Introduction

The Internet has redefined how people, including men who have sex with men (MSM), interact. Researchers have found that MSM are increasingly using the Internet to facilitate social and sexual relationships [1, 2]. With the increased use of the Internet for social and sexual relationships there has also been an increase in HIV risk among MSM, attributed in part to ease of access to sex partners online [3]. For this population of MSM, understanding the relationship between behavioral factors such as drug use or multiple sex partners and sexual risk can facilitate the development of effective interventions [3].

A study by Rosser et al. found that men who reported searching for sexual partners online also reported unprotected sex with multiple male partners in the past 12 months [1]. In this study, which compared online and offline sexual encounters participants reported engaging in unprotected anal intercourse (UAI) with twice as many men with partners met online than with partners met offline. Although another recent study found no increased risk of UAI with partners met online, it did find that two-thirds of study participants who had ever met partners online reported an increased number of sex partners since they began using the Internet to meet sexual partners [3]. Similarly in a recent study that assessed the Internet as an HIV/STD risk for MSM those with both on-line and off-line sex partners had a greater number of sex partners than those with only offline or only online sex partners [4]. Given the increased sexual risk for HIV among MSM who seek sexual partners online, effective strategies are needed to reduce high-risk behaviors in this population [5].

Online prevention has the potential to reach MSM who may be less accessible through traditional venues such as bars, clubs, or community-based organizations. In addition to offline prevention activities it may be an important component of a multi-pronged approach to reducing sexual risk-taking among MSM. The reach and effectiveness of online prevention are not fully known [5] and should be further examined. This report presents data on participation in online and/or offline prevention activities among a sample of NYC MSM who use the Internet to meet sex partners. It examines the relationship of participation in online and/or offline prevention to sexual risk among these MSM.

Methods

Study Design

As part of the Centers for Disease Control and Prevention’s (CDC) national Web-based HIV Behavioral Surveillance (WHBS), investigators at the New York City Department of Health and Mental Hygiene (NYCDOHMH) conducted an anonymous cross-sectional Internet survey from April through August of 2007 among MSM. Banner advertisements were posted and rotated through a variety of social networking websites including some that catered specifically to MSM. The advertisements allowed interested individuals direct access to the survey after they gave informed consent. Multiple advertisements were used, each containing images of attractive, semi-clothed (mostly shirtless) men with brief information about the study. Participating websites had different posting policies; hence, to view the clickable banner advertisement some websites required registration onto their website, while others did not. When an Internet user clicked on the banner advertisement, the participant was linked to the website hosting the survey. A unique study ID was then automatically generated within the system and the participant could begin the survey. The survey was voluntary and it offered no incentive for participation.

Eligibility criteria included being 18 years of age or older, born male, and a resident of the NYC metropolitan area. The survey used Secure Sockets Layer (SSL) encryption technology, a commonly-used protocol for managing the security of a message transmission on the Internet. The host website of the survey did not use cookies, it did not collect IP or email addresses, or any other identifying information. To help assure website security, the website and supporting technology underwent a rigorous certification and accreditation process. The study was approved by the institutional review boards of the NYCDOHMH and the CDC.

Survey Instrument

The questionnaire collected information on demographics (race/ethnicity, age, education), Internet usage behaviors (personal, work), sex and drug use behaviors (main and casual partners, group sex, injection and non-injection drug use), HIV and STD testing experiences, access to local HIV prevention services (individual or group counseling) and participation in HIV prevention activities (condom use at last sexual encounter, and Internet-based HIV prevention sessions). Participants could refuse to answer any survey questions. For those interested participants, the questionnaire contained links to HIV information and prevention resources. The survey took approximately 30 min to complete unless the respondent reported an extensive drug and/or sexual history, in which case additional questions were asked. For example, those reporting a history of injection drug use were asked about the use of clean needles and cookers, and those with extensive sexual histories were asked about their participation in group sex activities, their experience with performance enhancing drugs (Viagra, Levitra, or Cialis), and condom use when using drugs.

Sample

Approximately 1.3 million impressions (pop ups) of the WHBS banner advertisements appeared on monitors across the NYC metropolitan area and 4,760 men clicked on the banner advertisements and were directed to the survey (Fig. 1). Overall, 4,143 men were eligible while 617 were ineligible because they did not meet the eligibility criteria. Of those eligible, 2,046 (49%) completed the questionnaire. Of those completing the questionnaire, 1,700 (83%) reported sex with another man in the past 12 months, and 1,142 (67%) of these men reported meeting at least one sex partner online. Of these 1,142 men, 1,124 reported sex with a main or casual partner in the past 12 months and comprise the analysis sample.

Fig. 1
figure 1

WHBS sample recruitment characteristics

Full size image

Measurement

The WHBS questionnaire was developed by the CDC and had its roots in the National HIV Behavioral Surveillance (NHBS) project, which was developed as an ongoing U.S. surveillance system to monitor trends in HIV risk behaviors through venue-based recruitment. In contrast, WHBS was developed as a potential way to collect behavioral data online from MSM populations not reached through NHBS, including men who do not self-identify as gay or bisexual and gay-identified MSM who do not attend NHBS venues.

To measure participation in online prevention, the questionnaire asked, “In the past 12 months, how often have you visited a website for information about safer sex?” Participation in offline prevention was measured by asking respondents if they had participated in a 1-to-1 HIV prevention talk or group HIV prevention in the past 12 months. The following self-reported measures were included in the analysis: frequency of internet use (once a day vs. more than once a day), drug use in the past 12 months (yes/no), HIV testing history in the past 12 months (yes/no), HIV status (negative, positive, unknown), unprotected anal intercourse (UAI) in the past 12 months (yes/no), type of sexual partner (main or casual) at the last sexual encounter, and demographic information (age, race/ethnicity, and education).

Statistical Analysis

Analysis was conducted on the 1,124 sexually active MSM who reported meeting at least one sex partner online and had sex with a main or casual partner in the past 12 months. The outcome variable in this analysis was UAI. Bivariate analyses of categorical data were conducted using the Chi-square test and logistic regression to estimate odds ratios (OR). A simultaneous multivariate logistic regression model was used to analyze the association of UAI with prevention activities, adjusting for use of drugs, HIV testing history, sociodemographics, and frequency of Internet use. Though a conservative measure, for our model we used the type of sexual partner (main or casual) at the last sexual encounter, since WHBS did not ask questions about the type of sexual partner in the past 12 months. Criteria for including variables in the multivariate model included statistical significance (P < 0.05) in the bivariate analysis; socio-demographic variables were used as controls. The rationale for including frequency of Internet use in the model is that usage may be an indicator of the degree of risk—that is, frequent Internet users seeking sex might be at greater risk for sexually transmitted diseases [5]. HIV testing history was included in the model because it may be an indicator of susceptibility to or concern over becoming infected with HIV [6]. Data analyses were conducted using SAS 9.1 (SAS Institute, Cary, NC).

Results

The majority of the sample was white (63%); while blacks, Latinos, and other races/ethnicities represented 13, 18, and 6% of the sample, respectively (Table 1). Eleven percent of the men were between the ages of 18 and 19, 48% were in their 20s, 24% were in their 30s, 14% were in their 40s, and 3% were in their 50s or older. The median age was 27 years. The majority of respondents identified as homosexual (85%) or bisexual (12%). The sample was well-educated (98% had equal to or more than a high school education). Most (84%) were born in the United States. Nine percent self-reported being HIV-infected, 76% HIV-negative, and 15% had unknown serostatus. Eighty-six percent reported ever having tested for HIV. A lower percentage (58%) reported having tested for HIV in the past 12 months. UAI in the past 12 months was reported by 53% of the sample, with the same proportion (53%) reporting either unprotected receptive anal intercourse or unprotected insertive anal intercourse. Nearly half of the men (48%) reported use of non-injection drugs to get high excluding alcohol in the past 12 months. The majority of the participants (63%) identified as frequent Internet users—that is, they reported use of the Internet at least once a day for personal reasons though not necessarily for meeting sexual partners on line.

Table 1 Demographics and risk behavior characteristics among MSM participants in an online survey in NYC, 2007
Full size table

Overall, 56% of the respondents reported participation in online or offline prevention activities—that is, they had visited a website about safer sex or attended a 1-to-1 HIV prevention talk or a group HIV prevention session within the past 12 months. By prevention modality, 36% reported participation on-line only, 8% off-line only, and 12% both. In the bivariate analysis, there was a significant association between UAI and participation in either on-line or off-line prevention activities. Men who reported participation in either online or offline prevention activities were less likely to engage in UAI than those who did not (30% vs. 38%). Men who were under the age of 30 (36%) were more likely to engage in UAI compared to those who were older (29%) as were those who had a high school education or less (43%) versus those who had greater than a high school education. In the bivariate analysis there was no significant association between UAI and race/ethnicity, frequency of Internet use, or HIV testing history (Table 2).

Table 2 Bivariate analysis. Association of socio-demographic and other variables with UAI
Full size table

In the multivariate analysis the following covariates were included: age, race, drug use, prevention activity in past 12 months, partner type, frequency of Internet use, HIV status, and HIV testing history. Men who participated in online or offline prevention activities were significantly less likely to engage in UAI (adjusted odds ration [AOR] = 0.68, 95% CI 0.53–0.89). Men who reported sex with a main partner in their last sexual encounter were significantly more likely to engage in UAI than men who reported sex with a casual partner in their last sexual encounter (AOR = 2.18, 95% CI 1.66–2.86). Other variables significantly associated with engaging in UAI included, being under the age of 30 (AOR = 1.47, 95% CI 1.11–1.95), drug use in the past 12 months (AOR = 1.54, 95% CI 1.18–2), being a frequent Internet user (AOR = 1.46, 95% CI 1.10–1.92), and having a self-report HIV positive status (AOR = 2.62, 95% CI 1.66–4.12). Those with a high school education or less were marginally more likely to engage in UAI (AOR = 1.39, 95% CI 0.96–2.02). Race and HIV testing history were not significantly associated with UAI (Table 3).

Table 3 Multivariate analysis of unprotected anal intercourse
Full size table

Discussion

Findings from this anonymous cross-sectional Internet survey provide insight into participation in prevention activities and its association with UAI among MSM who meet sexual partners through the Internet. While the Internet is known for being used to facilitate risky sexual behavior [2], our data suggests that some MSM who use the Internet to acquire sex partners are also exposed to HIV prevention information. In fact, over half of our sample was exposed to prevention activities, and those who were exposed to prevention, whether online or offline, were significantly less likely than those that were not exposed to engage in UAI. Our findings in WHBS indicate an association and not causality. Other factors, such perception of risk may also explain this association.

Our data also indicate that MSM who reported frequently using the Internet, aged under 30, HIV positive, and using drugs in the past 12 months were more likely to engage in UAI. Online prevention efforts for MSM who meet sex partners online should focus on these groups. Future prevention planning should also develop engaging strategies such as interactive web-based tools. Innovative online video interventions for MSM have already been developed with success [7]. Chiasson et al. developed a 9-min online video intervention for MSM that included behavioral questionnaires at enrollment and 3-month follow-up [7]. The researchers found that in the 3 months after the online video intervention men were significantly more likely to disclose HIV status to partners and less likely to report a casual partner or UAI in their most recent sexual encounter. Engaging frequent Internet users may be the first step in drawing them into online prevention and may be a gateway into either online and offline prevention.

The association between younger age and UAI has been found in previous studies. For instance, in their online study to examine risk factors for sexually transmitted diseases and HIV among young adults who seek sex partners on the Internet, McFarlane et al. [8] reported that young adults using the Internet were at greater risk for sexually transmitted diseases. Further, recent updated figures from the CDC for all youth—which would include young MSM—indicate that youth under the age of 30 account for 34% of new infections [9]. Internet prevention among young adults and young MSM in particular should address health promotion such as correct and consistent condom use [8]. For these young adults, prevention efforts might include frank discussions about risky sexual behaviors such as engaging in unprotected anal sex.

The association between self-reported HIV positive status and UAI that we observed and that has been reported by others [3, 10] also needs further attention. Although only a small proportion of our sample self-reported being HIV-infected, it is an important population to reach via online and offline prevention efforts. In their case–control study of men recruited online Hirshfield et al. [11] found that a large proportion of the HIV-positive men who reported multiple sex partners also reported unprotected anal sex with partners who were HIV-negative or of an unknown status, potentially transmitting infection to others. Identifying this population creates an opportunity for education about sexual risks, especially among those who might be newly diagnosed [6].

Prevention efforts should also address the risk of drug use in the context of meeting sexual partners online. Our findings show a significant correlation between drug-using MSM and UAI. A study by Chiasson that compared online and offline sexual risk in MSM found that 25% reported use of drugs before sex [3]. In this same study, a predictor of UAI in multiple partner encounters included use of crystal methamphetamine before sex. In related findings of a sample recruited online, a study by Hirshfield et al. found that UAI, use of crystal methamphetamine, and multiple sex partners were strongly associated with an STD [11]. Recent findings like these and those reported here clearly underscore the need for prevention targeted toward drug-using MSM and, as previously discussed, HIV positive MSM.

Further, in our multivariate analysis we were able to determine that the odds of engaging in UAI in the past year were greater among respondents who reported that their last sexual partner was a main partner. As previously stated, in our analysis we were limited by the survey to using the type of sexual partner (main or casual) at the last sexual encounter, since WHBS did not address the type of sexual partner in the past 12 months. Having a main partner at the last sexual encounter is a conservative measure of this type of sexual partnership, which suggests that the association of UAI with having a main partner at the last sexual encounter is robust. Our findings indicate that the risk of HIV transmission from our sample may be greater among MSM with high-risk main partners than from casual partners. A recent finding from a multi-site study among MSM found that most HIV transmissions (68%) were from main sexual partners [12]. These transmissions were attributed to higher number of sex acts, more frequent receptive roles in anal sex, and lower condom use with main partners [12]. To fully understand this association, further research is needed to explore HIV risk among MSM whose last sexual partner was with a main partner.

Although we found no significant association between UAI and race/ethnicity in our multivariate model, the CDC [9] has found that overall blacks and Hispanics bear the highest burden of new HIV diagnoses. Among these populations are MSM. Hence, prevention planners should develop interventions that are appropriate for black and Hispanic MSM when they are developing prevention programs for MSM.

Limitations

WHBS data are not representative of all NYC MSM; specifically, findings can only be generalized to MSM who completed the survey among the WHBS-participating websites. Further, WHBS data are self-report and subject to bias. Notwithstanding the anonymity of the survey, the nature of some of the questions asked—illicit drug use, number of sex partners, group sex participation—might make responses prone to social desirability leading to underreporting. It should also be noted that there was under representation of non-White MSM, which may be related to differential access to the Internet, types of WHBS-participating websites (only 1 WHBS-participating website catered specifically to MSM of color), and types of models used in the recruitment banner advertisements (limited racial diversity). As with all Internet surveys, nothing prevented MSM from taking a survey multiple times; however, this is unlikely as there was no incentive given for survey participation. Finally, as with all cross-sectional studies, WHBS can only indicate associations, not causality. Hence, we cannot conclude that prevention activities caused any changes in sexual behavior; we can only suggest that associations exist between prevention activities and sexual risk behaviors.

Conclusion

In our sample of Internet-using MSM, the odds of engaging in UAI were lower for those exposed to online or offline prevention. The findings suggest that providing online prevention programs to MSM who use the Internet to meet sex partners can decrease sexual risk behavior. The data contribute to the growing literature on HIV prevention among MSM generally, and more specifically, on the potential of online prevention. Future research should address the efficacy of “combo” prevention—that is, a combination of online and offline prevention activities—in reducing UAI among MSM. This research should target drug-using MSM, MSM under the age of 30, and self-reported HIV positive MSM, three populations that in our study were most likely to report UAI.