Chiasson, Mary Ann DrPH*; Hirshfield, Sabina PhD*; Remien, Robert H PhD†; Humberstone, Mike BFA*; Wong, Tom MD, MPH‡; Wolitski, Richard J PhD§
Men who have sex with men (MSM) continue to account for many (51%) newly diagnosed HIV infections in the United States well into the third decade of the epidemic.1 From 2003 through 2004, there was a statistically significant 8% increase in the number of MSM diagnosed with HIV/AIDS.2 Increased HIV risk among MSM has been attributed to a number of behavioral factors, including but not limited to safer sex fatigue, HIV treatment optimism, crystal methamphetamine addiction, and easy access to sex partners through Internet hook-up sites.3-9 The relative contributions of each of these factors to HIV transmission is a matter of debate.
An understanding of the individual relations between these behavioral factors and high-risk sex and the interrelations among these factors is essential to the development of meaningful interventions. The role that the Internet plays is of particular interest, given the enormous societal changes associated with Internet use and the ready access that most Americans have to it (nearly two thirds of Americans are on-line).10 A meta-analytic examination of on-line sexual risk behavior found that 40% of MSM recruited off-line reported using the Internet to look for sex partners11 and MSM are significantly more likely than heterosexual men and women to have sex with partners they meet on-line.9,12
Nevertheless, many unanswered questions about whether the Internet facilitates high-risk sexual behavior among MSM remain. Some data show that men who use the Internet to meet sex partners may engage in more high-risk behavior than those who meet partners in other ways.11-15 To date, only 1 study in the United Kingdom has directly compared the behavior of men with on-line and off-line partners.16 Like most other studies, however, it examined behaviors over defined periods (ie, the past 3 months, 6 months, or 12 months) but cannot link individual behaviors to specific sexual encounters. We present the findings from a study of MSM recruited on-line that compares the behavior in the last sexual encounter of men who met their partners on-line with that of men who met their partners off-line.
The on-line questionnaire collected basic demographic information on age, gender, race/ethnicity, education, income, marital (partner) status, and country and state of residence. A brief lifetime sexual history was taken. Detailed questions focused on the last sexual encounter that occurred in the 3 months before the survey and included type of male sexual contact (anal or oral, insertive or receptive, with and without condoms), type of female sexual contact (vaginal, anal, or oral with and without condoms), partner type (steady or new/casual), number of partners in the encounter, method of meeting partners (eg, on-line, bar), drug (eg, crystal methamphetamine, cocaine, ecstasy, gamma hydroxy butyrate (GHB), ketamine, marijuana, nitrite inhalants, sildenafil) or alcohol use before sex, disclosure of HIV status (participant and partner[s]), sexually transmitted infections (STIs), HIV status and testing behavior, and use of antiretroviral therapy by those who were HIV-seropositive. Links to HIV/STI education, prevention, and treatment web sites and drug and alcohol treatment and mental health hotlines were included on the last screen of the survey. On average, it took between 10 and 15 minutes to complete the survey. The survey was offered in English and French. No participant incentives were provided.
A total of 6122 individuals clicked on the banner and consented to participate in the study. The analysis presented in this article is based on a subsample of 1683 men older than 18 years of age from the United States or Canada who reported sex with a new or casual male partner in their last sexual encounter in the 3 months before study participation. Participants were excluded for the following reasons: 2342 had missing or inconsistent data, 1221 reported sex with a main partner only, 388 lived outside the United States or Canada, 258 were women, 122 were younger than the age of 18 years, 74 reported no sex or sex only with a woman in their last encounter, 28 were transgendered, and 6 were duplicates. The remaining analytic group consisted of 1297 men reporting single new/casual partner encounters and 386 men reporting multiple-partner encounters, which sometimes included a main partner in addition to a new/casual partner(s).
Data analysis was conducted using SPSS 11.5 for Windows (SPSS, Chicago, IL). In bivariate and multivariate analyses, the primary outcome of interest was unprotected anal intercourse (UAI) in the last sexual encounter. Bivariate analyses were performed using the Pearson χ2 test to evaluate the statistical significance and the odds ratio (OR) as the measure of association. Independent predictors of UAI were assessed using a series of multivariate logistic regression analyses. Variables significant in the bivariate analyses, variables shown to be related to UAI in other studies, and variables likely to be confounders were included in the final model.
The 1683 men included in this analysis resided in all 50 US states and 10 Canadian provinces. They were predominantly white (80%) and college educated (51%), with a median age of 36 years (range: 18-85 years). Most men lived in big cities or their suburbs (68%). Of the 1298 men who reported ever having an HIV test, 11% (143) reported being HIV-seropositive. Although most men described themselves as gay (81%), nearly 20% described themselves as bisexual. Men were equally likely to have met their last sex partner on-line (51%) or off-line (49%), and nearly one quarter (23%) reported UAI in their last encounter. Most men meeting partners off-line met them at bars, dance clubs, parties, and other events where alcohol and drugs are likely to be available. Use of drugs and alcohol before sex was common (25% and 28% respectively).
Overall Comparison of Men Who Met Partners On-Line and Off-Line
Overall, 82% of the study participants “ever” had sex with someone they met on-line, and 66% of them reported that they had more sex partners after they started meeting partners on-line than they had had before. Men who met their last sex partner on-line and those who met their last partner off-line were similar demographically, with some exceptions. Men who met their last partner on-line were significantly more likely to be younger than 30 years of age compared with men who met their last partners off-line (37% vs. 29%, OR = 1.49, 95% confidence interval [CI]: 1.20 to 1.84; P < 0.001). They were also more likely to be white than men meeting their last partners off-line (83% vs. 77%, OR = 1.44, 95% CI: 1.12 to 1.84; P = 0.003). The same proportion (23%) of men reported UAI with partners met on-line and off-line. Drinking alcohol before sex was much more common among men who met their last partner off-line (38%), however, compared with those who met their last partner on-line (18%; OR = 2.8, 95% CI: 2.21 to 3.54; P < 0.001). Similarly, 30% of the men who met their partner off-line compared with 19% who met their partner on-line used drugs before sex (OR = 1.87, 95% CI: 1.48 to 2.37; P < 0.001).
On-Line and Off-Line Partner Comparisons
Comparisons of the characteristics of men meeting partners on-line and off-line and of the behaviors they reported in their last sexual encounter are presented in Tables 1 and 2. There were statistically significant differences (not shown) between the 1297 men reporting a single partner and the 386 men reporting multiple partners in their last encounter for every characteristic in Table 1 except education and for every behavior in Table 2 except HIV disclosure; therefore, data were analyzed separately for these 2 groups of men.
Men reporting a single-partner encounter with a partner met on-line were significantly more likely to be younger, to be white, to identify as bisexual, to have fewer lifetime male sex partners, and to have ever had sex with someone met on-line than men who met their last sex partner off-line. In contrast, the characteristics of those reporting multiple-partner encounters were similar whether they met their partners on-line or off-line, with 1 exception. Like men with a single-partner encounter, they were significantly more likely to have ever had sex in person (as opposed to cyber [virtual] sex) with a partner met on-line if they met their last partner on-line (see Table 1).
For men reporting single-partner encounters, behaviors in the last sexual encounter also differed according to whether the study participant met his partner on-line or off-line (see Table 2). Men meeting partners on-line were significantly less likely to report drug and alcohol use before sex than those meeting partners off-line. In contrast, they were significantly more likely to have anal sex with partners met on-line, but the prevalence of UAI among those who reported anal sex was nearly identical for those meeting partners on-line and off-line.
Fewer behavioral differences were evident with on-line and off-line partners in men reporting multiple-partner encounters (see Table 2). Although drug use before sex was more common overall in multiple-partner compared with single-partner encounters, it did not vary for partners met on-line and off-line, with one exception. Marijuana use was significantly more common when partners were met off-line compared with on-line. Alcohol use before sex was also significantly more common when partners were met off-line than on-line. No differences in the prevalence of anal intercourse, protected or unprotected, were seen for on-line and off-line partners. Nevertheless, HIV disclosure was significantly more likely to occur when partners were met on-line than off-line for men reporting single-partner and multiple-partner encounters (Fig. 1).
Predictors of Unprotected Anal Intercourse in Single-Partner Encounters
Predictors of UAI in men who reported anal intercourse in their last encounter were examined separately for men with single-partner and multiple-partner encounters (Table 3). Men who met partners on-line and off-line were equally likely to report UAI. For men with single-partner encounters, not having a college degree; being HIV-seropositive; and using crystal methamphetamine or sildenafil before sex were all significant predictors of UAI in the bivariate analysis. Only the absence of a college degree and using crystal methamphetamine before sex retained significance in the multivariate analysis, which also included age, manner of meeting partner, and alcohol use before sex.
Predictors of Unprotected Anal Intercourse in Multiple-Partner Encounters
Men reporting multiple-partner encounters were considered to have UAI if they reported UAI with any partner in the encounter. Significant predictors of UAI in men with multiple-partner encounters were similar to those for men with single-partner encounters in the bivariate analysis: being HIV-seropositive and using crystal methamphetamine or sildenafil before sex. Only being HIV-seropositive retained significance in the multivariate model, however, which included age; education; manner of meeting partner; and using crystal methamphetamine, sildenafil, or alcohol before sex (see Table 3).
Bivariate and multivariate analyses examining the relations between additional variables and UAI were also carried out but are not presented here in detail. HIV status disclosure by both partners (yes/no), race/ethnicity (white, black, Hispanic, or mixed/other), sexual identity, and use of drugs other than sildenafil and crystal methamphetamine were not significantly associated with UAI in any analysis.
To assess the potential for HIV transmission in single-partner encounters, the frequency of HIV serosorting was examined (Table 4). HIV-seropositive men were significantly more likely to report UAI with other HIV-seropositive men than with HIV-seronegative or unknown serostatus partners (OR = 2.81, 95% CI: 1.00 to 7.85; P = 0.04) in their most recent sexual encounter. Nevertheless, 26% (25 of 95) of HIV-seropositive men reported UAI with partners who were either HIV-seronegative or of unknown serostatus. In addition, 8% (72 of 902) of HIV-seronegative men reported UAI with a partner who was HIV-seropositive or of unknown serostatus. All 33 of the UAI encounters reported by the 182 men whose HIV status was unknown were potentially serodiscordant. Overall, 11% (130 of 1179) of the men with single-partner encounters reported UAI with men who were serodiscordant or of unknown serostatus. A similar analysis could not be performed for men reporting multiple-partner encounters, because HIV disclosure questions were not asked for each partner in the encounter.
Among men reporting single-partner encounters, the proportion of HIV-seropositive men reporting seroconcordant UAI compared with discordant or potentially discordant UAI was similar whether the partner was met on-line (30% [6 of 20 men]) or off-line (33% [5 of 15 men]) (OR = 0.86, 95% CI: 0.16 to 4.53; P = 0.83), although the numbers were small. In contrast, HIV-seronegative men were significantly more likely to report UAI with seroconcordant partners met on-line (66% [69 of 104 men]) than off-line (50% [37 of 74 men]) (OR = 2.0, 95% CI: −1.02 to 3.81; P = 0.03).
In this on-line study of 1683 MSM from the United States or Canada whose last sexual encounter in the prior 3 months was with a new or casual male partner, we directly compared demographic characteristics and substance-using and sexual behavior in that encounter of participants with partners met on-line with those of participants with partners met off-line. Bivariate and multivariate analyses showed that men reporting anal intercourse in their last encounter were equally likely to have UAI with partners met on-line and off-line, similar to findings reported by Bolding et al16 from the United Kingdom.
Our study design has several advantages compared with those of other studies examining the role of the Internet in high-risk sexual behavior in MSM. Previous studies have found that men who meet partners on-line engage in more risk behavior in general, but these studies did not collect detailed information comparing behavior in individual encounters with partners met on-line and partners met off-line.4,11-14 By examining behavior occurring just before or during a single recent sexual encounter, we were able to improve participant recall, make direct comparisons of behavior with partners met on-line and off-line, and eliminate the temporal sequence uncertainties introduced by the use of summary measures of behavior for 3- or 6-month periods as outcome and predictor variables.
Our event-based analysis also identified a rarely described but high-risk behavior, having more than 1 partner in a sexual encounter, which was reported by 23% of the study participants in their last encounter. This compares with a 27% 6-month prevalence of group sex reported by HIV-seronegative men participating in a seroincidence study.17 Multiple-partner encounters require further study, because men reporting them in our study were significantly more likely to have UAI than were men reporting single-partner encounters, and group sex was significantly associated with HIV seroconversion in the univariate analysis of the study by Buchbinder et al.17
As in other studies, drug use before sex,18-21 less education,22 and HIV seropositivity16,23,24 were found to be strong predictors of UAI among men in this study who reported anal sex. Crystal methamphetamine use and not having a college degree were significant predictors of UAI in the multivariate model for those with a single-partner encounter, whereas being HIV-seropositive was the only variable that retained significance in the multivariate analysis for those reporting a multiple-partner encounter.
We examined serosorting and serodiscordant sexual partnering for the 95 HIV-seropositive men reporting a single-partner encounter in our sample. HIV-seropositive men were more likely to report UAI with an HIV-seropositive partner, although approximately one quarter (26%) of the HIV-seropositive men reported UAI with an HIV-seronegative or serostatus unknown partner in their last encounter. If our questionnaire had collected information on more than a single encounter, this finding would probably be comparable to the report that nearly half (47%) of the HIV-positive men with nonmain partners recruited from a variety of community-based venues in New York and San Francisco reported UAI with a seronegative or serostatus unknown partner in the 3 months before study participation.25 It is considerably higher, however, than the 6% reporting serodiscordant unprotected insertive anal intercourse (no information on receptive anal intercourse presented) in the last encounter in a study of 1923 recently diagnosed HIV-seropositive men from 16 states.26
Such comparisons of the proportion of HIV-seropositive men reporting serodiscordant UAI across studies are problematic, because time frames, recruitment sites (clinical vs. nonclinical), and definitions of UAI differ. In addition, individual variation in high-risk behavior over time has been described in HIV-seropositive men, with periods of unsafe sex alternating with periods of abstinence and safer sex.27
Most study participants reported being seronegative and having seronegative sex partners. Because HIV testing was not performed as part of our study and nearly 1 in 5 men in this high-risk sample reported not knowing his HIV status, however, it is quite likely that there were more serodiscordant contacts than those acknowledged by the participants. Lack of knowledge of HIV status is common among MSM. In a 2003 HIV serosurvey of MSM from 5 US cities, the overall seroprevalence was 25% and nearly one half (48%) of the HIV-positive men were previously unaware of their infection.28
One potential advantage of meeting partners through the Internet is that on-line hook-ups may facilitate disclosure of HIV status by potential partners before they meet in person. Fear of rejection, an important deterrent to disclosure for HIV-seropositive men,29-31 may be less of an issue on-line, because partner preferences are often stated in user profiles and the Internet is viewed as a place where men can avoid abuse and discrimination.32 In our study, men reporting single-partner or multiple-partner encounters were significantly more likely to report disclosing their HIV status to partners met on-line compared with those met off-line. Similar findings were reported in a study of Latino MSM meeting partners on-line,33 although in that study, HIV-positive men were significantly less likely to disclose their serostatus compared with HIV-negative men.
HIV disclosure was not statistically related positively or negatively to UAI in our study. Because our study focused on the most recent sexual encounter, however, we did not ask about the consistency of behaviors over time or whether disclosure occurred before or after sex. Other studies have found that the association between disclosure and high-risk behavior is complex and involves consistency of disclosure34,35 and an explicit discussion of safer sex with the discordant partner.36 Nevertheless, our results suggest a cautionary note about the expectation that disclosure of serostatus would necessarily lead to safer sex.
Although our study and the study by Bolding et al16 found no increased risk of UAI with on-line partners, the increased number of sex partners and geographically expanded sexual networks associated with meeting partners on-line can amplify transmission of HIV and other STIs.37 Two-thirds of the participants in our study who ever had sex with someone they met on-line reported an increased number of sex partners since they began using the Internet to meet partners. In addition, men meeting partners on-line meet partners in many other venues.38,39 Given the large number of partners men reported in the 3 months before study participation in our study (13% of men with single-partner encounters and 23% of men with multiple-partner encounters reported more than 10 partners) and that approximately one quarter of the men reported UAI in their most recent encounter, the probability of exposure to HIV and other STIs is high in this group of MSM.
Internet-based behavioral surveys are becoming increasingly popular as the world moves on-line. The strengths of on-line surveys (rapid, inexpensive, and diverse sample recruitment and more accurate reporting of high-risk behavior) and their weaknesses (missing data, selection bias, and an inability to verify identity of the study participant) have been well described.37,40 We attempted to minimize some of these disadvantages by ensuring participant anonymity, not offering incentives for participation, and recruiting from 14 different gay-oriented web sites whose content varied from news and entertainment to sex. In addition, timing and placement of the recruitment banner advertisement differed from site to site, lessening the likelihood of systematic recruitment bias.
Like most other studies of MSM regardless of recruitment venue, our Internet-based sample is a convenience sample. Because neither the general population of MSM nor the population on-line has been fully enumerated or described,41 it is not possible to know whether the findings from our study are generalizable to all MSM using the Internet who have sex with new or casual partners. Nevertheless, the demographic characteristics of our study participants are similar to those of other MSM samples recruited on-line.13,14,42,43
Our study confirms the findings of previous studies that a large proportion of MSM meeting sex partners on-line are engaging in behavior that puts them at risk for HIV and other STIs. We also found that these men are no more likely to engage in UAI with partners met on-line than they are with partners met off-line. The presence on-line of large numbers of MSM who report multiple high-risk behaviors,11 the high frequency with which MSM seek HIV/STI information on-line,12,43 and the success that we and others have had in recruiting MSM into on-line surveys suggest that the Internet may be the ideal venue for behavioral interventions targeting this population.
The authors thank Claudette McKenzie and Jill Wasserman for their technical and administrative assistance.
1. Centers for Disease Control and Prevention. HIV/AIDS Surveillance Report, 2004. Vol. 16. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, 2005:1-46. Also available at: http://www.cdc.gov/hiv/stats/harslink.htm
2. Centers for Disease Control and Prevention. Trends in HIV/AIDS diagnoses-33 states, 2001-2004. MMWR Morb Mortal Wkly Rep
3. Wolitski R, Vadiserri R, Denning P, et al. Are we headed for a resurgence of the HIV epidemic among men who have sex with men? Am J Public Health
4. Elford J, Bolding G, Sherr L. Seeking sex on the Internet and sexual risk behaviour among gay men using London gyms. AIDS
5. Morin S, Vernon K, Harcourt J, et al. Why HIV infections have increased among men who have sex with men and what to do about it: findings from California focus groups. AIDS Behav
6. Sheon N, Crosby M. Ambivalent tales of HIV disclosure in San Francisco. Soc Sci Med
7. Morin S, Steward W, Charlebois E, et al. Predicting HIV transmission risk among HIV-infected men who have sex with men. J Acquir Immune Defic Syndr
8. Wolitski R. The emergence of barebacking among gay and bisexual men in the United States: a public health perspective. J Gay Lesbian Psychother
9. Elford J. Changing patterns of sexual behaviour in the era of highly active antiretroviral therapy. Curr Opin Infect Dis
10. Fallows D. How Women and Men Use the Internet
. Washington, DC: Pew Internet & American Life Project; 2005.
11. Liau A, Millet G, Marks G. Meta-analytic examination of online sex-seeking and sexual risk behavior among men who have sex with men. Sex Transm Dis
12. Rietmeijer C, Bull S, McFarlane M, et al. Risks and benefits of the internet for populations at risk for sexually transmitted infections (STIs). Sex Transm Dis
13. Benotsch E, Kalichman S, Cage M. Men who have met sex partners via the internet: prevalence, predictors, and implications for HIV prevention. Arch Sex Behav
14. Kim A, Kent C, McFarland W, et al. Cruising on the Internet highway. J Acquir Immune Defic Syndr
15. Whittier D, Seeley S, St. Lawrence J. A comparison of web- with paper-based surveys of gay and bisexual men who vacationed in a gay resort community. AIDS Educ Prev
16. Bolding G, Davis M, Hart G, et al. Gay men who look for sex on the Internet: is there more HIV/STI risk with online partners? AIDS
17. Buchbinder S, Vittinghoff E, Heagerty J, et al. Sexual risk, nitrite inhalant use, and lack of circumcision associated with HIV seroconversion in men who have sex with men in the United States. J Acquir Immune Defic Syndr
18. Marks G, Bingman C, Duval T. Negative affect and unsafe sex in HIV-positive men. AIDS Behav
19. Colfax G, Vittinghoff E, Husnik M, et al. Substance use and sexual risk: a participant- and episode-level analysis among a cohort of men who have sex with men. Am J Epidemiol
20. Hirshfield S, Remien R, Humberstone M, et al. Substance use and high-risk sex among men who have sex with men: a national online study in the USA. AIDS Care
21. Purcell D, Moss S, Remien R, et al. Illicit substance use, sexual risk, and HIV-positive gay and bisexual men: differences by serostatus of casual partners. AIDS
22. Xia Q, Osmond D, Tholandi M, et al. HIV prevalence and sexual risk behaviors among men who have sex with men. J Acquir Immune Defic Syndr
23. Mansergh G, Marks G, Colfax G, et al. “Barebacking” in a diverse sample of men who have sex with men. AIDS
24. Purcell D, Parsons J, Halkitis P, et al. Substance use and sexual transmission risk behavior of HIV-positive men who have sex with men. J Subst Abuse
25. Parsons J, Schrimshaw E, Wolitski R, et al. Sexual harm reduction practices of HIV-seropositive gay and bisexual men: serosorting, strategic positioning, and withdrawal before ejaculation. AIDS
26. Centers for Disease Control and Prevention. High-risk sexual behavior by HIV-positive men who have sex with men-16 sites, United States, 2000-2002. MMWR Morb Mortal Wkly Rep
27. Aidala A, Lee G, Garbers S, et al. Sexual behaviors and sexual risk in a prospective cohort of HIV-positive men and women in New York City, 1994-2002: implications for prevention. AIDS Educ Prev
28. Centers for Disease Control and Prevention. HIV prevalence, unrecognized infection, and HIV testing among men who have sex with men-five U.S. cities, June 2004-April 2005. MMWR Morb Mortal Wkly Rep
29. Serovich J, Mosack K. Reasons for HIV disclosure or nondisclosure to casual sexual partners. AIDS Educ Prev
30. Zea M, Reisen C, Poppen P, et al. Asking and telling: communication about HIV status among Latino HIV-positive gay men. AIDS Behav
31. Gorbach P, Galea J, Amani B, et al. Don't ask, don't tell: patterns of HIV disclosure among HIV positive men who have sex with men with recent STI practising high risk behaviour in Los Angeles and Seattle. Sex Transm Infect
32. Davis M, Hart G, Bolding G, et al. Sex and the Internet: gay men, risk reduction and serostatus. Cult Health Sex
33. Carballo-Dieguez A, Miner M, Dolezal C, et al. Sexual negotiation, HIV-status disclosure, and sexual risk behavior among Latino men who use the Internet to see sex with other men. Arch Sex Behav
34. Parsons J, Schrimshaw E, Bimbi D, et al. Consistent, inconsistent, and non-disclosure to casual sexual partners among HIV-seropositive gay and bisexual men. AIDS
35. Hart T, Wolitski R, Purcell D, et al. Partner awareness of the serostatus of HIV-seropositive men who have sex with men: impact on unprotected sexual behavior. AIDS Behav
36. Crepaz N, Marks G. Serostatus disclosure, sexual communication and safer sex in HIV-positive men. AIDS Care
37. Chiasson M, Parsons J, Tesoriero J, et al. HIV behavioral research online. J Urban Health
38. Mettey A, Crosby R, DiClemente R, et al. Associations between internet sex seeking and STI associated risk behaviours among men who have sex with men. Sex Transm Infect
39. Centers for Disease Control and Prevention. HIV transmission among black college student and non-student men who have sex with men-North Carolina, 2003. MMWR Morb Mortal Wkly Rep
40. Mustanski B. Getting wired: exploiting the Internet for the collection of valid sexuality data. J Sex Res
41. Black D, Gates G, Sanders S, et al. Demographics of the gay and lesbian population in the United States: evidence from available systematic data sources. Demography
42. Chiasson M, Hirshfield S, Humberstone M, et al. Increased high risk sexual behavior after September 11 in men who have sex with men: an internet survey. Arch Sex Behav
43. Rhodes S, DiClemente R, Cecil H, et al. Risk among men who have sex with men in the United States: a comparison of an internet sample and a conventional outreach sample. AIDS Educ Prev
© 2007 Lippincott Williams & Wilkins, Inc.