Introduction

The efficacy of dual drug combination (tenofovir and emtricitabine) oral pre-exposure prophylaxis (PrEP) for preventing acquisition of HIV has been extensively investigated and reported [1]. PrEP has been incorporated into combined antiretroviral guidelines for prevention and treatment by the World Health Organization (WHO); is featured in HIV prevention guidelines in several countries and regions; and has been approved by regulatory bodies in at least 15 countries [2]. In most countries currently implementing or piloting PrEP, the delivery is focused towards those at highest risk of acquiring HIV, following recommendations from mathematical modelling [3,4] as well as WHO guidelines.

The focus on those at highest risk stems from the relatively high cost of PrEP which would be mitigated by reducing new infections where they are highest, and from the need of these populations for new options to prevent HIV infection. Women in general are vulnerable, and in sub-Saharan Africa constitute 56% of all people living with HIV [5]. Female sex workers (FSWs) are a subset of this population, with some of the highest HIV prevalence and incidence rates [5]. While FSWs were not included in the oral PrEP efficacy studies as a specific focus population among high risk women [6,7], they are now the focus of many implementation studies in various phases of research across the globe[8]. In South Africa, FSWs make up an estimated 91% of the sex worker population [9], and experience HIV prevalence rates as high as 72% (found in the greater Johannesburg area) [10].

PrEP represents a new opportunity for protection among FSWs who often experience difficulties in ensuring condom use. Key factors affecting their vulnerability include ability to negotiate condom use, violence, poverty, and substance use [11–16]. These factors, in addition to the distinctive structural issue of criminalization of sex work in most countries, make FSWs both uniquely suited for PrEP but also potentially harder to reach with interventions. In this regard, ‘real-world’ studies for FSWs thus far have centred on using an implementation science approach to determine how to integrate PrEP into existing service structures. Beyond the practicalities of implementation, however, is the need to better understand FSWs’ individual perspectives of PrEP.

In these “early days” of PrEP introduction, end-user perspectives of actual use in ‘real-world’ settings are critical to the success of PrEP, and will ultimately determine the fate of future programming. The importance of qualitative research “to answer questions about experience, meaning and perspective”[17] from the position of PrEP users to ensure the relevance in intervention design and successful implementation of PrEP cannot be overstated. While quantitative data will help to characterize populations at highest risk, identify priority geographical settings in terms of HIV incidence for PrEP delivery [18], and evaluate the cascade of care from uptake to retention [19], qualitative data will explain why gaps in successful programming exist and point to how to close those gaps [20].

This paper presents findings from serial in-depth interviews (IDIs) conducted with FSW participants during the course of the Treatment And Prevention for Sex workers (TAPS) Demonstration Project in South Africa. This sample represents one of the first set of FSWs to take up PrEP and report on individual perspectives of actual use. The line of inquiry was situated within a holistic approach based on a social-ecological framework, denoting reflections on the real and practical, rather than theoretical or trial-based use [21]. This research provided the opportunity to examine motivations for and barriers to uptake, as well as what the women most valued about their use of PrEP. Such data will help to inform PrEP implementation scale-up, and in particular demand generation efforts, outreach to potential PrEP users, provider training around PrEP management and adherence support and design of differentiated PrEP services ideally integrated into existing platforms.

The overall aim of the interviews was to elicit feedback on the uptake and use of both an early antiretroviral treatment (ART) for HIV-positive and PrEP intervention for HIV-negative FSWs integrated into an existing, comprehensive health programme tailored for sex workers. This paper will specifically explore the lived experiences and perceptions of taking up and using PrEP among FSWs engaged in the TAPS Demonstration Project.

Methods

This qualitative research was conducted as a component of the TAPS Demonstration Project and included as part of the main study protocol [22]. TAPS was an implementation study designed to assess whether it was feasible, acceptable, safe and cost effective to roll out oral PrEP and early ART as part of an HIV intervention to FSWs in two urban clinics (Johannesburg and Pretoria) in South Africa [22]. TAPS was launched in March 2015 and completed in July 2017. In general, women had to be currently engaged in sex work, not pregnant or having a major health complication (such as multi-drug resistant tuberculosis) at the time of study enrolment, and had to be age 18 or over to be included in TAPS. The study protocol has been published as well as the primary study findings, where further details on participant inclusion and exclusion criteria as well as demographic characteristics for the entire study population can be found [19,22].

A selection of women enrolled in the study was invited to participate in the additional IDI component during their month 1 clinic visit. TAPS participants were selected using a random sampling algorithm set to select a sample of 15% of those enrolled with the aim of enrolling a sample of up to 10%. This was according to the original design which sought a total enrolled cohort of 400 women across two sites. Women who accepted were required to complete an additional informed consent process for participation in the IDIs at the month 1 visit, and were then scheduled for interviews to be held on or around their 3, 6, and 9 month clinic visits. Completed informed consent forms were kept in separate files at the clinic sites in fire-proof, limited access locked cabinets. After each completed IDI, participants were given R50 (~4 USD) as travel money, however there was no other monetary incentive given in connection with TAPS participation.

Initial interview guides were developed based on findings from a systematic review of qualitative evidence on women’s perspectives on use of female-initiated prevention products in sub-Saharan Africa [23,24], as well as results from formative research conducted in preparation for TAPS [25], including focus group discussions with potential users [26]. The premise and structure of the interview guides were based on the Modified Social-Ecological Model (MSEM) designed by Baral et al [27] to facilitate a holistic and broad inquiry. In practice, we took the levels illustrated in the MSEM, and built questions expanding from the individual to the broadest HIV epidemic level. The first interview guide was then adapted slightly for the second round of interviews, based on findings from the first round, as was the guide for the third round. In this way, themes were linked to follow stories of product use over time, but also adapted to changes observed to target relevant topics. In addition to questions about PrEP beliefs and experience, numerous questions were asked relating to sexual partners, both clients and regular or romantic. The gender of such partners was not explicitly asked as a direct question to participants, although no assumptions regarding gender identity were made during questioning.

The overall aim of the serial interview (often called Longitudinal Qualitative Research or LQR [28]) approach was twofold: 1) to see if and/or how women’s realities changed over time with use of PrEP, and 2) whether they may become more open with the interviewer over time and provide richer narratives of their lives than those presented in the first interview, hopefully reducing social desirability bias in responses [29]. While this method has been prevalent in cancer-related research [28], it has more recently been employed in qualitative components of PrEP-related trial research [30]. It should be noted, however, that the change over time is not the primary focus of this particular paper as it would have limited the exploration of themes to those in which change over time rather than allowing for a broader exploration of the data. It would be difficult to fully comprehend the change over time in this research without first looking at broader perspectives and positioning. Instead, this aspect will be examined in further detail in a future publication likely combining PrEP and early ART. Here, changes in perspectives will be presented as they relate to the key emerging themes. IDIs were digitally recorded, conducted in the participant’s preferred language, usually a combination of local vernacular and English, and lasted up to 1.5 hours. Interviews were later translated and transcribed simultaneously by research assistants into English. Interviews were conducted by research assistants (all women) who had been comprehensively trained in qualitative interview skills and had combined experience of this method of more than five years. No one else was present during the interviews.

Analysis of the data began with development of a coding framework created separately by two researchers and then compared. A final framework aligning codes was developed and employed in 100% coding by both researchers of all transcripts. The framework and transcripts were uploaded and coded in QSR NVIVO 10. Each round of interviews was coded in a separate database using the same coding framework.

After coding was complete, the two researchers compared findings and discussed any differences in coding to align results. This process was completed first, before the analysis of the serial aspect of the data. Thematic analysis was led by the first author, using the Braun and Clarke method [31]. This approach involves a six-phase process: familiarisation with the data, generation of initial codes, searching for themes, reviewing themes, defining and naming themes, and finally producing the report. In this regard, we adopted a contextualist position, where the analysis aimed to report on the experiences and realities of the participants as well as how the external social environment influenced how the participants make meaning of their realities in relation to PrEP.

All interviews were included in the analysis, even though there were not complete sets of three rounds for each participant. For those participants with second and third IDIs, change over time was explored using a trajectory approach, chosen since the aim was to take into account “individuals’ experiences over time” [32]. Coded data were compared as associated with the prominent themes emerging from the initial analysis across the time points for each participant. As there were fewer second and third round IDIs and the change over time was confined to the themes included for this analysis, it was possible to employ a simplified matrix approach and compare results worksheets (developed in Microsoft Word) for the three coded databases by hand. This was conducted independently by the first and second authors, and then findings were compared and found to have few discrepancies which were resolved.

Results

From those invited across the two sites, 18 women from the PrEP arm agreed to participate in the IDIs, 11 in Johannesburg and 7 in Pretoria. Following the first round, 10 women completed the second round of IDIs (4 in Johannesburg, 6 in Pretoria), and 6 completed the third round (1 in Johannesburg, 5 in Pretoria). This dropout relates to the overall rates of loss to follow up across the larger TAPS study [19]. In total, data from 34 interviews were coded and are included within this analysis. IDI participant demographics are shown below in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Qualitative research participant demographics.

https://doi.org/10.1371/journal.pone.0212271.t001

The ages of the PrEP IDI participants ranged between 23 and 40, with 10 between the ages of 23–30, and 8 between the ages of 31–40. None of the women were married, though nine of them had a steady partner, three of whom lived with their partners. A majority of the women were originally from Zimbabwe (n = 12), one from Lesotho, and the rest from Gauteng, Mpumalanga, and Eastern Cape provinces in South Africa. Most of the women had at least some high school education, with three only having completed primary school. Finally, most of the women worked in brothels (n = 12), while the rest worked on the street or in other less formal settings such as vacant lots.

At the start of the project, few of the women had heard about PrEP prior to learning about it during the peer educator outreach and recruitment efforts. With this in mind, the IDIs aimed to situate PrEP within women’s lives, including every day worries, relationships with family, friends, and partners, caring for children, and navigating their working situations. The data presented here fall into themes describing their perceptions of risks and responsibilities, adoption of PrEP, experiences of PrEP use, and interactions with people in their lives as related to PrEP. The themes inherently overlap and are intertwined, according to the complexities and intermingling of relationships and social dynamics of women’s lives interacting with HIV prevention, and PrEP in particular [33]. These data reflect both a practical implementation (or public health) perspective of services and lived experiences. Each participant quote is labelled with a pseudonym, the location (JHB = Johannesburg; PTA = Pretoria), and the IDI number (e.g. first, second, third round).

Discussion

This is one of the first reports of personal experiences and perspectives of actual use of oral PrEP among a group of FSWs outside of a clinical trial setting. This research provides insight into the risks and responsibilities that women perceive in their lives, the ways in which women adopted PrEP to mitigate their risk or ameliorate realities, and the characteristics of PrEP that they most valued, all of which are critical to consider within the context of implementation.

The findings produced from this research are grounded within the MSEM framework, where personal views and factors interact with and are influenced by social and sexual networks, community, societal and public policy, and the HIV epidemic itself. The components of this framework are visible within the themes where community perceptions of PrEP influenced personal beliefs and experiences of use, or where knowledge of the vast epidemic made it difficult to believe there were still HIV-negative FSWs left to take advantage of new prevention methods. The themes identified in this paper make clear the significant role that family, friends and peers play in shaping perceptions of HIV risk, of the efficacy and potential impact of PrEP and, ultimately, in PrEP uptake. The role of social networks in propagating accurate information about PrEP, and in shaping narratives of PrEP confidence and acceptability should not be underestimated.

Disbelief in the existence and/or efficacy of PrEP was a prominent theme, which according to the IDIs, affected the motivation of women to come to the clinic and also affected women trying to maintain use. When those around users are constantly questioning whether such a preventive product can be real, it may be difficult to continue use. This could speak to reasons for dropout in the larger TAPS study for which the primary, programmatic results have been presented [19]. As PrEP had not been widely introduced as a national policy when these interviews were undertaken, there was no generalised community education about its use which may have contributed to mistrust of the product. This disbelief, or distrust, was also a phenomenon seen in the VOICE-D sub-study and HPTN/ADAPT studies which found scepticism of PrEP efficacy to be a significant barrier to use [36,37]. Both Normalization Theory [38] and the Diffusion of Innovations [39] speak to the imperative of community assimilation of interventions for successful implementation. Indeed, without belief in and demonstration of efficacy on a personal level (e.g. repeated negative HIV tests with PrEP use), little engagement can be expected. This was underscored by the participants through the course of the IDIs, and often raised ethical quandaries around not being able to provide others with PrEP. The clinic did, however, allow and promote couples testing as desired by the participants.

The repeated discourse on disbelief or distrust in the existence and efficacy of PrEP also points to a need for concerted community based messaging beyond target populations. While mathematical modelling has suggested that focused PrEP delivery to those at highest risk will have the greatest impact in the most cost-effective manner [3,4,40], the findings in this study combined with those from VOICE-D and ADAPT, imply that those at highest risk may be reticent to engage with PrEP without community endorsement. Development of implementation strategies will necessitate an initial contextual review of community perceptions of PrEP and relevant education campaigns to balance messaging around HIV risk, PrEP use, and efficacy, keeping in mind the risk of stigmatizing the PrEP intervention.

Risk, and risk perception [41], have been extensively researched and debated in regards to translating into effective HIV prevention uptake [42–44]. Research published from the FEMPrEP efficacy trial found that women in the study who did not maintain PrEP use also did not accurately assess their own risk [45]. In the TAPS study, the primary results suggest that the population accessing PrEP are women who may be at self-acknowledged high risk, which was confirmed repeatedly through the IDIs [19]. However, it is unclear whether those who dropped out were accurately assessing their own risk, though one woman who dropped out of the PrEP arm early returned later with an HIV-positive result. The risks articulated in these IDIs, however, were directly linked to the dangers of sex work and closely intertwined with the acknowledgement of personal responsibilities. In addition, these personal responsibilities could also contribute to women’s abilities to maintain use if priorities for making money or going to school prevented them from making it to the clinic for visits and PrEP refills. At the same time, stress around condomless sex with main partners seemed to be reduced for some women who worried they might be at risk when partner fidelity was unclear.

While sex work can be a choice and many women have been empowered in their engagement in it [46], women in this sample of sex workers overwhelmingly felt that they currently had no options to transition out. Most transition programmes have failed because they don’t offer women options to make enough money to meet their perceived responsibilities [47]. In these IDIs, women had differing ideas about how to move on to other employment, with some asking for jobs in the clinic and others pursuing educations with the hope of empowering themselves to advance.

The notions of empowerment and control were underlying currents throughout the findings of these IDIs. While most of their worries remained constant throughout the progression of the interviews, there was also an alleviation of stress which occurred as PrEP efficacy became a personal reality. While PrEP is not a magic solution to all women’s problems, for some in this study it started to symbolize hope and future prospects, as well as peace of mind where safety was concerned, where women could view their business as business without as much fear, and sexual interactions with partners were less worrisome.

Interestingly, the findings in this paper at the thematic level are very similar to what has been published so far from qualitative research of PrEP use among MSM. Research published from an open-label PrEP study (following the iPrEX efficacy trial) described relief from stress around the possibility of HIV infection and a sense of security as additional benefits of using PrEP [48]. These feelings did not seem to change sexual behaviour, which from analyses conducted thus far and what has been reported in our qualitative research, is also what happened in TAPS [19]. Similar findings regarding the alleviation of fear around HIV were presented from another study of MSM in the United States [49]. Encouragingly, this research suggests that PrEP use may provide sources of critical, constructive, and empowering HIV protection for those at highest risk of HIV without subsequent risk compensation, as was widely feared [48,50–53].

Not as many women returned for second and third IDIs as hoped, which is a risk with any longitudinal research and in particular with more transient populations such as sex workers. It is difficult to assess why participants drop out over time when they do not provide specific reasons themselves. However, based on what we learned here and what was learned in earlier research preparing for the TAPS study [25,26], it is likely that a combination of social dynamics around belief in PrEP and its efficacy (including stigmatization of taking ARVs as a sign of illness), personal risk assessment which may fluctuate over time, personal responsibilities which supersede clinic visits and participation in research, and potential issues with substance use probably contributed to retention. For these reasons and to further unpack the dropout seen in TAPS to help support PrEP programming, additional research is being considered to potentially interview women who were lost to follow up, to determine the reasons.

Despite the limited number of second and third round IDIs, the process of serial interviews enabled us to examine progression of thinking and feeling around PrEP use. In most cases, the participants became more committed over time, and discovered personal strength in their commitment, even if adherence was not always perfect. Their willingness to be more open about issues with product use grew over time, which was a benefit of this method. The practicalities of use were not the central themes for any of the IDI participants, however, women were able to share how they managed adherence or cycled off and on PrEP over time. One prominent component of managing PrEP use, which became more apparent over time, was the development of supportive structures for pill taking. Some women found support from partners, but many more had family members or friends who were already HIV-positive and could provide inside knowledge of pill taking habits. In fact, the high prevalence of HIV in the community, and particularly among other women in sex work, could be another contributing factor for staying HIV-negative and a motivation for PrEP use. These notions reinforce the idea that synergies can be found among HIV-negative and positive populations, which may also help to destigmatise interventions.

Conclusion

PrEP represents a valuable new HIV prevention option for women in sex work who are able to remain in care. Risks and responsibilities were the main expressed drivers of PrEP uptake and use in these interviews, however, personal and community disbelief in the reality of PrEP efficacy challenged this. As PrEP programming is scaled-up, it will be important to ensure accurate, relevant, and widespread messaging is employed in communities to generate demand and support for PrEP use, even if use is to be focused for those at highest risk. PrEP could be one of the most significant game-changers in HIV prevention for the past 20 years, and requires this kind of in-depth, qualitative research to understand if and how it impacts on women’s lives and what motivates them to take it up and adhere and thus how services can be better designed to respond to these needs.

Acknowledgments

The authors thank the women who participated in the study and shared their experiences. The entire TAPS team was instrumental in the successful completion of the project and this work. Peter Weatherburn provided invaluable support for the analysis and writing of this manuscript as part of a PhD project. Co-investigators Prof Francois Venter and Dr Gabriela Gomez provided much needed technical and moral support for the entire TAPS project.