Research on HIV infection and prevalence among men who had sex with prostitutes

In this study, we systematically analyzed 87 population-based surveys conducted over 2 decades in SSA. From 2000 to 2020, 8% of sexually active men in SSA reported ever paying for sex, and this proportion was higher in urban areas. Men who paid for sex were 50% more likely to be living with HIV compared to men who did not pay for sex, and only 68% of men in the last decade reported using a condom during their last paid sex. Men who paid for sex had a slightly higher probability of having ever tested for HIV, but ARV use and VLS among men who paid for sex were similar to those among men who did not pay for sex, although the evidence for ARV use and VLS were limited to fewer countries.

We found important regional variations in reports of paid sex. A higher proportion of men reported ever paying for sex in Central and Eastern Africa compared to Western and Southern Africa, which is consistent with a 2006 systematic review [16]. The overall proportion of sexually active men who paid for sex in the past 12 months was 2.7%, which is lower than an estimate of 4.3% from a review of 2010–2016 DHS surveys [30]. Differences can be explained by the present study including more surveys, representing more participants and 8 more countries. Also, our main analyses use lifetime measures of paid sex, which may be less prone to underreporting and social desirability bias than measures of recent paid sex. However, our population size estimates for clients of sex workers are probably lower bounds because of potential non-disclosure of paid sex. For example, the proportion of adult women engaged in sex work is estimated to range from 0.4% to 4.3% [31]. Given these numbers, it is unlikely that our estimate of 8% of men paying for sex would be sufficient to sustain this number of women engaging in sex work, although a study from Rwanda reported similar numbers of female sex workers and clients [32]. All surveys included here used face-to-face interviewer-administered questionnaires. Responses could be affected by social desirability bias, and this bias could be greater than for alternative confidential survey methods such as polling booth surveys [22].

Consistent with regional variations in population HIV prevalence, HIV prevalence among men who paid for sex was highest in Southern Africa and lowest in Central and Western Africa [1]. Men who paid for sex were more likely to be living with HIV than men who did not pay for sex, which is consistent with a 2008 analysis [33]. A similar finding was also found for female sex workers, whose odds of living with HIV were 12 times higher than that of all women aged 15–49 years [5]. HIV prevalence among men who paid for sex was lower in surveys conducted from 2010 onwards compared to surveys conducted before 2010, but uncertainty was large, and we cannot rule out that prevalence among this group remained stable. Although HIV prevalence was lowest in Western Africa, this region had the highest HIV PR comparing men who pay for sex with those who do not. In these settings, adding interventions that focus on the unmet prevention needs of men who pay for sex may be more cost-effective than those focused on the general population [34].

Men who paid for sex were more likely to have ever tested for HIV across regions and time. The PR for lifetime HIV testing between men who paid for sex and men who did not was highest in Western Africa and lowest in Central Africa. Higher risk perception encouraging testing, or greater availability of testing in areas with higher HIV burden, may explain this result [35,36]. Men living with HIV who paid for sex were equally likely to have been tested as those who did not paid for sex, which could have implications for knowledge of HIV status in this group. A recent study found that diagnosis coverage and time to diagnosis in SSA have drastically improved over the last decade, but in 2020 the largest group of individuals unaware of their status was men [37]. Distribution of HIV self-tests to sex workers, who can then distribute the tests to peers, clients, and partners, may further improve knowledge of status among men who pay for sex [38]. This approach is preferable to interventions such as index testing that require sex workers to disclose the identity of their male clients, which could put the sex workers at risk of violence, loss of sex work income, or both [6].

As men who pay for sex often also have female partners not involved in sex work, they may disproportionally contribute to population-level HIV transmission if virally unsuppressed [9,15,39]. Our results suggest comparable ARV use and VLS levels among men who paid for sex and men who did not. However, these estimates are based on a small number of surveys from 2012–2017, highlighting important data gaps. For SSA as a whole, the 2020 estimates of ARV use and VLS remain below UNAIDS targets, and men may be less likely to initiate and adhere to ARV treatment than women [1,40,41]. Treatment access can be facilitated by services targeted to men who are more likely to frequent sex workers, such as migrant laborers, long-distance truck drivers, mine workers, and other men who travel for work [42]. Improving access to treatment for men who pay for sex is especially important as, from 2010 onwards, only 68% of men used a condom the last time they paid for sex. A recent analysis of 29 DHS surveys from 2010–2019 found that, among men who reported condom use at their last paid sex, 84% reported consistent condom use during paid sex [43], which is higher than our estimates. Since the survey instruments only asked about consistent condom use if men reported condom use at last paid sex, we would expect consistent condom use measures to be higher. Altogether, these results suggest that, when men pay for sex and use condoms, they tend to do so consistently. Nevertheless, clients of sex workers often have decisive power over condom use during paid sex, and global evidence suggests higher HIV prevalence among clients of sex workers who do not use condoms [9,44,45]. For these reasons, continued condom use promotion in this population is strongly warranted.

Our results should be interpreted considering several limitations. First, population-based surveys of sexual behaviors depend on self-reports, so estimates could be affected by recall and social desirability biases [46–48]. Use of a lifetime measure of paid sex may have alleviated underreporting, but measures could still be underestimated, and, in the case of measures of association, this underestimation could attenuate the effect sizes towards the null. Confidential measures like polling booth surveys or audio computer-assisted self-interviews may improve accuracy [22,49,50]. Second, survey instruments captured men who have “paid” or, in a few instances, “given money, gifts, or favors in exchange” for sex. As money can be exchanged for sex outside of sex work, we cannot be certain that all men in our population are clients of sex workers. There are many sex work typologies, and transactional sex that involves exchanging gifts or favors may not have been reported as paid sex. For instance, relationships between male “sugar daddies” and younger women may be an important type of transactional sex that is probably not entirely captured in our surveys [51]. This could partly explain the smaller population size estimates for the Southern Africa region. For example, 18% of men in 2 South Africa provinces reported “ever having sex with a woman in prostitution,” but 66% reported having had some type of transactional sex [52]. Third, most surveys do not specify the paid partner’s gender when asking about paid sex, and we cannot be certain that all men in our analytical sample have paid for sex with a woman. However, the proportion of men who have sex with men in this region is estimated to be small [53]. Fourth, the surveys included had slightly different questionnaires and sampling strategies. Nevertheless, questions were largely similar, and using these multiple data sources allowed us to integrate information from more countries and respondents. Finally, few surveys had information on ARV treatment and VLS, so these estimates may not be generalizable to all regions.

Strengths of this study include our exhaustive analysis of all available population-based surveys with information on men who ever paid for sex in SSA, without restriction to any survey type. We synthesized new information on the epidemiology of HIV and the HIV prevention and treatment cascades among men who pay for sex. Our large sample size allowed investigation of patterns by regions and over time, and we estimated adjusted PRs using standardization to control for the effects of age and urban or rural area of residence.