Is ‘bareback’ a useful construct in primary HIV-prevention? Definitions, identity, and research

Abstract

The terms bareback and bareback identity are increasingly being used in academic discourse on HIV/AIDS without clear operationalisation. Using in-depth, face-to-face interviews with an ethnically diverse sample of 120 HIV infected and uninfected men, mainly gay-identifying, and recruited online in New York City, this study explored respondents’ definitions of bareback sex, the role that intentionality and risk played in those definitions, and whether respondents identified as ‘barebackers’. Results showed overall agreement with a basic definition of bareback sex as condomless anal intercourse, but considerable variation on other elements. Any identification as barebacker appeared too loose to be of use from a public health prevention perspective. To help focus HIV-prevention efforts, we propose a re-conceptualisation that contextualises risky condomless anal intercourse and distinguishing between behaviours that are intentional and may result in HIV-primary transmission from those that are not.

Introduction

Some commentators have called the barebacking a failure of HIV prevention (Goodroad et al. 2000) and a ‘craze’ (Crossley 2002, p. 47). Others have pointed out that bareback behaviour and barebackers should be specifically targeted for renewed HIV-prevention efforts (Halkitis et al. 2005b, Wolitski 2005, Berg in press), given that ‘barebacking and its corresponding behaviours pose immediate public health risks’ (Halkitis et al. 2005b, p. S27). Furthermore, Grov et al. (2007, p. 534) warn about ‘the growth of barebacking-centred online communities and Web sites’. Yet, as a review of the literature reveals, bareback behaviour and identity are inconsistently operationalised terms that run the risk of being reified, creating confusion rather than scientific progress. Our study seeks to shed some light on this dilemma.

Bareback definition

The term ‘bareback’ appeared in the gay press in the mid-1990s. It initially referred to intentional condomless anal intercourse, mainly among HIV-infected gay men (Gendin 1997). However, by the time Silverstein and Picano published a new edition of the iconic The Joy of Gay Sex in 2003, bareback was defined simply as condomless gay sex. Early debates within the gay community focused on the risk of HIV transmission that the practice entailed for both HIV-seroconcordant and serodiscordant partners (see Carballo-Diéguez and Bauermeister 2004, for an earlier review; see also Junge 2002). This attracted the attention of researchers who saw in most cases the need to define the term. Some researchers chose ‘intentional condomless anal intercourse in HIV-risk contexts’ (Carballo-Diéguez and Bauermeister 2004, p.1; Suarez and Miller 2001), noting two key elements—intention and risk—that might distinguish the term from other less precise definitions. Other researchers defined it as: ‘intentional anal sex without a condom with someone other than a primary partner’ (Mansergh et al. 2002, p. 653; Berg in press); ‘intentionally seeking or engaging in unprotected anal sex among HIV-positive gay men’ (Elford et al. 2007, p. 93); intentional unprotected anal intercourse regardless of serostatus or partner type (Halkitis et al. 2003, Grov 2004, Tomso 2004, Bimbi and Parsons 2005); or any ‘sex that occurs without the protection provided by a condom’, not limited to gay men (Gauthier and Forsyth 1999, p.86).

Lacking a standard definition of, and consensus on, the role that intentionality of condomless sex and HIV-transmission risk (or lack of it) play in bareback sex, some researchers went back to the sources, i.e., asking gay men what bareback sex means. In most cases, brief surveys were administered online or to community samples asking respondents to define bareback sex or presenting scenarios with degrees of intentionality. Using this approach, Halkitis et al. (2005a) and Halkitis (2007) conclude that bareback meant anal intercourse with no condom use, including heat-of-the-moment cases in which there is temporary non-use. Similarly, Huebner et al. (2006), in a Central Arizona study, found that 64 percent of their respondents concur with Halkitis et al.’s participants, but add that this means with any sexual partner. These reports present data from quick surveys, sometimes no longer than five minutes, in which respondents were not asked to describe their understanding of the term. Furthermore, when brief scenarios were used to elicit information, they did not specify the HIV status of protagonists. As a result, proportions of respondents endorsing different definitions vary widely.

Many authors acknowledge problems in defining bareback. Yet, that acknowledgement has not led to any resolution, with authors choosing differing terms for different reasons. This definitional imprecision leads to difficulty in comparing findings andin developing evidence-based prevention responses.

Bareback identity

Researchers have also tackled a related topic: bareback identity. Yep et al. (2002, p. 4) argue that barebacking ‘may be viewed as reinforcement of a sexual identity, resistance to imposed behavioural norms, creation of a new sexual and political identity [emphasis added], or a continuation of practices unaffected by organized messages aimed at stopping such practices’. For Wolitski (2005, p. 11), barebacking may be ‘a cause for increased risk by providing a social identity for men who prefer unprotected sex, creating role models that celebrate the benefits of unprotected sex, changing social norms about protected and unprotected sexual practices, and establishing social and sexual networks of men who prefer unprotected sex’. Shidlo et al. (2005, p. 120) state that ‘a barebacker assumes an identity as someone who practises intentional UAI [unprotected anal intercourse] and experiences it as ego-syntonic, or consistent with his sense of self: “I have bareback sex because this is who I am”’. Halkitis et al. (2005b, p. S28) speculate that ‘barebacking behaviour and barebacking identity may be very different constructs, just as gay identity is not necessarily synonymous with same-sex behaviour’.

Based on the idea of a distinct bareback identity, studies have compared individuals who identify as barebackers with those who do not. Thus, Halkitis et al. (2005b, p. S33) found the former more likely to report alcohol and other drug use and sexual risk behaviour, and ethnic minority men less likely than white men to be both familiar with the term and identify as barebackers. Elsewhere, Halkitis (2007) asked participants to indicate level of agreement with a six-item ‘Barebacking Identity Scale’ with items such as ‘I describe myself as a barebacker’ and ‘Barebacking is an important part of who I am’. The analysis failed to show associations between this scale and number of bareback partners or sexual behaviours. Parsons and Bimbi (2007) used a 10 to15 minute survey administered at community events in New York City and found that about 12% of respondents identified as barebackers—yet, the meaning respondents attached to barebacking was not explored. These same authors suggest that bareback as an identity may relate to a ‘resistance’, not only to homophobic mainstream culture, but also to gay communities that regard barebacking with suspicion; there is textual evidence of this on some of the bareback websites we researched (Carballo-Diéguez et al. 2006). However, as we will see below, our respondents did not register this resistance. Therefore, while barebacker can be understood as a new sexual identity, how barebacker is operationalised as an identity in practise is no clearer from the literature.

The present study

Given this lack of consensus, we designed the Frontiers in Prevention study to explore, among other topics, from the perspective of men who report engaging in bareback sex, the meaning of bareback sex, which words besides bareback are used to name the practise, and whether respondents identify as barebackers. Furthermore, we sought to redefine the term within a conceptual model that might help orient future work in this area.

Methods

Sample

In the first phase of the study, we identified the six most popular free Internet sites used by men in New York City to meet other men interested in barebacking (see Carballo-Diéguez et al. 2006, and Dowsett et al. 2008). Next, between April 2005 and March 2006, we recruited men who met these eligibility criteria: (1) be at least 18 years old; (2) live in New York City or can commute; (3) report using the Internet to meet men at least twice per month; (4) self-identify as a barebacker or as someone who practises barebacking (‘Are you into bareback or do you consider yourself a barebacker?’); (5) have had intentional, condomless, anal intercourse with a man met over the Internet; and (6) use at least one of the Internet sites identified in the first phase of the study. The men were recruited exclusively through these Internet sites in approximately equal numbers of European Americans (EA), African Americans (AA), Latinos/Hispanics (LH), and Asian/Pacific Islanders (API). We also stratified the sample to include about two thirds who reported both being HIV-negative and having had unprotected receptive anal intercourse in the previous year. Individuals who qualified were scheduled for interviews in our research offices as soon as possible after the screening.

Procedure

Respondents underwent a consent process that explained that the study sought to explore ‘barebacking’ (no definition was given) and personal, interpersonal or other circumstances associated with barebacking, and which HIV-prevention methods were acceptable to people who do not use condoms. Then, each respondent underwent an audio-recorded, in-depth, face-to-face interview conducted by one of three clinical psychologists on our team. A structured assessment using a Computer Administered Self Interview (CASI) followed, which, besides demographics and HIV status, explored inter alia the respondent’s sexual identity and behaviour in the previous two months. The interviews lasted about two hours, and respondents were compensated with US$50 for their time. This study was approved by the Institutional Review Board at the New York State Psychiatric Institute.

In-depth interviews

A guide provided the structure for the in-depth interview. It started by asking about the respondent’s motives in volunteering for the study, then inquired about his use of the Internet to meet other men, his preferred Web sites and his strategies in meeting men. During this discussion, if the respondent did not use the word bareback, the interviewer would say: ‘Let’s talk about bareback sex now. If you had to give a “dictionary definition” of bareback sex for someone who never heard of it before, what would you say?’ If the respondent merely replied: ‘sex without condoms’, the interviewer would probe further: ‘If two people are using a condom and it breaks, would you call it bareback?’ and ‘If a couple is monogamous and HIV-negative, would you call the sex they have bareback if they don’t use condoms?’ This probing explored whether intentionality and HIV-transmission risk were present but not explicit in the respondent’s definition. Next, the interviewer asked: ‘What are other words people use for barebacking?’ Then, after discussion about the individual’s experience of bareback sex, the interviewer asked: ‘Do you consider yourself a barebacker?’

Data analysis

A commercial service transcribed the interviews. We verified the accuracy of transcripts by comparing them with the recording. We next created a preliminary codebook consisting of major themes, with the topic of each section from the interview guide as a first-level code (e.g., ‘bareback definition’ became a first-level code). The codebook included definitions, inclusion and exclusion criteria, and examples. We coded four transcripts independently and compared them to identify concurrence or lack of it in using the codebook. We modified the codebook where necessary and conducted independent coding using NVivo, a software programme for qualitative data analysis. Once our four trained coders reached an 80% ‘intercoder convergence’ (Hruschka et al. 2004), we coded the rest of the interviews. Finally, the first author identified modal responses, omitting recurrent or repetitious text. Quotes from respondents presented here are accompanied by study identity number, ethnicity, HIV serostatus and age to give the reader a sense of the diversity of respondents. The limited quantitative data used in this report was processed using SPSS.

Results

Demographics and sexual behaviour

The study’s respondents were 120 men, of whom 31 reported being HIV-infected. Their average age was 34, with HIV-infected men being, on average, seven years older and earning about US$10,000 less per year than uninfected men (US$30,000). Respondents had, on average, two years of college education. In terms of race/ethnicity, 35% were European American, 31% Latino/Hispanic, 28% African American and 17% Asia/Pacific Islander. Most of the respondents self-identified as gay. The men had had, on average, between 13 and 14 sexual partners in the previous two months and they practised unprotected anal intercourse frequently.

Defining bareback

Lack of condom use

I: If I were to ask you for a dictionary definition of bareback sex, what would you say?

R: Sex without condoms.

I: Is there anything else in the definition or just that?

R: Just that. (015, EA, HIV+, 43 years)

This lack of condom use was often the first and the only element mentioned. Sometimes, this was referred to as ‘flesh on flesh’ or ‘latex free’ sex. Some defined it without reference to any type of sex and encompassing both heterosexual and homosexual acts.

R: In the sexual context, this means, um, just not using a condom, and having intercourse, whether anal, vaginal or, or oral. Not using a condom. (093, AA, HIV−, 33 years)

At other times, definitions were restricted to anal intercourse without condoms and to sex among men only, for which some respondents considered that intrarectal ejaculation should also be included in the definition:

R: …a person who not only gets – has sex without using a condom but who also lets a guy come inside his anus or whatever you want to call it…That’s what the true barebacker [is] in my book. (082, LH, HIV−, 24 years)

Lastly, respondents sometimes noted the words ‘natural’ and ‘intimate’ as key to theirdefinition:

R:…[sighs] dictionary definition of bareback sex. Hmm. Wow. [laughter] Uh, the nitty gritty, the ultimate feeling of intimacy between two people…The way sex was meant to be…I consider it being truly intimate with somebody. There’s no barriers … It feels so much better than a condom. (050, AA, HIV-, 25 years)

Intentionality

A few individuals spontaneously included intention not to use condoms as a requiredelement of the definition.

R: I would say, the deliberate-less, the deliberate condom-less act of sexual intercourse, especially anal, um, what would you say? With the intent of enhanced pleasure. (033, EA, HIV+, 32 years)

However, most respondents did not mention intention. So, we then asked: ‘If two guys are having sex with condoms and the condom breaks, would you call that barebacking?’ About a third of the respondents rejected this and, instead, introduced into the definition intention to have condomless sex or lack of intention to protect (not necessarily the same thing). This was referred to, for example, as ‘prior agreement not to use condoms’, or ‘knowledge’ or ‘awareness that no condoms were used’. About one third of respondents, however, qualified the ‘condom breakage’ scenario:

R: No, that’s – no. Not if they don’t realize it or they’re, you know, I think if they realize it breaks and they choose to go without, then yeah, then it is barebacking. (019, EA, HIV+, 45 years)

Therefore, the intentionality was brought into the definition by the failure to stop having intercourse at the point of becoming aware that the condom broke. This nuance was not, however, without its reservations:

R: Technically, I guess, briefly, unless they don’t pull out or if they continue, then yeah. But barebacking is usually a conscious choice. But I understand that is kind of complicating the situation. But if the condom breaks, then no, I don’t think so, because I think that barebacking isn’t by accident. It’s a conscious choice, unless the person’s fucked up and doesn’t know what you’re doing. (113, EA, HIV−, 31 years)

Awareness of risk

Some respondents volunteered that bareback sex was risky:

R: Um. [pause] Condomless, unprotected sex, with knowing, with the knowledge that you could be having sex with men who have sexually transmitted diseases, or who don’t know their HIV status or who are HIV positive. (010, LH, HIV−, 25 years)

To explore the issue of risk further, we asked: ‘If two people are HIV negative and monogamous, and they decided that they’re going to have sex without condoms, would you call that bareback or not?’ Many respondents said they would consider this bareback sex. Other respondents stated that condomless sex between an uninfected monogamous couple should not be considered bareback sex.

R: That’s not bareback … (Why not?) It’s not because these two guys are in a relationship. They’re in a monogamous relationship. They love each other. They’re both HIV negative. They know their status. They’ve – it’s natural, I mean, for the gay world…But it’s just natural for them to have sex without a condom, if they know neither one has HIV or has an STD, or whatever, and they’re not sleeping around on each other. (153, AA, HIV−, 41 years)

In other words, were it not for HIV and AIDS, condomless sex would not only be the norm among gay men, it would not even be considered an issue.

Other words used to refer to bareback sex

We also asked what other words respondents knew that referred to bareback sex. Besides noting the initials, ‘BB’, the word most frequently mentioned was ‘raw’.

R: Well, they used to call it…back, like 94, 93…they would say ‘I like skin-to-skin sex’. But that term didn’t really catch on the way that the term bareback did…And ‘raw’, you still hear people say ‘raw’. (033, EA, HIV+, 32 years)

Some ethnic minority respondents said that bareback was a term used by ‘White folks’, whereas:

R: Oh, black guys, you know, mostly just say ‘raw’, you know, or ‘skin-to-skin’. [laughs]. (039, AA, HIV−, 29 years)

There were richer metaphors too:

R: ‘Raw dog’, ‘raw sex’, um, ‘I wanna flood your hole’, um, a lot of fusion of hip-hop language, you know, applied to, to the sexual terms, um…‘I’m going to give you nutt’, as opposed to, um, ‘I’m going to give you my load’. So, what ‘cum’ meant – well, rather, ‘cum’ and ‘load’ is analogous to ‘nutt’, n-u-t-t, for Black and Latino men. The word ‘nasty’ is used, you know. (010, LH, HIV−, 25 years)

Other words or phrases used were ‘uninhibited’, ‘natural’, ‘pig’, ‘unwrapped’, ‘uncovered’, ‘unprotected’, ‘raunchy’ and ‘down and dirty’. A few respondents used the word ‘freak’. One Latino respondent even used crudo, the literal translation of raw in Spanish.

‘Seeding’ and ‘breeding’ were terms that appeared to denote ejaculation:

I: What other words do you use for barebacking?

R: ‘Raw’. And ‘bareback’. They’re the only two now. I mean, you see guys using euphemisms, but they usually refer to transmission of HIV. ‘Breed me’, ‘seed me’, give me the ‘taint’. So they really go beyond that definition. Anything beyond ‘raw’ and ‘bareback’, it means fucking bareback with someone you know to be HIV positive. (025, EA, HIV−, 53 years)

This is an important distinction. ‘Bug chasing’ and ‘gift giving’ are terms related to bareback sex that have caused much media fervour. These terms did not arise spontaneously in the interviews and, when probed, many respondents did not know about them and simply inferred what they might mean.

R: Bug chasers! Ah, yes. Those poor, deluded people who romanticize getting HIV. (045, LH, HIV+, 35 years)

Bareback identity

When we asked respondents if they considered themselves barebackers, most replies focused on condomless sexual practices, and based on that men responded either affirmatively or negatively, or qualified how the term would apply to them. A few respondents queried whether having condomless sex sufficed to define someone as being one thing or another.

Yes, I am a barebacker

About a third of the respondents labelled themselves as barebackers—and more often it was HIV-positive than HIV-negative men.

I: Do you think of yourself as a barebacker?

R: Yes.

I: Is that an identity?

R: That’s an identity. That’s the truth. The truth…is the light. So I’m a barebacker, baby. And I ain’t going to sugar-coat [it] – I’m a barebacker [singing], I’m a barebacker! [laughter] OK?

I: That identity, is that, is that a private one? Is that something you –

R: I would want somebody to know? Yes, I’m a barebacker. I feel… it, it, it gives me a sense of empowerment, so to speak. I feel good about [that] shit. Yeah, I like the ass, I like to fuck, and I like to get fucked. You know, and I like to be explicit. And I can get to the exact nature of what I’m about, so it empowers me. Barebacker, huh? You know, that is that term. (012, AA, HIV+, 38 years)

No, I am not a barebacker

About one quarter of respondents rejected labelling themselves. At times, this wasrelated to the stigma associated with the behaviour, particularly for HIV-negative men.

I: What keeps you from considering yourself that?

R: I guess just the stigma attached to it. The stigma and – I’m not going to have bareback sex with every guy I meet. That’s why I don’t consider myself a barebacker. (078, API, HIV−, 26 years)

Others said they did not consider themselves barebackers because they ‘did try to use condoms’ or it was not ‘the biggest part of my sexual experience’, or because labelling oneself as such would make others think, ‘Oh, sure, he’s a barebacker, so he’ll accept my dick inside him’.

Maybe I am a barebacker

A few respondents said that the label applied to them only partially or sometimes. Some contradicted themselves in trying to explain why:

I: Do you consider yourself a barebacker?

R: Sixty percent of the times, yes, I do, mm-hmm. Yes I do. You know, because like I say I do…my best to practise safe sex, but once, you know, I meet a certain person or – it’s like – it’s like something that will go off in me that I’ll be, like wow, I would just love to feel him inside, you know? Or I would just love to run up in them and – stuff like that. (083, AA, HIV−, 29 years)

One uninfected Latino young man reasoned that he was only 40% a barebacker because that was the proportion of times he had condomless sex with strangers, whereas the other 60% of the times he did it with people he knew.

Discussion

Our methodology for this part of our study relied exclusively on qualitative interviews so as to gain a deeper understanding of the phenomenon of barebacking without imposing preconceived notions. Since we specifically recruited men who identified as barebacker or someone who practises barebacking, and furthermore, someone who had had intentional, condomless, anal intercourse with a man or men met over the Internet, we expected some consensus in their accounts. Instead, the diversity of responses reported here suggests the distinct possibility that the same or even a wider range of views on barebacking pertain to gay men broadly. Furthermore, some of the minority ethnic men’s responses may indicate diversity among other, non gay-identifying, men who have sex with men.

We did not define ‘bareback’ for respondents, but they were in broad agreement that bareback refers to intercourse without condoms. This would seem to support investigators who reported that the colloquial term originally used mainly for HIV-infected individuals may have lost its early specificity (Halkitis 2007). However, generalisation has not occurred; we found much variation among respondents in the interpretation of everything beyond this initial phrasing. Our findings reveal pitfalls in considering bareback as a simple reference to condomless sex. This leads us, and others, to question the validity of some research undertaken thus far. Berg (in press) states:

Although early research suggested there was relative congruity between MSM and health professionals in their understanding of the term barebacking, the term may have developed too fast at the community-level for researchers to keep abreast, undermining researchers’ understanding of the behaviour and possibly the validity of extant empirical research about barebacking.

Moreover, the reification of barebacker from one who practises a behaviour to one who has a specific identity has exacerbated the confusion. Our respondents eloquently argued for and against the barebacker label. For some, identifying as barebacker might be interpreted as defiance of mores that restrict sexual freedom, but it is doubtful that it functions as the organising principle for a sexual identity. For other respondents, the label was an uncomfortable one that they either rejected outwardly or accepted partially, with different rationalisations to explain their views. There was evidence that stigma associated with intentionally having unprotected anal sex affected HIV-negative men in different ways from HIV-positive men. Yet, it is more complex than that. In fact, there was no single definition embraced by all men; and assertions from researchers, or practitioners, that there is a prevailing community-held consensus on what bareback means are not supported by these men, the very ones practising bareback sex regularly enough to anticipate consensus. Nor was there evidence of an overwhelming uptake or positioning of barebacker as a dominant or functional identity. These findings pose a challenge for research and HIV prevention.

First, we focus on research implications. Researchers are not simply reflecting larger confusion existing among gay men or in gay community discourse; researchers have played a part in creating this confusion. As Junge (2002, p. 196) argues, noting Farmer’s (1996) call for a critical epistemology of emerging infectious disease:

Farmer (1996) has focused particularly on how publication of AIDS research in peer-reviewed scholarly journals provides a site of discursive production wherein choices in lexical representation may reinforce stigmatizing or essentializing conceptualizations of certain populations.

Researchers, in arguing for a certain take on barebacking, bolstered by certain data, are involved in a discursive practise that exercises constructive power over the phenomenon—no research merely ‘reports’. As Tomso (2004, p. 88–89) argues:

[B]ug chasing and barebacking exist as phenomena [emphasis in the original] largely because of what Foucault would call the constitutive, disciplinary operation of scientific, activist, and popular discourses about them. This is to say that those who are currently investigating and writing about these phenomena, as much so if not more than the men whose sexual lives are the subjects of these investigations, are epistemologically accountable for the emergence of bug chasing and barebacking as social ‘problems’.

Second, we must note the interaction between researchers and our findings, with those responsible for designing and delivering HIV prevention. For researchers are not the only ones responsible; other social actors are involved too—activists, advocates, educators, commentators, the media and anyone else who proclaims on the issue. Within this interaction, the men who bareback lose their voice and the right to constitute bareback as it pertains to their lives. This can lead, inter alia, to a barebacker becoming a stigmatised ‘other’, an outcome that conceals rather than reveals the nature of the phenomenon and related prevention complexities. For, as Clatts and Mutchler (1989), also cited by Junge (2002, p. 196), state:

AIDS and the ‘dangerous and anti-social other’…fix our attention on a relatively small range of possible vectors of this disease…and direct our attention away from other possible factors of etiology and spread.

Supporting evidence-based prevention is a central purpose of this project—called Frontiers in Prevention. Given our results and literature review, we believe there is need to re-conceptualise the idea of bareback to focus public health discourse and inform its practice. However, Huebner et al. (2006, p. 70) warn:

Studies that fail to define barebacking for participants might be inquiring about any number of behaviours, depending on each participant’s individual understanding of the term. Additionally, even studies that do define the expression might encounter problems among participants who understand barebacking differently and ignore researcher definitions when responding. Even if participants can be compelled to suspend their own understanding of the term and to report about barebacking as defined by researchers, the external validity of such research is questionable given that definitions are constructed in the study that might not exist in the real world.

Yet, if no attempt is made to clarify this definition of barebacking, then confusion will continue, research incomparability will grow and evidence-based prevention will be even less possible.

Therefore, as a first step, there is a need to separate from other sexual events a category of condomless acts between two (or more) men, whether unintentional or intentional that do not constitute a risk for a HIV infection, either because both individuals are seroconcordant (both either positive or negative), and irrespective of whether the participants call their behaviour barebacking or call themselves barebackers. We think it useful to distinguish intentional condomless intercourse in HIV risk contexts (our preferred definition of bareback) from other condomless sex that is accidental, unintentional or non-consensual, and from condomless sex that is intentional but risk-free, such as between monogamous seronegative individuals or those non-monogamous couples who practice ‘negotiated safety’ (Kippax et al. 1993). We concur with Wolitski (2005, p. 25) that: ‘Differentiating the risks associated with barebacking from those associated with carefully reasoned risk reduction strategies that include unprotected sex between uninfected primary partners is critical’. We make this distinction for a number of reasons, not least being that each kind of condomless event in this first category requires a targeted HIV prevention strategy and, therefore, different health promotion approaches to assist the men involved to assess and reduce risk, e.g., educating to achieve successful negotiated safety. Also, in terms of primary HIV infection, intentional condomless sex between two (or more) HIV positive men, often referred to as ‘sero-sorting’ (Truong et al. 2006), should join this category. We recognise concerns about super-infection and STI risk, but we argue these warrant prevention strategies that are not central to primary HIV prevention.

Next, we need to clarify a second category: condomless anal intercourse that is unintentional but may result in primary HIV transmission. This would include heat-of-the-moment sex, condom breakage, forced unprotected intercourse between possibly serodiscordant partners, and what our respondents labelled ‘technical barebacking’—when unprotected sex occurred without intention between possibly serodiscordant partners. Here too, sexual positioning may render some behaviours riskier than others. This category does include those who may, in retrospect, label themselves barebackers after the event, even if this was not a label or shorthand term they applied to themselves beforehand. For many of these potential transmission scenarios, HIV science has long had inconsistent names, incommensurable theories, and incompatible data sets and findings, that to render these a single research object is nonsense. Furthermore, to call all these behaviours barebacking, either in research or for prevention purposes, confuses a complex set of sex events that do not have the same determinants or factors and cannot be addressed by singular prevention strategies.

The third category comprises condomless anal intercourse that is intentional and may result in primary transmission of HIV. The category includes ‘strategic positioning’ (whether the infected partner takes the receptive role with a partner of the opposite status) irrespective of intrarectal ejaculation, which may further qualify the risk. The issue of awareness of appreciable risk is central to this category—the risk calculus is done and condomless sex proceeds anyway. It is important not to assume ignorance or lack of knowledge always plays a part here. Lastly, this includes condomless anal intercourse that is not called barebacking by some of those who practise it, that which is called barebacking, and that by those who also identify as barebackers.

The usefulness of this typology lies in its capacity to focus attention where the epidemiological importance lies. Behaviours in the first category carry significantly less epidemiological importance from the perspective of primary HIV transmission than behaviours from the second and third categories, since the first, by definition, will not result in new HIV cases. It is not the case that HIV prevention education is not needed for this first category—mistakes can be made, judgement can be poor—but it needs a different kind of education. This includes strategies that stimulate correct and consistent condom use, as well as strategies for dispensing with condoms in certain situations and after clear precautions, e.g., sequential HIV testing, negotiated safety, sero-sorting, and HIV-seroconcordant monogamy.

The second category also includes a well-known cluster of prevention problems that HIV prevention researchers and educators have uncovered and grappled with as the epidemic progressed, and which have enhanced epidemiological importance. Contextual factors pertain too, e.g., knowledge about individual viral load, background viral load, and community prevalence. Adam (2005, p. 334) adds other circumstances to this category:

[B]arebacking is distinguishable from the wider range of unplanned, episodic, unprotected sexual encounters that men in interviews attribute to a variety of circumstances such as: a resolution to erectile difficulties experienced with condoms, through momentary lapses and trade-offs, out of personal turmoil and depression, or as a by-product of strategies of disclosure and intuiting safety.

There is increased risk of primary infection here, but the risk arises from a very different set of circumstances from the other two categories, and warrant their own prevention agenda. Educating men to understand these and develop strategies to deal with potential infection, e.g., by the use of post-exposure prophylaxis, may require different services as well.

This leaves the third category as a clear, although not entirely new, target (see Gold and Skinner 1993), in which the condomless anal intercourse with appreciable risk that may result in new infection is intentional. This is the category we call bareback, and we believe the use of that term should be restricted to this category, first by researchers and, subsequently, by practitioners. It is this specificity that will lead, we believe, to targeted prevention responses, ones that may not always be condom-focused, such as risk reduction (Suarez and Miller 2001), strategic positioning (Kippax and Race 2003), microbicide use when it becomes available (Carballo-Diéguez et al. 2007), pre-exposure prophylaxis (Nodin et al. 2008), or others.

Barker et al. (2007) argue that there is a ‘non-equivalence’ between barebacking and unprotected anal intercourse in HIV epidemiology. After all, we have known that not all condomless sex is risky for a long time—this was first discussed in the literature 15 years ago (Kippax et al. 1993, and based on data collected first in 1986). We recognise these are clearly issues that merit further research and discussion. However, we argue that it behoves researchers not to muddy the waters for both research and for the development of targeted HIV prevention or related health promotion with imprecise or inapplicable usage of the term bareback for all sex acts in all three categories.

Our typology acknowledges that some gay men regard all condomless anal intercourse as bareback, while other gay men say not all condomless intercourse is bareback, and there are those who use the term merely as shorthand. Nonetheless, our typology prevents the confounding of prevention education targeting in deploying any overarching notion of barebacking that cannot offer sufficient specificity. It recognises that while all who call themselves barebackers practise barebacking, not all who engage in barebacking call themselves barebackers. In recognising that confusion of usage, most importantly, our typology maintains a focus on contextualised sexual behaviour by focusing on the many kinds of ‘relational nexus’ (Riggs 2006) that are negotiated by men in sex (sometimes correctly, sometimes incorrectly in terms of HIV), rather than reifying risk as a single characteristic of individual personalities or psyches. After all, as Junge (2002) points out, a couple having condomless sex may include one partner who does it intentionally and another who is unaware of the situation or is responding to pressure. It is the sexual relations between people and how these influence behaviour—rather than the terms adopted or identity—which may, or may not, facilitate primary HIV infection. Such sexual events may be understood as belonging in different categories at the same time and so require different approaches to prevention.

There is heuristic value in focusing scientific inquiry on bareback Internet sites, on sex clubs that sponsor bareback events or offer bareback rooms, and on sero-sorting networks that facilitate condomless intercourse. However, in developing evidence-based prevention focused on the terms barebacking or barebackers, the target of these efforts would be less than specific and quite dispersed unless there is definitional rigour and the definition is restricted as our typology suggests. Moreover, if norms concerning condom use relax, e.g., with the increased effectiveness of highly active antiretroviral therapies (Crepaz et al. 2004), and as the use and meaning of the term is unevenly distributed among different ethnic gay men or different groups of men who have sex with men as we and others have suggested, the word bareback may fade from being useful in understanding what is driving the sexually transmitted epidemic in the USA—and maybe elsewhere. Until that happens, the focused usage we suggest may be the most helpful way out of the current confusion.

Our study has some limitations. It focused on a moving target—the use of a vernacular term—but one that is also discursively constructed in research. Of course, our conclusions are likely to be time-bound by where the debate is to date and should be considered with caution accordingly. The use of qualitative methods precludes generalisations to all gay men. Our findings may also be affected by sample specificity, and may not reflect what is currently happening outside the USA or amongst all men who engage in intentional condomless sex. Clearly, there is more to discover about the contribution of serostatus and ethnicity, and of ‘bug-chasing’ and ‘gift giving’, but space limitations have prevented us exploring that here. Nevertheless, our findings emphasise the importance for researchers, as a start, to define and operationalise the terms we use carefully to reduce confusion, and, in this case, to specify barebacking narrowly and assist thereby in producing better focused research. This is important if the evidence offered by researchers is to be useful to prevention workers in developing well-targeted programmes where bareback sex may be implicated.