Condom Use Among Male Sex Workers With Their Male Clients in Mombasa, Kenya: Results of a Sub-analysis of a Pilot, Multilevel, Structural, HIV Risk-Reduction Intervention

Abstract

Despite the advent of PrEP, consistent condom use continues to be recommended since PrEP does not prevent sexually transmitted infections (STIs). This is important for high risk populations (e.g., male sex workers; MSW) in low resource, high HIV/STI prevalence settings, such as the Mombasa region in Kenya. This study aimed to examine the relationship between MSW’s condom use, and their knowledge, beliefs, and attitudes about condoms. MSW (N=158) completed surveys on their sexual behaviors/practices/attitudes. We used multiple regressions to identify associations between condom use, HIV knowledge/attitudes, and self-efficacy. Three-quarters of participants reported always using condoms in the past week and 64.3% reported always using condoms in the past month with male clients. Mean scores for knowledge and attitudes/self-efficacy toward condoms/safer sex were positively associated with condom use. Interventions to build self-efficacy, such as condom negotiation, and/or bringing up condom use with clients may be useful for Kenyan MSW.

Enthusuiasm over existing oral (tenofovir disoproxil fumarate/emtricitabine; TDF/FTC), injectable (cabotegravir) and emerging delivery modes of HIV pre-exposure prophylaxis (PrEP; e.g. vaginal rings, implants) should not undermine initiatives to promote consistent condom use. Specifically, the U.S. Centers for Disease Control and Prevention (CDC)’s website states, “Since PrEP only protects against HIV, condom use is still important for the protection against other sexually transmitted infections (STIs). Condom use is also important to help prevent HIV if PrEP is not taken as prescribed.” (CDC, 2021a). This is critical, as there was a 30% increase in chlamydia, gonorrhea, and syphilis infections between the years 2015 and 2019 in the United States alone. Further, in 2019, there were 2.5 million reported cases of these aforementioned sexually transmitted infections (STIs) in this geographic context; this represents an all-time high for the sixth consecutive year (CDC, 2021b). This trend extends to men who have sex with men (MSM). Specifically, since TDF/FTC was approved for use as PrEP, condom use among MSM has decreased (Holt et al., 2018), and incidence of chlamydia, gonorrhea (Smith et al., 2015), and syphilis has increased (Traeger et al., 2019). Condoms remain a key factor in preventing HIV transmission (Eakle et al., 2014), and are the only barrier method to prevent sexually transmitted infections (STIs). Thus, research to promote condom use and condom use education remains urgent and important.

Research to facilitate condom use is important for many regions in sub-Saharan Africa (SSA), where prevalence of HIV and STIs remain high, and access to PrEP and STI screening/treatment may be limited (UNAIDS, 2018). Specifically, a 2014 meta-analysis indicated an estimated 11.9% global pooled HIV prevalence across 88 studies, 36.3% of these infections were in SSA (Oldenburg et al., 2014). While data summarizing pooled prevalence of STIs are not available, the limited amount of work on STIs in SSA suggests that the pooled rates for chlamydia (17.9%) and gonorrhea (15.8%) for men who have sex with men (MSM) seeking PrEP across multiple West Africa sites (e.g., Cote d’Ivoire, Mali, Togo, Burkina Faso) are high (De Baetselier et al., 2019).

MSM who exchange money for sex, i.e., male sex workers [MSW], in SSA are at increased risk for HIV and other STIs (Muraguri et al., 2015; Vuylsteke et al., 2012). In addition, MSW may face challenges in accessing biomedical HIV prevention and other sexual health services due to a constellation of factors, including stigma and discrimination in healthcare settings (Scorgie et al., 2013) and the illegality of homosexuality in some sub-Saharan African countries (Itaborahy, 2012; Kibicho, 2003). Thus, it is critical to understand condom use among MSW in SSA, since this may be the most consistently available form of HIV and STI prevention for these men, and one that is not dependent on medical prescriptions.

In Kenya, sex work is a primary driver of the HIV epidemic (DHS, 2003; FHI360, 2022) and MSW in this context may also face higher STI risk (Muraguri et al., 2015). HIV prevalence among MSW is an estimated 33.2% (Oldenburg et al., 2014). Further, these men may act as a bridge to other populations, since 59% of MSW in coastal Kenya reported also having female partners (Smith et al., 2015). Mombasa county was selected as the site for the current study because it is a “hot spot” for male and female sex work in Kenya, and therefore may share commonalities with other settings known for sex work on the African contienent, which would increase the generalizability of our findings. Specifically, existing estimates show that MSM (including prison inmates) in Mombasa account for 20.5% of new HIV infections (Kenya, 2010; UNAIDS, 2009) in the Coast Province; prevalence of new STIs in this setting are unknown. For both Kenyan men and women who engage in sex work, PrEP and non-HIV STI knowledge may still be suboptimal (Embleton et al., 2016; Restar et al., 2017), accessing sexual health services complicated (Restar et al., 2017), and compounded by the illegality of homosexuality and sex work (Itaborahy, 2012; Kibicho, 2003). Consistent condom use is still difficult for many MSW (e.g., when male clients request condomless sex and/or offer more money in exchange for it (George et al., 2019), when some men believe that condoms can be uncomfortable) (Giano et al., 2020). Given these circumstances and the aforementioned CDC recommendations, condoms remain a critically important part of protecting the sexual health of MSW in this and similarly vulnerable settings. Thus, to improve condom use among MSW in Kenya and elsewhere in SSA, we must first understand MSW’s beliefs and perceptions around this behavior.

Extant research shows that generally, knowledge, beliefs, and attitudes about condoms impact condom use (D’Anna et al., 2012; Klein, 2011). In SSA, studies have shown that pro-condom attitudes, supportive social norms, communication about condoms, self-efficacy, perceived threat of HIV, practical knowledge about condom use, and knowledge of HIV prevention mechanisms influence safer sex practices in diverse populations (Eggers et al., 2014; Kalolo & Kibusi, 2015; Matovu & Ssebadduka, 2013). Understanding the ingredients that help to construct thought processes about condom use is critical to designing condom use interventions (Eggers et al., 2014). However, to our knowledge, few studies look at this in Kenyan MSW.

The present study fills this unique gap in the literature on the association between MSWs’ condom use behaviors with male clients, and their knowledge and perceptions of condoms. Specifically, because of the illegality of same-sex sexual contact and sex work, teasing out the beliefs and behaviors about condoms within the context of same-sex transactional sex encounters is extremely difficult. Using a cross-sectional survey among MSW in Mombasa county, Kenya, we examine MSW’s ideas around condom use and the relationship between condom use behavior with male clients and knowledge, beliefs, and attitudes about condoms or condom use with clients.

Methods

Study Design, Setting, and Participants

The present analysis is drawn from baseline interviews with N=158 MSW. These men participated in a baseline survey (December 2014-May 2015) prior to implementation of Project Boresha (Improve), a pilot multi-level structural HIV risk-reduction intervention for male and female sex workers and male clients delivered in bars and clubs in Mombasa, Kenya. Specifically, Project Borasha provided on-site peer education on condom use and HIV risk reduction, as well as condoms, lubricants, and sexual health services (e.g., HIV and STI testing, counseling, and care), delivered in sex work settings. Three bars/clubs in the greater Mombasa area were assigned to the aforementioned intervention condition and three to control condition (e.g., no HIV risk reduction services were delivered). The intervention was delivered over a 7-month period between October 2016 and April 2017. Other results from Project Boresha have been previously published (Restar et al., 2017; Valente et al., 2022; Valente et al., 2019).

Participants were recruited from six bars/clubs in Mombasa identified as venues where male and female sex workers engaged clients. These entertainment venues were also patronized by people who were not buying or selling sex. MSW work independently in the aforementioned establishments and are paid by clients rather than as employees of the venue. Bars/clubs were pair-matched, based on the following characteristics: size, demographic profile of patrons, and the proportion of sex workers who patronize the establishment. Then, members of the pairs were assigned to intervention or control conditions (3 intervention, 3 control) for a seven-month period (October 2016-April 2017).

Eligibility criteria were ascertained via self-report. MSW enrolled in this study had to report: coming to one of the six bars/clubs at least four times in the past month, anal intercourse at least once with a male client met in one of the venues in the last three months, being at least 18 years old, and being of Kenyan nationality. All participants had to be visibly sober at the time of interview to ensure they had the cognitive capacity to participate in an interview.

Ethical Considerations

The study protocol and procedures were approved by the New York State Psychiatric Institute-Columbia University Department of Psychiatry Institutional Review Board (PSF 6862), the Kenyatta Hospital-University of Nairobi Research Ethics Committee (P807/12/2015), and the University of Ghent Ethics Committee (B670201628199). Written informed consent was obtained prior to commencement of the interview.

Measures

All measures used in this study were translated from English to Swahili, and then back translated to English by a member of the research team fluent in both languages. Our outcome measures included two single items: “Always used condoms with male clients in the past week” and “Always used condoms with male clients in the past month;” answers were dichotomized as yes/no. Table 1 describes the potential correlates of condom use used in this analysis: demographics (i.e., age, age at first commercial sex encounter, number of male clients in the past month, education (last grade completed), and whether or not they are HIV-positive (yes/no)). Table 2 describes: (1) an adapted version of the 6-item HIV Prevention Knowledge measure (e.g., “It is safe to use a condom more than once”), with true/false responses (KNBS, 2010); and (2) an adapted version of the 10-item Safer Sex and Condom Use Self-Efficacy measure assessing participation in contexts, settings, and situations that may make it difficult to practice safer sex (e.g., sex in public environments, sex under the influence of drugs or alcohol, sex to escape negative emotions, sex within interpersonal relations of unequal power, situational experiences of sexual dysfunction, and sex with partners who resist condom use). Items were assessed along four- or five-point Likert scales, with responses along these points varying by item statement, and a higher score reflecting greater self-efficacy (Brien et al., 1994; Diaz et al., 2004).

Table 1.

Demographic Characteristics and Sexual Behaviors among Male Sex Workers (N=158)

Demographic Characteristics
Variable	Mean [SD]
Age	26.2 [5.8]
Age at which participant had first commercial sex encounter	18.0 [4.6]
	N (%)
Education  None, or only literacy  Primary incomplete  Primary complete, no secondary  Secondary incomplete  Secondary complete  Technical college complete  University complete	7 (4.4%) 30 (19.0%) 36 (22.8%) 25 (15.8%) 43 (27.2%) 11 (7.0%) 6 (3.8%)
Participant is living with HIV (yes)	21 (13.3%)
Participant has had an STI in the past year (yes)	41 (26.1%)
Sexual Behaviors
	Mean [SD]
Number of male sex work clients in the past month	6.0 [10.4]
Number of commercial sex acts in the past 7 days	3.9 [3.8]
Number of commercial sex acts in the past month	16.5 [15.1]
Number of men with whom the participant had anal sex in the last month (paying and non-paying partners)	11.7 [10.3]
Participant’s perception of HIV risk  No risk at all for HIV  Risk of HIV is small  Risk of HIV is moderate  Risk of HIV is high	9 (6.8%) 48 (36.4%) 23 (17.4%) 48 (36.4%)
	N (%)
a Always used condoms with male (clients) in the past month (yes)	101 (64.3%)
a Always used condoms with male (clients) in the past week (yes)	118 (75.2%)
Number of weeks per month participant visits bars or nightclubs for sex work  About 4 weeks per month  About 3 weeks per month  About 2 weeks per month  About 1 week per month	137 (86.7%) 12 (7.6%) 8 (5.1%) 1 (0.63%)
Sex partners and types of sexual practices  Participant had a female partner    Participant had vaginal sex with female partner    Participant had anal sex with female partner    Participant had both vaginal and anal sex with female partner   Participant had a male partner    Participant had receptive anal sex with male partner    Participant had insertive anal sex with a male partner    Participant had both receptive and insertive anal sex with a male partner	27 (17.1%) 11 (7.0%) 0 (0.0%) 1 (0.6%) 157 (99.4%) 90 (57.3%) 23 (14.7%) 44 (28.0%)
Reasons for not using condoms with a paying partner (n=56 reported condomless sex with a paying partner)  Client refused  Participant trusts the client  Participant was drunk or using drugs  Client offered to pay more for condomless sex  No condoms were available  Client was drunk or using drugs  Participant prefers sex without a condom  Participants knows that the client is not living with HIV  Participant feels that condoms decrease sexual pleasure  Other	21 (37.5%) 18 (32.1%) 14 (25.0%) 13 (23.2%) 11 (19.6%) 5 (8.9%) 5 (8.9%) 2 (3.6%) 1 (1.8%) 3 (5.4%)
Participant reports that it is easy to obtain male condoms	154 (97.5%)
Participant reports that it is easy to obtain lubricant	119 (81.0%)

^aDesignates dependent variables

Table 2.

HIV prevention knowledge and attitudes and aelf-efficacy toward condoms and safer sex

HIV Prevention Knowledge
Variable	N (%)
It is safe to use a condom more than once (% correct answer)	131 (82.9%)
It is safe to have sex without a condom if it is with your regular partner (% correct answer)	136 (86.1%)
Pulling out before the male ejaculates prevents transmission (% correct answer)	119 (75.3%)
As long as both partners wash themselves after sex, it is not necessary to use a condom (% correct answer)	142 (90.5%)
If someone has HIV, you can see that straightaway (% correct answer)	139 (88.0%)
What do you think has the highest risk of HIV transmission?  Vaginal intercourse (% incorrect answer)  Anal intercourse (% correct answer)  Both the same (% incorrect answer)  Don’t know (% don’t know)	30 (19.0%) 90 (57.0%) 33 (20.9%) 5 (3.2%)
MEAN SCORE [max=1; min=0] = mean [SD] – Adjusted for items included after PCA [scale reliability coefficient = 0.70]	0.80 [0.20]
Attitudes and Self-Efficacy towards Condoms and Safer Sex	Mean [SD]
Condoms go against my values or religious beliefs [1=strongly disagree; 5=strongly agree]	2.4 [1.3]
Many of my peers would not use condoms with clients [1=strongly disagree; 5=strongly agree]	3.2 [1.4]
Most people who are important to me think I should use condoms when I have sex with clients [1=strongly disagree; 5=strongly agree]	4.4 [0.9]
How pleasurable would it be for me to always use condoms during receptive anal intercourse [1=very unpleasurable; 4=very pleasurable]	2.8 [0.9]
How likely is it that you will always use a condom when you have sex with a client [1=very unlikely; 4=very likely]	3.5 [0.7]
How likely is it that you would not use a condom when a client pays more/How likely is it that you would pay more to have sex with a male sex worker without a condom [1=very unlikely; 4=very likely]	2.0 [1.1]
How likely is it that you will talk with clients about condoms [1=very unlikely; 4=very likely]	3.5 [0.6]
How likely is it that you will always us a condom when you have sex with a regular client [1=very unlikely; 4=very likely]	3.3 [0.8]
How difficult would it be to talk about condom use with clients [1=very difficult to do; 4=very easy to do]	3.2 [0.8]
How difficult would it be to protect yourself from getting HIV when you have sex [1=very difficult to do; 4=very easy to do]	3.2 [0.8]
MEAN TOTAL – Adjusted for items included after PCA [scale reliability coefficient = 0.80]	3.1 [0.4]

Statistical Analyses

Power analyses showed that a sample size of N=160 MSW has sufficient power and will exceed the minimum required prevalence threshold to detect differences in our outcome variables across multiple time series surveys. We include N=158 in our present analysis. This would still meet this threshold, particularly because our analysis uses cross-sectional data from baseline only (rather than multiple surveys in a timed series). Descriptive statistics were generated for demographic variables (means and standard deviations for continuous variables and frequencies for categorical variables). Separate Principal Component Analyses (PCA) were conducted on both the HIV Prevention Knowledge and Safer Sex and Condom Use Self-Efficacy measures to identify items that should be retained in each scale for analyses. The items presented in each of these two scales in Table 3 summarize items that were grouped together following the PCA after rotated factor loadings formed our new potential scales. Cronbach’s alphas were calculated to assess subscale reliability for each of the two potential scales. If the scale had an alpha ≥ 0.7, it was retained and scale means were calculated.

Table 3:

Results of simple and multiple logistic regressions for consistent condom use with clients in the past week” and in the past month

Demographic Characteristics
	Always used condoms with clients in the past week		Always used condoms with clients in the past month
	Simple Reg. OR [95%CI]	Multiple Reg. aOR [95%CI]	Simple Reg. OR [95%CI]	Multiple Reg aOR [95%CI]
Age	1.01 [0.94, 1.07]	1.03 [0.95, 1.11]	1.01 [0.96, 1.07]	1.03 [0.95, 1.11]
Age at first commercial sex	0.94 [0.86, 1.02]	0.96 [0.88, 1.05]	0.92 [0.85, 0.99]*	0.91 [0.84, 1.00]
Number of male sex work clients in the past month	1.03 [1,00, 1.06]	1.03 [0.99, 1.07]	1.03 [0.99, 1.06]	1.03 [1.00, 1.07]
Education  None, or only literacy  Primary incomplete  Primary complete, no secondary  Secondary incomplete  Secondary complete  Technical college complete  University complete	[REF] 0.95 [0.15, 5.94] 1.1 [0.18, 6.57] 0.48 [0.07, 3.37] 0.76 [0.13, 4.52] 0.94 [0.11, 7.73] 0.50 [0.03, 7.45]	[REF] 1.77 [0.16, 19.30] 1.59 [0.15, 16.95] 0.51 [0.04, 6.64] 1.03 [0.10, 10.88] 1.50 [0.11, 21.30] 0.95 [0.04, 21.12]	[REF] 1.08 [0.20, 5.73] 0.75 [0.15, 3.90] 0.63 [0.11, 3.49] 0.52 [0.10, 2.66] 0.76 [0.11, 5.28] 1.33 [0.15, 11.93]	[REF] 2.01 [0.25, 15.86] 0.60 [0.08, 4.80] 0.59 [0.07, 5.19] 0.41 [0.05, 3.22] 0.91 [0.09, 9.49] 1.65 [0.12, 22.69]
HIV Prevention Knowledge
Significant scale items identified through PCA	OR [95%CI]	OR [95%CI]	OR [95%CI]	OR [95%CI]
Pulling out before the male ejaculates prevents transmission	1.40 [1.03, 1.89]*	---	1.56 [1.04, 2.34]*	---
As long as both partners wash themselves after sex, it is not necessary to use a condom	1.18 [0.76, 1.83]	---	1.48 [0.83, 2.65]	---
Mean score for questions identified through PCA	1.45 [0.96, 2.18]	---	1.93 [1.02, 3.65]*	2.62 [1.06, 6.44]*
Attitudes and Self-Efficacy toward Condoms and Safer Sex
Significant scale items identified through PCA	OR [95%CI]	OR [95%CI]	OR [95%CI]	OR [95%CI]
How likely is it that you will always use a condom when you have sex with a client	0.35 [0.20, 0.62]*	---	0.24 [0.13, 0.45]*	---
How likely is it that you will talk with clients about condoms	0.49 [0.25, 0.97]*	---	0.31 [0.16, 0.61]*	---
How likely is it that you will always use a condom when you have sex with a regular client	0.36 [0.22, 0.60]*	---	0.27 [0.16, 0.48]*	---
How difficult would it be to talk about condom use with clients	0.51 [0.33, 0.81]*	---	0.46 [0.30, 0.72]*	---
How difficult would it be to protect yourself from getting HIV when you have sex	0.59 [0.38, 0.91]*	---	0.54 [0.36, 0.83]*	---
Mean score for questions identified through PCA	0.23 [0.11, 0.48]*	0.21 [0.09, 0.48]*	0.15 [0.07, 0.33]*	0.15 [0.06, 0.34]*

^*Designates statistically significant variables

Simple and multiple logistic regressions were performed to identify potential associations with each of the two condom outcomes. For individual items in the two scales assessing HIV Prevention Knowledge, and Attitudes and Self-Efficacy toward Condoms and Safer Sex, only simple regressions were performed (since responses are reflected in the mean score calculation, and inclusion of individual items in multiple regressions would be duplicative). Multiple logistic regressions were computed separately for each outcome variable. Multiple models included variables that were significant in the simple regressions, and were controlled for demographic factors (e.g., age, age at first commercial sex, number of male sex work clients in the past month, education). Since missing data for both models were <3%, listwise deletion was used, meaning that cases missing any data were dropped.

Results

Table 1 summarizes the demographic characteristics and sex behaviors among MSW study participants (N=158). Participants had a mean age of 26.2 years, were an average 18 years old when they had their first commercial sex encounter and reported an average of six male clients in the past month. Over a quarter of participants (41;26.1%)) reported having a sexually transmitted infection (STI) in the past year, and 21 (13%) were living with HIV. Three-quarters of participants reported always using condoms in the past week (118;75.2%)) and 101 (64.3%) reported always using condoms in the past month with male clients.

Table 2 presents descriptive statistics on the HIV prevention knowledge and attitudes and self-efficacy towards condoms and safer sex among the MSW in this study. Specifically, participants had a mean score of 0.8 [min=0; max=1] on the HIV Prevention Knowledge measure; for all questions, the proportion of correct responses ranged between 90 (57.0%) and 142 (90.5%). For the Attitudes and Self-Efficacy toward Condoms and Safer Sex Scale, participants had a mean score of 3.1 [min=1; max=4].

Table 3 summarizes the results of the simple and multiple regression analyses. Apart from “age at first commercial sex,” no demographic characteristics included in any of the models were associated with either outcome in simple or multiple regressions. The mean score for the Attitudes and Self-Efficacy toward Condoms and Safer Sex Scale was associated with always using condoms with clients in the past week [aOR=0.21; 95% CI=0.09–0.48], and past month [aOR=0.15; 95% CI=0.06–0.34]. All five items identified in the PCA (the analysis used to identify items that should be retained in each of the two scales for analyses) were significantly associated with both outcomes in the simple regression analyses. These included how likely participants would: always use a condom when they have sex with a client; talk to clients about condoms; always use a condom when they have sex with a regular client; and, how difficult it would be for participants to talk about condom use with clients, and protect themselves from getting HIV when they have sex (see Table 3 for ORs).

On the other hand, the mean score for the HIV Prevention Knowledge measure was only associated with always using condoms with clients in the past month [aOR=2.62; 95% CI=1.06–6.44]. Of the two items identified in the PCA, only one (Pulling out before the male ejaculates prevents transmission) was significantly associated with the two outcome variables in the simple analyses (see Table 3 for ORs). Table 2 shows that 119 (75.3%) of participants answered this item correctly.

Discussion

This study sheds light on the psychological activities that lead to condom use in MSW in Mombasa, Kenya – an area with high HIV prevalence among MSW. Specifically, we identified relationships between MSW’s knowledge, beliefs, attitudes, and self-efficacy about condoms or condom use and their self-reported condom use behavior with male clients (past week/past month). Few prior studies have aimed to understand the psychological activities that promote consistent condom use with male clients in male sex work environments, particularly those environments located on the African continent. However, some literature has identified the specific types of interventions that may facilitate condom use with sex (e.g., peer outreach, normalizing same-sex experiences) (Geibel et al., 2012; Okal et al., 2009). Thus, the subsequent discussion outlines our findings’ utility to enhance existing, evidence-based condom promotion strategies, education initiatives, or research studies for MSW in sex work settings on the African continent.

Implications for Future Research and Interventions

In our sample, condom use with male clients was inconsistent and self-reported STI and HIV were high. These results, coupled with CDC recommendations to use condoms in tandem with PrEP (or alone, if PrEP is unavailable), underscore the importance of intervention development and implementation to promote condom use, especially among MSW who work in bars and clubs. Our findings suggest that interventions to build self-efficacy skills, such as condom negotiation, and/or bringing up condom use with clients may be particularly relevant for this population. Specifically, items identified in our PCA analysis (and subsequently included in the aggregated scores applied in regressions) were those that pertained to participants’self-efficacy around condom use behaviors. These questions relate to participants’ self-controlled behavioral intentions and perceived difficulty related to intentions to perform self-controlled HIV prevention tasks. Thus, it could be beneficial to target these areas in future interventions. Further, we found some evidence that initiating sex work at a younger age may result in less frequent condom use, indicating the need to target younger MSW for involvement in condom use interventions. (Chemnasiri et al., 2010).

However, constructing future projects to facilitate condom use may prove difficult since research in MSW populations on this topic remains understudied in all global settings, despite the salience of this issue (Qiao et al., 2019). We were unable to find MSW-focused condom promotion programs implemented on the African continent, even though some programs focused on MSM in this context included MSW. Of the recent scientific work that does exist, including two intervention studies in the United States and one in the United Kingdom to reduce MSW’s exposure to HIV through anal sex, all had little impact on condom use behavior (Ballester-Arnal et al., 2014). Some of these studies offered recommendations for future interventions. These include greater engagement with MSW, keeping interactions with study participants as brief as possible (e.g., to encourage participation), and implementation of interventions in sites that MSW typically frequent to facilitate follow-up (Ballester-Arnal et al., 2014).

Implications for Educational Initiatives

Participants could benefit from gains in overall HIV prevention education. Specifically, the proportion of correct responses for individual knowledge items ranged from 57.0% to 90.5%. In simple regressions, the only item positively associated with the two outcome variables was correctly answering the True/False statement: “pulling out before the male ejaculates prevents transmission.” This indicates that knowledge about multiple basic HIV prevention strategies is variable in this extremely high-risk population. Building HIV prevention knowledge is a critical first step that must be incorporated into any condom use interventions.

However, delivering HIV prevention information is fraught in our particular context. Specifically, given the illegality of sex work and stigmatization of same-sex sexual contact in multiple settings on the African continent, some MSW (as well as other MSM) may delay or forego HIV/STI prevention and treatment behaviors (Fay et al., 2011; Kennedy et al., 2013), complicating efforts to deliver this crucial education. Thus, efforts to educate MSW about condom use, HIV/STI prevention, and other safer sex behaviors likely require structural and practical changes. For example, these changes could include: training a lay (including MSW peers) health workforce to deliver these messages; promoting community sensitization initiatives to protect MSW and other MSM from stigmatization, discrimination, and other forms of unfair treatment; and preparing Africa-based healthcare providers to respond to the needs of sexual minority communities in an affirming way. (van der Elst et al., 2015). Training a lay workforce of MSW peer educators to deliver this information may be particularly effective, since existing research with female sex workers (FSW) shows that FSW engagement and empowerment is an especially effective prevention and education tool (Beyrer et al., 2015). In addition to this, offering safer sex education to the male clients of MSW could also prove beneficial in this context.

Strengths and Limitations

This study provides some insight on how Kenyan MSW in the greater Mombasa area may make decisions about condom use with clients. This information can be incorporated into future condom use interventions targeting these men, since condoms are necessary to prevent the increasingly prevalent non-HIV STIs, and HIV in situations when oral pills and injectable PrEP are not preferred or available. Understanding this topic may be of particular importance within the context of COVID-19, since these men may now face increased economic insecurity, and disruptions in access to health services. Thus, it is important to continue to further our understanding of condom use behavior in this population, since condoms will remain recommended until biomedical forms of STI prevention are developed and made available to Kenyan MSW.

Some study limitations need to be noted. Data presented in these analyses are cross-sectional and do not capture changes in condom use knowledge and self-efficacy over time. Therefore, we are unable to document the effects that the evolving contexts and settings in which MSW work have on factors that influence condom use; this includes the ongoing COVID-19 pandemic. Further, we were unable to estimate the potential bias generated by the possible cluster effect of participants recruited from the same venues sampled in this research. For this reason, confidence intervals of the aORs may be underestimated. Additionally, our power analysis was designed to estimate the sample needed to capture changes in participant data over time, rather than cross-sectionally, which was the design of this analysis. Therefore, we may have somewhat overestimated the number of participants needed to detect differences in our outcome variables in the present paper. In addition, we can only assume that study participants are sufficiently representative of similar types of bars patronized by MSW in Mombasa, and it is unknown whether other types of MSW (e.g., home-, street- or internet-based MSW) have similar views on condoms and HIV prevention strategies. Lastly, this was, to our knowledge, the first time our measures were used with Kenyan MSW, and thus their precise applicability to our target population is unknown. However, given the scale reliability coefficients presented in Table 2 for the knowledge (0.70) and acceptability (0.80) scales, it is likely that our iterations of the scales were appropriate. Despite these limitations, given the dearth of information on this topic, we believe that this analysis makes an important contribution to the literature on HIV prevention in MSW.

Conclusions

Choices in HIV prevention methods are important since PrEP discontinuation in high-resource settings, such as the US, is high (Serota et al., 2020; Spinelli & Buchbinder, 2020). Additionally, no forms of PrEP protect against STIs. Furthermore, given that PrEP is not equally available to all people in all geographic settings, condoms remain an important part of HIV and STI prevention efforts. These inequities will continue to persist as new drugs and delivery modalities (e.g., injectable drugs, such as cabotegravir) continue to roll onto the market. This is especially true in high poverty settings, where male sex work is prevalent, and same-sex sexual contact is illegal or highly stigmatized. Thus, it is important to continue to develop condom-focused interventions in these settings, such as Kenya, to ensure that MSW have a viable means to protect themselves from HIV and other STIs. Based on our findings, future research and interventions could focus on building self-efficacy around condom use – particularly around self-controlled condom use behaviors, and perceived difficulty related to performing self-controlled HIV prevention tasks. Future education efforts could focus on building HIV/STI prevention knowledge among MSW – which will require structural changes such as training MSW as peer health workers, community destigmatization efforts, and training healthcare workers to address MSWs’ (and other MSMs’) needs in an affirming way.

Key Considerations.

• Future interventions targeting male sex workers (MSW) in Mombasa, Kenya (and other sites on the African continent) should focus on self-efficacy around condom use.

• These interventions should also focus on building on condom use knowledge.

• Training a workforce of lay peer health workers (e.g., other MSW) could help to contribute to these efforts.

• Efforts to promote condom use in this population are particularly important; STI rates continue to rise globally, and condoms remain the only HIV prevention method for many vulnerable populations.