PrEP Discontinuation and Prevention-Effective Adherence: Experiences of PrEP Users in Ugandan HIV Serodiscordant Couples

Abstract

Background:

Optimal adherence to oral pre-exposure prophylaxis (PrEP) for HIV prevention involves aligning consistent PrEP use with periods of risk to achieve prevention-effective adherence. Prevention-effective adherence is predicated on individuals discontinuing PrEP during periods without expected risk. For stable, serodiscordant couples, ART adherence by the HIV-positive partner markedly decreases HIV transmission risk, potentially obviating the need for continued PrEP use; yet little is known about actual lived experiences of discontinuing PrEP.

Methods:

In-depth qualitative interviews were carried out with HIV-uninfected PrEP users in serodiscordant couples taking part in the Partners Demonstration Project at IDI-Kasangati, Kampala, Uganda. Open-ended interviews elicited information on the partnered relationship; understandings of PrEP; prevention strategies; and experiences of PrEP discontinuation. An inductive, thematic, content-analytic approach was used to analyze study data.

Results:

Uninfected partners experienced PrEP as a valued resource for preventing HIV acquisition. Despite ongoing ART use by HIV-positive partners for a period of time consistent with viral suppression, discontinuation of PrEP was experienced as a loss of protection and a corresponding increase in risk of HIV acquisition. Uninfected partners responded with strategies aimed at offsetting this subjective sense of increased risk, specifically: (1) changing sexual practices; (2) prioritizing fidelity in the relationship; (3) increasing reliance on condoms; and (4) seeking evidence of partners’ ART adherence.

Conclusions:

These experiences highlight the challenges PrEP users in serodiscordant couples face in discontinuing PrEP for prevention-effective adherence. Flexible interventions that support individuals during this transition may increase comfort with discontinuing PrEP when alternative prevention strategies provide protection, such as a partner’s consistent adherence to ART.

Introduction

Oral pre-exposure prophylaxis (PrEP) with antiretroviral medication is highly efficacious in preventing the acquisition of HIV infection [1-3]. How well PrEP works in practice, however, depends upon how well users adhere [4].

Adherence to PrEP for HIV prevention differs from adherence to ART. While adherence to ART should be consistent for life, optimal PrEP adherence involves alignment of PrEP use with periods -- or “seasons” -- of risk. Users should take PrEP when they are at risk of HIV acquisition, then discontinue when they are no longer exposed to HIV. This pattern of PrEP use has been characterized as “prevention-effective adherence” [5-6].

Prevention-effective use of PrEP means individuals must be willing not only to take PrEP, but also to stop taking it. PrEP uptake and adherence have attracted considerable research interest [7-11]. Uptake and adherence to PrEP are important research areas [7], and have been shown to correspond to relational dynamics such as stress and trust between partners [10], and to perceived levels of HIV risk [8]. However, almost nothing is known of users’ experiences of discontinuing PrEP [12].

To date, research addressing users’ willingness to start and discontinue PrEP as their risk changes has been based on hypothetical or anticipated behavior. Namey et al. (2016) interviewed Kenyan and South African women at higher risk of HIV to gauge likely PrEP use, had it been available to them. Almost all indicated they would have used PrEP; half estimated they would have suspended it at some point [13]. Elsesser et al. (2016) evaluated interest in using PrEP for short periods of increased HIV risk on the part of men who have sex with men (MSM). In this sample of 7305 individuals, 92.6% characterized daily dosing as a barrier to PrEP use, but 74.3% indicated they would use it for short periods [14].

The Partners Demonstration Project (2012-2015, Clinicaltrials.gov ) evaluated an integrated strategy of delivering PrEP and ART to mutually disclosed, heterosexual serodiscordant couples in East Africa. The strategy offered time-limited PrEP to uninfected partners until HIV-positive partners had used ART for six months, and were considered virally suppressed and not at risk of transmitting HIV [15-20]. Counseling offered as part of Partners Demonstration Project visits included information on the use of PrEP as an HIV prevention tool. Counselors informed couples that unlike ART, PrEP use is not intended to be life-long. Rather it is to be used during periods of high risk of exposure to HIV infection” [21]. After six months of ART, counselors and clinicians encouraged PrEP users to discontinue the medication. Counseling at ART initiation included information about the prevention benefits of ART.

The Partners Demonstration Project represents a particular set of circumstances in which PrEP users had knowledge of their partner’s HIV status and ART use, and PrEP discontinuation was timed to coincide with viral suppression. These circumstances combined to minimize transmission risk, potentially reassuring uninfected partners that PrEP could be safely discontinued.

Drawing upon qualitative data, we investigate PrEP discontinuation in a subsample of HIV-negative partners in serodiscordant couples using PrEP as part of the Partners Demonstration Project. The goal is to develop concepts representing potential real-world consequences of, and responses to, PrEP discontinuation for future users practicing prevention-effective adherence.

Methods

Adherence Measurement in the Partners Demonstration Project

As part of the Partners Demonstration Project, adherence to PrEP was measured using electronic adherence monitoring (medication event monitoring system, or MEMS, WestRock, Switzerland) [22]. Overall median adherence to PrEP was estimated at 88% (interquartile range [IQR] 64–99%). In light of uncertainty at the time about dosing levels sufficient for prevention of HIV transmission, prevention-effective PrEP adherence was operationally defined in two ways for quantitative analysis: (a) as an average of ≥4 doses/week during periods of high risk; and (b) as an average of ≥6 doses/week during periods of high risk. At ≥4 doses/week, prevention-effective adherence was observed for 88% of follow-up visits at which uninfected partners were defined as being at high risk. At ≥6 doses/week, prevention-effective adherence was observed for 75% of follow-up visits at which uninfected partners were defined as being at high risk.

The Qualitative Study

Sampling and Recruitment

Participants in the qualitative study were HIV-uninfected Partners Demonstration Project participants receiving PrEP at the IDI-Kasangati, Uganda site. Purposeful sampling was used to identify a single group of individuals with varying experiences of PrEP adherence and discontinuation [23]. We included men and women in roughly equivalent numbers, and identified individuals who had been taking PrEP for different lengths of time. Sampling was carried out prospectively, to promote variation by contrasting time on PrEP among potential participants with those already in the sample. The range of times on PrEP represented in the study sample is 1-22.5 months.

Data Collection

Participants in the qualitative study took part in a single, in-depth individual interview. Interviews were conducted in the local language (Luganda) by Ugandan research assistants trained in qualitative data collection methods. Interviews lasted about one hour, and took place in private settings of the participants’ choosing. Most interviews took place in participants’ homes, outdoors in quiet area, or in a private room at the clinic. Interviews were audio recorded, with permission. Each interview was simultaneously transcribed and translated into English by the research assistants shortly after completion.

Interviews were carried out using a semi-structured interview guide that addressed the following topics: (1) adherence challenges; (2) dosing routines; (3) understandings of and attitudes toward PrEP; (4) the partnered relationship; (5) risk behaviors; (6) prevention strategies used in addition to PrEP; (7) reproductive desires and intentions; and (8) experiences of discontinuing PrEP. The semi-structured interview guides were flexible enough to ensure that interviewers could probe where necessary in order to elicit richly detailed descriptions of the thoughts and experiences of each interviewee.

Data Quality

Transcripts were reviewed by the third author as they were completed in order to monitor data quality. Continuous feedback through weekly calls and emails was provided by the third author to the research assistants to improve interviewer technique, and ensure interview content remained focused on study goals.

Analysis of Data

Analysis of interview data used an inductive, thematic, content analytic approach to develop conceptual categories representing user responses to discontinuing PrEP [24]. The process of review began as each transcript was produced. Each transcript was thoroughly reviewed in its entirety, and the content of the complete dataset was used to inform the analytic process of category construction.

The second and third authors developed, tested, and revised the codebook. The third author and a research assistant coded the entire qualitative dataset using the qualitative management software Dedoose. The third author trained the research assistant on the codebook and its application. Both coders then applied the codebook to a subset of interviews, compared the results, and resolved differences to ensure a consistent approach to the coding process. The third author provided continuous oversight of the coding process to ensure accuracy.

Once coding was complete, the first author undertook an inductive approach to analyzing the coded dataset with the goal of formulating an initial set of categories that represent how participants responded to discontinuing PrEP. The first author developed labels and in-depth descriptions of the initial categories, and provided quotes as supporting evidence for the content of each category. Initial categories were reviewed by the first, third and final authors. The first author revised the categories through an iterative process that involved revisiting the dataset and providing additional depth and detail to the category descriptions.

The final product of this analytic process is a set of five categories. Categories represent concepts and experiences that were of deep importance to the interviewees, and which occurred with regularity throughout the dataset. The first category that emerged from our inductive analysis describes how participants perceived – or anticipated how they would perceive - their risk of HIV infection following discontinuation of PrEP (Category A). Categories B-E describe the specific responses –both enacted and planned - to this perceived heightened risk. We develop and present each category by including different dimensions of the concept, and add explanatory depth by detailing the variety of meanings of the concept for participants. Categories are presented in Results, below (see also Figure 1).

Figure 1:

PrEP Use and PrEP Discontinuation in Relation to the Qualitative Interview among Ugandan PrEP Users Participating in a Qualitative Study: [July 2014-May 2016]

Ethical Approval

The study was reviewed and approved by institutional review boards of the University of Washington, Partners Healthcare/Massachusetts General Hospital, and the National AIDS/HIV Research Committee of the Ugandan National Council for Science and Technology (UNCST).

Results

Participant Characteristics

Seventy-five individuals took part: 40 women and 35 men. Median age was 28 years. Almost all (97%) reported being married. Median years of education was 7. Median time participants reported living together with their partners at enrollment in the Partners Demonstration Project was 3 years. Median time since learning they were in a serodiscordant relationship was 1 month. Two-thirds (68%) of study participants had discontinued PrEP at the time of the qualitative interview.

Overall median adherence to PrEP among qualitative study participants was 89%. Defined as ≥4 doses/week during periods of high risk, prevention-effective adherence was observed for 91% of follow-up visits for qualitative study participants. Defined as ≥6 doses/week during periods of high risk, prevention-effective adherence was observed for 80% of follow-up visits for qualitative study participants (see Table 1).

Table 1:

Characteristics of Ugandan PrEP Users Participating in a Qualitative Study: [July 2014-May 2016] (N=75)

	Median (IQR) or N (%)
Age, Years, at enrollment in the Partners Demonstration Project	28 (23-35.5)
Gender, Female	40 (53%)
Married, at enrollment in the Partners Demonstration Project	73 (97%)
Education, Years, at enrollment in the Partners Demonstration Project	7 (5.5-11)
Length of Time Since Discovery of Serodiscordance, Months, at enrollment in the Partners Demonstration Project	1 (1-4)
Length of Time Cohabitating, Years, at enrollment in the Partners Demonstration Project	3 (0-5)
Length of Time in Partners Demonstration Project, at Qualitative Interview, Months	14.5 (8-18)
Discontinued PrEP, at Qualitative Interview	51 (68%)
Overall PrEP Adherence	89% (67%-99%)
Prevention Effective PrEP Adherence w/ high HIV risk
a.>4 doses per week	91%
b.> 6 doses per week	80%

Qualitative Results: Conceptual Categories

A. Discontinuing PrEP Raises Perceived Risk of HIV Infection

Uninfected partners in the qualitative study experienced PrEP as a valued resource for the prevention of HIV transmission. PrEP was associated with an improved sense of security: individuals explained that PrEP use put them on the “safe side” – affording extra protection during sexual relations. “If that medicine was not available,” one participant commented, “I would be like a soldier in a battlefield expecting to be shot at any time” (Table 2, A, 1).

Table 2:

Quotes illustrating content of descriptive categories for Ugandan PrEP Users Participating in a Qualitative Study [July 2014-May 2016]

Summary Statement	Elaboration	Data Excerpts
A. Discontinuing PrEP Raises Perceived Risk of HIV Infection	1. PrEP affords uninfected partners an improved sense of security from HIV infection	“I take [PrEP] to be like a security person. The more I take, I get many security guards in my body who can fight HIV even if it forces itself out of him into me. Such a security can eat up HIV once it gets into my body so that it cannot multiply within my body. That is what I personally know.” Female, Age 27
	2. Losing extra protection associated with PrEP caused distress	“The risk is too much… I am always worried. Previously I had something [PrEP] that was protecting me, but now I am not using anything… They told me that now the virus in my husband’s body has weakened and he cannot infect me but I told them I do not accept that and I told them, ‘just help me. If [PrEP] is finished from the clinic and if you can sell it to me then tell me to buy it. Because I will be able to save money and buy it, but I want it.’” Female, Age 24 “I used to feel so good [on PrEP] and it hurt me when they told me to stop taking it….What made me feel bad was that, personally, it was saving my life. My life! But when they told me ‘you are not going to take PrEP again’ then I said, ‘maybe I will acquire the virus.’ [I was] getting worried! I was worried because I am no longer taking PrEP. If it is only my husband taking ARVs, I think I have more chances [of infection] compared to before when I was taking PrEP.” Female, Age 21
B. Reliance on Condoms for Protection Against HIV Transmission	1. Condom use alleviated worry associated with PrEP discontinuation	“I first got scared a lot. I said, ‘I have stopped taking [PrEP], now how am I going to live?’ I said, ‘The pills I have been taking have been protecting my life but I may acquire the virus now since I have stopped taking them.’ However later I became strong and what was strengthening me was the condoms that we continued using.” Male, Age 37 “Now that I have stopped PrEP, I am at risk. And if I do not… have a condom with me for my partner to put it on, I am at risk. Being at risk is being tied up when both of us have not put on a condom. However, when he puts on a condom I get relieved.” Female, Age 46
	2. Association between condoms and sex outside the partnership	“He kept quiet for some time… when [I asked] again for the second time, I told him ‘Let us use a condom.’ He said, ‘It seems you have other partners. Why do you like the condom so much?’” Female, Age 47
	3. Condom use is linked to authority in the home	“I was at his home and that is what he wanted. I was forced…he decided not to use a condom that day. In my mind I had it that I can acquire HIV since we had not completed the dose. I knew it very well but I had nothing [I could] do. I could not resist sex because I am a woman and he has authority in his home.” Female, Age 27
	4. Requests for condom use could heighten risk of forced sex	“I told him to use condoms and he refused. I told him, ‘sleep alone in your bed and I will also sleep alone because you do not want to use a condom, you want to infect me.’ What will come out of that is this. The man is more powerful, and what will happen is he will rape me…you cannot be equally powerful like a man.” Female, age 47 “I: What would happen when you refused to have sex? R: We would quarrel … then you say, ‘Oh my God.’ Then later, after some time, you accept to have sex with him. I: Did refusing to have sex with him cause him to force you to have sex? R: Yes. He wanted to force me because I refused to have sex with him.” Female, Age 38
	5. PrEP as important source of protection for women exposed to forced sex	“If he rapes you when you are taking PrEP, it helps you. But if he rapes you when you had stopped taking PrEP, you have to get this virus… [PrEP] helps not to get the virus.” Female, Age 32 “You know men, when you tell him ‘use a condom because you are infected,’ he can refuse. So if you do not use a condom and also you do not swallow the medicine…it is not good.” Female, Age 39
C. Prioritizing Fidelity in the Partnered Relationship	1. Can’t control infidelity; can control PrEP	“He had started taking his ARVs by then, so when he got healed… he was going out for partying. He was going out partying with others and this was not done once or twice. I endured because there are always problems in a marriage, but then I noticed that he was not changing his behaviors. He was not changing. I endured and stayed with him, taking my PrEP because I want my life.” Female, Age 24
	2. Fidelity to the partner known to be on ART	“You can go and sleep around with another partner whom you do not know whether he is infected or not. If he is infected, does he take ARVs or not, you never know. God could have saved you this time round and then you get the virus from the extramarital partner because you do not know his viral load. When it comes to this partner, you know that he takes his ARVs.” Female, Age 40 “He is taking ARVs and they say that when one is taking ARVs, he cannot infect a partner. So I cannot go to another partner when I do not know his HIV status, whether he takes ARVs or no[t].” Female, Age 47
	3. Seeking out prevention methods beyond ART in face of infidelity	“She is adulterous. Remember, she is infected and she is contracting more. That thing looks very bad to me and that is why I use those ‘guards’; the condom and the medicine. When there are no condoms [and] no time for me to look for a condom I do not do it. I do not have sex because of protecting myself.” Male, Age 50
D. Changing Sexual Practices to Reduce Risk	1. Reduce frequency of sexual encounters	“I :Ever since you stopped taking [PrEP], what changes have there been there in the way you protect yourself from acquiring HIV? R: The way I have been protecting myself is because the number of times we have sex had reduced.” Female, Age 24 “I: Tell me about the changes that came up in your sexual relationship after you stopped taking PrEP. R: The change that came up was that we reduced the number of times we have sex. Before we would have sex about four times in a week but after, the number of times reduced to three times. I: Why did you decide to reduce the number of times you have sex? R: It was because of getting worried after this had happened. I: What were you worried about? R:I got worried about acquiring the virus in case we have unprotected sex and at the same time I was worried about contracting the virus when I do not have PrEP that I used to take before. I was not worried so much when I was taking PrEP.” Male, age 27
	2. Minimize “bruising” during sex	“We have been protecting ourselves such that we do not hurt each other during having sex. I have been preparing her well such that by the time I am ready she is also ready too and I insert in without hurting her such that no one gets wounds and no one has wounds.” Male, Age 23 “[Now] I will not be swallowing the medicine [PrEP] and he might infect me any time… I will make sure there is no damage [to] our bodies when we want to have unprotected sex. The damage will make me to get infected very fast.” Female, Age 39
	3. Resisting unwanted sexual activity and perceived heightened risk of bruising	“You know that men have a lot of authority at home. A lot. He is stronger than me even if I refuse he can decide…he can decide to forcefully have sex with me. The reason why I had to accept to have sex with [him] although I did not want it is because I knew he could use force to have sex with me. Yet using force is of great danger to my life. It is dangerous because I will not be ready. So I had to settle down and make sure I accept to have sex so that he does not do it with force…. There are a lot of chances for one to be damaged in one’s private parts once forced to have sex. … It is easy for the private parts to tear, and blood can move from one person to another, which can transmit HIV.” Female, Age 27
E. Establishing Confidence in the Infected Partner’s ART Adherence	1. Perceived risk bound to partner	“What will [stopping] mean to my chance of acquiring the virus? When I stop taking PrEP, I think my chances will depend on his chances.” Female, Age 29
	2. Direct observation of ART dosing	“What strengthened me when I stopped taking PrEP was that the clinic staff gave her ARVs, and she took them. Because she used to take them at night at 10 pm when I am around.” Male, Age 23 “R: I was confident enough because I saw her when she was taking them and I am the one reminding her to take. I: You said that you were confident. Confident about what? R: Confident that she has taken a treatment that prevents me also because when she has taken ARVs, I know that my chance [of acquiring HIV] is not high but it is low.” Male, Age 42
	3. Sharing test results increases trust in protection	“They have tested his CD4 count, and he tells me…Depending on the medical notes/records, they tell him ‘the viral load is this much and the CD4 count in your body has increased.’ There and then I get to know that the ARVs work.” Female, Age 22
	4. Partner’s absence diminishes trust in protection	“He travels a lot. If he was working nearby and he was coming back home every day and you know that he takes ARVs at night or in the morning when he is leaving, there you can know that he has taken ARVs. But most times he is not around, and you cannot know whether he took ARVS or not.” Female, Age 20 I: “Would do you want to take PrEP again in case there is another opportunity for you to take it or not? R: Yes. I: Why? F: It’s because for me I know that I have taken PrEP. But for him, I cannot know whether he has taken ARVs. I know that I have taken PrEP today, tomorrow and even the following day. I know that I am at least safe. But for him - I do not know whether he has missed ARVs, or he didn’t take it or he is lying to me.” Female, Age 20

Study participants transitioning off PrEP following ≥ 6 months of ART use by their partners feared that discontinuation raised their risk of HIV infection. They experienced the transition as loss of the “extra protection” they had previously enjoyed. The resulting sense of increased vulnerability to infection could cause considerable distress (Table 2, A, 2).

When individuals discontinued PrEP, they grappled with how to respond to their perceived increased risk. Four types of responses emerged from qualitative study participants’ descriptions of their efforts to manage their sense of heightened risk. They are represented as categories B-E, below.

B. Reliance on Condoms for Protection Against HIV Transmission

Some participants reported relying on condoms to mitigate the heightened sense of risk they associated with discontinuing PrEP. Regular condom use reduced these participants’ concerns that they could become HIV-infected through sexual encounters with their regular partners (Table 2, B, 1).

Condoms offset couples’ sense of increased transmission risk following PrEP discontinuation when partners were mutually invested in their use, and ready access could be assured. However, some individuals –both men and women -- who endeavored to use condoms following PrEP discontinuation were unable to persuade their partners. Uninfected participants explained that their requests for condom use were interpreted as evidence of sexual activity outside the relationship, and consequently, met with resistance and suspicion (Table 2, B, 2).

Gender-based imbalances in sexual negotiating power within couples also inhibited condom use. While women viewed condoms as a resource in mitigating infection risk after PrEP discontinuation, they reported that socially accepted norms of “authority” within households left them unable to enforce condom use within the relationship (Table 2, B, 3).

Consistent condom use was challenging for women who feared requests to introduce condom use would provoke arguments, or even heighten the risk of forced, unprotected sex in the relationship (Table 2, B, 4).

Women who experienced forced sex found PrEP to be a particularly valuable resource. They could not control unwanted sexual advances from their partners, but PrEP allowed them to control the means of protection against infection. Gaining access to PrEP helped reduce the fear that unwanted sexual encounters with their partner would result in HIV infection (Table 2, B, 5).

C. Prioritizing Fidelity in the Partnered Relationship

Unfaithful partners living with HIV represented a source of perceived risk that was largely outside of the control of the uninfected partner. When a participant was powerless to ensure fidelity, PrEP emerged as a reliable source of protection that he or she could control (Table 2, C, 1).

Study participants who had discontinued PrEP felt that remaining faithful to their partners reduced risk of HIV transmission. Interviewees eschewed pursuit of extramarital partners because the HIV status of these partners was “unknown.” Having sex with an HIV-positive partner known to be taking ART was thought to pose less risk than a partner of unknown HIV status and unknown treatment status (Table 2, C, 2).

Sexual activity outside the primary relationship was seen as undermining the trustworthiness of protection from ART use. To compensate for this undermining of ART-conferred protection, uninfected partners sought additional means of protection. Some turned to condom use, or tried to delay or forego sexual intercourse when an infected partner was suspected of having sex outside the relationship (Table 2, C, 3).

D. Changing Sexual Practices to Reduce Risk

Reducing the frequency and/or duration of sex was a risk reduction strategy couples initially adopted following discovery of serodiscordance. This strategy became particularly important upon discontinuing PrEP, when couples sought to adopt sexual practices they believed might offset the loss of a valued source of protection (Table 2, D, 1).

A common theme among participants was concern about the risks posed by tears in the skin occurring during sex. Interviewees worried that these tears – which they termed “bruising” – placed them at greater risk for infection. Altering sexual practice to minimize bruising was cited as another way of compensating for increased risk after discontinuing PrEP (Table 2, D, 2).

The strategy of “avoiding bruising” required negotiation with one’s sexual partner. Women whose partners respected their wishes perceived themselves as being at lower risk for infection from bruising. In contrast, women who were forced to have sex when they were unwilling or not ready considered themselves to be vulnerable; it was preferable to reluctantly accept a partner’s unwanted sexual advances rather than to attempt to resist them and accrue the risk associated with bruising (Table 2, D, 3).

E. Establishing Confidence in the Infected Partner’s ART Adherence

Effective prevention of HIV transmission through ART requires viral suppression, which in turn requires good adherence. In light of this, following discontinuation of PrEP, uninfected partners’ sense of risk became bound up with their confidence in their partner’s consistent use of ART (Table 2, E, 1).

Those in a position to monitor their partners’ ART use could be more confident. Living together as a couple made it relatively easy to observe or have reliable knowledge of a partner’s dosing behavior. Direct observation of ART dosing helped uninfected partners feel reassured that adherence conferred protection (Table 2, E, 2).

Being informed of the results of routine HIV lab tests measuring viral load and CD4 counts was another form of monitoring evident in the data. Uninfected individuals whose partners shared results of viral load testing felt reassured by having the information. This in turn increased their trust in ART as an effective prevention tool (Table 2, E, 3).

Some interviewees insisted that if they could not see their partner take their medicine, they could not believe that ART would protect them. For example, dosing behavior could not be observed when a partner was away from home. Individuals who were not in a position to monitor their partner’s adherence felt unable to place confidence in ART use as prevention. This left them feeling at heightened risk for infection (Table 2, E, 4).

Discussion

This paper describes real experiences of discontinuing PrEP by a group of HIV uninfected men and women in Ugandan serodiscordant couples who participated in the Partners Demonstration Project. Despite being informed of the prevention benefits of ART, and with at least six months of ART use by HIV-positive partners, these PrEP users experienced discontinuation of PrEP as a loss of protection against HIV and a corresponding increase in risk of HIV acquisition. They responded with strategies aimed at offsetting this subjective sense of increased risk. These strategies included reliance on condom use, prioritizing fidelity in partnered relationships, changing sexual practices, and establishing confidence in good ART adherence by partners living with HIV. These insights highlight and define the challenges that individuals face in managing PrEP discontinuation as part of a prevention-effective adherence strategy.

Prevention-effective adherence refers to periodic use of PrEP aligned with periods of HIV risk over the life course. The concept is based on a “seasons” model that characterizes HIV risk as dynamic, and waxing and waning over time. To derive optimal benefit from a prevention-effective adherence approach, prospective PrEP users must be able to assess their changing risk accurately.

Accurate assessment of changing HIV risk may be challenging for PrEP users. Corneli et al. (2014) investigated risk perception in relation to PrEP adherence in the FEM-PrEP clinical trial, concluding that perceived risk may be one explanation for poor adherence [8]. Many participants in this study population of women at high risk, particularly in South Africa, perceived their risk to be low, underestimating actual risk. Our data suggest the opposite may also occur. Uninfected partners in serodiscordant relationships in our qualitative sample appear to have overestimated their risk of HIV acquisition following discontinuation of PrEP, as they reported feeling at risk despite being protected by their partner’s long-term ART use.

The counseling couples received as part of the Partners Demonstration Project may have shaped participants’ perceptions of risk. Standardized messages covering HIV serodiscordance, ART use, PrEP use, and use of time-limited PrEP as a “bridge” to long-term prevention with ART were organized into a framework used to guide counseling sessions at every follow-up visit. Counselors explained the rationale for discontinuing PrEP, but could also endorse continuing the medication in certain circumstances. At the time, the concept of ART as prevention was relatively new; this may have impacted the clarity and confidence with which this information was presented [21]. Despite counseling efforts, uninfected partners in this qualitative sample felt more vulnerable to acquiring HIV after discontinuing PrEP.

Typically, participants followed the integrated strategy and discontinued PrEP when advised to do so, but some chose to continue, for a variety of reasons. These included a desire to conceive, sexual partners outside the primary relationship, and concern about partner’s ART adherence and viral suppression. Others discontinued independently, when the primary relationship ended, or when they simply lost interest.

The capacity to adopt the risk reduction strategies described here depended on uninfected partners’ ability to exert control in their sexual relationships. While a few male qualitative participants experienced difficulty with condom use, it is women who reported experiencing an inability to negotiate sexual practices with their male partners. The concept of “coercive control” distinguishes efforts to establish and maintain dominance over an intimate partner from situationally engendered discrete acts of physical violence. Gender inequality sets the stage for coercively controlling partnerships in which men are empowered to subordinate their female partners by systematically restricting their personal autonomy [25-26]. Women’s inability to reduce their perceived risk following PrEP discontinuation by declining risky sex may be understood as a manifestation of coercive control.

The impact of male partners on women’s use of female-controlled HIV prevention methods, including their participation in investigational research, has been repeatedly documented [27-30]. Male dominance in decision-making about HIV prevention and sexual practices akin to what has been described here has been previously reported in research on perceptions of PrEP among serodiscordant couples [31]. PrEP use has also been characterized as having a beneficial effect on serodiscordant relationships, strengthening the partnership [10, 32]. Close analysis of PrEP and ART dosing behavior by serodiscordant couples reveals reliance on joint approaches to adherence [33]. Taken together, this work suggests adherence to PrEP will be more thoroughly understood when examined in relational context; future research on PrEP adherence will be well-served by adopting a dyadic approach [34].

Examples of dyadic approaches to conceptualizing health behavior do exist to inform and guide future work. Lewis et al [35] envisage health behavior change for couples as stemming from interdependence in the relationship. They propose that interdependence can result in the partners adopting a cooperative, or “communal” approach to making changes in favor of health-promoting behavior. Focusing on HIV prevention specifically, Karney et al. [34] offer a dyadic conceptualization that specifies contextual influences at multiple levels of analysis (e.g. interpersonal, structural), to depict influences on cooperation between partners in the interest of safer sex.

The significance of fear and worry about acquiring HIV for uninfected, at-risk individuals merits greater recognition for its detrimental effect on quality of life. HIV-related anxiety received some early attention as a barrier to HIV testing [36-37]. More recently, researchers have documented the negative emotional impact of anticipated stigma resulting from HIV infection for HIV-negative gay and bisexual men [38]. Relief from fear of infection is emerging as a major benefit of PrEP for users across key population groups [39-40].

Our participants identified and described a set of challenges that shaped how they managed PrEP discontinuation within the context of prevention-effective adherence. Supporting future PrEP users as they work to align PrEP use with variable seasons of risk is vital to facilitating an effective scale up of prevention-effective adherence across the region. Our results suggest several avenues for intervention. Up-to-date information about risk reduction that includes emphasis on the preventive effects of ART treatment may help uninfected persons assess their risk, and thus their need for PrEP, more accurately. The Undetectedable=Untransmissable (U=U) campaign may be particularly useful in communicating in clear, unambiguous terms that an individual who is living with HIV but virally suppressed cannot infect his or her partner [20].

Counseling for couples that sets standards for behavior in successful partnered relationships may increase women’s ability to negotiate sex, alleviating the felt need to rely indefinitely on PrEP for HIV prevention. Encouragement from counselors to share information about ART adherence and the results of routine monitoring may increase uninfected partners’ confidence in protection afforded by ART treatment, facilitating discontinuation of PrEP. The varied responses to PrEP discontinuation identified here, and the clear importance of the relationship context in shaping these responses, point to the value of a tailored approach to providing formalized support to individuals discontinuing PrEP.

This paper has characterized PrEP discontinuation for users in Ugandan serodiscordant couples. PrEP discontinuation is a key component of prevention-effective adherence, in which PrEP use is aligned with periods of high risk. It is of interest to consider what prevention-effective adherence might mean for other populations of PrEP users, for whom patterns of change in risk levels and capacities for continuous assessment of risk might differ. Prevention-effective adherence in varied populations of high risk groups is an important topic for future research.

We acknowledge the following limitations of this analysis. First, as qualitative research, we do not claim or seek generalizability of results. Rather, we draw insights from interview data aimed at deepening understanding of PrEP users’ inclination to “practice” prevention-effective adherence. Second, the data represented here reflect the experiences of one particular group of PrEP users: HIV-uninfected individuals in serodiscordant relationships living in a single location and participating in a demonstration project of PrEP. The experiences of other types of PrEP users will not be the same. Finally, we acknowledge the possibility that the data may be subject to social desirability bias. However, we believe this risk is offset by the fact that participants were not reporting on a behavior they expected to be judged on by others.

Conclusions

As PrEP scale up moves forward in sub-Saharan Africa and users begin to practice prevention-effective adherence, it will be important to understand and address the real-life challenges presented by a potentially cyclic model of medication use. This analysis takes a first step-- detailing the sense of loss, the perception of increased risk of HIV infection, and the varying efforts to offset perceived increased risk that accompanied discontinuation of PrEP for a sample of PrEP users in Ugandan serodiscordant couples. Results suggest discontinuing PrEP will be challenging. Interventions that are sensitive to the acute concerns that individuals face during this transition may facilitate accurate assessment of changing risk, and help users feel confident in using PrEP as circumstances dictate for prevention-effective adherence and HIV prevention.

Figure 2:

Responses to Feeling at Increased Risk of HIV Infection Following Discontinuation of PrEP for Ugandan PrEP Users Participating in a Qualitative Study [July 2014-May 2016]