Secondary Logo

Journal Logo


Factors Associated With Self-Reported Unprotected Anal Sex Among Male Sex Workers in Mombasa, Kenya

Geibel, Scott MPH*; Luchters, Stanley MD, MSc; King'ola, Nzioki MA; Esu-Williams, Eka PhD; Rinyiru, Agnes BA; Tun, Waimar PhD§

Author Information
Sexually Transmitted Diseases: August 2008 - Volume 35 - Issue 8 - p 746-752
doi: 10.1097/OLQ.0b013e318170589d
  • Free

HIV PREVENTION PROGRAMS FOR men who have sex with men (MSM) and men who sell sex to men (i.e., male sex workers who sell sex to men) are an important component of a comprehensive HIV public health strategy.1,2 The inclusion of MSM or male sex workers in HIV programs in sub-Saharan Africa, however, has largely been neglected by governments and other HIV service organizations, because of denial of the existence of African homosexuality, high levels of community stigma and discrimination, and the criminalization of same-sex behavior in many countries.3

Available data on the prevalence and characteristics of same-sex sexual activity in Africa is limited. Within the past 5 years however, a limited yet growing body of literature has highlighted the existence of MSM in sub-Saharan Africa and their vulnerability to HIV. Researchers reported an HIV prevalence of 21.5% among a snowball sample of 442 MSM4 in Senegal, where national HIV prevalence is 1.6%.2 A separate study of MSM risk behavior in Senegal reported a large proportion of their sample had received money for sex,5 indicating that transactional or commercial sex may play an important role in some African MSM networks. A population-based study in South Africa reports that among males aged 15 to 26 years, having had sex with a man was significantly associated with HIV seropositivity.6

Researchers modeled the Kenyan HIV epidemic, and estimated 4.5% of new infections in Kenya in 2005 occurred through men having sex with men.7 In Nairobi, Kenya, MSM who self-identified as sex workers were significantly more likely to have experienced discrimination or violence than other MSM in the sample.8 Results from a study in Kilifi, Kenya, report HIV prevalence of 24.6% among a cohort of 285 MSM around the Mombasa area, many of whom are sex workers.9 A capture-recapture enumeration estimated that 739 male sex workers who sell sex to men were active in Mombasa District, Kenya,10 confirming the existence of a highly vulnerable population insufficiently addressed in national and local HIV programming.

Mombasa is located on the coast of Kenya, with a population of approximately 1 million people residing in the district area. The district covers approximately 100 square miles, and the main industries are shipping and tourism. Both Kenyan and foreign tourism has been linked to the male sex trade in Kenya's Coast region,11 but same-sex sexual relationships also exist in Mombasa that do not involve foreign partners, tourists, and/or sex work.12

Because of the evidence of large numbers of male sex workers in the Mombasa area, study investigators coordinated an intervention study in consultation with the Coast Provincial Medical Office and the National AIDS Control Council, which cites MSM as a target group for behavior change communication in the Kenya National HIV/AIDS Strategic Plan.13 Time-location sampling has been conducted successfully among MSM venues in the United States14 and Thailand,15 where application of this methodology to an MSM subgroup of male sex workers has also been documented.16 We present here results from the first behavioral survey of male sex workers who sell sex to men in Africa using time-location or venue-based, sampling procedures.


Sampling and Procedures

This study benefited from a collaborative effort in May 2006 to map locations where male sex workers were believed to seek clients, identify male sex workers, and enumerate them via a capture-recapture procedure described previously.10 This activity yielded a database including 65 locations within Mombasa District containing estimates of male sex workers who sell sex to men present at each venue and the peak hours when they might be present. These locations served as the primary sampling units or clusters, and included bars, nightclubs/discos, beach areas/bars, public streets and parks, and other private brothels, businesses, and estates. All clusters were selected at the first stage of sampling, and second stage selection was based on probability proportional to the number of male sex workers found at each venue in the enumeration database. This sample design was assumed to be self-weighting, as the high mobility of the target population make it impossible to estimate fixed numbers at most venues, and participants were recruited through a “take all” approach at the second stage until quotas were reached at each venue.

Twelve MSM key informants, all of whom were familiar with male sex work in Mombasa, were trained to assist with the identification and recruitment of male sex workers at the sampling venues. These MSM peer mobilizers were instructed by the study coordinator on which venues to go, and on how many “contacts” they were to confirm as male sex workers who sell sex to men. The confirmation process involved an initial visual identification, following which the peer mobilizer would approach and start casual conversation with the suspected male sex worker contact. Once the peer mobilizer was confident that the contact had a history of having sex with men, he would ask the contact if he had recently sold, and is currently willing to sell sex in exchange for money and/or goods. If the contact was confirmed as an male sex worker who sells sex to men, and was 16 years of age or older, the peer mobilizer then gave a brief description and justification of the research activities, and offered to escort the contact to a central location where they were introduced to the study coordinator and interview team. Male sex workers who rejected the offer to participate were recorded as refusals.

Respondents were compensated 300 Kenya Shillings (KSH; about US $4.50) for their time and return transportation. Concerns about confidentiality and safety were paramount to the recruitment process, hence the decision to conduct the interviews at central locations and not at the individual venues. Interviews were conducted person-to-person using handheld computers, and no personal identifications or measurements were recorded on the mobilizer forms or questionnaires.

The interview team was also trained to assess survey respondents, with the assistance of the handheld software, for possible need of medical treatment or HIV counseling. These respondents were given information about local clinics, and a research clinic located north of Mombasa where they could receive an initial needs assessment. Respondents who elected to engage in this needs assessment were given the choice of referral to linked community services or to participate in an MSM study cohort providing free comprehensive services. Free condoms were also offered to the survey participants following the interview.

Ethical Approval

The Kenyatta National Hospital Ethics and Review Committee reviewed and provided local ethical approval to the study, which was supplemented by ethical review and approval by the Population Council's Institutional Review Board and administrative approval by the Coast Provincial Medical Office. All participants provided individual informed consent to participate.

Measures and Statistical Analysis

Data were collected on socio-demographic characteristics, sexual behaviors, prevention knowledge and practices, reported STI symptoms, discrimination, violence, and health service usage. The questionnaire adapted and combined indicators from both the female sex worker and MSM surveys developed by Family Health International for Behavioral Surveillance Surveys.17 Survey interviews were administered using Perseus MobileSurvey 7 (Perseus Development Corporation, Braintree, MA) software on Dell Axim X3 handheld computers. The data were analyzed using Stata version 9.2 (StataCorp, College Station, TX). Variables from the data were initially examined using frequencies and descriptive techniques, and then chi-square tests were used to detect associations among categorical variables and the Wilcoxon rank sum test was used to detect differences between medians.

Condoms are more effective for HIV prevention when used consistently, and people who never or inconsistently use condoms are at higher risk.18,19 Therefore, the dependent variable used for this analysis is self-reported unprotected sex, that is, inconsistent or no condom use with male clients in the past 30 days. Condom use was measured using a 3-point scale (“How often did you use a condom for anal sex with all of your male paying clients in the past 30 days: Always, sometimes, or never?”). This variable was dichotomized into reported inconsistent or no use (sometimes or never) versus consistent (always) condom use with male clients. One respondent stated he “did not know,” and was excluded from the analysis.

To determine factors associated with unprotected sex, independent categorical variables were assessed for distribution and content then recoded into dichotomous variables where applicable, and continuous variables were dichotomized based on their distributions at the median. Factors significant at the (<0.05) level in the binary logistic regression analysis were selected for inclusion in the initial multivariate logistic regression model. Variables not significant at the (<0.05) level were systematically removed from the model using the backwards stepwise method, after which model subset statistics were evaluated and tested for “fit”20 and a final model identified. Standard errors in both the binary and multivariate logistic regression analyses were adjusted using Stata's svy procedures,21 to account for cluster effects among males contacted at the same venues.


Sample Characteristics

A total sample of 510 male sex workers were identified and contacted at the venues, and 425 survey interviews were completed, resulting in an 83.3% acceptance rate. Interviews were conducted from October to December 2006. As summarized in Table 1, most male sex workers were Kenyan (98.4%), with 6 respondents from Tanzania and 1 from Uganda. The majority of respondents were less than 30 years old, and approximately two-thirds of the sample had completed primary school. Religion was divided between Muslim respondents (48.0%) and the Christian Catholic (29.6%), and Protestant (19.3%) faiths. Sex work was the only source of occupational income for many (41.7%), whereas other income sources included street vending and casual labor. The median amount earned from the last clients was KSH 1000 (about US$ 15.00). At the same time, over one-half were currently providing financial support to partners, family, or other people. Over half of male sex workers reported typically drinking alcohol 2 days or more per week.

Characteristics of Male Sex Workers Who Sell Sex to Men in Mombasa District (n = 425)

Most of the respondents (87.3%) reported having a male client within the past 7 days, and the median number of male clients reported in the past 7 days was 2. The majority of last paying male clients were Kenyan (80.9%). Over half of respondents reported selling sex outside Mombasa in the past year and identified 39 distinct locations within Kenya (Fig. 1), and 9 other locations in the Africa region, and 3 locations overseas. Noncondom use during anal sex with last male clients was reported by 42.1%, whereas 64.0% reported inconsistent (47.5%) or no (16.5%) use of condoms with all male clients in the past 30 days. Slightly more than one-half of respondents (57.6%) reported being only an insertive partner for anal sex with their last client. Physical abuse was reported by 12.2% of all respondents over the past 12 months, whereas 9.9% were victims of sexual violence or rape.

Fig. 1
Fig. 1:
Locations in Kenya where male sex workers have sold sex to male clients outside Mombasa in the past 12 months (n = 233, multiple responses allowed). This figure uses mapping data courtesy of the Africover Initiative, Food and Agricultural Organization, United Nations, 2000.

One-quarter of respondents (25.2%) reported currently living with a male sexual partner. Sexual activity with women was also reported by a number of male sex workers; 67.1% had ever had sex with a woman, 14.6% had a female paying client in the past 30 days, and 25.4% reported a nonpaying female sexual partner in the past 30 days. Additionally, 4.0% were currently living with a female wife, and 8.2% were living with a nonmarital female sexual partner. Some respondents were living concurrently with various partners, as overall 28.8% reported living with either a male or female sex partner.

Self-reported sexual identities were given in English and Kiswahili terminologies, including basha, king, shoga, queen, bisexual, gay, kuchu, and homosexual. Bashas or kings made up 51.3% of the sample, and were found to be characteristically and qualitatively distinct from the other groups. Bashas or kings were significantly more likely to be the insertive partner during anal sex with their last male client than those identified otherwise (97.7% vs. 14.8%, P <0.001). They were also more likely to have had a nonpaying female sex partner in the past 30 days (44.7% vs. 27.8%, P = 0.004), to be providing financial support to someone (61.9% vs. 51.2%, P = 0.026), and to have had first sex with a man at a later age (median: 19 vs. 16 years, P <0.001).

Although a large majority of respondents knew that consistent condom use prevents HIV transmission (86.6%), only 64.7% of male sex workers in Mombasa knew that HIV can be transmitted through unprotected anal sex. Slightly over one-half of male sex workers (55.8%) reported having ever received HIV counseling and/or testing. Some respondents also reported accessing MSM-friendly services at a drop-in centre or clinic in the past 12 months (22.1%). Approximately, one-third of respondents reported having burning urination in the past 12 months, whereas those reporting penile discharge (21.9%) and anal discharge (8.3%) were less.

Use of Oil-Based Lubricants

Only 21.2% of respondents correctly knew that a water-based lubricant should be used with latex condoms. Vaseline or petroleum jelly was used by 36.2%, and 15.3% reported using a water-based lubricant during anal sex with their last male client, whereas 28.6% reported using no lubrication. Among respondents who use condoms (n = 357), male sex workers who used an oil-based lubrication with their last male client were significantly more likely to have ever experienced condom breakage (48.4% vs. 35.6%, P = 0.015).

Factors Associated With Unprotected Sex

In Table 2, we analyzed the bivariate associations between unprotected sex in the past 30 days and selected background characteristics, sexual identity and behaviors, reported STI symptoms, exposure to discrimination and violence, and exposure to HIV counseling and services. Key characteristics were associated with unprotected sex; male sex workers who engaged in frequent alcohol use (3–7 days per week), who self-identified as basha or king, and who received less money from their last male client (KSH 1000 or less) were significantly more likely to engage in unprotected sex with male clients. Incorrect knowledge about both HIV transmission via anal sex and consistent condom use as a prevention strategy were significantly associated with unprotected sex, as was self-reporting of burning urination or penile discharge in the past 12 months. Respondents who had never accessed key health services, including HIV counseling and testing or visiting an MSM-friendly drop-in centre or clinic, were also significantly more likely to report unprotected sex.

Bivariate Logistic Regression

Table 3 summarizes the results of multiple logistic regression of factors associated with unprotected sex after adjusting for all factors associated in bivariate analysis. Four variables were significant: not knowing HIV can be transmitted via anal sex [adjusted odds ratio (AOR), 1.92; 95% confidence interval (95% CI), 1.16–3.16]; drinking alcohol 3 or more days per week (AOR, 1.63; 95% CI, 1.05–2.54); self-report of burning urination within the past 12 months (AOR, 2.07; 95% CI, 1.14–3.76); and having never been counseled or tested for HIV (AOR, 1.66; 95% CI, 1.07–2.57). The factors in this model correctly classified consistent condom use or unprotected sex in the past 30 days for 67.0% of respondents, and the model was statistically significant (F = 5.95, P <0.001).

Final Multivariate Model


This study characterizes the HIV risk behaviors of a mobile and highly vulnerable population. We documented high rates of unprotected sex and low levels of basic HIV understanding, which emphasize the critical need to provide the most basic of HIV prevention messages and services for male sex workers in an African setting. Thirty-five percent of male sex workers in Mombasa were not aware that HIV can be transmitted via anal sex.

After the survey was completed, male sex workers who participated in a focus group discussion rationalized this belief in 2 ways: anal sex was perceived to be a dry environment through which the HIV virus cannot move, and that anal sex as a risk behavior was never mentioned in Kenyan media campaigns or by health educators. Indeed, most HIV prevention programs in Africa operate under the assumption that vaginal sex is the only mode of HIV sexual transmission. This may be a concern for African women as well, as studies have documented anal sex practice among female sex workers in Kenya,22,23 and among young South African males in heterosexual relationships.24

Male sex workers self-reported high levels of recent anal discharge, penile discharge, and burning urination—the latter significantly associated with unprotected anal sex. Furthermore, respondents reporting no exposure to HIV counseling and/or testing were significantly more likely to report unprotected sex. To improve access to STI treatment and counseling and testing programs, we recommend increasing outreach to male sex workers through peer education, and training health care providers to provide sensitive and appropriate counseling and care. These proposed interventions may help connect more male sex workers to services, despite high levels of community stigma.

We found that more frequent alcohol use was associated with unprotected sex and recommend further assessment of harmful or hazardous drinking among male sex workers, strengthening linkages to existing substance abuse programs in Mombasa, and establishing peer support groups. It is also possible that low self-esteem and depression may be cofactors influencing both unprotected sex and substance abuse. This study did not systematically assess psychosocial disorders, and is recommended as a future priority for research with MSM in Africa.

Male sex workers in Mombasa are generally mobile and some sell sex throughout the country. A majority of their clients are Kenyan citizens, suggesting a sizable local population sexually attracted to men without foreign influence. The demographic and behavioral differences among self-described sexual identities highlight the importance of understanding identity to inform effective programming. Data from qualitative discussions conducted with male sex workers after this survey, however, suggest that some sexual behaviors and individual motives for engaging in sex work vary within identity subgroups. Further analysis of the relationships between identity, sexual desire, and financial motives among male sex workers in Kenya is needed.

Limitations of this study include, but are not limited to, differing skills of MSM as mobilizers may have led to variation in correct and efficient identification of men who sell sex to men, and the frequencies of individual respondent venue attendance were not accounted for in sampling or analysis, which may bias the results toward male sex workers who attend particular venues on a more frequent basis.25 A research clinic 20 km north of Mombasa was actively providing risk reduction counseling, referral for treatment, and peer education activities for MSM during this research,9 which may have resulted in a more well-informed survey population than that which may exist in other African cities.

Inconsistent or no condom use by male workers cannot be attributed to supply issues, as most reported condoms are very affordable and available at 10 KSH (packet of 3; about US$0.15) or for free. Many respondents, however, reported using oil-based lubricants such as petroleum and baby lotion, and less than a quarter of respondents correctly knew that only water-based lubricants should be used with latex condoms. Water-based lubricants are available mainly at pharmacies and select supermarkets in Mombasa, and are prohibitively expensive for many male sex workers. This study found a significant association between the use of oil-based lubricants and condom breakage, supporting findings of previous studies26–28 and underscoring the need for education and provision of water-based lubricants to MSM as part of an integrated condom distribution strategy in Africa.

In summary, lack of knowledge of anal sex as an HIV risk behavior, frequent alcohol use, self-reported burning urination in the past year, and never having received HIV counseling or testing were significantly associated with unprotected anal sex among male sex workers who sell sex to men in Mombasa District. The development of basic services, including the development of specifically tailored counseling, prevention, and treatment programs for this vulnerable population focusing on anal transmission are strongly recommended.


1.Office of the United States Global AIDS Coordinator. The President's Emergency Plan for AIDS Relief: U.S. Five-Year Global HIV/AIDS Strategy. Washington, DC: Office of the United States Global AIDS Coordinator, 2004.
2.UNAIDS. 2006 Report on the Global HIV/AIDS Epidemic. Geneva: UNAIDS, 2006.
3.Johnson CA. Off the Map: How HIV/AIDS Programming is Failing Same-Sex Practicing People in Africa. New York: International Gay and Lesbian Human Rights Commission, 2007.
4.Wade AS, Kane CT, Diallo PAN, et al. HIV infection and sexually transmitted infections among men who have sex with men in Senegal. AIDS 2005; 19:2133–2140.
5.Niang CI, Tapsoba P, Weiss E, et al. ‘It's raining stones’: Stigma, violence and HIV vulnerability among men who have sex with men in Dakar, Senegal. Cult Health Sex 2003; 5:499–512.
6.Jewkes R, Dunkle K, Nduna M, et al. Factors associated with HIV sero-posivity in young, rural South African men. Int J Epidemiol 2006; 35:1455–1460.
7.Gouws E, White PJ, Stover J, et al. Short term estimates of adult HIV incidence by mode of transmission: Kenya and Thailand as examples. Sex Transm Infect 2006; 82:51–55.
8.Onyango-Ouma W, Birungi H, Geibel S. Understanding the HIV/STI Risks and Prevention Needs of Men Who Have Sex With Men in Nairobi, Kenya. Washington: Population Council, 2005.
9.Sanders EJ, Graham SM, Okuku HS, et al. HIV-1 infection in high risk men who have sex with men in Mombasa, Kenya. AIDS 2007; 21:2513–2520.
10.Geibel S, van der Elst EM, King'ola N, et al. ‘Are you on the market?’: A capture-recapture enumeration of men who sell sex to men in and around Mombasa, Kenya. AIDS 2007; 21:1349–1354.
11.Kibicho W. Tourism and the sex trade: Roles male sex workers play in Malindi, Kenya. Tourism Rev Int 2003; 7:129–141.
12.Roscoe W, Murray SO. Boy-Wives and Female Husbands: Studies in African Homosexualities. New York: Palgrave, 1998.
13.National AIDS Control Council. Kenya National HIV/AIDS Strategic Plan 2005/6-2009/10: A call to Action. Nairobi: Office of the President, 2005.
14.MacKellar DA, Gallagher KM, Finlayson T, et al. Surveillance of HIV risk and prevention behaviors of men who have sex with men—A national application of venue-based, time-space sampling. Public Health Rep 2007; 122(suppl 1):39–47.
15.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 2005; 19:521–526.
16.Centers for Disease Control and Prevention (CDC). HIV prevalence among populations of men who have sex with men—Thailand, 2003 and 2005. MMWR Morb Mortal Wkly Rep 2006; 55:844–848.
17.Behavioral Surveillance Surveys. Guidelines for Repeated Behavioral Surveys in Populations at Risk of HIV. Arlington, VA: Family Health International, 2000.
18.Pinkerton SD, Abramson PR. Effectiveness of condoms in preventing HIV transmission. Soc Sci Med 1997; 44:1303–1312.
19.Davis KR, Weller SC. The effectiveness of condoms in preventing heterosexual HIV transmission. Fam Plann Perspect 1999; 31:272–279.
20.Long JS, Freese J. Regression Models for Categorical Dependent Variables Using Stata. College Station, TX: Stata Press, 2003.
21.StataCorp. Stata Statistical Software Release 9: Survey Data. College Station, TX: StataCorp, 2005.
22.Schwandt M, Morris C, Ferguson A, et al. Anal and dry sex in commercial sex work, and relation to risk for sexually transmitted infections and HIV in Meru, Kenya. Sex Transm Infect 2006; 82:392–396.
23.Fonck K, Kaul R, Kimani J, et al. A randomized, placebo-controlled trial of monthly azithromycin prohpylaxis to prevent sexually transmitted infections and HIV-1 in Kenyan sex workers: Study design and baseline findings. Int J STD AIDS 2000; 11:804–811.
24.Lane T, Pettifor A, Pascoe S, et al. Heterosexual anal intercourse increases risk of HIV infection among young South African men. AIDS 2006; 20:123–125.
25.Xia Q, Tholandi M, Osmond DH, et al. The effect of venue sampling on estimates of HIV prevalence and sexual risk behaviors in men who have sex with men. Sex Transm Dis 2006; 33:545–550.
26.Voeller B, Coulson AH, Bernstein GS, et al. Mineral oil lubricants cause rapid deterioration of latex condoms. Contraception 1989; 39:95–102.
27.Rosen AD, Rosen T. Study of condom integrity after brief exposure to over-the-counter vaginal preparations. South Med J 1999; 92:305–307.
28.Steiner M, Piedrahita C, Glover L, et al. The impact of lubricants on latex condoms during vaginal intercourse. Int J STD AIDS 1994; 5:29–36.
© Copyright 2008 American Sexually Transmitted Diseases Association