HIV treatments optimism and sexual behaviour among gay men in Sydney and Melbourne

Received: 25 May 1999; revised: 27 July 1999; accepted: 17 August 1999.

Objectives: To investigate associations between gay men‚s optimism and sexual behaviour in the context of new HIV treatments.

Methods: Cross-sectional surveys (using anonymous, self-completed questionnaires) were conducted in Sydney during February 1998 (n = 2200) and in Melbourne during January 1998 (n = 1891). Gay men were recruited at social and sex-on-premises venues, clinics and fair days/carnivals.

Results: In a multivariate analysis, unprotected anal intercourse with casual partners (UAIC) was associated with being recruited at a sex-on-premises venue (rather than a clinic or fair day/carnival), HIV positivity, having been tested for HIV less than 6 months ago (rather than over 2 years ago or never having been tested), and not having a regular partner. Over and above these factors, UAIC was associated with agreement with the statements ‚An HIV-positive person who is on combination therapy is unlikely to transmit HIV‚ and ‚I‚m less worried about HIV infection than I used to be‚.

Conclusions: The data reveal a significant relationship between UAIC and certain aspects of optimism in the context of new HIV treatments. Whereas the direction of causality cannot be specified, there is a clear need for HIV and sexual health education programmes to clarify issues of viral load, new and drug resistant strains of HIV, and other infectious agents.

Introduction

‚HAART breeds complacency on safer sex‚ declared a headline of Issue No. 4 of The Bridge, the official daily newspaper of the 12th World AIDS Conference in Geneva. However, a close examination of 14 Geneva presentations which included data on the relationship between sexual practice and knowledge/perceptions concerning combination antiretroviral therapies revealed a more complex picture.

A number of studies presented in Geneva found few or no associations between attitudes towards combination therapies and sexual practices. Men who have sex with men in Montreal did not believe that the new drug treatments encouraged HIV-positive persons to abandon their safe sex practices[¹]. For men who have sex with men in San Francisco, the success of new antiviral treatments had not reduced concern about infection among the majority of men and most had not changed their sexual behaviour as a result of recent treatment successes[²]. Among men who have sex with men in Seattle, there was no obvious substantial increase in unprotected anal intercourse during the years 1996-1997 when more effective treatment for HIV became available[^3,4]. Based on periodic surveys of gay men in Sydney, ideas about combination therapy and viral load were generally unrelated to findings of increased unprotected anal intercourse with casual partners (UAIC) during the period 1996-1998[^5,6]. More broadly, among the general public in Germany there was no evidence that reports of the new therapies had led to a decrease in protective behaviours[⁷].

Some researchers reported rather mixed and limited findings. Among gay men in the United Kingdom, there was some evidence of increased sexual risk taking because of perceptions that HIV is now a less serious condition[⁸]. Among gay men in London, those who were less worried about HIV infection in the light of improved treatments were more likely to report unprotected anal intercourse, but beliefs about infectivity were not associated with unprotected anal intercourse[⁹]. Based on a cohort of serodiscordant male couples in New York City, there was a relationship between optimism about new treatments and unprotected anal sex for HIV-negative men but not for HIV-positive men[^10,11].

Some of the presentations in Geneva provided evidence of associations between HIV treatments optimism and sexual practice, albeit often based on small subsets of optimistic persons within much larger groups of more cautious persons. In a sample of homosexual and bisexual men in a midwestern US city, a small minority of men indicated that they practised safer sex less often since the introduction of new AIDS treatments[^12,13]. A small subset of gay and bisexual men in Chicago reported recent increases in unsafe sex which correlated with their reduced concern with unsafe sex[¹⁴]. Among gay and bisexual men in Los Angeles, higher levels of optimism regarding antiretroviral agents were associated with an increase in unsafe sexual practices[¹⁵]. Among HIV-positive men who had engaged in risk behaviours with other men in New York and San Francisco, there was an association between taking antiviral therapy and/or undetectable viral load and perceptions of risk[¹⁶]. In periodic surveys of gay men in France and Switzerland, although there was no change in the overall level of protection between 1994-1995 and 1997, a very small minority of men believed that new treatments encouraged them to protect themselves less than before[¹⁷].

To add to the complexity, one presentation in Geneva provided evidence that optimism about the new treatments may have protective outcomes. Among HIV-infected adults in Australia, whereas the belief that new combination antiretroviral treatments are effective had no impact on condom use with regular serodiscordant partners, such belief was significantly associated with more consistent condom use with casual partners[¹⁸].

It is not possible to draw a neat conclusion from the research presented at the Geneva conference and in related publications. There appears to be no straightforward relationship between optimism surrounding the success of new HIV therapies and sexual risk behaviour. This lack of consistency may be due to a number of factors. It may indicate different sexual cultures in various parts of the world. It may stem from previous inadequacies in the types of questions asked; knowledge rather than attitudinal questions. Moreover, none of the previous analyses attempted more than straightforward univariate tests of association. It is possible, for example, that optimism and sexual behaviour are associated with differences in socio-economic status, HIV status and other important demographic and social variables.

Clearly then, there is a need for a closer examination of the issues and a more thorough analysis of the possible relationship between reduced concern about HIV transmission and how this may relate to sexual practice. To this end, we examined data from cross-sectional surveys of gay men in Australia‚s two largest cities, Sydney and Melbourne. We were particularly interested to see - in a large sample with sufficient numbers of HIV-positive and status unknown men to make the analysis feasible - whether attitudes towards viral load testing, HIV therapies and HIV infection were related to sexual risk practice after accounting for other possible explanatory effects.

Methods

Participants

Men were included in the survey if they had had sex with another man during the past 5 years. In addition, and so as to exclude any interstate or overseas visitors on a one-off or occasional trip, they had to live in either Sydney or Melbourne or to have participated regularly in the respective city‚s gay community. Men who did not meet these two criteria did not complete a survey.

Questionnaire

An anonymous, self-completed questionnaire was used. It took approximately 10 min to complete. Items covered some demographics, aspects of gay community attachment, sexual identity and relationships, sexual practices with regular and/or casual partners, and HIV testing and results. The use of these types of questions and instruments has been validated by way of triangulation in previous studies of gay men[¹⁹].

Participants responded to six statements about viral load testing, HIV therapies and concern about HIV infection (see Table 1). The introduction to these was, ‚The next six statements are about viral load testing and current combination therapies for HIV‚, and there were three response categories: true; false; unsure.

Table 1

Procedure

The cross-sectional surveys were conducted during February 1998 in Sydney and during January 1998 in Melbourne. Gay and homosexually active men were recruited at the Sydney Gay and Lesbian Mardi Gras Fair Day and at the Melbourne Midsumma Carnival as well as at gay social (bars, night-clubs and a sports venue), sex-on-premises venues, and at sexual health clinics in each city during the subsequent week. Diverse venues and clinics frequented by gay men were selected to attain a broad sample of Sydney and Melbourne gay and homosexually active men.

A combination of paid (in the main) and volunteer staff was used for recruitment. All men attending the gay venues and clinics were approached with the aim of recruiting 100% of those in attendance. At the Sydney Fair Day and the Melbourne Carnival, recruiters were positioned at strategic stalls and alleyways and they approached as many men as possible.

Data analysis

Recruitment sites were collapsed into four categories: sex-on-premises venues; social venues; clinics; and fair day/carnival. Number of gay friends and amount of free time spent with gay friends were summed to give an index of attachment to the gay community. Retaining the whole sample as our base, anal intercourse with casual partners was collapsed into a binary variable, any UAIC during the previous 6 months (any UAIC) versus no UAIC during the previous 6 months (no UAIC). The latter category incorporated all three strategies which a man may adopt to avoid HIV transmission to/from a casual partner: no casual partners; casual partners but no anal intercourse per se; casual partners but all anal intercourse was protected. Note that UAIC is the appropriate sexual risk indicator in this context given the history of Australian sexual health promotion amongst gay men, including education programmes around ‚negotiated safety‚ for men in HIV-negative seroconcordant regular relationships[²⁰].

Responses to the statements about viral load testing, HIV therapies and concern about HIV infection were recoded for optimism (true) and otherwise (false/unsure). This was not the only scoring method we investigated but it proved to be the most parsimonious and theoretically plausible one in accounting for the relationship between attitudes and sexual practice.

The data were analysed initially using summary statistics and χ2 tests of association. Subsequently, a logistic regression was conducted to determine whether the statements about viral load testing, HIV therapies and concern about HIV infection were independently associated with any UAIC after controlling for all other variables included in the analysis. The order of fit of families for logistic regression was: recruitment (city; recruitment site); education; HIV status; time since last HIV test; social environment variables (attachment to gay community; sexual identity; relationship status); and lastly, responses to the statements about viral load testing, HIV therapies and concern about HIV infection.

Results

A total of 4091 men were recruited, 2200 in Sydney (overall response rate, 77.0%; range across sites, 59.2-92.3%) and 1891 in Melbourne (overall response rate, 75.5%; range across sites, 70.9-100%). From the various sites, 2342 (57.2%) were recruited at the fair day/carnival, 753 (18.4%) at sex venues, 586 (14.3%) at social venues and 410 (10.0%) at clinics. The men ranged in age from 15 to 88 years (median, 33 years). They were well educated with 1745 (43.7%) having attended university, 781 (19.6%) having undertaken a trade certificate or diploma, 856 (21.4%) having completed 6 years of high school and 610 (15.3%) having completed only 4 years of high school or less. (Discrepancies in n throughout are due to small amounts of missing data.)

Most of the men were gay-identified (3714, 91.7%). Mostly gay-identification accords with the finding that for the majority of the men ‚most or all‚ of their friends were also gay men (2344, 57.5%) and ‚a lot‚ of their free time was spent with gay men (2141, 52.5%). In terms of sexual partners, 2595 (63.4%) had a regular partner during the previous 6 months and 2969 (72.6%, not mutually exclusive) had casual partners during that interval.

Asked when they last had an HIV test, 1745 (43.7%) reported less than 6 months ago, 584 (14.6%) 7-12 months ago, 621 (15.6%) 1-2 years ago, 632 (15.8%) over 2 years ago, and 409 (10.2%) had not been tested. Most of the men reported that they were HIV antibody negative (2941, 73.4%), 564 were seropositive (14.1%) and 501 stated that they did not know their serostatus (12.5%).

Responses to the statements about viral load testing, HIV therapies and concern about HIV infection are presented in Table 1. For most of the statements, relatively few men concurred. Only for the last two statements (5 and 6) did a sizeable proportion of the men agree with the statement.

For the first three statements (1, 2 and 3) there was no association between responses and sexual behaviour with casual partners. However, men who had any UAIC were significantly more likely to concur with statements 4 and 6 than their counterparts who had no UAIC (see Table 2). A similar association, albeit a trend, pertained to statement 5.

Table 2

Variables that were not significant in the full logistic model were: city (Sydney versus Melbourne); education; the social variables of attachment to the gay community and sexual identity; responses to statements 1, 2, 3 and 5; and first-level interactions. (Note that the variables city, education and responses to statement 5 were significantly associated with any UAIC at the univariate level. Further details are not presented here as the primary focus is on multivariate modelling for independent associations.) Variables that remained in the logistic model appear in Table 3. (Note that each of these variables was also significant at the univariate level.)

Table 3

As shown in Table 3, men recruited at the clinics and at the fair day/carnival, were less likely to have any UAIC than those recruited at sex-on-premises venues. HIV-negative men were less likely to have any UAIC than HIV-positive men. Those tested for HIV over 2 years ago, and those never tested, were less likely to have any UAIC than those tested less than 6 months ago. Men who had a regular partner were less likely to have any UAIC than men without a regular partner.

Over and above these differences (i.e., after controlling for any effects of recruitment site, HIV status, etc.), any UAIC was associated with two of the attitudes toward HIV therapies and HIV infection (statements 4 and 6). Men who had any UAIC were more likely to concur with ‚An HIV-positive person who is on combination therapy is unlikely to transmit HIV‚ than their counterparts who had no UAIC. Similarly, men who had any UAIC were more likely to endorse ‚I‚m less worried about HIV infection than I used to be‚ than those who had no UAIC.

Discussion

A key finding is that relatively few gay men in Sydney and Melbourne are optimistic about HIV treatments. Whereas a majority of men believe that combination therapies are effective in preventing serious illness (statement 5), few men believe that HIV-positive persons on combination therapies are unlikely to transmit HIV (statement 4) or those with undetectable viral loads cannot pass on the virus (statement 1). Likewise, rather few men believe that, if taken early enough, combination therapies can cure HIV infection (statement 2). Statement 2 could be interpreted as post-exposure prophylaxis or as early treatment per se. If the statement was interpreted as referring to post-exposure prophylaxis, the results indicate that men do not feel optimistic of its success. Alternatively, the finding may be explained by the limited Australian public health promotion of post-exposure prophylaxis.

The current data reveal a significant relationship between Sydney and Melbourne gay men‚s sexual risk taking (operationalized as UAIC) and social and sexual environment (recruitment at a sex-on-premises venue; not having a regular partner), positive HIV status and recent HIV testing. Over and above these factors, certain aspects of optimism in the context of new HIV treatments are also associated with UAIC. This is an important finding because even after controlling for the possible extraneous influences of HIV and relationship status (e.g., positive-positive sex [²¹] and negotiated safety [²⁰]), men who are more optimistic about HIV treatment advances are more likely to engage in UAIC.

Our call for a more thorough, multivariate analysis of the possible relationship between treatments optimism and sexual behaviour was justified. Whereas previous, more straightforward analyses indicated few associations between ideas about combination HIV therapy and sexual behaviour[^5,6], the current analysis was able to uncover significant relationships. Interestingly, at the multivariate level (that is, after controlling for other differences between the cities) there were no differences between Sydney and Melbourne. The question remains whether or not differences in gay sexual culture are to be found in other parts of the world.

Whereas there is an association between UAIC and optimism in the context of new HIV treatments, the current data do not permit us to disentangle the direction of causality. On the one hand, it may be the case that men who are optimistic about improved HIV treatments are more prepared to engage in UAIC on the basis of altered perceptions of risk. On the other hand, and equally plausible, it may be that men who have UAIC subsequently rationalize or account for their behaviour in terms of perceived lower HIV infectivity and the availability of more advanced HIV treatments. In the former scenario, HIV treatment advances would drive sexual risk taking. In the latter case, having UAIC (and possibly avoiding HIV transmission which serves to further affirm the effectiveness of the new treatments) is explained or justified by the success of the treatments. Additional close-focus work is needed to analyse these possibilities although one such study among Australian gay men found scant evidence that advances in HIV medicine were driving risk taking. Rather, desire for intimacy and the role of romance played a pivotal role in sexual practice and assessments of risk[²²].

Our finding of a significant relationship between HIV treatments optimism and sexual risk taking among gay men corroborates various other work[^12-17]. An important aspect of our findings, and that of others, is that the relationship is driven in the main by a very small minority of optimistic men, an even smaller subset of whom actually engage in sexual risk practices[^12-14,17]. Nonetheless, these men comprise an important group within a relatively high HIV prevalence population. These results signal the need to explain to HIV-positive and HIV-negative men alike just what the new treatments mean in terms of possible transmission of HIV. Men need to be provided with up-to-date information regarding drug resistance, co-factors of infection, the implications of viral load test results, and the fact that information needs to be updated regularly. They possibly also need reassurance that at least for the time being the caution with which the vast majority of men view the new treatments and the sustaining of safe sex is appropriate; to do otherwise will be to run the risk that the very emergence of the new HIV treatments which hold so much promise for the quality of life of HIV-positive persons will lead to a resurgence of HIV infections.