Article Text

Download PDFPDF

Original article
Location-based HIV behavioural surveillance among MSM in Auckland, New Zealand 2002–2011: condom use stable and more HIV testing
  1. Peter J W Saxton1,
  2. Nigel P Dickson1,
  3. Anthony J Hughes2
  1. 1AIDS Epidemiology Group, Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand
  2. 2Research Unit, New Zealand AIDS Foundation, Auckland, New Zealand
  1. Correspondence to Dr Peter Saxton, AIDS Epidemiology Group, Department of Preventive and Social Medicine, University of Otago, PO Box 913, Dunedin 9054, New Zealand; p.saxton{at}


Objectives Over the last decade, annual HIV diagnoses among men who have sex with men (MSM) in New Zealand increased, then stabilised in 2006 and have not increased further. The aim was to examine trends in behaviours in order to better understand this pattern and inform community-based prevention.

Methods From 2002 to 2011, we conducted five repeat cross-sectional behavioural surveillance surveys among MSM at community locations in Auckland (fair day, gay bars, sex-on-site venues; n=6091). Participation was anonymous and self-completed. Recruitment methods were consistent at each round.

Results Overall, the samples became more ethnically diverse and less gay community attached over time. Condom use during anal intercourse was stable across three partnering contexts (casual, current regular fuckbuddy, current regular boyfriend), with a drop among casual contacts in 2011 only. In the 6 months prior to surveys, there was a gradual decline over time in the proportion reporting >20 male partners, an increase in acquiring partners from the internet and increases in engagement in anal intercourse in some partnering contexts. HIV testing in the 12 months prior to surveys rose from 35.1% in 2002 to 50.4% in 2011, mostly from 2008.

Conclusions This first indepth examination of trends in HIV-related behaviours among five consecutive large and diverse samples of MSM in New Zealand does not suggest condom use is declining. However, subtle changes in sexual networks and partnering may be altering the epidemic determinants in this population and increasing exposure.

  • HIV
View Full Text

Statistics from


New Zealand has a concentrated mature HIV epidemic that predominantly affects gay, bisexual and other men who have sex with men (MSM).1 ,2 After a drop in the mid-1990s, as in many developed countries3 ,4 there was a resurgence in new HIV diagnoses in this group from 2002.2 The number in New Zealand then stabilised from 2006 and has not increased further.5 HIV prevalence studies conducted in clinic6 and community7 settings indicate that the prevalence of infection remains low relative to many developed countries, as is the proportion with undiagnosed infection.7 Of considerable interest is whether behaviours influencing new diagnoses such as condom use, potential exposure to HIV and HIV testing have changed over the last decade. Such data are available for a number of gay communities internationally,8–11 but trends in New Zealand's experiences have not been extensively examined.

The aim of this paper is to report trends in behaviours among MSM in New Zealand. We report findings from a large and diverse location-based HIV behavioural surveillance programme between 2002 and 2011.


We conducted five cross-sectional surveys in offline locations (gay community fair day, gay bars and sex-on-site venues) in Auckland between 2002 and 2011. Auckland is the New Zealand city with the largest concentration of MSM, the highest HIV prevalence and contributes around 60% of domestically-acquired HIV infections diagnosed among MSM nationwide.2 ,6 ,12 Eligibility criteria were being male, aged 16 and over and having any sexual contact with another man in the previous 5 years. Those who had taken part in previous survey rounds were not excluded. Participation was anonymous and self-completed onsite with the recruitment approach (venue selection and study processes) kept consistent at every round. Details regarding the Gay Auckland Periodic Sex Survey (GAPSS) methodology have been published previously.13

Respondents reporting insertive or receptive anal intercourse with a given partner type (casual, current regular) were invited to indicate the frequency of condom use for that modality using a five point scale (‘always’, ‘almost always’, ‘about half the time’, ‘very rarely’ or ‘never’) in the 6 months prior to survey. Current regular partners were those with whom the participant had had sex four times or more over this period, and were further differentiated by those described as a ‘boyfriend, long-term lover, life partner, civil union partner’ (hereafter ‘boyfriend’-type) or a ‘fuckbuddy, friend you have sex with’ (hereafter ‘fuckbuddy’-type). These questions were asked in the same way over the five survey rounds.

We monitor four main sets of indicators. These are (i) condom use always or almost always (high condom use); (ii) any unprotected anal intercourse (anyUAI) among those engaging in anal intercourse, stratified by partner type (casual, fuckbuddy-type, boyfriend-type); (iii) anyUAI as a proportion of the total sample and stratified by partner type; and (iv) HIV testing ever and in the previous 12 months among non-previously diagnosed respondents.

For analysis of respondent characteristics, sexual partnering and behaviour variables, we conducted a χ2 test for heterogeneity and for trend across the five surveys. For the four main indicators, we conducted χ2 tests for trend over the whole period 2002–2011 and the subperiods 2002–2006 and 2006–2011. Adjusted ORs with 95% CIs (AOR 95% CI) controlling for age and recruitment site examined change between successive surveys. All analyses were conducted using Stata V.11.

Instead of response rates we calculated an ‘acceptance rate’ which divides the number of completed surveys by the sum of declined and accepted approaches. This is because it is not possible to record the outcome of approaches made to unique individuals when participation is anonymous, and multiple approaches to the same individual by different recruitment staff are likely during intensive small-community fieldwork.13 As the denominator in the acceptance rate is the number of approaches not the number of individuals, it will underestimate the true response rate.


We collected 6091 responses from MSM across five survey rounds. Acceptance rates were 80%, 69%, 67%, 53% and 49%, respectively. Over time, proportionately more respondents were recruited from the community fair day and proportionately fewer from sex-on-site venues (table 1). The age and sexual identity profile of successive samples fluctuated significantly but with no clear trend, and samples became significantly more ethnically diverse (fewer European and more Asian) and spent less free time with other gay men. A similar proportion at each round had previously been diagnosed HIV positive (average 4.7%). Disaggregating the samples by recruitment site revealed diverging trends in age profiles (the fair day samples became progressively younger, gay bar and sex-on-site samples progressively older, p trend=0.001 for each site), and the fair day samples became less European (p trend <0.001) and spent less free time with other gay men (p trend=0.003) (latter not shown in table 1).

Table 1

Sample characteristics by survey round; location-based behavioural surveillance of men who have sex with men in Auckland, New Zealand (percentage)

There was an increase in the proportion having sex with a man met online but a gradual reduction in the proportion reporting over 20 male partners (table 2). Engagement in any sex with a casual, current fuckbuddy or current boyfriend-type partner remained stable over time. Of respondents who reported a casual or fuckbuddy partner, an increasing proportion over time reported anal intercourse (from 68.2% in 2002 to 76.4% in 2011, p trend<0.001 for casual partners; from 76.8% in 2002 to 85.4% in 2011, p trend=0.062 for a fuckbuddy partner).

Table 2

Sexual partnering in previous 6 months by survey round; location-based behavioural surveillance of men who have sex with men in Auckland, New Zealand (percentage)

High condom use was reported by over 80% of MSM engaging in anal intercourse with casual partners in the previous 6 months, over two-thirds of MSM during anal intercourse with a fuckbuddy partner and over a third of MSM during anal intercourse with a boyfriend partner at every survey (table 3). Trends in high condom use were steady for all partnership categories, with the exception of condom use with casual partners which declined in 2011 (p trend for period 2006–2011=0.015) but was otherwise stable.

Table 3

Main indicators by survey round; location-based behavioural surveillance of men who have sex with men in Auckland, New Zealand (percentage, AOR 95% CI vs previous survey round)

Employing a denominator of respondents engaging in anal intercourse with each partner type, trends in anyUAI reverse-mirrored findings for high condom use (table 3). Rates of anyUAI with boyfriend and fuckbuddy partners were flat, as they were generally for casual sex, although there was an increase in unprotected intercourse in 2011 with casual partners (p trend for period 2006–2011=0.050).

Expanding the denominator to the total sample revealed increasing trends in anyUAI in the context of sex with casual partners (from 15.1% in 2002 to 19.9% in 2011, p trend for period 2002–2011=0.002) and with a boyfriend partner (from 22.5% in 2002 to 26.3% in 2011, p trend for period 2002–2011=0.045).

HIV testing was stable from 2002 to 2006 and then increased in 2008 and 2011 (table 3). Testing in the previous 12 months rose from 35.1% in 2002 to 50.4% in 2011 (p trend <0.001) among non-previously diagnosed positive respondents. Rates of lifetime testing increased from 74.9% in 2002 to 81.6% in 2011 among all respondents (p trend <0.001).

The trends identified in condom use, anyUAI and HIV testing remained statistically significant when tested for change between consecutive surveys after adjusting for age and recruitment site (table 3).


We found that condom use in repeat cross-sectional location-based samples of MSM in Auckland, New Zealand, has been mostly sustained across multiple partner types over the period 2002–2011, although this was lower in the 2011 survey in the context of casual sex. We observed significant rises in recent and therefore not surprisingly lifetime HIV testing, in particular from the 2008 survey. We also observed proportional increases in the practice of anal intercourse with casual and boyfriend partners, and an increase in the proportion of respondents reporting a current boyfriend. This meant that whereas condom use itself remained steady during anal intercourse, these secular trends elevated the overall proportion who may have been exposed to HIV through engagement in any UAI.

GAPSS is a repeat cross-sectional study conducted in community settings, and the sample characteristics fluctuated modestly across surveys as well as exhibiting trends towards more ethnic diversity and less gay community attachment. Without routinely collected representative data on gay men, we cannot know whether these observed trends are most likely due to underlying shifts in the Auckland general population, socio-demographic shifts within the gay community in Auckland, attrition of certain MSM away from venues where sampling occurs or the introduction of biases in our sampling approach over time. Changes in gay communities participating in behavioural surveys have been noted in other countries.14 ,15 We have elsewhere discussed ‘sample drift’ among MSM from offline to online venues and its implications for interpreting findings from repeat cross-sectional surveys.13 Our results should therefore be compared with MSM recruited online for a more complete appraisal of behavioural trends.

This is the first paper to formally examine trends in condom use and HIV testing among MSM in New Zealand, which has had a relatively successful history of HIV control. Strengths of the analysis include the use of large and diverse samples that have been collected in consistent ways over a decade and coincident with a rise, stabilisation and then fall in HIV diagnoses in this country. As HIV prevalence is the greatest in Auckland, these behavioural trends are likely to drive the pattern of HIV diagnoses across New Zealand. We also examined condom use stratified by three sexual partnership types, reflecting the conditional nature of this practice, using two alternative measures (high condom use and anyUAI) and two denominators (those engaging in anal intercourse and the total sample).

A limitation is that we present an aggregated analysis. It will be important to examine the experiences of subpopulations of respondents who might play a disproportionate role in onward transmission, such as HIV positive MSM and MSM with high numbers of sexual partners, or whose behaviour appears to be patterned, such as according to partners' perceived HIV status or by modality of anal intercourse. We restricted the main indicators to condom use and HIV testing, but trends in other factors that can drive HIV epidemics such as sexual concurrency and sexual mixing ought to be explored.16 Likewise, HIV testing in the 12 months prior to survey is a commonly used indicator that helps standardise reporting between countries, but a measure that better relates recent testing practices to recent HIV exposure should be investigated.

Survey acceptance rates reduced over time from 80% in 2002 to 49% in 2011. As the denominator counts the number of approaches not unique individuals, this may be due to improved recruitment efforts over time (recruitment staff increased from 10 in 2002 to 40 in 2011) and underestimates the true response rate, but it could also reflect an actual decline in willingness to participate. Response rates have been declining generally in survey research17 and have been documented in location-based surveys of gay men elsewhere.18 Changes in response rates will not bias findings unless this occurs non-randomly; information on non-responders is not collected in our study and is often impractical in these settings. Thus, it is not possible to know whether there is any underestimation or overestimation of behavioural outcomes over time, nor the nature of this. Informal support for the survey continues to be high in the gay community and the recruitment regularly achieves over 1000 MSM respondents over 1 week in a city of 1.4 million inhabitants.

In the context of otherwise stable condom use, the drop in 2011 with casual partners (but not fuckbuddy or boyfriend partners) is a concern. A potential bias is that 2011 was the first time participants were invited to provide oral fluid specimens for anonymous HIV testing.7 In all, 80% provided these, and those who did were more likely to report anyUAI with a casual partner (42.3% vs 29.9%, p=0.018). As the rate of anyUAI among those not providing a specimen in 2011 (29.9%) was most similar to the previous survey round's finding (31.2%), it is possible that provision of a specimen improved recollection or disclosure of unprotected episodes in the previous 6 months. Corroboratory evidence is available from MSM recruited on online dating sites in Auckland in a 2011 companion survey, who were not asked to provide oral fluid specimens. That should be examined and future offline surveys conducted to better gauge whether the 2011 offline data point was an anomaly. A general limitation to repeat surveys of sexual practices is that they rely on self-reported behaviours that cannot be verified, even though our anonymous study protocols ought to have minimised deliberate misreporting.

The stable proportion of respondents with diagnosed HIV infection is consistent with estimates of diagnosed HIV prevalence among MSM from national surveillance1 and from sexual health clinics.6 This is in contrast to a rising number of MSM living with diagnosed HIV,1 many of whom will now be ageing. As behavioural surveys will oversample younger gay community members, a stable diagnosed HIV prevalence is an indication of ongoing incidence, reduced mortality due to antiretroviral therapy (ART) and retirement from the sexually active population. The sexual behaviours and condom use of MSM with diagnosed and undiagnosed HIV infection in Auckland have been reported previously,7 and a repeated bio-sampling in future years would enable trends in these important groups to be monitored.

These findings add to our understanding of HIV prevention locally and internationally. New Zealand's approach has been based on early, gay-affirmative, community-led and policy-supported responses from the mid-1980s, with an emphasis on primary prevention and in particular a continued focus on valuing, promoting and facilitating condoms for anal intercourse. That men have remained as willing to use condoms over the last decade demonstrates that this remains possible even in a low prevalence mature epidemic, given the determination of prevention agencies and the resources to do so. Furthermore, free community-based rapid HIV testing was introduced and promoted in Auckland from 2008 and recent testing rates have increased since then. However, as rapid tests still account for a small proportion of all testing, caution needs to be applied before attributing the rise to this alone.

New Zealand witnessed a rise in per-capita rates of new HIV diagnoses among MSM prior to 2006 in common with many industrialised countries with mature epidemics, albeit at a lower level than the UK, Australia and the USA.2–4 The universality of this resurgence probably reflected a rise in HIV incidence, which was often preceded by rises in UAI.19–22 Our data show stable trends in condom use from 2002, arguing against ongoing condom erosion being the leading cause of New Zealand's increase in HIV diagnoses, which then stabilised and have not increased. If a major change in condom use had occurred prior to 2002 we would not have detected this and it could have resulted in the increased incidence. However, 34% of MSM in a 1996 nationwide study reported anyUAI during anal intercourse with casual partners,23 consistent with the rate of 33.5% in the first 2002 GAPSS survey.

We found no synchronicity between rates of recent testing and annual HIV diagnoses. Our data therefore agree with research concluding that increases in new diagnoses among MSM are unlikely to have been driven by increases in HIV testing alone.4 Diagnosing more incident infections could have reduced the force of infection from undiagnosed HIV infected MSM and helped transmission rates level off, as new diagnoses have done in New Zealand recently, but no change in average CD4 count at time of diagnosis has been recorded among MSM in New Zealand which would indicate this.24 Diagnosing and linking more latent infections into treatment could also reduce the population HIV viral load and contribute to less transmission; approximately 70% of known HIV positive participants in 2011 were currently on ART (unpublished data), but no data on trends of ART uptake have been published for New Zealand. That self-reported HIV prevalence among respondents has remained steady even though HIV testing rates have increased may be due to the comparatively low proportion of HIV infected MSM who were undiagnosed in Auckland in 2011 (21%),7 or that the increase in testing is not occurring among those most at risk.

Instead, it is plausible that a combination of factors is exerting a multiplicative effect on population-level transmission risk, and that small shifts in and between sexual networks are pushing the reproductive rate of HIV near or beyond the epidemic threshold.16 ,25 This is supported by other findings reported here, for example, the increase in partner acquisition online between 2002 and 2004, and an increase in anal intercourse within sexual partnerships from 2006. If these shifts in the sexual marketplace have established new values for the epidemic determinants among MSM, it is possible that condom use needs to be even higher than it already is in order to further constrain HIV transmission opportunities, and/or infected individuals diagnosed earlier and offered treatment if appropriate.

While trends in Auckland's HIV epidemic drive the overall New Zealand pattern, the use of Auckland-based behavioural data to interpret the nationwide HIV diagnosis trends must also be mindful that a small proportion of annual diagnoses occur in other cities, mostly Wellington and Christchurch. Trends in behaviours among MSM in those cities could also be examined to validate these possible explanations. Last, the implications of our data need to be interpreted in the context of a relatively low HIV prevalence country with a mature concentrated epidemic, low levels of undiagnosed infection and broad support for condom use within HIV prevention organisations26 and gay communities27 alike.

Our findings support the use of regularly conducted, comprehensive, location-based, behavioural surveillance among MSM to monitor trends in prevention uptake, inform control strategies and interpret HIV epidemiological data.28 ,29 These systems can also inform control strategies for other sexually transmitted infections among MSM, which are emerging as substantial public health problems. A particular challenge raised by these data is how best to respond to stable adherence to condoms among MSM in New Zealand, but a small and steady rise in absolute exposure due to changes in sexual partnering and practices.

Key messages

  • Condom use can be sustained among men who have sex with men (MSM) in a concentrated mature low prevalence HIV epidemic.

  • Increases in condom use and screening may still be required to offset possible changes in epidemic determinants over the last decade, such as changes in sexual practices, and evident by re-emerging sexually transmitted infections.

  • Trends in behaviours of online-recruited MSM are needed to corroborate those of offline-recruited MSM.


The authors would like to thank the men who participated in the study and the site operators for allowing us access.


View Abstract


  • Handling editor Jackie A Cassell

  • Contributors PS, ND and TH designed the study. PS project managed data collection, conducted the analysis and wrote the first draft. All authors contributed to the manuscript and approved the final version.

  • Funding PS conducted the research during tenure of a Health Sciences Career Development Award of the University of Otago. Study funding was provided by the Ministry of Health, the AIDS Epidemiology Group at the University of Otago Medical School and the Research Unit at the New Zealand AIDS Foundation.

  • Competing interests None.

  • Ethics approval Ethics approval was obtained from the Northern X Regional Ethics Committee of the Ministry of Health.

  • Provenance and peer review Not commissioned; externally peer reviewed.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.