Perspectives of US women participating in a candidate PrEP study: adherence, acceptability and future use intentions

Abstract Introduction Limited data exist on acceptability of candidate pre‐exposure prophylaxis (PrEP) regimens among US women. We evaluated PrEP experiences, attitudes and future use intentions among sexually active women who completed the US‐based HIV Prevention Trials Network 069/AIDS Clinical Trials Group 5305 study. Methods Women participated in the study between March 2013 and November 2015. We analysed computer‐assisted self‐interview (CASI) surveys among 130 women and conducted in‐depth interviews among a subset of 26 women from three sites. Interviews were conducted in mid/late‐2015. Results Most women (57%) reported very good/excellent PrEP adherence on CASI, although 21% acknowledged over‐reporting adherence at least some of the time. Commitment to preventing HIV infection, a sense of ownership of the study, and keeping pills stored in a visible location facilitated adherence. Adherence barriers included “simply forgetting” and being away from home. Most women interviewed did not intend to use PrEP in the future because of lack of perceived need due to their own (as opposed to their partners’) low‐risk behaviour and concerns about affordability – but not because of side effects or other characteristics of the regimens. Discussion Improving HIV prevention options for US women will require access to affordable PrEP as well as expanding women's understanding of relationship‐ and community‐level factors that increase their risk of acquiring HIV.


| INTRODUCTION
Although the number of new HIV infections in the United States (US) has declined over the past decade, as of 2015, women constituted almost one quarter (24%) of all people living with HIV [1]. Until recently, HIV prevention strategies available to women at risk of heterosexual transmission were largely limited to approaches that required male partner cooperation (i.e. female or male condoms). Products that are safe, acceptable, discreet and can be controlled by women are needed. Oral preexposure prophylaxis (PrEP) may be one such strategy. Randomized controlled trials (RCTs) have demonstrated that daily oral tenofovir disoproxil fumarate/emtricitabine (TDF/FTC), when used correctly and consistently, is highly effective for preventing HIV infection among men and transgender women who have sex with men (MSM) [2][3][4], HIV-seronegative partners in serodiscordant relationships [5], and heterosexual men and women [6]. However, two PrEP RCTs in sub-Saharan Africa that exclusively enrolled women [7,8] failed to demonstrate efficacy, largely due to study drug non-adherence. A review of oral PrEP efficacy trials in women found that drug levels consistent with daily pill-taking were associated with protection [7], suggesting that adherence is the critical factor influencing protective outcomes.
Adherence, as well as attitudes and beliefs about PrEP, among MSM in the US has received considerable attention in the literature [8][9][10][11][12][13][14][15][16][17][18][19][20][21][22][23]. However, few studies have been published regarding attitudes towards use of PrEP among women who have accumulated real experiences taking candidate PrEP products [24][25][26][27]. Moreover, few PrEP demonstration projects among US women have been planned [28], leaving a substantial at-risk population in the US largely under represented. Despite PrEP's promise as an HIV prevention tool for US women, the lack of data on product acceptability and adherence in this population (especially in the setting of inconsistent adherence demonstrated in international clinical trials) may challenge its successful implementation for HIV prevention among women in the US.
In response to this paucity of data, we collected quantitative and qualitative data from women participating in HIV Prevention Trials Network (HPTN) 069/AIDS Clinical Trials Group (ACTG) 5305 study [29] to better understand reasons for adherence and/or non-adherence to study drug, attitudes towards PrEP and intentions to use PrEP after the end of the study. Given the few PrEP trials involving US women and demonstration projects seeking to involve them, the results of this study may help develop approaches for supporting adherence in the rollout of current and future PrEP technologies.  [29]. As described in detail in the presentation of primary outcomes [29], study regimens consisted of three pills once-daily: MVC 300 mg, FTC 200 mg and TDF 300 mg, with matching placebos. HIV-seronegative women who reported a history of condomless vaginal or anal intercourse with >1 HIV-seropositive male partner or man of unknown serostatus within 90 days, and had adequate safety laboratory parameters were enrolled at 12 US-based study sites (Baltimore, MD; Boston, MA; Chapel Hill, NC; Cleveland, OH; Los Angeles, CA; Newark, NJ; New York City, NY; Philadelphia, PA; Pittsburgh, PA; San Juan, PR; Seattle, WA; and Washington, DC). Participants received randomized study regimens for 48 weeks with follow-up visits at weeks 2, 4 and 8, and then every eight weeks through week 48. At each study visit, interval history, physical examination, safety laboratories, blood plasma for drug levels, HIV testing and adherence counselling were performed at all study visits. Electronic drug monitoring was used throughout the trial via a single pillbox (Wisepill TM , Wisepill Technologies, Cape Town, South Africa) containing the three study medications.
The study protocol and the procedures for the qualitative substudy were approved by institutional review boards at the local institutions affiliated with each study site. A written informed consent was obtained from all participants prior to conduct of any study-related activities.

| Survey data
Detailed behavioural risk assessments and HIV risk perception were assessed via computer-assisted self-interview (CASI) at baseline and serially (every eight weeks). Attitudes towards study drug, barriers and facilitators to adherence, and intentions to use PrEP after the study's conclusion were measured at week 48. Adherence-related facilitators and barriers were identified by women using a list where they could select all that applied.
We reviewed CASI data specific to beliefs, attitudes and adherence among the 130 women who were on study drug at week 48. As reported in the presentation of the primary outcomes for women in the parent study (HPTN069/ACTG5305) [29], of 188 women enrolled, 160 (85%) completed the study follow-up, and of them, 30 participants had permanently discontinued the study regimen prior to week 48. Thus, we focused on CASI data from women with recent experiences with the study drugs (N = 130 women on study medication). Although the study used biological markers of adherence in a subset of women, the current evaluation focused on selfreport data and did not seek to distinguish between those with and without detectable drug at week 48.

| Analyses
Data analysis was performed for all enrolled women who completed the parent trial and had not discontinued their assigned study regimen. CASI data were analysed using SAS 9.4 (SAS Institute, Cary, NC). Descriptive statistics were performed to provide a demographic profile of study participants and summarize survey responses.

| In-depth interviews
A subset of 26 women from three sites (New Jersey, New York and North Carolina) were interviewed within two weeks of study completion (week 48). Convenience sampling was used, as the qualitative protocol was adopted late in the parent study and there was no opportunity to randomly select from the small pool of women who remained in the study. Site staff invited study participants to join the qualitative substudy as they completed their final study assessments. Interviews were conducted with a total of twenty-six women, which included twenty-three who were on study drug at the end of the study and three who had discontinued prior to study end. In contrast to CASI collected data, we retained the three who had discontinued study drug in the interview subset of women because interviewers could provide the opportunity to probe and unpack experiences with study drug over the course of the study.
Interview domains in the structured interview guide included experiences using a candidate PrEP regimen, attitudes towards the study drug, factors contributing to product adherence or non-adherence, HIV risk perception and intention to use PrEP after study completion. Interviews were conducted by a trained interviewer identified at each site and lasted an average of 1.5 hours. Participants received $75 as compensation for their time and were reimbursed for transportation. All activities were approved by the Institutional Review Boards at Rutgers New Jersey Medical School (NJ), Weill Cornell Medicine (NY) and the University of North Carolina at Chapel Hill (NC).

| Analyses
Interviews were transcribed, coded and analysed in Dedoose 7.0.23 (SocioCultural Research Consultants, Los Angeles, CA) using a thematic framework analysis approach [30][31][32]. Five trained analysts reviewed the transcribed interviews and applied coding "frames" based on the in-depth interview (IDI) guide. Each frame was iteratively reviewed by several members of the coding team in order to identify and document key themes across the interview. Through this process, a codebook was developed with detailed and nuanced definitions of key themes, supporting themes and example quotes. Thematic codes were applied to transcripts using Dedoose. Intercoder agreement was assessed at various points in the analysis process. Coding discrepancies were discussed by the analysis team, the codebook revised accordingly and recoding performed when necessary to ensure consistent application of codes. To identify the most salient themes, code frequency reports and coded text reports were generated.

| RESULTS
3.1 | Participant characteristics (Table 1) As detailed in Table 1, the median age of female respondents who completed the parent trial on their assigned drug regimen (N = 130) was 39 years (range: 18-61). Most were not married (90%), had completed high school but less than a four-year college degree (82%) and were unemployed (59%). Almost threequarters of participants (74%) identified as non-White and 13% self-reported Hispanic/Latina ethnicity. At baseline, women reported an average of two sexual partners (SD: 9.0) in the past month and one partner with whom a condom was not used (SD: 2). Women completing the trial on study drug were largely comparable to those who discontinued early, with the exception of being older (39 vs. 32 years of age). Women in the interview subset did not significantly differ from the full sample of women in demography or behavioural characteristics.

| Self-reported drug adherence
Participants reported generally high adherence at week 48. As presented in Table 2, more than three-quarters of women reported taking study drug most of the time (35%) or always (44%) in the past 30 days. Most women similarly reported their ability to take daily study medications as very good (18%) or excellent (40%). Most women (75%) reported that they never over-reported their adherence to study team members; 21% reported overestimating their actual adherence intentionally some to all of the time.

| Barriers and facilitators to product adherence
The most commonly cited challenges to adherence (Table 3) were "forgetting" (45%), being away from home (39%) and not having pills available when they were scheduled to be taken (22%). Few participants reported HIV stigma (4%), concern about disclosing participation in the study (4%) or concerns about side effects (8%).
Respondents identified commitment to preventing HIV infection (63%), a sense of "ownership" of the study (38%), and making sure pills are available (28%) and stored in a location that was easily visible to them (34%) as factors that made adherence easier (Table 3). About a third of participants integrated pill-taking into existing routines (31%), developed reminder strategies (23%) or kept track of dosing on their pill boxes (13%). Twenty-six per cent of women reported concern about HIV acquisition as facilitating adherence.

| Attitudes towards PrEP
Women estimated the efficacy of the study drugs (how good the study drugs were at preventing HIV infection), on average, to be 69% out of 100%. Most participants had an overall positive opinion of PrEP. As presented in Table 4, over 75% would recommend PrEP to others, and most (60%) believed PrEP would be good for "anyone. " Almost half of participants (48%) reported an increase in HIV preventive behaviour while on study ("Since getting the study pills, I do more to protect myself against HIV than I did before"). In response to items asking about which prevention modalities women felt they had the most "control" over, half of women (50%) reported feeling they had more control over a partners' condom use than daily PrEP; more than a quarter of women  (29%) perceived an equal amount of control and only about 20% reported they felt more personal control over PrEP use.

| Experience using study product
Two main themes were identified concerning women's experiences using study products: (1) unrealized anticipated side effects and (2) appreciation of the benefits of participating in a study. Women anticipated side effects over time, which were often not realized or dissipated within the first weeks of the study. A number of participants also described the importance of education and counselling by study staff in tempering their concerns about possible side effects Participant: I had a little bit of diarrhea. That's it. No vomiting, no nausea, none of that. Just one day total. So it really wasn't bothersome at all in terms of, "I don't want to take this medicine because of it. " Participant: At first, the study nurse said, "You might get some side-effects, " and I was like, "Wait a minute. What am I getting myself into?" and then, once I started taking it and a week went by, I said, "Oh, okay, I see now, " and I would take my medicine every day, every day.
Interviewer: So you were a little nervous at first.
Participant: At first, yeah, because I didn't know how, because she said some people, it didn't make them feel right. I said, "Well, I wonder how it's going to affect me?". She said, "It might not affect you the way it affect them, " so I started taking it, and I was fine. I was okay.
Participants also described experience with the study that were appreciated and described as beneficial, including regular HIV testing and health monitoring.
Participant: I liked the fact that I got checked for -I got my blood checked and everything, HIV tested.  Participant: I liked the fact that they gave you condoms. And even though they don't -I wouldn't -I didn't want to get paid for it but they was like "Well we pay you anyway. " I was like "Ok. That's fine. " I get the money, the bus tickets and then like I get the information. I get schooled on some things. I really like the fact that I could learn from it.

| Barriers and facilitators to product adherence
Four themes concerning barriers and facilitators to study drug adherence were identified (Table 5). Three themes described barriers to study product adherence: (1) the impact of competing demands (e.g. family, employment, unrelated health issues, unexpected events); (2) lack of established routine; and (3) believing one should not dose if the "prescribed" dose time was missed. Participants did not describe any concerns about side effects or perceived HIV-related stigma as factors in nonadherence.
One overarching theme was identified as facilitating adherence: consciously building adherence habits (e.g. using cell phone alerts, keeping pills visible, engaging their social network to help with reminders). In addition, almost all of the women who were interviewed valued the care they received as research participants and expressed a desire to contribute to HIV prevention research, although participants did not link these factors to study drug adherence.

| Future use intentions
Of the twenty-six women interviewed, only nine expressed clear intentions of using PrEP after the study ended. The most commonly cited reasons for not planning to use PrEP were (1) low perceived HIV risk and (2) uncertainty about PrEP access and efficacy (Table 6).
Women perceived themselves to be at low risk of HIV acquisition, citing their own monogamy, their partner's low risk of acquiring or having HIV, and condom use. A few women reported discontinuing sexual partnerships during the course of trial, and did not feel that they would need PrEP until they were once again sexually active. Valuation of risk appeared largely focused on women's own (vs. partner) behaviour, with being in a monogamous relationship often noted as evidence for no HIV risk.
Women also expressed uncertainty about where to access PrEP and whether it would be covered by insurance; some participants were unaware that oral PrEP with TDF/FTC was approved by the US Food and Drug Administration for at-risk women in the US. Women were also unsure how PrEP would fit into their current HIV prevention strategies, whether it would be efficacious enough to replace condoms or whether it should be part of a combination of interventions for HIV prevention.

| DISCUSSION
Although most new HIV infections in the US are among men, more than 7000 women received an HIV diagnosis in 2016 [1]. We used mixed methods (CASI and IDIs) to explore acceptability of, adherence to, and intentions concerning future use of PrEP among US women who completed a 48-week Phase II prospective, randomized, double-blinded, multisite, safety and tolerability study of four candidate PrEP regimens comprised of combinations of TDF, FTC, and MVC. Women found the study regimens acceptable, but among those interviewed a lack of perceived risk or need and concerns about affordable access limited enthusiasm for using PrEP after the end of the study.
Women reported high levels of adherence, with most reporting taking their study drugs most or all of the time during the previous month; however, nearly a quarter of women acknowledged over-reporting the proportion of study medications they had taken at times. Previous studies have linked intentional non-adherence and over-reporting adherence to study teams as related to concerns about safety and side effects [33-37]; however, few women in our study cited side effects as barriers to adherence or discussed safety concerns in interviews. Participants discussed initial apprehension about side effects that were never realized, which may have decreased concerns about safety over time.
Lapses in drug adherence were often explained as unintentional and due to a lack of established routines, reminder tools

Lack of Established Routine
Women discussed a lack of established routine, including changes in daily schedule, employment-related events and long-term travel or moving. Although these events often resulted in only one or two missed pills, some women reported that unanticipated events

Not Dosing if Dose Time was Missed
Some women reported not dosing when they missed their specific dose time, suggesting that some did not understand that they could take pre-exposure prophylaxis any time during the day Participant: Oh shit. I can't take them now it's after 12 o'clock.

Consciously Building Adherence Habits
Discussion that supported this theme centred around developing new routines and pill-taking reminders. This included using technology (cell phones, alarm clocks), keeping pills in a visible location as well as developing routines for pilltaking individually or in conjunction with others Participant: But you could set your timer on the phone, but when you're not used to taking pills, it's kinda hard. You forget. But once you start being in the study for a long time, you'll start taking them more regularly. Participant: I knew that I had to take them. So, when I get up I eat and I'll just take the pills -I knew I had to take it because I was -I'm in the research, so I knew I had to take it. So, I'd just wake up in the morning and eat and just take them. Participant: Yeah, and sometimes my fianc e said, "Did you take your medicine this morning?" and I was like, "Yes, I took it, " because he has diabetes so he takes his medicine, so I would take mine. He'd be like, "Take your medicine, " and I was like, "I'm taking it, " so he would take his medicine and I would take mine. and competing demands. Several women also reported the (erroneous) belief that a dose could or should not be taken if the exact "prescribed" dose time was missed. In CASI and interviews, facilitators for taking pills included developing habits, such as keeping pills visible, and a commitment to HIV research and to the study. The impact on adherence of developing habits and using reminder tools is well described in the ART (and other) adherence literature [38][39][40][41], and further underscored by our findings. These results highlight the importance of verifying patients' understanding about the timing of medications and what to do when doses are inadvertently missed or taken late.
Women had generally positive impressions of PrEP after participating in the study and believed PrEP would be a useful HIV prevention tool and "good for anyone, " with a majority indicating they would recommend it to others. However, some women in the interviews expressed uncertainty about using PrEP in the future because they were unsure about how PrEP would fit into their current prevention strategies and whether it would be accessible through insurance. In fact, several women seemed unaware that an FDA-approved PrEP regimen was available to at-risk women in the US. Intentions for PrEP uptake may have been higher if the study actively transitioned participants to PrEP prescribers in the local area. But the major reason for women's relative lack of interest in using PrEP after the study was their perception that they were not at risk for HIV acquisition. Discourse concerning participants' valuation of risk was focused almost exclusively on the participant's behaviourwith little attention to the influence of partner characteristics and community HIV infection rates. Although women in our study may have accurately assessed their likelihood of acquiring HIV, their assessments could also have missed important external factors that elevated their risk, such as community-level HIV risk. We do note that women, generally, did appear to report generally low HIV risk (condomless sex, number of partners) behaviours; however, future research should carefully consider associations between reported behaviour and valuation of HIV risk for women specifically. A more holistic discourse about HIV risk and risk perception that includes partner, community and structural influences may be particularly important for women's uptake of PrEP in communities that are highly impacted by HIV. Although our study was conducted in the US, this recommendation may generalize widely; women's unique vulnerabilities to HIV are worldwide [42] and campaigns that focus on individual risk factors may promote an underestimate of HIV risk for women.
Surprisingly, women did not appear to consider PrEP as a prevention strategy that was more under their control than condoms. Because condom use often requires women to negotiate with partners or otherwise rely on partner initiated behaviours and PrEP is self-directed and can be administered without partner awareness [43], we had anticipated that most women would rate PrEP as more in their immediate control than condom use. The most commonly selected response to this item was perceiving more control over condom use. This may have been influenced by the study context which involved taking a study drug under evaluation; where as they can visually see a condom being used and are confident in its function, they cannot see the study drug "working" and do not know if it is actually effective. Additionally, women were counselled on use of condoms throughout the study, with more than half reported that participation in the study increased their use of prevention strategies over time. Thus, this finding should be re-examined in the context of open-label (known to be effective), easily accessible, PrEP projects.
Findings should also be interpreted in the light of several limitations. First, only a small number of study participants at three sites were asked to participate in IDIs, thus may not be representative of the women in the overall study. Second, HPTN069/ACTG5305 trial participants were healthy volunteers in a blinded trial; women not engaging in research may differ from those who do in concerns about safety and challenges with intentional non-adherence. Additionally, while eligibility criteria required participants to be at-risk for HIV (i.e. condomless vaginal or anal intercourse with ≥1 HIV-infected or unknown-serostatus man within 90 days of enrolment), they were not necessarily at high risk for HIV infection. Finally, study drug adherence and related attitudes towards the drug and the regimen may dramatically differ from adherence and attitudes towards drugs known to be effective and widely available outside of a study context. Study participants completed interviews in late 2015. Efforts to increase PrEP awareness and access for women has increased in a number of regions in the US since that time, although PrEP uptake remains generally low. Nonetheless, women in our study may differ in terms of PrEP knowledge, awareness and attitudes from more contemporary samples [44].
Strengths of this study include a study population that is racially and ethnically similar to that of women living with HIV in the United States, with significant representation from Black and Hispanic women. Additionally, study sites were broadly representative of US major cities and geographically reflect US regions most affected by the HIV epidemic.

| CONCLUSION
Women's valuation of their personal risk for HIV infection will be critical to PrEP's success as an HIV prevention strategy for women. Building women's demand for PrEP may be challenged by their personal risk assessments in settings where individual-level risk factors appear minimal but contextual and partner-level factors confer elevated risk for HIV. More holistic thinking about women's HIV risk will be needed to maximize the effectiveness of PrEP as an HIV prevention strategy for women in the US.

F U N D I N G
The parent study was supported by the Division of AIDS, National Institute of Allergy and Infectious Diseases, National Institutes of Health, through the HIV Prevention Trials Network (UM1-AI068619, UM1-AI068613 and UM1-AI068617) and the AIDS Clinical Trials Group (UM1-AI-068636). Gilead Sciences and ViiV Healthcare provided study drugs. The qualitative substudy was supported by Gilead Sciences.