Differences in human immunodeficiency virus (HIV) prevalence have been reported between black, white, and Asian men who have sex with men (MSM), mainly in studies from the United States of America (USA) and the United Kingdom (UK).1–6 HIV prevalence is generally higher among black and lower among Asian MSM when compared with white MSM. The factors that underpin these differences, however, are not fully understood.1,3,4
Previous studies among ethnic minority MSM in the UK have tended to focus on “black” and “Asian” MSM.6,7 These broad categories, however, are made up of a diverse range of ethnic groups.8 “Black” MSM include men of black Caribbean as well as black African ethnicity, whereas “Asian” MSM include men of Indian, Pakistani, and Bangladeshi origin, as well as other countries in the region. Consequently, ethnic minority MSM in Britain merit further examination, to explore their diversity and risk behaviors in greater depth.
Since 2004, 10 Central and Eastern European countries have joined the European Union, leading to an influx of people, including MSM, from these countries to the UK.9 In addition, some community groups have noted an increased visibility of MSM from South and Central America, which may be a consequence of recent migration to the UK.10,11 Because their numbers appear to have increased in recent years, MSM who have moved to Britain from these countries clearly merit consideration.
The aim of the Men and Sexual Health project (MESH) was to examine the sexual health of ethnic minority and migrant MSM living in Britain by means of an online survey.12 In this article, we explore differences in the uptake of HIV testing and self-reported HIV seropositivity between MSM from a number of ethnic and migrant groups in Britain. We also examine whether differences in HIV seropositivity can be explained by individual risk factors for HIV, such as sexual behavior or recreational drug use.
For the MESH project, we recruited a national sample of ethnic minority MSM using both “online” (through the Internet) and “offline” (e.g., through sexual health clinics or gay venues) methods. We also recruited MSM who had migrated to Britain from South and Central America (SCA) or from Central and Eastern Europe (CEE) (i.e., Bulgaria, Czech Republic, Estonia, Hungary, Latvia, Lithuania, Poland, Romania, Slovakia, and Slovenia); in this article, we refer to these men as “key migrant” MSM. In addition, we recruited a comparison group of white British MSM who were born in the UK. Key migrant and white British MSM were recruited primarily “online,” i.e., through the Internet. Ethnic minority MSM were recruited both “online” and “offline.” All men were asked to complete a questionnaire online, which took 20 to 30 minutes. The methods have been described in detail elsewhere.12
Our question on ethnicity was based on the 2001 census for England and Wales in which each person in the household was asked “What is your ethnic group?”13 In the census, respondents could tick one of the following: white (British, Irish, Other); black (Caribbean, African, Other); Asian (Indian, Pakistani, Bangladeshi, Other); Chinese or other ethnic group; mixed (black Caribbean and white, black African and white, Asian and white, any other mixed background). We modified this classification slightly. Instead of “Asian and white,” we created a classification “Indian, Pakistani, or Bangladeshi (IPB) and white”, as we were particularly interested in this subgroup. We omitted “any other mixed background” (such respondents could tick “other ethnic group”). In accordance with the expansion of the ethnic group categories in the 2011 census, we also included “Arab.”14
In Britain, a question on ethnicity was first included in the 1991 census. The census classification relies on how an individual defines himself or herself and reflects the group they see themselves belonging to. It does not necessarily reflect their place of birth; some people who belong to an ethnic minority may be born in the UK, others may be born elsewhere.12 The census classification is used extensively by research groups in Britain for collecting data on ethnicity.
A question on country of birth allowed us to identify men who were born in SCA or CEE (“key migrant” MSM). The majority of men born in Central and Eastern Europe described their ethnicity as “white other.” Most of the men born in SCA described themselves as “white other” or had ticked “Other ethnic group” on the questionnaire. A small number of men born in SCA or CEE ticked one of the ethnic minority categories (e.g., black Caribbean, black African, etc). To avoid double counting, men born in SCA or CEE who ticked one of the ethnic minority categories were classified as key migrants for the purpose of this analysis (mentioned in Results section).
Between August 2007 and April 2008, we promoted the MESH project using banner advertisements on community, health promotion, and social networking Web sites used by ethnic minority MSM. We also promoted the project on Gaydar, the most popular gay dating site in the UK. In February 2008, we sent an e-mail describing the MESH project to people who managed the e-mail lists of different community groups and asked them to forward the e-mail to their list members. The banner advertisements and e-mails contained a direct link to the free-standing MESH project Web site and online questionnaire.
To recruit ethnic minority MSM “offline,” we advertised the project in sexual health clinics, bars and clubs in the 15 British towns, and cities with the largest ethnic minority populations: (in alphabetical order) Birmingham, Bradford, Brighton, Bristol, Cardiff, Glasgow, Leeds, Leicester, Liverpool, London, Luton, Manchester, Newcastle, Nottingham, and Sheffield.15 Between October 2007 and February 2008, sexual health clinics in these towns and cities promoted the MESH project among ethnic minority MSM by displaying posters and postcards in their waiting areas and providing information verbally if the opportunity arose during a clinic consultation. HIV prevention and health promotion organizations distributed posters and postcards as part of their outreach work in bars, clubs, drop in centers, and other gay venues. In London, we distributed postcards at black gay pride events as well as in clubs and venues known to attract a large number of ethnic minority or South American MSM. We also placed advertisements in the London, Manchester, and Newcastle gay press, and postcards were included in the Freshers' pack sent to all university Lesbian, Gay, Bisexual and Transgender (LGBT) societies in the UK by Gaydar. Men recruited “offline” were asked to complete the questionnaire using the Internet, that is, online.
Men were asked to provide information on their sociodemographic characteristics (age, ethnicity, country of birth, place of residence, employment, and education), sexual identity and behavior, HIV test history, HIV status, recreational drug use, HIV treatment optimism, and use of the Internet for seeking sex. All information was self reported.
If men reported unprotected anal intercourse (UAI) in the previous 3 months, we asked about the HIV status and ethnicity of their partner(s). UAI was classified as either concordant (only with a partner of the same HIV status) or nonconcordant (with a partner of unknown or different HIV status). Men reporting both concordant and nonconcordant UAI were assigned to the group of greatest risk for HIV transmission, that is, nonconcordant UAI. UAI reported by men who had never been tested for HIV was classified as nonconcordant. Being unaware of their own HIV status, they were not able to establish concordance with a sexual partner.16,17
We calculated self-reported HIV seropositivity by dividing the number of men who said they had ever received a positive HIV test result by the number who said they had ever had an HIV test. HIV seropositivity derived in this way has been used in other UK studies as a proxy for HIV prevalence where the collection of biologic samples was not feasible.6,18
For financial reasons, the questionnaire was only in English. All questionnaires were anonymous and confidential. Men were not compensated for participating in the study. Consequently, there was no financial incentive for completing a questionnaire more than once.
Data were analyzed using STATA software (version STATA/SE8.2 for Windows). Men who were more than the age of 18 years, lived in the UK, and reported ever having had sex with a man were eligible for inclusion. Because of small numbers, for some of the analyses, ethnic groups were combined as follows: “black MSM” (comprising black Caribbean, black African, black other, black Caribbean and white, and black African and white men); “South Asian MSM” (comprising Indian, Pakistani, Bangladeshi, and Indian, Pakistani, Bangladeshi and white men).
In descriptive analyses, χ2 tests were used to compare different groups. We used univariable logistic regression models to examine the crude association between ethnic or key migrant group and (i) sexual behavior, (ii) uptake of HIV testing, and (iii) HIV seropositivity. These associations were further examined in multivariable logistic regression models controlling for individual risk factors, specifically age, place of residence, place of birth, education, employment, HIV treatment optimism,16 recreational drug use19 as well as HIV status (sexual behavior model only), or sexual behavior (HIV seropositivity model only). All analyses were based on self-reported data.
In the text and tables, data are presented (i) for all ethnic minority MSM combined (except in Table 5), (ii) for black and South Asian MSM separately, and (iii) for the 13 individual ethnic groups described earlier in the text. Data are also presented for key migrant MSM (i.e., men born in CEE, and men born in SCA) and for the comparison group of white British men born in the UK. All men were living in Britain at the time of the survey.
More than 19,000 people clicked through to the homepage of the MESH online questionnaire and gave their consent to take part in the survey. Of these, 17,425 matched the inclusion criteria (i.e., they were male, aged >18 years, lived in the UK, and reported ever having had sex with a man); 1241 men described themselves as ethnic minority, 173 were born in SCA, 243 were born in CEE, whereas 13,717 were white British (15,374 in total). The remaining 2051 men described themselves as white Irish or white other (excluding men born in SCA or CEE). Most of the “white other” men were born in Western Europe, USA, Canada, South Africa, or Australia. The white Irish and white other men were excluded (unless they were born in SCA or CEE), as migration from these other countries was not the focus of our analysis.
Of the 15,374 ethnic minority, key migrant and white British MSM who matched the inclusion criteria,13,278 men completed the entire questionnaire and provided full information on their age, ethnicity, HIV status, and UAI in the previous 3 months. These men were included in our analysis; 991 ethnic minority MSM, 136 men born in SCA, 207 men born in CEE, and 11,944 white British MSM (Table 1).
Of 136 men born in SCA, 85 described themselves as “white Other,” 37 ticked “other ethnic group,” and 14 described themselves as belonging to a specific ethnic minority group (black Caribbean, n = 2; black other, n = 2; black African and white, n = 6; black Caribbean and white, n = 3; and Chinese, n = 1). Of the 207 men born in CEE, 203 described themselves as “white Other,” 2 as black Caribbean, and 2 were of unknown ethnicity.
The majority of men were recruited online, through Gaydar or other Web sites: white British MSM, 96%; men born in CEE, 95%; men born in SCA, 86%; ethnic minority MSM, 72%.12
Of the 15,374 respondents who matched the inclusion criteria, 13,278 were included in the analysis, whereas 2096 were excluded because of incomplete data on age, ethnicity, HIV status, or UAI. Compared with men who were included in the analysis, those who were excluded were younger, more likely to identify as bisexual, to report HIV treatment optimism, and to have tested positive for HIV (full data available from authors on request).
Ethnic minority and key migrant men were younger than white British men (P < 0.001) and more likely to live in London (P < 0.001) (Table 1). Half the ethnic minority MSM was born in the UK, ranging from 13% for Chinese men to 89% for black Caribbean and white men (P < 0.001). By definition, all the key migrants were born outside the UK, whereas all the white British men were born in the UK. Ethnic minority and key migrant MSM were more likely to say they were students than white British men (P < 0.001) and more likely to have some form of higher education (P < 0.001).
Overall, half the men said they had used recreational drugs in the past 12 months, although this varied between groups (P < 0.01) (Table 2). In general, ethnic minority and key migrant men were more likely than white British men to believe that new HIV treatments made people with HIV less infectious (P < 0.001) (Table 2). The majority of respondents in all groups had used the Internet to look for sexual partners in the previous 12 months.
Sexual Identity and Behavior
In all groups, the majority of men described themselves as gay or homosexual ranging from 59% for Bangladeshi men to 94% for CEE men (P < 0.001) (Table 3). Overall, ethnic minority men were more likely to describe themselves as bisexual than white British men (18% vs. 13%, P < 0.001). However, there was substantial variation between ethnic groups (P < 0.01). Although the percentage of black African, Indian, Pakistani, Bangladeshi, and Arab men who identified as bisexual was elevated (range, 20%–41%), the percentage for black Caribbean, black other, black Caribbean-and-white, and Chinese men was little different from that for white British men (range, 7%–16%) (Table 3). Very few respondents described themselves as heterosexual (range, 0%–1%, data available on request). Most respondents said they had only had sex with a man (or men) in the previous 12 months but this varied between ethnic and migrant groups (range, 76%–96%, P = 0.05) (Table 3). In general, ethnic differences in sexual behavior reflected differences in sexual identity.
Unprotected Anal Intercourse
Over a quarter of respondents (27%) reported UAI with a partner of unknown or discordant HIV status in the previous 3 months (i.e., nonconcordant UAI), with no statistical evidence of differences between ethnic minority, key migrant, and white British MSM (P > 0.85) (Table 4). In multivariable analysis, there was no statistical evidence of differences between any of the individual groups in the percentage reporting nonconcordant UAI nor between “black men” and “South Asian men” (Table 4).
Uptake of HIV Testing
Nearly two-thirds of white British MSM (65.0%) said that they had ever tested for HIV. For ethnic minority and key migrant MSM, this percentage ranged from 59.7% (Pakistani MSM) to 90.4% (MSM born in CSA) (P < 0.001) (Table 5). In multivariable analysis, the odds of ever having had an HIV test were elevated for black, Chinese, and “other” MSM as well as for men born in CSA and CEE. The odds were also elevated for each individual group of black men (black Caribbean, black African, etc), but the 95% confidence intervals were wide in some cases. There was no evidence of a difference in uptake of HIV testing between South Asian and white British MSM (Table 5).
In our study, the uptake of HIV testing varied between white British, key migrant, and ethnic minority men (Table 5). As we did not know the prevalence of HIV among men who had never had an HIV test in our study, only men who had ever had an HIV test were included in the HIV seropositivity analysis.
Among men who had ever had an HIV test, self-reported HIV seropositivity was 13.1% for white British men. For ethnic minority and key migrant MSM, HIV seropositivity ranged from 3.8% for Chinese men to 18.7% for men from SCA (P < 0.001) (Table 5).
The odds of self-reported HIV infection for MSM from all South Asian groups combined were lower than for white British MSM in univariable and multivariable analyses (adjusted odds ratio: 0.43, 95% confidence intervals: 0.23, 0.79, P = 0.007) (Table 5). The odds of HIV infection were also reduced for each individual South Asian group, but confidence intervals were wide. Compared with white British MSM, self-reported HIV infection was also less common in MSM of Chinese ethnicity, among key migrants from CEE countries, and among MSM from Other Asian ethnic groups, although this association was slightly attenuated in the multivariable model. Men from SCA had higher odds of HIV infection than white British men in univariable but not in multivariable analysis.
The odds of self-reported HIV infection in MSM from all black ethnic groups combined were similar to those of white British MSM in both univariable and multivariable analyses (adjusted odds ratio: 1.17, 95% confidence interval: 0.79, 1.73, P = 0.3) (Table 5).
In this study of MSM living in Britain, we found differences between white, ethnic minority, and key migrant men in the uptake of HIV testing and in the percentage who had tested positive for HIV. Compared with white British men, HIV seropositivity was lower for South Asian and Chinese MSM and for men who were born in CEE, but did not differ for MSM from black ethnic groups or from countries in SCA.
Our findings are at variance with earlier studies conducted in the USA, which found that black MSM have higher HIV prevalence than white men, although they did report that Asian men have relatively low prevalence.1–5 In the US studies, HIV prevalence was measured directly from blood samples. While self-reported HIV seropositivity is not the same as directly measured prevalence, the lack of a marked difference between black and white MSM in our survey is notable. The higher uptake of HIV testing by black compared with white MSM in our study suggests that this is unlikely to be due to ethnic differences in undiagnosed infection.
The UK Gay Men's Sex Survey (UKGMSS) found a 2-fold difference in the odds of HIV infection between black and white British MSM surveyed in 2002,6 and in subsequent years,20–23 based on self-reported HIV seropositivity. In 2007, however, the UKGMSS reported a smaller difference between black and white British men compared with earlier surveys.24 The 2007 UKGMSS was conducted just a few months before the MESH survey, and the self-reported HIV seropositivity estimates for black and Asian MSM in the 2 studies are comparable. Taken together, the 2 studies suggest that the ethnic patterning of HIV infection among gay men in Britain may be fluid and could change over time. Differential selection bias could also account for differences between the 2 surveys in the ethnic distribution of HIV infection.
The low HIV seropositivity among men from CEE in our survey reflects patterns of infection among MSM in their region of origin.25 A study conducted among MSM from CEE living in the UK in 2009 also reported low HIV seropositivity (5%) in this group.18
A fundamental question is whether the differences in HIV seropositivity between ethnic minority, key migrant, and white British MSM seen here can be explained by individual risk factors for HIV. Compared with white British men in our study, ethnic minority and key migrant MSM were, in general, younger, more likely to believe that new treatments for HIV made people with HIV less infectious and more likely to live in London. These factors are all associated with HIV infection.6,16 On the other hand, there were no differences between ethnic minority, key migrant, and white British men in the percentage who reported nonconcordant UAI in the past 3 months. After adjusting for confounding risk factors in a multivariable model, the odds of HIV infection remained lower for South Asian or Chinese MSM as well as for migrants from CEE countries, compared with white British MSM. A number of studies conducted in the USA have also found that in multivariable analyses, individual risk factors cannot explain differentials in HIV infection between ethnic groups.1,3,4 In the USA, however, it is the elevated prevalence of HIV among black MSM that cannot be explained by individual risk factors.3,4 In our study, it is the lower HIV seropositivity among men of South Asian or Chinese ethnicity or among migrants from CEE that cannot be explained in this way.
We also found marked differences in the uptake of HIV testing between ethnic minority, key migrant, and white British MSM, as has been reported elsewhere.6,18 Taken together, these studies suggest that HIV testing services in the UK are accessed by men from different ethnic backgrounds, born in the UK and elsewhere. Nonetheless, it appears that men from a South Asian background are less likely to test for HIV than men from other ethnic backgrounds suggesting that targeted messages about the benefits of HIV testing may be warranted for this group.
Our study throws into sharp focus the diversity of ethnic minority MSM in Britain. In our sample, black Caribbean respondents were more likely to be born in the UK than black African respondents, reflecting different patterns of migration from the Caribbean and Africa to Britain in the second half of the 20th century.26 For many variables (e.g., sexual identity), there were differences between black Caribbean and black African men or between Indian and Pakistani men. These differences are concealed when men from these ethnic groups are classified as “black” or “Asian.” For some of our analyses, we had to use these broad groupings because of small numbers. Nonetheless, we have been able to highlight important differences between individual ethnic groups as well. Furthermore, our study has alerted us to some important differences between men of mixed ethnicity (e.g., black Caribbean and white) and men of single ethnicity (e.g., black Caribbean).
The strengths of our study are its large size and national coverage. The study has, to the best of our knowledge, the largest sample of ethnic minority MSM surveyed in Britain to date and the first to examine this population in depth in this country. The broad characteristics of the ethnic minority MSM here reflect those of the ethnic minority population recorded in the census. For example, in the 2001 census, the ethnic minority population was younger than the white British population and more likely to live in London.26,27
The main limitation of the study is that it relied on convenience samples, as is often the case for research among MSM.28,29 Consequently, we cannot claim to have recruited a representative sample of ethnic minority, key migrant, or white British MSM. Some men who met the inclusion criteria were excluded from the analysis because they did not provide full information on key variables including UAI. The excluded men were more likely to report risk factors for UAI than the men who were included. Consequently, we may have underestimated the overall level of UAI in our analysis. However, there is no evidence that this underestimation varied systematically between ethnic minority, key migrant, and white British MSM.
The questionnaire was only in English, which would have prevented men with limited knowledge of the language from participating. A further limitation is our reliance on self-reported information and the absence of biologic samples, which could introduce both selection and measurement biases.
Recruiting men through the Internet does not allow us to calculate a response rate. On the other hand, the Internet allowed us to reach a large number of men across Britain including men from an ethnic minority background. A recent study found that ethnic minority groups in Britain were just as likely to have access to the Internet at home as the rest of the population. In 2006, just over half of those surveyed said they had access to the Internet at home.30 This percentage will have increased since then. The authors noted that “there is evidence that ethnic minority groups are enthusiastic users of communications services.” This suggests that using an online survey and promoting the MESH project on a variety of Internet sites was an effective way of reaching ethnic minority MSM. Nonetheless, the number of men in some of the ethnic groups was small, highlighting the challenges of recruiting men who are a minority within a minority.
In conclusion, self-reported HIV seropositivity was lower among men of South Asian and Chinese ethnicity and among migrants from CEE living in Britain compared with white British men. These differences could not be explained by corresponding differences in sexual behavior or other individual risk factors for HIV. How then can these differences be explained? One possibility is that there may be sexual networks of MSM based on ethnicity, which could place men in some ethnic groups at greater or lower risk of HIV infection than men in other groups.3,4 This has been examined in the USA2,31 but to date has not been explored in Britain. We will consider this possibility in a future article. This study highlights the importance of health promotion targeting MSM from all ethnic and migrant groups in Britain, as their prevailing patterns of high-risk sexual behavior do not appear to differ.