Introduction
Based largely on the apparent success of using antiretroviral drugs to prevent seroconversion following occupational exposures to HIV [1], some have proposed using such drugs after sexual exposures. Concerns have been raised, however, that offering such treatment to individuals will lead to population-wide [2] increases in sexual risk behavior because of the perception that a `morning after pill' is available [3] to retroactively resolve risks taken during sexual contact. Although there are some data indicating that optimism surrounding the perceived success of newer anti-retroviral therapies is not related to riskier sexual behavior [4], evidence from recent studies concerning HIV treatment optimism in the wake of protease inhibitor availability suggests that there may be an association between sexual risk behavior and belief in new treatment efficacy [5-8]. In addition, data from one study indicate that risk behavior increased in some HIV vaccine trial participants, [9] perhaps because of a similar perception of technological protection from the virus. Some researchers have also recently observed an increase in both self-reported sexual risk behavior [10,11] and incidence of rectal gonorrhea [12] in gay men since highly active antiretroviral therapy has become available, leading to further speculation that risk behavior may be increasing among gay men because of new treatment optimism.
In San Francisco, the first federally-funded study of the feasibility and consequences of HIV post-exposure prevention (PEP) following sexual and parenteral exposures was recently completed [13]. Over 300 individuals enrolled, nearly all of whom were gay men presenting after possible sexual exposures to HIV. Although PEP is not yet offered in most localities and the study in San Francisco was the first of its kind, it is becoming more widely available through other research studies and community clinics. Given the recent data indicating that risk behavior may be increasing among gay men and that there may be an association between treatment optimism and risk behavior, it is important to examine any possible deleterious effects that PEP availability may have on the success of already existing forms of behavioral primary prevention. The present study was conducted as an initial exploration of the relationship between knowledge of PEP and sexual risk behavior among gay men over time as PEP became widely available in San Francisco.
Method
We conducted two street outreach interviewer-administered surveys to assess San Francisco gay men's familiarity with PEP as well as to explore if knowing about PEP was related to more sexual risk behavior (first survey in June 1998, Time 1; second survey in January 1999, Time 2). All male pedestrians passing strategic locations in the city's predominantly gay Castro neighborhood were approached by gay male interviewers (eight interviewers in total, three African-American, four Caucasian, one Latino, of various of ages) and asked to participate in a brief anonymous survey. Approximately 85% of individuals approached agreed to participate, which required 5-10 min. Those who completed the survey were given a debriefing form that provided contact information for the study investigators and information about the institutional review board approval.
Community outreach campaign
In between Time 1 and Time 2, there was a community-wide outreach campaign to raise awareness about PEP and to let San Francisco Bay Area residents know that the research study was underway and was offering free PEP to those with significant exposures to HIV. Although PEP had been available in the city during Time 1, it was not as widely known to be available. Two different outreach messages were selected by a series of focus groups of gay men and were used to publicize the research study. The first, which appeared widely throughout the Bay Area on billboards and posters had a picture of two men of color laying in bed together with the words `We know what to do' written above them in large print. Smaller text then read: `We try to have safer sex, but we've had accidents. We didn't panic. There's a free treatment that might prevent HIV infection, but you've only got 72 hours to start. We called right away.' A phone number was then listed to contact the study. A second campaign that was targeted to gay bars, sex clubs, and bath houses showed a broken condom with the words: `Oh Sh*t!' in large letters above it, followed by smaller text that read: `The condom broke. You've only got 72 hours to start a free treatment that might prevent HIV infection.' Again, the study contact phone number was listed at the bottom of the poster. Both campaigns were also reprinted on small cards that were made available in various locations throughout the Bay Area, including HIV testing sites. Participants in the second survey at Time 2 were shown these cards and asked if they had seen either of the campaigns.
Participants
At Time 1, before the community outreach campaign began, we surveyed 295 gay men (67% white, 9% African-American, 9% Asian-American, 9% Latino, 6% of another ethnicity). At Time 2, using six of the same eight interviewers (two African-American, three Caucasian, one Latino), we interviewed 234 gay men (74% white, 7% African-American, 4% Asian-American, 9% Latino, 6% of another ethnicity). The mean age of participants was 36.6 years at Time 1 and 35.3 years at Time 2. With respect to HIV status, 24% self-identified as HIV-positive at Time 1 and 28% did so at Time 2. At Time 1, almost three-quarters of the sample (74%) had graduated from a 4-year college or had an advanced graduate degree, 19% had a 2-year college/associate's degree or some technical school training; and 7% had a high school education or less. At Time 2, 75% reported having at least a 4-year college degree, 15% had a 2-year college/associate's degree or some technical school training; and 10% had a high school education or less. There were no statistically significant differences with regard to self-reported ethnicity, age, HIV status, or education level between Times 1 and 2.
Assessing the impact of PEP availability on risk behavior
Three analyses were used to examine the impact of PEP on sexual risk behavior. First, before being asked about knowledge of PEP, respondents were asked about sexual risk behavior (unprotected insertive or receptive anal intercourse in the past 3 months) and their HIV serostatus. After participants were asked if they had heard of antiretrovirals, they were asked two questions to assess familiarity with PEP: `Have you heard of these drugs being used by HIV-negative people who think that they have recently been exposed to HIV. Some people call this `PEP' or post-exposure prevention/prophylaxis' with a follow-up of Some people also call this the AIDS `morning after pill' to prevent HIV infection. Have you heard of that? The relative levels of sexual risk behavior were then compared between those who had heard of either PEP or the `AIDS morning after pill' and those who had not.
Second, participants who had heard of either PEP or the `AIDS morning after pill' were asked if they were aware that PEP was available in San Francisco to HIV-negative people who believed that they had been recently exposed to HIV. The relative levels of sexual risk behavior were then compared between those who knew that PEP was available in San Francisco and those who did not.
Third, participants who had heard of either PEP or the `AIDS morning after pill' were asked a series of follow-up questions with the stem `Has knowing about PEP made you more or less likely to. . .' followed by sexual behaviors and cognitions and responses ranging from `less likely' to `more likely'. The set of behaviors and cognitions included: `perform unprotected oral sex to ejaculation in your mouth', `have unprotected anal sex as a top [insertive partner]', `have unprotected anal sex as a bottom [receptive partner]', and `worry about condom breakage'. These questions were designed to assess participants' direct impressions of whether their personal level of sexual risk behavior had been affected by knowing that PEP is available. This third approach was used as a more direct (although potentially more biased) method of assessing the impact of PEP on sexual risk behavior.
Results
Knowledge of PEP and PEP availability
Almost all participants at both Time 1 (97%) and Time 2 (96%) had heard of antiretroviral drugs such as zidovudine and lamivudine. Forty-six percent of the sample at Time 1 had heard of `PEP' and 54% had heard of the `AIDS morning after pill' in general. In total, 172 (58%) men had heard of either `PEP' or the `AIDS morning after pill' at Time 1. Of these 172, 97 (56%, 32.3% of the entire sample) knew that PEP was available in San Francisco. At Time 2, more respondents had heard of PEP: 127 (54%) had heard of `PEP', 149 (64%) had heard of the `AIDS morning after pill', and 163 (70%) had heard of either term. Of the 163 who had heard of either term at Time 2, 104 (64%, 44% of the entire sample) knew that PEP was available in San Francisco. Thus, more participants at Time 2 had heard of PEP in general and more knew that it was available in San Francisco. Over half (57.4%) of the participants who had heard of PEP at Time 2 recognized one or both of the outreach campaign messages when they were shown the small cards and asked if they had seen them before.
Sexual Behavior
At Time 1, 163 men reported any anal intercourse in the past 3 months (55% of the sample). Of these, 19 (6% of the overall sample, 12% of the 163) had experienced condom breakage and 73 (25% of the overall sample, 45% of the 163) had anal intercourse without a condom at least once. At Time 2, 149 (64%) of the sample reported any anal intercourse in the last 3 months. Among these 149 men, 11 (5% of the overall sample, 7% of the 149) had experienced condom breakage and 76 (33% of the overall sample, 51% of the 149) had anal intercourse without a condom at least once. Men in the Time 2 sample were significantly more likely than men in the Time 1 sample to report any anal intercourse in the past 3 months (χ2, 3.82;P = 0.05) as well as any unprotected anal intercourse (χ2, 3.86;P = 0.05). Men were equally likely to have experienced condom breakage at the two time points.
Relationship between PEP availability and risk behavior
Is knowledge of PEP related to sexual risk behavior?
Among the men in the Time 1 sample who had heard of `PEP' or the `AIDS morning after pill' (n = 172), 24% reported any unprotected anal intercourse in the previous 3 months (see Table 1). Among those men at Time 1 who had not heard of either term, this number was 26%. These percentages were not statistically different (χ2, 0.21;P = 0.64) suggesting that there was no association with having heard of PEP and reporting risk behavior at Time 1.
The same comparison among only the 163 men who reported any anal intercourse in the previous 3 months indicated that 40% of those who had heard of either `PEP' or the `AIDS morning after pill' had unprotected anal intercourse at least once whereas 53% of the men who had not heard of either of these terms had done so. The comparison of these percentages was also not statistically significant (χ2, 2.32;P = 0.13) indicating no association between sexual risk behavior and having heard of PEP among the men reporting any anal intercourse at Time 1.
At Time 2, however, the comparison of the reported level of risk behavior between men who had heard of `PEP' or the `AIDS morning after pill' and those who had not yielded a significant difference (χ2, 4.06, P = 0.03). As shown in Table 1, more unprotected anal intercourse was reported among men who had heard of PEP (37%) than among men who had not heard of PEP (23%). As a follow-up to this analysis, this same comparison was done excluding the 66 men who self-identified as HIV-positive to assess if there was a significant relationship between knowledge of PEP and reported risk behavior among men in the sample who were more likely to view PEP as a `safety net' for themselves (i.e. they themselves could use PEP because they believed they were HIV-negative or were not sure of their HIV status). Among these men, 31% of those who had heard of PEP reported unprotected anal intercourse compared to 24% of those men who had not heard of PEP (χ2, 0.88, non-significant).
The same comparison among only the 149 men who had engaged in any anal intercourse at all in the previous 3 months also did not yield a significant difference. Among these men at Time 2, having heard of PEP was not related to reporting having engaged in any unprotected anal intercourse (52% of those who had heard of PEP and 47% of those who had not heard of PEP, χ2, 0.27;P = 0.60).
Is knowing that PEP is actually available in San Francisco related to sexual risk behavior?
As shown in Table 1, at Time 1, among the men who had heard of PEP or the `AIDS morning after pill' (n = 172), men who knew that PEP was available in San Francisco (n = 97) were not significantly more likely to report any unprotected anal intercourse than men who did not know that PEP was available in San Francisco (29% versus 18%, respectively, reporting any unprotected anal intercourse; χ2, 2.78;P = 0.10).
A similar result was found at Time 2 (see Table 1). Among the 163 men who had heard of PEP or the `AIDS morning after pill', 39% of those who knew that PEP was available in San Francisco (n = 49) reported any unprotected anal intercourse compared to 37% of those who did not know that PEP was available in San Francisco (χ2, 0.04;P = 0.84).
Do gay men self-report that knowing about PEP has made them more sexually risky?
The vast majority of the participants who had heard of PEP reported little or no impact of PEP on their sexual behavior at either Time 1 or Time 2. For example, a small percentage (3.7%) reported that knowing about PEP had made them `more likely' or `a little more likely' to have unprotected receptive anal intercourse at Time 1; this cumulative total was also small at Time 2 (1.8%). For all behaviors asked about, approximately 75% of men who had heard of PEP at Time 1 or Time 2 reported that it had no impact on their sexual behavior. Interestingly, however, 19.9% of those at Time 1 and 19.8% of those at Time 2 indicated that knowing about PEP had made them either `less likely' or `a little less likely' to worry about condom breakage.
Conclusions
Our study suggests that being familiar with PEP was not associated with increased sexual risk behavior among gay men in San Francisco before a community outreach campaign was conducted to publicize the availability of PEP in the city. After a community outreach campaign was undertaken, however, an association was detected between having heard of PEP and participants' self-reported levels of unprotected anal intercourse. Participants who had heard of PEP at Time 2 were more likely than those who had not heard of PEP to have engaged in unprotected anal intercourse. When men of unknown or negative HIV status at Time 2 were examined separately, however, this difference was not significant.
To explore the possibility of whether it was the knowledge that participants could actually access PEP in San Francisco that was related to their risk behavior, we compared the relative levels of risk behavior between men who knew PEP was immediately available in San Francisco and those who did not. Among those men who had heard of PEP at both time points, those who knew that it was actually available in San Francisco were not more sexually risky than those men who did not know that it was available. This finding lessens the likelihood that it was knowledge of PEP availability per se that caused the gay men in the Time 2 sample to have engaged in more risk behavior.
Finally, participants' own direct estimates of the impact of PEP on sexual risk behavior generally did not suggest that participants believed that the availability of PEP was causing them to engage in more risk behavior. A sizable number of men did, however, indicate that the availability of PEP had made them worry less about condom failure. Perhaps PEP is being viewed by these men as a `safety net' for when condoms fail or `slip-ups' occur.
These results should be interpreted with caution. Although all male pedestrians were approached, these men certainly are not a random sample of gay men. There are populations of gay men that were less likely to be reached with this sampling method (e.g. men living in other neighborhoods, more `closeted' men who avoid the main gay neighborhood). In addition, only a relatively small number of individuals in each of the ethnic minority groups was sampled.
There are, however, some important implications of these findings. For the men sampled in this study at Time 2, one analysis indicated that there may be an association with knowledge of PEP and engaging in risk behavior. Given the larger historical context of HIV/AIDS treatment optimism, it is also possible that some men are more risky because they believe in the efficacy of medical treatments - including when they are used for primary prevention. On the other hand, it is also possible that the findings of the current study reflect the fact that the community outreach campaign was targeted to venues such as gay bars and sex clubs where more sexually risky gay men are likely to be found. Thus, over time, it is likely that these riskier men would have been more likely to become familiar with PEP. The higher levels of risk observed in these men may therefore have already been present when the PEP campaign reached them. The lack of a comparison city where PEP had not become available makes it difficult to judge if the higher risk observed at Time 2 is a general historical trend, or a change that may have something to do with the availability of PEP in San Francisco. The finding that knowledge of actual PEP availability in San Francisco was not related to risk behavior supports the conclusion that it was not PEP availability per se that is causing more risk behavior among gay men. It will, however, be important to monitor this possible association over time as PEP becomes more widely accessible, whether through formal research studies, individual physicians, or informal friendship networks. Such data may prove important in deciding how and when PEP can be most effectively offered as for sexual exposures in public health interventions.
References
1. Cardo D, Culver D, Ciesielski C. et al. A case-control study of HIV serconversion in health care workers after percutaneous exposure. N Engl J Med 1997, 337: 1485 -1490.
2. Katz M, Gerberding J. Postexposure treatment of people exposed to the HIV virus through sexual contact or injection-drug use. N Engl J Med, 1997, 336: 1097 -1100.
3. Shriver M. Postexposure prophylaxis: A community member's perspective. In Psychosocial and Public Health Impacts of new HIV Treatments. Edited by Ostrow D, Kalichman S. New York: Plenum; 1999.
4. Wells H, Sherr L, Norton J, Miller R, Johnson MA, Elford J. The new HIV therapies: Implications for HIV prevention. AIDS 1998, 12: 117 -118.
5. Dilley JW, Woods WL, McFarlad W. Are advances in treatment changing views about high-risk sex? N Engl J Med 1997, 227: 501 -502.
6. Ekstrand ML, Stall RD, Bein E, Osmond D, Coates TJ. Treatment optimism is associated with sexual risk taking among gay men in San Francisco. Presented at the Fourth International Conference on Biopsychosocial Aspects of HIV Infection. Ottawa, July 1999.
7. Sherr L, Elford J, Bolding G, Maguire M. Combination therapies for HIV and sexual risk behavior among gay men in London. Presented at the Fourth International Conference on Biopsychosocial Aspects of HIV Infection. Ottawa, July 1999.
8. Kippax S. Risk practice and new treatments: optimism among gay men: a mixed message. Presented at the Fourth International Conference on Biopsychosocial Aspects of HIV Infection. Ottawa, July 1999.
9. Chesney MA, Chambers DB, Kahn JO. Risk behavior for HIV infection in participants in preventive HIV vaccine trials: A cautionary note. J Acquir Immune Defic Syndr 1997, 16: 266 -271.
10. Ekstrand M, Stall RD, Paul J, Osmond D, Coates TJ. Gay men report high rates of unprotected anal sex with partners of unknown or discordant HIV status. AIDS 1999, 13: 1523 -1533.
11. Van de Ven P, Prestage G, French J, Knox S, Kippax S. Increase in unprotected anal intercourse with casual partners among Sydney gay men in 1996-1998. Aust N Z J Public Health 1998, 22: 814 -818.
12. Centers for Disease Control and Prevention. Increases in unsafe sex and rectal gonorrhea among men who have sex with men - San Francisco, California 1994-1997. MMWR 1999, 48 :45-48.
13. Kahn JO, Bamberger J, Roland M, et al. HIV postexposure prevention after sexual or needle sharing exposure: The San Francisco PEP Project. Presented at the 12th World AIDS Conference. Geneva, June-July 1998.
© 2000 Lippincott Williams & Wilkins, Inc.