TRADITIONAL STRATEGIES to meet sex partners (e.g., going to bars, meeting through friends) have been augmented by the practice of soliciting sex partners via the Internet. This medium is emerging as an environment of potential risk for acquiring or transmitting sexually transmitted disease (STD), including HIV, the virus that causes AIDS.
Various Internet venues (e.g., chat rooms, bulletin boards, virtual communities) provide an excellent means of initiating sexual contact without initial face-to-face interaction. For example, adolescents who prefer to hide sexual activity from parents or peers may initiate contact on the Internet. Men who have sex with men (MSM) but who do not openly identify with homosexual men may use the Internet as an "invisible" means of acquiring sex partners.
Recent Internet statistics indicate that there are 129 million English speakers with Internet access.1 Several researchers have studied the Internet to understand the nature of the social system that users create. Most notably, Parks and Roberts2 administered a survey to users of MOOs (Multiple User Dimensions, Object Oriented), which are real-time, text-based Internet environments similar to chat rooms. They found that more than 90% of respondents had formed personal relationships with MOOs, approximately one third of which had resulted in face-to-face meetings. Romantic relationships comprised more than one quarter of the relationships described, 84% of which were opposite-sex relationships.2
There are journalistic rather than scientific data to support the theory that the Internet is used by MSM to facilitate sexual contact. A report published in the October 12 issue of www.salon.com entitled "You've got Male" outlines findings suggesting that America Online (AOL), an Internet service provider, has become a popular alternative to gay bars and bathhouses for MSM to solicit sex partners.3 Shaw reports that MSM find Internet Relay Chat to be an alternative- indeed, the only alternative-to gay bars.4 In August of 1999, six men with syphilis in San Francisco reported meeting male partners through an America Online user-created chat room. One additional man with syphilis had met his partners through a different online chat venue. The seven infected men listed a total of 99 sex partners from the previous 3 months; five of the seven men were HIV positive.5
This research was undertaken in to understand the nature and social context of Internet-initiated sexual contact and the implications of this contact for public health. We define the sexual Internet as the collection of Internet venues specifically created for the purposes of facilitating sexual contact between users. We do not examine other Internet interest groups, such as hobby-oriented chat rooms, which may also facilitate relationships or face-to-face contact.
Participant observation was used for the research.6 Although the participant-observation method frequently combines detailed observation in social settings with in-depth interviews, this effort focused exclusively on observation. Because little is known about how sexual contact is initiated via the Internet, we wanted to limit introduction of bias into social interactions by not acknowledging our presence as researchers, thereby avoiding the potential of altering an interaction.
We first sought to identify Web sites that would be appropriate for observation (i.e., sites with available chat rooms targeted at groups who may be at increased risk for STD or HIV transmission). These included gay sites, sites for swingers or persons interested in group sex, and sites for heterosexuals interested in meeting partners. Because of the enormous number of such sites available on the Web, we limited our search to sites with themes that would suggest person would seek meetings via Web contact.
We began by conducting keyword searches using the Alta Vista search engine on the World Wide Web. Our searches revealed thousands of possible sites available for exploration, but we chose to limit exploration to the top 20 sites from each individual and combined keyword search to identify sites for further observation.
A site was chosen for observation based on two criteria: (1) if the site attracted or targeted gay men or nongay-identified MSM, heterosexuals interested in meeting (including teens), or heterosexual couples interested in sexual encounters with other couples; and (2) if the site had chat rooms for persons logging in to the site. A number of sites meeting these criteria charged a fee for accessing their site. A number of other sites were accessible only through membership in a company that verifies individual age as 18 years or older by charging a monthly fee to a credit card in exchange for access to sites with adult material. We learned that many of the sites oriented toward heterosexual couples seeking liaisons with other couples required age verification, so we purchased this service for a number of months to observe these sites. We limited observations to chat rooms only and did not observe any pornographic sites.
The Institutional Review Board of the Centers for Disease Control and Prevention exempted this research from human subjects approval. Observations took place between February and June of 1999. Each observation session lasted for a minimum of 30 minutes, and in early observations sessions lasted between 45 minutes and 2 hours. In most chat rooms, observers and participants were required to log in with a screen name before being allowed access to a room. For the purposes of this study, we logged in as "Sean," a name selected because it is gender ambiguous. We offered no further personal information, and did not participate in the conversation occurring in the chat room under observation.
Three observers participated in the observation process; each observer took the lead in observing a room. After one observer completed the initial observation of a room, another observer would follow up and complete a number observations of that same room to facilitate interobserver reliability.
Once logged in to a chat session, we were able to access information on simultaneously logged-in chatters through their "profiles" (i.e., information they offer about themselves to other chatters). When available, we documented gender, age, sexual orientation, and ethnicity from these profiles. When not available, we were sometimes able to document these demographic data from chat text directly.
We developed a semistructured guide for field notes to facilitate documentation of behaviors relevant to STD and HIV prevention. This guide offered observers the opportunity to document evidence of demographic details and any discussion of behaviors considered risky for STD and HIV transmission, including evidence that chatters were arranging face-to-face meetings.
Observers noticed that the information obtained through field notes and chat text from a room became repetitious after multiple observations. When no new chat topics were emerging from observations, the observer ceased to observe the room.
Field note data were entered into EZText (Centers for Disease Control and Prevention, Atlanta, GA), a text analysis software.7 Using a content-analysis methodology, we analyzed data with an iterative coding process, beginning with open coding to identify themes and ideas expressed by observers in each section.8 One person was responsible for all coding in this process, so there was no need for documentation of interrater reliability. However, this person consulted fellow investigators regarding codes assigned to their field notes to ensure that the codes matched their observations. For each question on the field note form, observer responses were assigned multiple codes to reflect the range of topics addressed in the observation.
Subsequent to the open-coding process, the lead author sought consistent patterns or themes expressed by codes for each field note segment, and compared these patterns across multiple groups of observations. Stratifying observations by STD and HIV risk categories including MSM, heterosexuals, and couples (swingers) provided the most compelling differences in patterns of behavior or suggestion of behavior. In addition, stratification by type of site (i.e., AOL versus non-AOL sites) revealed illuminating patterns regarding access to the Internet that have important implications for STD and HIV risk. Therefore, results are presented to highlight these stratified patterns.
This research includes participant observation of 175 chat-room discussions including both AOL (n = 63) and non-AOL (112) rooms. We will present findings related to the demographics of chatters and the opportunity for risk behaviors generated by chatters online.
Of 175 observations completed, 81 (46%) were in rooms oriented toward gay men or MSM, 61 (35%) were in rooms targeting heterosexuals, and 33 (19%) were in rooms targeting couples. The target audience was determined by the site, floor, or room name. We completed 63 observations (36%) in AOL rooms; these rooms targeted MSM, heterosexuals, or couples and are included in the percentages reported previously. MSM rooms comprised 32% of the AOL observations (heterosexual sites, 35%; swinger sites, 32%). Comparison of these data should take into account that the observations are not mutually exclusive.
Table 1 shows detail related to profiles observed online. Chatters use profiles and handles or screen names to present personal information about themselves to chatters online. Some sites have opportunities to present detailed profile information, others have limited profile information, and others offer only opportunities to present screen names or handles. As mentioned previously, profiles could present demographic information, including gender, age, sexual orientation and race or ethnicity, but frequently they did not contain any or all of these details. Rather, profiles were more frequently used to present information about personal likes and dislikes. Profiles included revealing risk-related information about what a chatter sought in a sex partner, what types of sexual activities the chatter preferred, and where the chatter was located.
Table 2 is a quantitative summary of the demographics and risk behaviors observed stratified by type of room. Observations are of chat room conversation and profile detail. Given the large number of subscribers to AOL, we looked at the frequency in AOL rooms to see how much opportunity for risk is available in these rooms compared with non-AOL rooms. Because of the relatively even distribution of type of risk group observed in AOL rooms, our AOL findings do not weight a particular behavior or item observed in these rooms toward any one risk group.
Demographic and Personal Information Revealed
Chatters more frequently reveal their age in MSM rooms (54%) compared with straight (28%), swinger (6%), or AOL rooms (47%). When present, information about the age of chatters was present suggested that many were young (44% age 30 years or younger) and middle aged (26% age 31-45 years). Detail regarding ages was primarily available from profiles (32%) and chat text (27%).
Other personal information offered by chatters includes geographic location, which was mentioned most frequently in AOL rooms (71%) and MSM rooms (59%). Examples include, "I'm in NYC," "Young Hispanic in Orlando." Sometimes chatters indicated where they were willing to go or whether they were looking for persons in a particular geographic area (e.g., "Looking for men in NYC," "wanted, college chicks in NC"). This geographic positioning could be a way of introduction or it could be a way to facilitate a meeting by suggesting convenience for potential partners (e.g., "[I am in a] downtown Minneapolis hotel, you want some of what I got," from a room called Gay-ParaChat on Chat Planet). In other cases we observed the technique used by persons who may wish to travel distances to make contact. For example, a man in an AOL discussion entitled SFM4M indicated, "[I am traveling] to SF 8/20 to 8/30, to Phoenix 9/24 to 9/29 and Dallas 10/8 to 10/11." Frequently, a chatter offered a link from the chat discussion to a personal Web site (16% of MSM rooms, 7% of straight and AOL rooms). Less common was the exchange of e-mail addresses.
References Made to STD Prevention
We were able to document evidence of concern for safe sexual behavior through the mention of condoms or, as was frequently the case, persons who indicated that they are "disease and drug free," (D+D−) and are seeking the same. Other indicators of concern with safe sex behavior were evident through statements such as "seeking fun and safe sex," or "he's a bottom looking for an HIV - for a top." Also documented were statements such as "looking for clean fun," and "[I am a] bottom, safe." We observed a room on AOL created by members called healthHIV that had no references to condom use, but chatters discussed abstinence as a way to promote HIV prevention. Another AOL site created by members called HIVMSM had profiles with information such as "Toppin HIV+ but healthy," "HIV+ but healthy, like to bottom," "Top dad 4 U HIV-Pos, friendly, safe."
References Made to Sexual Risk Behaviors
Chatters in MSM rooms mentioned anal sex (looking for a "bottom" or offering sex as a "top" or "bottom" partner) in 26% of sessions observed, and in 27% of AOL rooms observed. For example, "anyone for a nice bottom?" from DenverM4M and "Accord 1234 is a bottom male looking for top," from barelylegalm4olderm.
References to seeking or having multiple partners occurred most frequently in the swinger rooms observed (21%) compared with AOL rooms (15%), MSM rooms (10%), and was not common in straight rooms (1%). Examples include "[we are] looking for 10 couples who want to swing not talk...all of the women will be bi," from the room Hottub on Swingersmeetingplace.com and "I have a different man for every occasion," from a chatter on www.gayweb.com.
Chatters seeking a sex partner or partners was observed frequently (36% of swinger rooms, 27% of AOL rooms, 17% of MSM rooms, 9% of straight rooms). Examples of these observations include, "looking for casual encounters," from the Lobby on www.adultfriendfinder.com, a site for straight adults. Also, Kerby from AOL adultf4moffline was "looking for any ladies to have hot sex." Couples sought couples with language such as "we like to play with either biguys/gals or couples," (from AOL bicouples). MSM sought sex with chat such as "looking for a top daddy," and "I can be an insatiable bottom; I'm looking for guys who are looking to play now," both from AOL MSM rooms.
We documented many cases in which chatters discussed having met persons through online contact (9% in MSM rooms, 15% in swinger rooms, 8% in AOL rooms). Negotiations to meet were also carried out in chat rooms (11% in MSM rooms, 5% in straight rooms, 3% in swinger rooms, 17% in AOL rooms). "Has anyone ever hooked up from this room?" was a query on an AOL site for couples, and chatters responded, "I have," and "yeah, we have." One observation revealed plans to meet among two gay men in a room called J/O Action Now on www.gay.com:
W: "I'm also changing planes for two hours at O'Hare in a week."
M: "So I can come and bl** you...I'm free after two that day."
W: "I should be done with business at 3."
M: "Let's do it then."
Data presented here have some limitations. Our observations were not comprehensive, in that we were not able to observe chat discussions in every chat room on the Internet. Furthermore, our observations represent limited snapshots of a small proportion of chat discussions that occurred during the period that observational data were being collected. Data are not quantitative, and therefore cannot present findings with statistical inferences. However, because we targeted our observations to specific risk groups and followed a systematic observation process, we were able to document important information concerning the risk that persons incur by participating in online chat discussions.
The data show that sexual risk is possible through contact in Internet chat rooms. The extent and prevalence of risk behaviors that result from an meeting online is currently unknown. These data from qualitative observations can only establish the type and context of risk. We have shown that rooms targeting specific groups (i.e., MSM and swingers) offer more evidence of chatters who engage in sexually risky behavior and offer chatters more opportunities to meet through online contact. We have also shown that AOL offers persons multiple opportunities to meet via chat rooms, and news released in August 1999 revealed that such opportunities resulted in the transmission of syphilis.5 In several instances, we have been able to document actual evidence that people have met after participating in an online chat room together. There is also evidence that arrangements to meet may be among the objectives for the initiation of private-message discussions.
The mention of risky behavior is common online, and there is evidence that persons use this medium to find sex partners and continue to solicit partners online. Why should we be concerned about this finding? We acknowledge that persons who are going to engage in unsafe behaviors will continue to do so with or without the Internet, minus an effective intervention. However, we cannot ignore the implications that the Internet influence may have on sexual risk and disease transmission. A person seeking a clandestine liaison pre-Internet would be forced to go to some effort to seek out partners, through bars, dating services, or other methods. A person who is not inclined to frequent bars for casual sex partners may have been dissuaded from looking. Couples seeking group sex would have to rely on word of mouth for interested partners. The possibility for anonymous sex-partner solicitation has, therefore, been possible through use of local newspaper personal advertisements and 900-number telephone lines. Hamers et al9 demonstrated that personal advertisements used for partner solicitation have increased emphasis on mention of HIV serostatus as a criterion for partner selection. Additional work in this area has been done by Hatala.10 We know of no other assessment of personal advertisements or telephone solicitation that offers comparable consideration of the role of media in sexual risk taking. The data presented here suggest that the Internet has a greater and more instantaneous reach than any other medium to facilitate encounters that result in sexual activity. It has affected sex partner solicitation in the same way it has our business and home lives: it is faster and easier. This may translate into faster, easier, and more efficient transmission of disease and infection.
Mention of safer sex on the Internet is not absent. The evidence of persons seeking disease-free and drug-free partners suggests that persons are taking some precaution to screen potential partners, even though we know nothing of specific precautions assumed once a meeting takes place. Evidence of safer-sex discussions, however, indicates that the Internet may be an appropriate and useful tool for the promotion of healthy sexual behaviors, including condom use. To address the concerns about the potentially more rapid transmission of disease and infection, we must develop and test messages for prevention that can be targeted to chat room users for the purpose of increasing knowledge, skills, attitudes, and behavior regarding sexual risk reduction. This can be accomplished in multiple ways: health educators can offer advice to chatters about prevention, negotiations with Web sites for STD and HIV-specific topic chats can be initiated, and care providers can offer their e-mail address to chatters as a method of obtaining more in-depth and personalized information regarding risk reduction. Other methods for increasing STD and HIV prevention using the Web are already in place through the comprehensive and detailed sites designed by local and national organizations dedicated to infectious disease prevention. These sites can be evaluated for their efficacy in promoting reduction of risky sexual behaviors.