Share this article on:

High-risk sexual behaviour increases among London gay men between 1998 and 2001: what is the role of HIV optimism?

Elford, Jonathana; Bolding, Grahama; Sherr, Lorraineb

EPIDEMIOLOGY & SOCIAL

Objective To examine whether HIV optimism (i.e. optimism in the light of new HIV drug therapies) can account for the recent increase in high-risk sexual behaviour among London gay men.

Methods Gay men (n = 2938) using London gyms were surveyed annually between 1998 and 2001. Information was collected on HIV status, unprotected anal intercourse (UAI) in the previous 3 months, and agreement with two statements concerning the severity of and susceptibility to HIV infection . Those who agreed were classified as ‘optimistic'.

Results Between 1998 and 2001, the percentage of men reporting high-risk UAI (i.e. UAI with a casual partner of unknown or discordant HIV status) increased: HIV-positive men 15.3–38.8%; HIV-negative men 6.8–12.1%; never-tested men 2.1–7.7%; (P < 0.01). Overall, less than a third were optimistic. In cross-sectional analysis, optimistic HIV-positive and -negative men were more likely to report high-risk UAI than other men (P < 0.05). However, the increase in high-risk UAI between 1998 and 2001 was seen in those who were optimistic and those who were not (P < 0.05). In multivariate analysis, the modelled increase in high-risk UAI over time remained significant after controlling for HIV optimism (P < 0.01), with no significant interaction between optimism and time.

Conclusion Among London gay men, no difference was detected between those who were optimistic and those who were not in the rate of increase in high-risk sexual behaviour between 1998 and 2001. Our findings suggest that HIV optimism is unlikely to explain the recent increase in high-risk sexual behaviour in these men.

From the aCity University, London, UK; and bRoyal Free and University College Medical School, University College London, UK.

Correspondence to: Professor Jonathan Elford, City University, Institute of Health Sciences and St Bartholomew School of Nursing and Midwifery, 20 Bartholomew Close, London EC1A 7QN, UK Tel: +44 020 7040 5702; fax: +44 020 7040 5717; e-mail: j.elford@city.ac.uk

Received: 20 July 2001;

revised: 8 March 2002; accepted: 25 March 2002.

Sponsorship: This work was funded by the Camden and Islington Health Authority.

Back to Top | Article Outline

Introduction

Since 1997, an increase in high-risk sexual behaviour, sexually transmitted infection, and the incidence of HIV has been reported among gay men living in Europe, Australia, the United States and Canada [1–8]. This increase has coincided with the availability of highly active antiretroviral therapy (HAART) [9–11]. According to some researchers the increase may be a direct response to HAART [2,12]. Gay men, it is postulated, are now more optimistic about HIV in the light of new drug therapies; and this optimism may have created complacency around safer sexual practice, resulting in an increase in high-risk sexual behaviour.

A number of studies have found an association between HIV optimism (as it is known) and high-risk sexual behaviour in gay men [3,4,8,13–19]. Being cross-sectional such studies were unable to establish causality. Optimism in the light of new drug therapies may have triggered sexual risk taking or may have been used as a post-hoc justification [15]. Nor could these studies examine separately changes in sexual risk behaviour over time among men who were optimistic and those who were not. This differentiation is crucial, because if the increase in high-risk sexual behaviour were caused solely by HIV optimism, we would expect to see the rise predominantly, if not exclusively, among gay men who are optimistic. Among those who do not share their optimism there should be little, if any, increase in high-risk sexual behaviour.

The objectives of this investigation were to examine: (i) trends in sexual risk behaviour among gay men in London between 1998 and 2001; (ii) the univariate association between HIV optimism and sexual behaviour during this period; (iii) the contribution of HIV optimism to changes in sexual behaviour that occurred between 1998 and 2001.

Back to Top | Article Outline

Methods

Between January and March each year from 1998 to 2001, gay men using gyms in central London were asked to complete a confidential self-administered questionnaire as part of an annual behavioural surveillance programme [20,21]. Five gyms were surveyed in all years; a sixth gym was added in 2000 and a seventh in 2001. Questionnaires were distributed in each gym over a one week period. Men were asked to provide information on their sociodemographic characteristics, HIV status, sexual risk behaviour, and optimism in the light of new HIV drug therapies. From 1999 onwards, respondents were also asked whether they had completed a questionnaire in a previous year. We were not able to link individual questionnaires over time.

Back to Top | Article Outline

Sexual risk behaviour

In each survey, men were asked whether they had had unprotected anal intercourse (UAI) in the previous 3 months, and if so, the type (main or casual) and HIV status of their UAI partner(s). UAI was classified as either seroconcordant (i.e. with a partner of the same HIV status) or non-concordant (with a partner of unknown or discordant HIV status) [22]. For men who reported more than one UAI partner in the previous 3 months, UAI was only classified as seroconcordant if they said all their partners were the same status as themselves. Consequently, concordant and non- concordant UAI were mutually exclusive categories. UAI reported by men who had never been tested for HIV was classified as non-concordant because, being unaware of their own HIV status, they were not able to establish concordance with their sexual partner. Men who reported UAI only with a main partner were analysed separately from those who reported UAI with casual partners.

Back to Top | Article Outline

HIV optimism

Drawing on the health belief model and protection motivation theory [23,24] we identified two dimensions of HIV optimism that relate to (i) severity and (ii) susceptibility. On the one hand, now that treatments have improved, the consequences of becoming infected with HIV may appear to be less serious than before. That is to say, there may have been a decline in the perceived severity of HIV-related disease. On the other hand, gay men may believe that because HAART reduces viral load, it will also make individuals with HIV less infectious. In other words, there may have been a decline in the perceived susceptibility to infection after exposure to HIV. Two single-item scales were created to reflect these different dimensions of optimism [15]. In each year, respondents were asked how much they agreed with the following statements:

`I am less worried about HIV infection now that treatments have improved’ (reduced-severity optimism referred to throughout this paper as optimism 1)

`I believe that new drug therapies make people with HIV less infectious’ (reduced-susceptibility optimism referred to as optimism 2)

In 1998 and 1999, men were asked to respond to each statement on a five-point linear scale: ‘not at all'; ‘a bit'; ‘somewhat'; ‘quite a lot'; and ‘a lot'. Men who responded ‘a bit', ‘somewhat', ‘quite a lot', or ‘a lot’ were classified as ‘optimistic'. The remaining men (`not at all') were classified as ‘not optimistic'. In 2000 and 2001, to harmonize our scales with those used in other countries, the responses were changed to: ‘strongly disagree'; ‘disagree'; ‘agree'; ‘strongly agree’ [25]. Men who said they agreed or agreed strongly were classified as ‘optimistic', the remainder were classified as ‘not optimistic'.

In 2000 we included an additional multi-item HIV optimism scale previously validated among gay men in Australia [25]. Four items concerned optimism 1 (`new HIV treatments will take the worry out of sex'; ‘I am less worried about HIV than I used to be'; ‘HIV/AIDS is a less serious threat than it used to be because of new treatments'; ‘if every HIV-positive person took the new treatments, the AIDS epidemic would be over'), whereas a further four concerned optimism 2 (`a person with undetectable viral load cannot pass on the virus'; ‘people with undetectable viral loads do not need to worry so much about infecting others with HIV'; ‘it's never safe to fuck without a condom regardless of viral load (reverse scored)'; ‘because of new treatments fewer people are becoming infected with HIV'). Including this multi-item scale in the year 2000 questionnaire allowed us to validate each of our single-item optimism scales against the equivalent four items in the Australian scale. For validation, men's responses to each item were scored: strongly disagree = 1, disagree = 2, agree = 3, strongly agree = 4. We then compared the respondent's score on each single-item gym scale with their mean score on the equivalent four items on the Australian scale.

Back to Top | Article Outline

Statistical analysis

Logistic model

The associations between UAI and the year of survey, optimism 1 and optimism 2 were examined in a logistic model by HIV status. In this model, the independent variables were year of survey (entered as a dummy variable 1998 = 0, 1999 = 1, 2000 = 2, 2001 = 3) and being optimistic (versus not being optimistic). The dependent variable was UAI (versus no UAI). Odds ratios for the year of survey measured the increase in risk from one year to the next. For example, an odds ratio of 1.4 meant the risk of reporting UAI increased by 40% over a 12 month period. Odds ratios for optimism described the risk of UAI among optimistic men compared with the risk among the remaining men. Separate analyses were conducted for concordant UAI and non-concordant UAI, broken down by partner type (main only or casual). All P values quoted in the paper were derived from the logistic models.

Back to Top | Article Outline

Univariate analysis

The univariate association between UAI and the year of survey was initially examined. If a significant association (P < 0.05) between a given category of UAI and the year of survey emerged, the univariate associations between that category of UAI and both optimism 1 and 2 were then examined.

Back to Top | Article Outline

Multivariate analysis

When a significant univariate association was found between UAI and both the year of survey as well as optimism (1 or 2), the multivariate association was examined by simultaneously entering the year of survey, optimism 1 and optimism 2 into a logistic model. To identify potential confounding factors in this model, the relationships between the year of survey, optimism 1 or optimism 2, and variables known to be associated with UAI were examined. These variables were age, being in a relationship, and steroid use [20,26]. The model excluded in 1999, 2000 or 2001 men who had completed a questionnaire in an earlier year to ensure the independence of samples. An interaction term was added to the model to examine whether UAI risk increased differentially for men who were optimistic compared with those who were not.

Back to Top | Article Outline

Observed and expected

We multiplied the UAI rates for 1998 by the number of men surveyed in 1999, 2000 and 2001, stratified by HIV status, to estimate the number of men who would have reported UAI between 1998 and 2001 had there been no change in risk during that period. Comparing the observed number of men reporting UAI between 1998 and 2001 with the number expected assuming no change in risk allowed us to estimate the increase in the number of men reporting UAI in that time.

Back to Top | Article Outline

Validation of optimism scale

To validate the optimism scales in 2000, a paired t-test was used to examine whether, at an individual level, the score on the single-item London scale was significantly different from the mean score on the equivalent four items on the Australian scale.

Back to Top | Article Outline

Results

Sample

Completed questionnaires were returned by 3319 respondents; 1018 gay men in 1998, 720 in 1999, 792 in 2000, and 789 in 2001. The estimated response rate in each year was 50–60% [15,20–22,26–28]. The majority of men were young (median age 33 years), white (90%), employed (90%), and university educated (70%). These characteristics did not vary significantly from year to year (data available from authors). The analysis was restricted to 2938 men who provided complete information on HIV status, sexual risk behaviour, HIV optimism, and potential confounders (Table 1). Of the men surveyed in 1999, 2000, and 2001, approximately half had completed an earlier questionnaire. After excluding these men, 1963 respondents were available for the independent-samples analysis (Table 1). Of the men surveyed in 2000 or 2001, 13% said they had also completed a questionnaire in a bar or club 2 months earlier as part of a parallel behavioural surveillance programme [1].

Table 1

Table 1

Back to Top | Article Outline

Year of survey and sexual risk behaviour, univariate analysis

Between 1998 and 2001, HIV-positive, -negative and never-tested men all reported an increase in non-concordant UAI with a casual partner (usually an individual of unknown rather than discordant status) (P < 0.01). HIV-positive men also reported an increase in UAI with a seroconcordant casual partner (P < 0.01) (Table 2).

Table 2

Table 2

There was no significant change in the proportion of men reporting UAI with a main partner alone (concordant or non-concordant) between 1998 and 2001 (P > 0.1).

Back to Top | Article Outline

HIV optimism

For all years combined (1998–2001), a quarter of the men agreed with the optimism 1 statement, whereas one-fifth agreed with the optimism 2 statement. These men were classified as optimistic (Table 3, all years column). HIV-positive men were more likely to agree with these statements than HIV-negative or never-tested men (P < 0.05). The two dimensions of optimism were not highly correlated: 463 men expressed agreement with optimism 1 alone, 261 with optimism 2 alone, whereas 276 agreed with both (Spearman correlation coefficient 0.4). There was no evidence of an increase in the proportion of men classified as optimistic between 1998 and 2001. The decline in the overall percentage of men classified as optimistic in 2000 and 2001 may have reflected the modification of response categories described in the Methods section (see Discussion).

Table 3

Table 3

Back to Top | Article Outline

Validation of optimism scale

For optimism 1, in the year 2000, there was no significant difference between the score on the single-item London scale and the mean score on the four equivalent items from the Australian scale for HIV-positive men [mean scores 1.9 (London) versus 1.8 (Australian) scale], HIV-negative men (1.7 versus 1.7) or never-tested men (1.6 versus 1.7) (paired analysis, all P > 0.1).

For optimism 2, no significant difference was seen between the mean scores on the London and Australian scales for HIV-negative men (1.6 versus 1.5) or never-tested men (1.4 versus 1.5) (P > 0.1). HIV-positive men had a higher score on the London scale (1.8 versus 1.5, P < 0.001), although the majority of HIV-positive men disagreed or disagreed strongly on both scales.

Back to Top | Article Outline

HIV optimism and sexual risk behaviour, univariate analysis

Non-concordant unprotected anal intercourse with a casual partner

For HIV-positive and -negative men there was a significant univariate association between each dimension of optimism and non-concordant UAI with a casual partner (P < 0.05) (Table 4). Optimistic men were more likely than other men to report high-risk sexual behaviour. This association was seen for all years combined as well as in most years individually.

Table 4

Table 4

For never-tested men, however, there was no significant association between either dimension of optimism and non-concordant UAI with a casual partner (P > 0.1).

Back to Top | Article Outline

Concordant unprotected anal intercourse with a casual partner

Among HIV-positive men, there was no significant association between optimism and concordant UAI with a casual partner (P > 0.1).

Back to Top | Article Outline

Year of survey, HIV optimism and sexual risk behaviour, multivariate analysis

In univariate analysis, the year of survey, optimism 1 and 2 were all significantly associated with non-concordant UAI with a casual partner among HIV-positive and -negative men. These univariate associations were, therefore, explored in a multivariate model. There were no confounding variables in this model because age, steroid use, and being in a relationship, all known to be associated with the dependent variable UAI, were not associated with any of the independent variables (year of survey, optimism 1 and 2).

Back to Top | Article Outline

HIV negative

For HIV-negative men in the multivariate logistic model, both the year of survey and optimism 1 (but not optimism 2) were significantly associated with non-concordant UAI with a casual partner (P < 0.001) (Table 5). There was no significant interaction between the year of survey and optimism 1 (P > 0.1).

Table 5

Table 5

Back to Top | Article Outline

HIV positive

For HIV-positive men in the multivariate model, both the year of survey and optimism 2 (but not optimism 1) were significantly associated with non-concordant UAI with a casual partner (P < 0.05) (Table 5). The interaction between the year of survey and optimism 2 was non-significant although borderline (P = 0.07); exponential of the regression coefficient for the interaction term, 1.72, 95% confidence interval 0.95, 3,10.

Restricting the analysis to the five gyms that were included in the survey each year did not alter these findings.

Back to Top | Article Outline

Observed and expected

Between 1998 and 2001, 315 men reported non-concordant UAI with a casual partner (Table 2) compared with an expected number of 207 had there been no change in risk behaviour during this time. This represents an increase of 108 men (HIV-positive 32, HIV-negative 54, never-tested 22) being 52% more than expected if no change in risk had occurred (108/207).

From the multivariate regression equation, we estimated that were there an interaction between optimism and the year of survey among HIV-positive men (P = 0.07), optimism could have accounted for an increase of 15 HIV-positive men reporting high-risk sexual behaviour between 1998 and 2001. This represents nearly half the increase among HIV-positive men (15/32) but only 14% (15/108) of the overall increase in the number of men in the whole study group reporting non-concordant UAI with a casual partner between 1998 and 2001.

Back to Top | Article Outline

Discussion

Between 1998 and 2001, there was a substantial increase in high-risk sexual behaviour among gay men surveyed in central London gyms. The proportion of men reporting UAI with a casual partner of unknown or discordant HIV status doubled during that time. This increase was seen among HIV-positive, -negative and never-tested men. An increase in high-risk sexual behaviour has also been reported among gay men using London bars and clubs [1]. Is this increase the result of HIV optimism?

Among HIV-negative and -positive men (but not never-tested men), those who were optimistic were indeed more likely to report high-risk sexual behaviour. However, cause-and-effect could not be established [15]. Interestingly, the dimension of HIV optimism associated with high-risk behaviour varied by status. For HIV-positive men in multivariate analysis, high-risk sexual behaviour was associated with a belief that new drug therapies make individuals with HIV less infectious. On the other hand, for HIV-negative men high-risk behaviour was associated with being less worried about HIV because treatments have improved.

The key question, however, is not whether HIV optimism is associated with high-risk behaviour in univariate analysis, but whether HIV optimism can explain the increase in UAI over time. In multivariate analysis, the modelled increase in high-risk sexual behaviour between 1998 and 2001 remained significant after controlling for HIV optimism. Nor was there a significant interaction between the year of survey, optimism and risk. In other words, no difference was detected between men who were optimistic (the minority) and men who were not (the majority) in the rate of increase in high-risk sexual behaviour between 1998 and 2001. However, if the increase in high-risk behaviour had been caused solely by HIV optimism, we would have expected to see the rise predominantly, if not exclusively, among gay men who were optimistic, with little if any increase among other men. That is to say, there would have been a significant interaction between year, optimism, and risk. This was not the case. Our data suggest, therefore, that HIV optimism is unlikely to explain the recent increase in high-risk UAI.

The strongest evidence supporting this line of reasoning is for HIV-negative men, because the interaction term in the multivariate model was indisputably non-significant (P = 0.5). Furthermore, an increase in high-risk sexual behaviour was also seen among never-tested men, even though UAI was not associated with optimism in this group. For HIV-positive men, however, the evidence is equivocal because the interaction term although non-significant was borderline (P = 0.07). This raises the possibility that the increase in high-risk behaviour among HIV-positive men may have been greater for those who were optimistic than for those who were not, although an increase was seen in both. The impact of this differential, however, was too small by itself to account for the overall increase in high-risk behaviour at a community level.

An increase in high-risk sexual behaviour among gay men has been reported since 1997 in Australia [4], the USA [5,6], and Europe [3,8,29]. HIV optimism has often been invoked as the driving force behind this worldwide increase, on the basis of reports of a cross-sectional association between optimism and UAI. All the studies have two features in common. First, as in our own study, only a minority of men (usually 10–20%) were classified as optimistic [13–17,19]. Most gay men were realistic rather than optimistic about new drug therapies for HIV. Second, and somewhat surprisingly, levels of optimism in such studies have remained remarkably stable since 1997, despite the increasingly evident benefits of HAART [10,30,31]. This alone weakens the case for optimism explaining the increase in UAI. If there were a direct causal relationship, one might expect the increase in high-risk sexual behaviour in these countries to have been accompanied by a corresponding increase in optimism over time, in the same way that an increase in tobacco consumption accompanied and explained the subsequent increase in lung cancer in the first half of the 20th century [32]. However, this has not been the case.

How then can we account for the upward trend in high-risk behaviour? The increase in high-risk sexual behaviour has coincided not only with the availability of HAART but also with other changes, such as increasing access to the Internet [33,34]. Several studies have shown an association between seeking sex on the Internet and high-risk sexual behaviour among gay men [22,35,36], raising the possibility that the Internet has become a new risk environment, although the underlying processes are not fully understood. Other contributory factors in London could be the increased opportunity in recent years for meeting sexual partners and having risky sex in saunas and backrooms (M. Maguire, Camden and Islington Community Health Services NHS Trust, London, personal communication). It is also possible that gay men have become habituated to the risk of HIV infection now that two decades have passed since AIDS was first reported [37]. Whatever factors are driving the increase in high-risk sexual behaviour, they appear to impact equally on men who are optimistic and men who are not.

Our analysis was conducted at an aggregate rather than an individual level using a series of cross-sectional samples from 1998–2001. This approach requires homogeneity of samples over time as well as the standardization of core variables. There was little variation in the characteristics of the study group between 1998 and 2001; in all years, the majority of men were young, white, employed and university educated. The dependent variable, self-reported UAI, was measured in precisely the same way in all years. The response categories of the independent variable (HIV optimism) changed, however, in 2000 to come in line with those used elsewhere. It is possible, therefore, that before 2000 we slightly overestimated the level of optimism among men, because the percentage classified as optimistic declined after the introduction of the new response categories. However, it is unlikely that this discontinuity has introduced a major source of bias. In all years, the majority of men were classified as not being optimistic, and a significant increase in high-risk sexual behaviour was evident in this group throughout the study. Our single item scales proved to be robust when compared with a validated multi-item optimism scale from Australia in 2000 [25]. There may have been some overlap in the meaning of the two optimism statements, however, so caution should be exercised in interpreting the finding that the dimension of optimism associated with high-risk sexual behaviour varied by HIV status.

The response rate of 50–60% is comparable with response rates reported in recent behavioural studies among gay men in the United States [38–40], although lower than in some UK studies [41,42]. The demographic and behavioural characteristics of the gym sample were similar to those of men surveyed annually in London gay bars, clubs and genitourinary medicine clinics [1]. HIV status was self-reported, which may have resulted in some misclassification of UAI, although it is unlikely that this varied over time [15].

Back to Top | Article Outline

Conclusion

Between 1998 and 2001 there was a substantial increase in high-risk sexual behaviour among gay men surveyed in central London gyms regardless of HIV status. It is unlikely that HIV optimism can explain this increase. These findings have important implications for health promotion and HIV prevention. Challenging HIV optimism is unlikely, by itself, to reverse the recent upward trend in high-risk sexual behaviour among gay men. Indeed, focusing on optimism as the driving force may draw attention away from other factors that underlie the recent increase in high-risk sexual behaviour. Identifying these factors should now be given the highest priority.

Back to Top | Article Outline

Acknowledgements

The authors would like to thank the managers and members of the gyms for their support and participation in the project; Richard Morris, Fiona Lampe, Jonathan Emberson and Alison MacFarlane for statistical advice; and the two anonymous referees for their invaluable comments on the paper.

Back to Top | Article Outline

References

1. Dodds JP, Nardone A, Mercey DE, Johnson AM. Increase in high-risk sexual behaviour among homosexual men, London, 1996–8: cross sectional, questionnaire study. BMJ 2000, 320: 1510–1501.
2. Stolte IG, Dukers NH, de Wit JB, Fennema JS, Coutinho RA. Increase in sexually transmitted infections among homosexual men in Amsterdam in relation to HAART. Sex Transm Infect 2001, 77: 184–186.
3. Adam P, Hauet E, Caron C. Recrudescence des prises de risque et des MST parmi les gays. Resultats de l'enquete press gay 2000. Paris: L'Institut de Veille Sanitaire; 2001.
4. Van de Ven P, Prestage G, Crawford J, Grulich AE, Kippax S. Sexual risk behaviour increases and is associated with HIV optimism among HIV-negative and HIV-positive gay men in Sydney over the 4 year period to February 2000. AIDS 2000, 14: 2951–2953.
5. McFarland W, Schwarcz S, Kellogg T, Hsu L, Kim A, Katz M. Implications of highly active antiretroviral treatment for HIV prevention: the case of men who have sex with men (MSM), San Francisco, 2000. In:13th International AIDS Conference. Durban, 2000 [Abstract MoPpD1127].
6. Wolitski RJ, Valdiserri RO, Denning PH, Levine WC. Are we headed for a resurgence of the HIV epidemic among men who have sex with men? Am J Pub Health 2001, 91: 883–888.
7. Calzavara L, Burchell A, Major C, Remis R, Corey P, Myers T. Increasing HIV incidence among MSM repeat testers in Ontario, Canada, 1992–1998. In:13th International AIDS Conference. Durban, 2000 [Abstract ThOrC718].
8. Bochow M. Schwule Manner, AIDS und safer Sex: neue Entwicklungen. Berlin: Deutsche AIDS-Hilfe e.V; 2001.
9. Hammer SM, Squires KE, Hughes MD. et al. A controlled trial of two nucleoside analogues plus indinavir in persons with human immunodeficiency virus infection and CD4 cell counts of 200 per cubic millimeter or less. N Engl J Med 1997, 337: 725–733.
10. Mocroft A, Vella S, Benfield T. et al., for the EuroSIDA Study Group. Changing patterns of mortality across Europe in patients infected with HIV-1. Lancet 1998, 352: 1725–1730.
11. Hogg R, O'Shaughnessy M, Gataric N. et al. Decline in deaths from AIDS due to new antiretrovirals. Lancet 1997, 124: 199–205.
12. Editorial Note. Increases in unsafe sex and rectal gonorrhea among men who have sex with men – San Francisco 1994–97.MMWR 1999, 48:45–48.
13. Dilley JW, Woods WJ, McFarland W. Are advances in treatment changing views about high-risk sex? N Engl J Med 1997, 33: 501–502.
14. Kelly JA, Hoffman RG, Rompa D, Gray M. Protease inhibitor combination therapies and perceptions of gay men regarding AIDS severity and the need to maintain safer sex. AIDS 1998, 12: F91–F95.
15. Elford J, Bolding G, Maguire M, Sherr L. Combination therapies for HIV and sexual risk behavior among gay men. J Acquired Immune Defic Syndr 2000, 23: 266–271.
16. Van de Ven P, Kippax S, Knox S, Prestage G, Crawford J. HIV treatments optimism and sexual behaviour among gay men in Sydney and Melbourne. AIDS 1999, 13: 2289–2294.
17. Misovich SJ, Fisher JD, Fisher WA. Belief in a cure for HIV infection associated with greater HIV risk behaviour among HIV positive men who have sex with men. Can J Hum Sexuality 2000, 8: 241–248.
18. Vanable PA, Ostrow DG, McKirnan DJ, Taywaditep KJ, Hope BA. Impact of combination therapies on HIV risk perceptions and sexual risk among HIV-positive and HIV-negative gay and bisexual men. Health Psychol 2000, 19: 134–145.
19. Kravcik S, Victor G, Houston S. et al. Effect of antiretroviral therapy and viral load on the perceived risk of HIV transmission and the need for safer sexual practices. J Acquired Immune Defic Syndr 1998, 19: 124–129.
20. Elford J, Bolding G, Maguire M, Sherr L. Sexual risk behaviour among gay men in a relationship. AIDS 1999, 13: 1407–1411.
21. Elford J, Sherr L, Bolding G, Maguire M, Serle F. Peer-led HIV prevention among gay men in London (the 4 gym project). Intervention and evaluation. In: Researching health promotion. Watson J, Platt S (editors). London: Routledge; 2000.
22. Elford J, Bolding G, Sherr L. Seeking sex on the Internet and sexual risk behaviour among gay men using London gyms. AIDS 2001, 15: 1409–1415.
23. Fisher JD, Fisher WA. Theoretical approaches to individual-level change in HIV risk behavior. In:Handbook of HIV prevention. Peterson JL, DiClemente RJ (editors). New York: Kluwer Academic; 2000.
24. Rosenstock I, Strecher V, Becker MH. The health belief model and HIV risk behaviour change. In:Preventing AIDS, theories and methods of behavioral interventions. DiClemente RJ, Peterson JL (editors). New York: Plenum Press; 1994.
25. Van de Ven P, Crawford J, Kippax S, Knox S, Prestage G. A scale of optimism-scepticism in the context of HIV treatments. AIDS Care 2000, 12: 171–176.
26. Bolding G, Sherr L, Maguire M, Elford J. HIV risk behaviours among gay men who use anabolic steroids. Addiction 1999, 94: 1829–1835.
27. Sherr L, Bolding G, Maguire M, Elford J. Viagra use and sexual risk behaviour among gay men in London. AIDS 2000, 14: 2051–2053.
28. Elford J, Bolding G, Sherr L. Peer education has no significant impact on HIV risk behaviours among gay men in London. AIDS 2001, 15: 535–538.
29. Dukers N, Goudsmit J, de Wit JBF, Prins M, Weverling G-J, Coutinho R. Sexual risk behaviour relates to the virological and immunological improvements during highly active antiretroviral therapy in HIV-1 infection. AIDS 2001, 15: 369–378.
30. Moore RD, Chaisson RE. Natural history of HIV infection in the era of combination antiretroviral therapy. AIDS 1999, 13: 1933–1942.
31. Sparano JA, Anand K, Desai J, Mitnick RJ, Kalkut GE, Hanau LH. Effect of highly active antiretroviral therapy on the incidence of HIV-associated malignancies at an urban medical center. J Acquired Immune Defic Syndr 1999, 21 (Suppl. 1) : S18–S22.
32. Surgeon General. Smoking and health.Report of the Advisory Committee to the Surgeon General of the Public Health Service. Washington: US Department of Health, Education and Welfare; 1964.
33. Internet statistics and demographics. A Library of Congress Internet resource page, http://lcweb.loc.gov/global/internet/inet-stats.html. 2000. Site accessed: 22 April 2002.
34. Internet growth. http://www2.aus.us.mids.org/growth/internet. Site accessed: 10 June 2002.
35. McFarlane M, Bull SS, Rietmeijer CA. The Internet as a newly emerging risk environment for sexually transmitted diseases. JAMA 2000, 284: 443–446.
36. Bull SS, McFarlane M. Soliciting sex on the Internet.What are the risks for sexually transmitted disease and STD. Sex Transm Dis 2000, 27: 545–550.
37. Sepkowitz KA. AIDS – the first 20 years. N Engl J Med 2001, 344: 1764–1772.
38. Koblin BA, Torian LV, Guilin V, Ren L, MacKellar DA, Valleroy LA. High prevalence of HIV infection among young men who have sex with men in New York City. AIDS 2000, 14: 1793–1800.
39. Valleroy LA, MacKellar D, Karon JM. et al. HIV prevalence and associated risks in young men who have sex with men. JAMA 2000, 284: 198–204.
40. Mansergh G, Marks G, Miller L, Appleby PR, Murphy S. Is ‘knowing people with HIV/AIDS’ associated with safer sex in men who have sex with men? AIDS 2000, 14: 1845–1851.
41. Nardone A, Dodds JP, Mercey DE, Johnson AM. Active surveillance of sexual behaviour among homosexual men in London. Communic Dis Pub Health 1998, 1: 197–201.
42. Hart GJ, Flowers P, Der GJ, Frankis JS. Homosexual men's HIV related sexual risk behaviour in Scotland. Sex Transm Infect 1999, 75: 242–246.
Keywords:

Gay men; HIV optimism; sexual behaviour; time trends

© 2002 Lippincott Williams & Wilkins, Inc.