIDS 1995: 9;306–308.
23. Cohen J, Cohen P, West S, et al. Applied multiple regression/correlation analysis for the behavioral sciences, 3rd ed. Hillsdale, NJ: Lawrence Erlbaum Associates, 2003.
24. MacKellar D, Valleroy L, Secura G, et al. Repeat HIV testing, risk behaviors, and HIV seroconversion among young men who have sex with men: A call to monitor and improve the practice of prevention. J AIDS 2002; 29:76–85.
25. Millett GA, Peterson JL, Wolitski RJ, et al. Greater risk for HIV infection of black men who have sex with men: a critical literature review. Am J Public Health 2006; 96:1007–1019.
26. Valleroy LA, MacKellar DA, Karon JM, et al. HIV prevalence and associated risks in young men who have sex with men. JAMA 2000; 284:198–204.
27. Centers for Disease Control and Prevention. Trends in HIV/AIDS diagnoses among men who have sex with men—33 states, 2001–2006. MMWR Morb Mortal Wkly Rep 2008; 57:681–686.
© Copyright 2009 American Sexually Transmitted Diseases Association
28. Marks G, Crepaz N, Janssen RS. Estimating sexual transmission of HIV from persons aware and unaware that they are infected with the virus in the USA. AIDS 2006; 20:1447–1450.