Men who have sex with men (MSM) have been disproportionately affected by the HIV pandemic due to a combination of structural, biological, social, and behavioral factors that increase their vulnerability to HIV infection.1-9 Even in countries with generalized HIV epidemics, the prevalence of HIV infection is considerably higher among MSM compared with men in the general population.2,10,11 Recent studies in several African countries confirm that HIV prevalence among MSM is high relative to the general population.6,9,12,13 In South Africa, little is known about the demographics, HIV prevalence, or risk factors for HIV among MSM in the local context, and current HIV policies and programs are largely unresponsive to the needs of MSM.14,15
The Johannesburg/eThekwini Men's Study (JEMS) was conducted as a rapid assessment to describe the HIV epidemic among MSM in South Africa and to plan for a larger multicity study on HIV and MSM. Although the study had 3 components: key informant interviews; focus group discussions (FGDs); and a survey, this article focuses primarily on the survey component of the study. The objectives of the survey were to determine HIV prevalence and risk factors for HIV infection among MSM in 2 cities. Secondary objectives were to pilot the use of respondent-driven sampling (RDS) for recruiting MSM and to pilot a questionnaire developed by the researchers.
The survey was conducted in Johannesburg and the municipality of eThekwini (Durban). Johannesburg is the largest city in South Africa, and Durban is the largest city in KwaZulu-Natal, the province with the highest HIV prevalence and the greatest number of people living with HIV infection.
Before the study started, a consultative meeting was held with key stakeholders, including representatives of Lesbian, Gay, Bisexual, Transgender (LGBT) organizations, to obtain input on protocol development and to ensure that the study was responsive to stakeholder needs. A steering committee was established to provide advice and guidance. The study was approved by the Research Ethics Committee of the South African Human Sciences Research Council and the Human Research Ethics Committee of the University of the Witwatersrand, Johannesburg.
A project manager was appointed and was responsible for overall quality assurance, and each site had a dedicated site manager. A field site was established in each city, based on accessibility, and having adequate space for study needs, including confidential interviews, specimen collection, and HIV counseling and testing. In Johannesburg, the site was provided free of charge, it was private and close to major transport routes. In Durban, difficulties with finding an affordable and appropriate site led to the team taking up the offer of the city's gay and lesbian community centre as the field site. The majority of survey staff appointed was drawn from the LGBT community, whereas 2 retired nurses were recruited and trained to collect the dry blood spots for anonymous HIV testing and to conduct voluntary counselling and testing. A detailed JEMS procedure manual was developed. All study staff received initial 2-day training, follow-up staff refresher training, and a dry-run was done in each city shortly before the first participants were recruited. Ongoing on-site supervision was provided at the Johannesburg site, but on-site supervision in Durban was intermittent because none of the investigators were based in Durban. In addition, despite numerous interventions by the investigators, there were performance problems both with overall project and Durban site management.
Formative research was conducted and consisted of semistructured interviews with 32 key informants and 18 FGDs with MSM in Johannesburg, Durban, and Cape Town. The latter was included because of access to 2 groups of HIV-positive and Muslim MSM. Key informants were drawn from 3 main categories as follows: LGBT activists; policy-makers and HIV program managers from government; and researchers and experts on MSM and HIV. The key informant interviews focused on the following: challenges faced by MSM in South Africa; HIV-related services for MSM; knowledge of other research on HIV and MSM; and advice for planning the FGDs and the survey.
The FGDs explored the following: the meaning of same-sex behavior; colloquial terms used to describe same-sex relationships; sexual behaviors and practices; knowledge of HIV prevention, treatment, care, and perceived needs; and suggestions for an appropriate research incentive, survey planning, and health service improvements.
Sample Size Calculations
As no estimates were available on HIV prevalence among MSM in South Africa, Actuarial Society of South Africa estimates of HIV prevalence among adult males were used to estimate sample size requirements.16 Assuming an HIV prevalence of 20%, a 95% confidence interval, with a precision of ±10 percentage points, a power of 80%, and a design effect of 1.25, the required sample size was 173 participants per city. We adjusted the sample size for a 10% nonresponse rate to set the target sample size of 200 men in each city.
Recruitment took place between July and December 2008. Participants were recruited using RDS,17-20 a method that had been used to recruit MSM in surveys in other countries, including several African countries.13,21,22 Eligible participants were required to be 18 years or older; live, work, or socialize in Johannesburg or Durban; be biologically male (irrespective of gender identity); and to report having sex with another male in the previous 12 months. Volunteers of differing race, age, and socio-economic status, who met the eligibility criteria, were recruited to serve as “seeds” to initiate recruitment chains. In Johannesburg, the survey started with 5 seeds, and in Durban, the survey started with 4 seeds. Each seed, and subsequent participant, was provided with 3 recruitment coupons. Additional seeds were added if seeds did not bring in participants. A further 9 seeds were added in Johannesburg, and 11 seeds were added in Durban.
Participants were given 40 South African rands (ZAR) (about $5 US) and a store voucher worth ZAR 40 to reimburse them for their travel expenses and to compensate them for their time and effort. Participants were given a further ZAR 40 shopping voucher for each additional participant recruited.
The questionnaire developed for the survey drew on elements from other questionnaires with input from steering committee members. Using a 12-month recall period, the questionnaire included sections on demographics, relationships with men and women, sexual behavior including the use of condoms and lubricants, drug, and alcohol use; HIV-related knowledge, HIV testing history and risk perception, and health service use.
After providing written informed consent, each participant completed a questionnaire that took about 1 hour. The questionnaires were piloted, both in an FGD and with a small number of MSM. Questionnaires, available in English, isiZulu, seTswana, and Afrikaans, were either self-administered in the presence of the trained interviewer who ensured that the participant understood the questions, or administered by the interviewer, depending on participant preferences. Each questionnaire was checked for completion by the interviewer in the presence of the participant before these were handed to the site manager who did an additional check.
HIV testing for the survey and HIV testing to inform participants of their HIV status were done separately to achieve a high HIV testing response rate in the survey, while providing participants with the choice of knowing their HIV status. All study participants were asked to provide finger-prick dried blood spot specimens for anonymous HIV testing in a laboratory. HIV testing was done in series, using combination of 3 licensed ELISA tests, in accordance with a standard World Health Organization algorithm.23 All participants were offered free confidential HIV counseling and testing and risk reduction counseling. Participants were also given information on the locations of HIV counseling centers for future reference.
The data were analyzed using Stata version 11.24 HIV prevalence was calculated separately for each city. Failure to reach the target sample size in Durban and failure to recruit a demographically diverse group of participants mirroring the demographics of the local population led to the decision not to adjust the results for the sampling method as is usually done in RDS studies,18 as this would not have achieved the desired effect of determining prevalence estimates that approximate those of a probability sample. To identify factors associated with being HIV positive, the data from both cities were combined and odds ratios were calculated using logistic regression, adjusting for city, age, educational level, and whether the participant had a main male partner. For ordinal variables, the Mann-Whitney test was used to assess whether the median value differed significantly by HIV status.
In Johannesburg, 204 participants were recruited in 15 waves over 21 weeks; and in Durban 81 participants were recruited in 11 waves over 17 weeks. In Durban, great difficulty was experienced in recruiting participants, largely because of the location of the study site and insufficient ongoing on-site supervision. Efforts to increase recruitment in Durban by enrolling additional seeds and adding a satellite site had limited success.
Participants had a median age of 22 years; a median age at sexual debut with a man of 17 years (Table 1); and were predominantly black Africans (88.3%) who identified as either homosexual/gay (76.9%) or bisexual (18.9%) (Table 2). HIV-positive participants were significantly older than HIV-negative participants and reported significantly more partners in the previous year, but age at sexual debut with a male and number of male partners in the previous year did not differ significantly by HIV status (Table 1). HIV-positive participants were more likely to have disclosed their sexual orientation to family and friends and were less likely to have ever had sex with a woman (Table 2).
Sixty-nine of 81 (85.2%) participants in Durban and 202 of 204 (99.0%) of participants in Johannesburg provided specimens for HIV testing, but 5 of the HIV test results were either missing or were unable to be linked to the participant questionnaire due to labeling errors. Among Johannesburg participants the HIV prevalence was 49.5% (95% CI: 42.5% to 56.5%), whereas among Durban participants, the HIV prevalence was 27.5% (95% CI: 17.0% to 38.1%). HIV prevalence differed by sexual orientation, being higher among gay-identified MSM (52.2%; 95% CI: 45.3% to 59.1%) than among those who identified as bisexual (13.7%; 95% CI: 4.3% to 23.2%). None of the 5 straight-identified MSM were HIV positive.
Behavioral and Other Risk Factors for HIV-Infection
More than half of the participants reported having unprotected anal intercourse (UAI) in the previous 12 months, with 40.6% reporting receptive UAI, and 17.7% reporting insertive UAI (Table 3). Around 30.6% of participants reported having unprotected sex with someone with an unknown HIV status, whereas 3.9% reported having unprotected sex with someone known to be HIV positive. Almost 3-quarters of participants reported having sex while under the influence of alcohol, but only a minority reported having used recreational drugs (16.9%), or having sex although under the influence of drugs (10.6%). Sixty-one percent of participants reported having had a sexual partner 10 or more years older. Many participants reported using condoms with substances that reduce the protective effect of condoms including petroleum jelly (32.3%) and lotion (28.3%). Despite high rates of reported high-risk behavior, only 36.2% of participants considered themselves to be at high risk of acquiring or transmitting HIV infection. The survey findings were supported by the formative qualitative research that found a high prevalence of HIV risk behaviors and practices, including multiple sexual partners, UAI, inconsistent condom use, and condom breakage or slippage.
HIV-positive men reported more high-risk practices than HIV-negative men (Table 3). HIV-positive participants were significantly more likely to report having receptive UAI; having UAI with someone known to be HIV-positive; and having an sexually transmitted infection (STI) diagnosed. There were no significant differences, however, by HIV status with respect to having sex while under the influence of alcohol or drugs. HIV-positive participants were more likely to consider themselves to be at high risk of HIV infection, but participants were not asked to disclose their HIV status, so it could not be established which HIV-positive participants were aware of their HIV-positive status.
The survey found a high prevalence of HIV among MSM in Johannesburg and Durban. In view of the smaller sample size and the greater number of refusals to provide a specimen for HIV testing in Durban, the HIV prevalence estimate for Durban is less reliable than the estimate for Johannesburg. The HIV prevalence among MSM in both cities was considerably higher than HIV prevalence among men in the general population (Fig. 1).16,25 Our study findings are similar to those from other African countries with generalized HIV epidemics,6,9,12 and 2 other South African studies of MSM, one in Soweto26 and the other in Cape Town27 (Fig. 1).
The higher HIV prevalence estimates in JEMS compared with those for MSM in Soweto,26 and MSM in Cape Town,27 can be attributed to regional variations in HIV prevalence; different sampling methods; differences in the characteristics of the study participants, particularly with respect to sexual orientation and socio-economic status; and bias due to refusal of HIV testing. Our study had a lower HIV testing refusal rate (4.9%) than the Soweto Men's Study (19.3%)26 and a lower proportion of straight-identified MSM (1.8%) than either the Soweto Men's Study (33.7%)26 or the Cape Town study (5.2%).27 Similarly, the higher HIV prevalence estimate in our study, compared with the 2008 South African national household HIV survey,25 could be because men may have been reluctant to disclose homosexual behavior in a household survey and because of lack of precision due to the small sample size (n = 86).25
Our study found a higher HIV prevalence among gay men than among bisexual men, similar to the other South African studies.26,27 We did not find any HIV infections among straight MSM, but the numbers were too small (n = 5) to draw conclusions.
The finding of a strong association between HIV infection and receptive UAI, unprotected sex with a person known to be HIV positive, and a history of an STI, is consistent with the findings of other studies.5,6,12,26,28 As with all cross-sectional surveys, the temporal sequence between HIV infection and risk factors was not determined, leading to uncertainty as to whether practices were causally-related to HIV infection or whether these practices were a consequence of being aware of one's positive HIV status. For example, the finding of an association between HIV infection and having UAI with someone known to be HIV positive may be attributable either to men becoming infected through UAI with an HIV-positive person or those who were aware that they were HIV positive seeking out other HIV-positive men as partners (serosorting), as reported in other studies.29-31
The lack of association between HIV infection and having sex under the influence of alcohol in this study is surprising given that other studies among both heterosexual populations32 and MSM populations,33,34 including the Soweto Men's Study,26 found an association between HIV infection and alcohol use. Failure to find an association between HIV and alcohol use in this study may be attributable to participants not being asked detailed questions about their drinking habits and sex under the influence of alcohol being pervasive among participants.
There is no standard recall period used in surveys of HIV risk behavior and practices. A recall period of 12 months was used in this study, in common with the recall period used to measure United Nations General Assembly Special Session indicators35 and similar to that used in the Cape Town study.27 Other studies have used recall periods of 1 month,13 3 months,36 6 months,12,26 or a variety of recall periods.9 This makes it difficult to make comparisons of the frequency of risk practices among participants in this study relative to those of other study populations.
There are several limitations to the study. The most important limitation is the failure to reach our sample size target at the Durban site due to site location and staff performance problems. Another limitation is the lack of demographic diversity among survey participants, particularly the small numbers of non-black African men, men over 25 years of age, and men of medium and high socio-economic status. Participants were not asked to disclose their HIV status. This meant that it was not possible to determine which participants were aware of their HIV-positive status and analyze how this knowledge had influenced their behavior and practices. Although there was piloting of the questionnaire, some questions were not well understood, such as questions that required participants to draw a distinction between main partners and casual partners. The lead-in question about recreational drug use (taken from a widely used MSM questionnaire) was awkwardly worded and may have led to an under-reporting of drug use.
Nevertheless, the study strengths are numerous. The geographical focus of JEMS is novel, and it is one of the first RDS studies completed among MSM in South Africa, and is the first to publish data from Johannesburg and Durban. The HIV prevalence estimates provide a basis for future comparisons of HIV among MSM in the 2 cities. Recruiting participants from the community rather than through venues is a strength and provides evidence that HIV infection is not restricted to visible MSM subgroups such as club goers, visitors at Gay Pride events, and health service users. The use of a dual HIV testing strategy that provided HIV testing and counseling as a service separate to the HIV testing for epidemiological purposes resulted in a high HIV testing rate, yet gave participants the choice of receiving their HIV results. Providing participants with the option of self-administering the questionnaire, albeit supervised, provided greater privacy and is likely to have led to more accurate reporting of practices that are subject to social sanction. The concurrent measurement of HIV and associated risk factors enabled the identification of the demographic, behavioral factors, and risk practices most strongly associated with HIV infection among MSM in South Africa, thus providing useful information for planning HIV prevention and testing programs targeting MSM. The lessons learnt from study execution have been used subsequently to plan for a multicity HIV study among MSM.
Further research is needed to assess the utility of other methods of sampling hidden populations, such as time-location sampling,37,38 for conducting HIV surveillance among MSM in South Africa. Particular attention needs to be given to finding ways of increasing racial, socio-economic, and age diversity, and expanding geographic coverage. It may be necessary to use a variety of recruitment strategies in future surveys to obtain a demographically diverse sample of MSM that provides a representation of the diversity of MSM in the South African population.
The study findings underscore the need to implement HIV surveillance among MSM and to expand the availability of HIV prevention programs targeting MSM. These activities would give effect to the goals and targets contained in the South African National Strategic Plan on HIV&AIDS and STIs, approved in March 2007.39
This study was funded by the United Kingdom Department for International Development. This study would not have been possible without the support of many people, particularly Robin Gorna, Loraine Townsend, Yanga Zembe, Adrian Puren, and Nonhlanhla Mkhize. The authors thank Matthew Chersich for reviewing an earlier draft of this paper and for providing useful comments.
1. Baral S, Sifakis F, Cleghorn F, et al. Elevated risk for HIV infection among men who have sex with men in low- and middle-income countries 2000-2006: a systematic review. PLoS Med
2. Cáceres CF, Konda K, Segura ER, et al. Epidemiology of male same-sex behaviour and associated sexual health indicators in low- and middle-income countries: 2003-2007 estimates. Sex Transm Infect
3. Colby DJ. HIV knowledge and risk factors among men who have sex with men in Ho Chi Minh City, Vietnam. J Acquir Immune Defic Syndr
4. Girault P, Saidel T, Song N, et al. HIV, STIs and sexual behaviours among men who have sex with men in Phnom Penh, Cambodia. AIDS Educ Prev
5. Go VF, Srikrishnan AK, Sivaram S, et al. High HIV prevalence and risk behaviors in men who have sex with men in Chennai, India. J Acquir Immune Defic Syndr
6. Sanders EJ, Graham SM, Okuku HS, et al. HIV-1 infection in high risk men who have sex with men in Mombasa, Kenya. AIDS
7. Sifakis F, Hylton JB, Flynn C, et al. Racial disparities in HIV incidence among young men who have sex with men: the Baltimore young men's survey. J Acquir Immune Defic Syndr
8. Van Griensven F, Thanprasertsuk S, Jommaroeng R, et al. Evidence of a previously undocumented epidemic of HIV infection among men who have sex with men in Bangkok, Thailand. AIDS
9. Wade AS, Kane CT, Diallo PA, et al. HIV infection and sexually transmitted infections among men who have sex with men in Senegal. AIDS
10. Jaffe HW, Valdiserri RO, De Cock KM. The re-emerging HIV/AIDS epidemic in men who have sex with men. JAMA
11. Joint United Nations AIDS Programme, World Health Organization.2009 AIDS Epidemic Update
. Geneva, Switzerland: UNAIDS & WHO; 2009.
12. Baral S, Trapence G, Motimedi F, et al. HIV prevalence, risks for HIV infection, and human rights among men who have sex with men (MSM) in Malawi, Namibia, and Botswana. PLoS One
13. Dahoma M, Johnston LG, Holman A, et al. HIV and related risk behavior among men who have sex with men in Zanzibar, Tanzania: results of a behavioral surveillance survey. AIDS Behav
14. Rispel LC, Metcalf CA. Breaking the silence: South African HIV policies and the needs of men who have sex with men. Reprod Health Matters
15. Reddy V, Sandfort TGM, Rispel LC, eds. From Social Silence to Social Science: Same-Sex Sexuality, HIV and Gender in South Africa
. Cape Town, South Africa: Human Sciences Research Council; 2009.
17. Heckathorn DD. Respondent-driven sampling: a new approach to the study of hidden populations. Social Problems
18. Heckathorn DD. Respondent-driven sampling II: deriving valid population estimates from chain-referral samples of hidden populations. Social Problems
19. Johnston LG, Khanam R, Reza M, et al. The effectiveness of respondent-driven sampling for recruiting males who have sex with males in Dhaka, Bangladesh. AIDS Behav
20. Yeka W, Maibani-Michie G, Prybylski D, et al. Application of respondent-driven sampling to collect baseline data on FSWs and MSM for HIV risk reduction interventions in two urban centres in Papua New Guinea. J Urban Health
. 2006;83(suppl 1):60-72.
21. Kajubi P, Kamya MR, Raymond HF, et al. Gay and bisexual men in Kampala, Uganda. AIDS Behav
22. Malekinejad M, Johnston LG, Kendall C, et al. Using respondent-driven sampling methodology for HIV biological and behavioral surveillance in international settings: a systematic review. AIDS Behav
. 2008;12(4 suppl):S105-S130.
23. Sato PA, Maskill WJ, Tamashiro H, et al. Strategies for laboratory HIV testing: an examination of alternative approaches not requiring Western blot. Bull World Health Organ
24. StataCorp. Statistical Software: Release 11
. College Station, TX: Stata Corporation; 2009.
25. Shisana O, Rehle T, Simbayi LC, et al. South African National Prevalence, Incidence, Behaviour and Communication Survey 2008: A turning tide among teenagers?
Cape Town, South Africa: Human Sciences Research Council; 2009.
26. Lane T, Raymond HF, Dladla S, et al. High HIV prevalence among men who have sex with men in Soweto, South Africa: results from the Soweto Men's Study. AIDS Behav
. 2009;DOI 10.1007/s10461-009-9598-y.
27. Burrell E, Mark D, Grant R, et al. Sexual risk behaviors and HIV prevalence among urban men who have sex with men in Cape Town, South Africa. Sex Health
28. Bautista CT, Sanchez JL, Montano SM, et al. Seroprevalence of and risk factors for HIV-1 infection among South American men who have sex with men. Sex Transm Infect
29. Eaton LA, Kalichman SC, O'Connell DA, et al. A strategy for selecting sexual partners believed to pose little/no risks for HIV: serosorting and its implications for HIV transmission. AIDS Care
30. Zablotska IB, Imrie J, Prestage G, et al. Gay men's current practice of HIV seroconcordant unprotected anal intercourse: serosorting or seroguessing? AIDS Care
31. Golden MR, Stekler J, Hughes JP, et al. HIV serosorting in men who have sex with men: is it safe? J Acquir Immune Defic Syndr
32. Fisher JC, Bang H, Kapiga SH. The association between HIV infection and alcohol use: a systematic review and meta-analysis of African studies. Sex Transm Infect
33. Woolf SE, Maisto SA. Alcohol use and risk of HIV infection among men who have sex with men. AIDS Behav
34. Lane T, Shade SB, McIntyre J, et al. Alcohol and sexual risk behavior among men who have sex with men in South African township communities. AIDS Behav
. 2008;12(suppl 4):S78-S85.
35. UNAIDS. Monitoring the Declaration of Commitment on HIV/AIDS: Guidelines on Construction of Core Indicators
. Geneva, Switzerland: UNAIDS; 2009.
36. Ruan Y, Li D, Li X, et al. Relationship between syphilis and HIV infections among men who have sex with men in Beijing, China. Sex Transm Infect
37. Kendall C, Kerr LR, Gondim RC, et al. An empirical comparison of respondent-driven sampling, time location sampling, and snowball sampling for behavioral surveillance in men who have sex with men, Fortaleza, Brazil. AIDS Behav
. 2008;12(4 suppl):S97-S104.
38. Semaan S, Lauby J, Liebman J. Street and network sampling in evaluation studies of HIV risk-reduction interventions. AIDS Rev
39. Department of Health. HIV & AIDS and STIs National Strategic Plan for South Africa, 2007-2011
. Pretoria, South Africa: Department of Health; 2007.
Keywords:© 2011 Lippincott Williams & Wilkins, Inc.
HIV; men who have sex with men; risk factors; South Africa; surveillance