Crosby, Richard PhD*†; Holtgrave, David R. PhD‡§; Stall, Ron PhD∥; Peterson, John L. PhD¶; Shouse, Luke MD, MPH#
From the *College of Public Health, University of Kentucky, Kentucky; †Rural Center for AIDS/STD Prevention at Indiana University, Indiana; ‡Department of Behavioral Sciences and Health Education, Rollins School of Public Health; §Emory Center for AIDS Research, Atlanta, Georgia; ∥University of Pittsburgh Graduate School of Public Health, Pittsburgh, Pennsylvania; ¶Department of Psychology, Georgia State University; and #Georgia Department of Human Resources, Division of Public Health, Atlanta, Georgia
Correspondence: Richard Crosby, PhD, College of Public Health at the University of Kentucky, 121 Washington Ave., Lexington, KY 40506-0003. E-mail: Crosby@uky.edu.
Received for publication September 22, 2006, and accepted February 20, 2007.
IN THE UNITED STATES, men who have sex with men (MSM) are a priority population for HIV prevention efforts.1–3 Among MSM, those identifying as African American experience a disproportionate risk of HIV infection or diagnosis of AIDS.2,4,5 This disparity has also been documented for specified age groups, i.e., young MSM2,6 (under age 25 years) and older MSM (over age 50 years).7 Several studies8–12 have examined racial differences in HIV sexual risk behaviors to explain racial disparities in HIV infection among MSM. Most studies have found comparable or lower rates of HIV risk behaviors among black MSM.
To extend this research, we analyzed behavioral surveillance data from MSM residing in an urban center of the South to examine whether black MSM engage in significantly greater levels of HIV risk behaviors than white MSM. Data were collected as part of a national surveillance study conducted by the Centers for Disease Control and Prevention.13 In addition, we tested the null hypothesis that black and white MSM would not differ with respect to having unprotected sex with a serodiscordant (or potentially serodiscordant) partner. Finally, a stratified analysis was conducted to examine whether associations between race and HIV risk behaviors differed based on men’s self-reported HIV serostatus.
Materials and Methods
From October 2003 until October 2004, men residing in Atlanta, GA, were recruited to participate in an anonymous, cross-sectional, venue-based survey. A sampling frame of more than 100 venues was generated through formative research and random selection was used each week to determine several venues for recruitment. The day and time of recruitment from the various venues were also determined by random selection. Project staff recruited men in gay-identified venues serving predominately white men, predominately black men, and both populations. One thousand two hundred thirty-three men were asked to complete a brief anonymous survey, 1145 of those were eligible to participate (92.9%), and 1006 of those agreed to participate and provided a complete interview (87.9%). Nine hundred eighteen men reported having had sex in the past year, and 846 of these identified as either African American or as white; these men comprised the analytic sample.
Trained interviewers used personal digital assistants to record answers provided by men. Interviews occurred in or near the recruitment sites and typically lasted 30 minutes. Study procedures were approved by the Institutional Review Board at Emory University. Interview questions were created by staff at the Centers for Disease Control and Prevention. The interview was designed as part of an ongoing HIV behavioral surveillance project.
The primary outcome used was unprotected anal sex with a partner of unknown or known discordant serostatus. Using the last time sex occurred with a main male partner (within the past 12 months) as a recall period, men were asked if the partner had been tested for HIV and, if so, was the result positive. Men having unprotected sex with men believed to have discordant serostatus (or with men of unknown serostatus) were compared with all remaining men who provided information about their serostatus, their last main partner’s serostatus, and their use of condoms at last sex with a main partner. The same procedure was repeated for men having sex with men whom they considered as nonmain partners.
Ten variables were assessed as a representation of men’s HIV-associated risk behavior. Each used the past 12 months as the recall period. Seven of these measures were dichotomous by nature. Whether men engaged in any unprotected anal sex with other men was assessed relative to both main and nonmain partners. Additionally, engagement in unprotected anal-receptive sex for both main and nonmain male partners was assessed as well as engagement in unprotected anal-insertive sex. Also, whether men had ever injected drugs was assessed. The remaining 3 measures were continuous and pertained to the number of anal and oral sex partners men reported in the past 12 months. The total number of partners as well as the number of main and nonmain partners was assessed.
Differences between black and white MSM were assessed by t test for continuous outcomes and χ2 tests for dichotomous outcomes. Before conducting t tests the distributions of the corresponding outcome measures were examined for extreme values. For the distribution representing total number of sex partners, 3 values fell far beyond the remainder of the distribution and were thus truncated at 100 each. Similarly, for the distribution representing total number of nonmain sex partners, 2 values fell far beyond the remainder of the distribution and were thus truncated at 100 each.
To control for possible confounding effects of age and education we then applied linear regression (for the outcomes assessed at a continuous level) and logistic regression (for outcomes assessed dichotomously). Next, we used contingency table analysis to determine whether differences existed between white and black men relative to having sex with serodiscordant and potentially serodiscordant male partners. Finally, we repeated the bivariate tests separately for men indicating they had last tested negative for HIV and for those indicating they had last tested positive for HIV.
Men ranged in age from 18 to 72 years. The mean age was 35.7 years (standard deviation = 9.9). Thirty-five percent (n = 306) of the men identified as being African American, with remaining 540 men identifying as white. Nearly one-half of the men (48.7%) reported having at least a bachelor’s degree; another 31.6% reported attending college. The majority (84.9%) of the men identified as gay, with 13.1% identifying as bisexual, 0.7% identifying as heterosexual, and 1.3% identifying as “other.”
Only a small portion of the men indicated they had never been tested for HIV. Differences in rates of never having a test between black men (5.8%) and white men (6.0%) were not significant (P = 0.92). Nearly 1 of every 5 men (19.2%) reported they were HIV-positive. Among black men 20.1% reported being HIV-positive compared with 18.6% among white men. This difference was not significant (P = 0.79).
Significant racial differences were not found with respect to the number of anal/oral sex partners reported in the past 12 months. The overall average number of anal/oral sex partners was 9.72 (standard deviation = 18.5; median = 4.0). black men reported an average of 8.41 (SD = 17.9) partners versus 10.46 (SD = 18.8) partners for white men (t = 1.56; df = 844; P = 0.12). Similarly, racial differences were not found regarding the number of nonmain male anal sex partners between the 2 groups. The overall average number was 5.42 (SD = 13.0). Black men reported having anal sex with an average of 5.36 (SD = 14.0) nonmain partners compared to 5.46 (SD = 12.5) nonmain partners among white men (t = 0.09; df = 631; P = 0.93). However, differences were found between the 2 racial groups in the number of main male anal sex partners (past 12 months). The overall average number was 1.12 (SD = 0.83). Black men reported having anal sex with an average of 1.3 main partners compared to 1.0 main partner among white men (t = 3.73; df = 584; P = 0.0001).
Table 1 displays descriptive findings and prevalence ratios, PR for 6 measures of unprotected sex, and the measure of injection drug use. All percentages shown in this and the other tables are based on the number of men indicating presence or absence of the risk factor, and these values are also displayed in the tables. As shown, black men reported lower levels of risk for all but one of these measures (any unprotected anal-insertive sex with nonmain partners), which did not differ between the 2 racial groups. These measures included any unprotected anal sex with main and with nonmain partners, any unprotected anal-receptive sex with main and with nonmain partners, and any unprotected anal-insertive sex with main partners. Table 1 also displays findings indicating that black men were significantly less likely than white men to report ever having injected drugs.
In regression analyses adjusted for age and education, because of skewness, the distribution representing education was dichotomized by a median split (at least some college versus high school education or less) and significance changed for only one association. The previously significant difference relative to having any unprotected vaginal sex with a main partner became nonsignificant (P = 0.08) in the presence of age (age was directly associated with this outcome, adjusted odds ratio = 1.05, 95% confidence interval, CI = 1.02–1.09, P = 0.003).
Serodiscordant (and Potentially Serodiscordant) Unprotected Sex
Of 449 men who provided information about their serostatus and their belief about the serostatus of their most recent main sex partner, 39 (8.7%) reported they had unprotected anal sex with a serodiscordant (or potentially serodiscordant) partner. This risk behavior was not significantly different between black (5.9%) and white (10.6%) men (PR = 0.56; 95% CI = 0.29–1.10). For most recent sex with a nonmain partner in the last year (n = 361), 54 (15.0%) men reported that they had unprotected anal sex with a serodiscordant (or potentially serodiscordant) partner. This risk behavior was not significantly different between black (14.0%) and white (15.6%) men (PR = 0.90;95% CI = 0.54–1.50).
Table 2 displays findings pertaining to the dichotomously assessed risk behaviors for men indicating they were HIV negative. Among these participants, more white men compared with black men engaged in risk behavior for HIV acquisition. This was the case for each of the 5 significant associations.
Table 3 displays findings pertaining to the dichotomously assessed risk behaviors for men indicating they were HIV positive. Among these participants, only one difference was found: black men were significantly more likely than their white counterparts to report ever injecting drugs.
Table 4 displays findings from the independent groups t tests that compared white with black men and were stratified by serostatus. Of these 3 measures pertaining to the number of sex partners reported by men over the past 12 months, one produced differential findings in the stratified analyses. Among those self-reporting as HIV negative, black men reported a greater mean number of main partners than white men.
Our findings were remarkably consistent with previous studies8–12 finding that black MSM have similar, if not lower, levels of sexual risk behavior compared to white MSM. The exception was that black men reported a significantly higher average number of main, anal sex, partners in the past 12 months than white men. Also important, racial differences in HIV sexual risks between black and white men were found only for men who reported that they were HIV negative (with the exception being whether men had ever injected drugs). This finding reveals that differences in HIV risk behaviors between black and white men were essentially nonexistent when considering only those men who reported being seropositive.
Interpretation of these findings requires that HIV acquisition behaviors be distinguished from HIV transmission behaviors. Among those men who reported they were HIV positive, the findings suggest that black MSM are no more likely than white MSM to engage in HIV risk behaviors that could cause HIV infection. Among those reporting an HIV-positive serostatus, only one difference was observed (ever injecting drugs), and no significant differences were observed relative to HIV-associated sexual risk behaviors between black and white MSM. The lack of differences in risk behaviors makes the descriptive findings even more important. For example, it is noteworthy that men reported having anal sex with an average of 6.6 partners in the past year and that 27.4% reported having unprotected anal sex with nonmain partners. The proportion of men reporting unprotected anal sex with main partners was quite high, perhaps not unexpectedly. However, it is also noteworthy that less than 10% recently engaged in unprotected anal sex with a known or potentially serodiscordant male partner. Moreover, finding that differences favoring greater protective behavior for black men occurred essentially among men believing they were HIV-negative was important. Further research is needed to explain why black men who are HIV-negative may be more likely to engage in protective behaviors than white men who are HIV-negative.
Findings are limited by the validity of men’s self-reported sexual behavior and by the use of a convenience sample. The use of the last time sex occurred as a recall period might also limit the study findings in that this may not adequately represent men’s sexual risk behavior. In particular, it should be noted that the analyses stratified by HIV serostatus are also subject to self-report bias. Also, in the stratified analysis, the single finding indicating greater risk for black, HIV-negative, men may be an artifact of potential differences regarding how white versus black men classify partners as main or nonmain. It is also critical to note that missing data were a common occurrence in the assessment of risk behaviors. Although much of the missing data are a product of men not engaging in a risk behavior (i.e., not having anal sex), other missing data are a function of men’s decision not to answer given questions. A particular concern regarding missing data is the large number of men not reporting the serostatus of their sex partners. If systematic differences existed between white and black men relative to this underreporting these differences may have biased the study findings. Further, (due in part to missing data) the observed null findings may be an artifact of low statistical power rather than a lack of true differences. This may be especially likely with respect to serodiscordant (or potentially serodiscordant) partners who reported unprotected anal sex with their most recent main male partner. Finally, it should be noted that the venues used in the sampling frame may have inadvertently yielded sample bias that may have influenced the study findings. Unfortunately, whether the sample overrepresented low-risk black men, for example, cannot be determined.
This study supports prior research that reported less HIV risk behavior for black MSM than white MSM, which suggests racial disparities in HIV infection are not explained merely by racial differences in HIV risk behaviors. Instead, alternative explanations have been offered that include the possible effect of higher STD rates, community viral load (i.e., greater HIV viremia among black MSM),14 limited awareness of HIV status, or lower HIV testing, among black compared to white MSM. The absence of significant racial differences for HIV-positive men, but significant differences for HIV-negative men, both of whom are aware of their HIV status, may provide further explanations that need to be considered. Specifically, the beneficial effects of awareness of HIV status among black MSM may depend more on whether they are HIV negative, than HIV positive, if this awareness is to substantially reduce their racial disparity in HIV infection.
1. Centers for Disease Control and Prevention. Need for sustained HIV prevention among men who have sex with men. Available online at www.cdc.gov/hiv/pubs/facts/msm.htm
. Accessed on November 19, 2001.
2. Centers for Disease Control and Prevention. HIV incidence among young men who have sex with men—Seven U.S. cities, 1994–2000. MMWR Morb Mortal Wkly Rep 2001; 49:440–443.
3. Sullivan PS, Chu SY, Fleming PL, et al. Changes in AIDS incidence for men who have sex with men, United States 1990–1995. AIDS 1997; 11:1641–1646.
4. Centers for Disease Control and Prevention. (2000). HIV/AIDS among racial/ethnic minority men who have sex with men—United States, 1989–1998. MMWR Morb Mortal Wkly Rep 2001; 49:4–11.
5. Blair JM, Fleming PL, Karon JM. Trends in AIDS incidence and survival among racial/ethnic minority men who have sex with men, United States, 1990–1999. JAIDS 2002; 31:339–347.
6. Koblin BA, Torian LV, Guilin V, et al. High prevalence of HIV infection among young men who have sex with men in New York City. AIDS 2000; 14:1793–1800.
7. Dolcini MM, Catania JA, Stall RD, et al. The HIV epidemic among older men who have sex with men. J Acquir Immune Defic Syndr 2003; 33:S115–S121.
8. Harawa NT, Greenland S, Bingham TA, et al. Associations of race/ethnicity with HIV prevalence and HIV-related behaviors among young men who have sex with men in 7 urban centers in the United States. JAIDS 2004; 35:526–536.
9. Heckman TG, Kelly JA, Bogart LM, et al. HIV risk differences between African American and white men who have sex with men. J Natl Med Assoc 1999; 91:92–100.
10. Peterson JL, Bakeman R, Stokes J, Community Intervention Trial for Youth Study Team. Racial ethnic patterns of HIV sexual risk behaviors among young men who have sex with men. J Gay Lesbian Med Assoc 2001; 5:55–62.
11. Stokes JP, Vanable PA, McKirnan DJ. Ethnic differences is sexual behavior, condom use, and psychosocial variables among black and white men who have sex with men. J Sex Res 1996; 33:373–381.
12. Torian LV, Makki HA, Menzies IB, et al. HIV infection in men who have sex with men, New York City Department of Health sexually transmitted disease clinics, 1990–1999: A decade of serosurveillance finds that racial disparities and associations between HIV and gonorrhea persist. Sex Transm Dis 2002; 29:73–78.
13. Sanchez T, Finlayson T, Drake A, et al. Human immunodeficiency virus (HIV) risk, prevention, and testing behaviors—United States, National HIV Behavioral Surveillance System: Men who have sex with men, November 2003–April 2005. MMWR 2006; 55(SS06):1–16.
14. Millett G, Peterson JL, Wolitski R, et al. Why are black men who have sex with men (MSM) at greater risk for HIV infection than other MSM? Am J Public Health 2006; 96:1007–1019