Oral PrEP pause and restart among adolescent girls and young women (AGYW) who initiated PrEP in an HIV prevention program in Kampala, Uganda: a qualitative study

Abstract

Introduction

There is interest in PrEP uptake amongst young people at high risk of HIV. However, there is a significant decline in PrEP starting from three to six months after PrEP initiation, thus affecting PrEP persistence and continuation due to pausing and stopping of PrEP, with a few restarting it. Understanding why AGYW pause PrEP use and what motivates the few to restart is critical for enhancing PrEP adherence and continuation. We explored pausing of PrEP and restarting it amongst AGYW in an HIV prevention program, in Kampala, Uganda, to inform future HIV health education interventions aimed at strengthening PrEP continuation.

Methods

Between November 2023 and March 2024, a qualitative study using 17 follow up interviews was carried out in Kampala, Uganda at month six of PrEP initiation. Participants were purposively sampled from AGYW aged 14 to 24-years-old who were HIV-negative and at high risk of acquiring HIV. Data were transcribed verbatim, coded in Nvivo 14, analyzed thematically using iterative categorization, and interpreted using the Health Belief Model.

Findings

PrEP side effects were the major reason for missing and eventually pausing. Mobility of AGYW led to pausing PrEP, as some reported forgetting to travel with their pills or travelling with inadequate pills, and lack of access to PrEP in new locations. Anxiety about using PrEP during pregnancy and having other new health conditions requiring long-term oral medication led to pausing of PrEP. However, when they pause, they mitigate the risk by using other HIV prevention methods such as condoms. The major reasons for PrEP restart were high HIV risk awareness, including having multiple sexual partners of unknown status. Inconsistent condom use, due to men refusing or removing condoms and higher pay for condomless sex during sex work, motivated restarting PrEP. Lastly, future marriage prospects and getting into long-term relationships motivated AGYW to remain HIV-negative through PrEP use.

Conclusion

Some AGYW who pause PrEP intend to restart in the future and, during the period of pausing PrEP, they mitigate HIV risk by using other HIV prevention methods. Therefore, tailored comprehensive HIV-prevention messages should be implemented to normalize challenges in the PrEP journey, emphasizing the importance of restarting PrEP and of using other HIV prevention methods during the pause period to enhance HIV prevention persistence and continuation.

Introduction

Adolescent girls and young women (AGYW) in sub-Saharan Africa remain at high risk of acquiring Human Immunodeficiency Virus (HIV) [1, 2], accounting for approximately 25% of new HIV infections globally, and they have large unmet needs for HIV prevention [3]. In 2021, approximately 250,000 AGYW were infected with HIV worldwide, and six out of seven cases of HIV infections among adolescents (aged 15–19 years) in East and Southern Africa occurred among girls [4]. 77% of AGYW who acquired HIV in 2023 live in sub-Saharan Africa (SSA), including almost two thirds (60%) in Eastern and Southern Africa [5]. AGYW in SSA are thrice as likely to be living with HIV than their male peers and they may acquire HIV five to seven years earlier than their male peers [6, 7]. In Uganda, AGYW are thrice as likely to acquire HIV infection than young men [4], accounting for 29% of new HIV infections nationally [3].

Given this great burden of HIV in SSA, there is a need for prevention strategies to minimize the risks of HIV acquisition amongst AGYW. The World Health Organization (WHO) recommends oral pre-exposure prophylaxis (PrEP) to be offered as a prevention choice for people at substantial risk of HIV infection [8–10]. Oral PrEP is highly effective in preventing HIV infection if adherently used [1, 11], thus holding promise for HIV prevention among AGYW. Demonstration projects conducted from 2016 to 2019 in South Africa, Kenya and Zimbabwe have shown high PrEP uptake, with 93–95% of AGYW accepting PrEP the first time it is offered to them [2, 12]. However, data show that only one-third of AGYW persist with PrEP use beyond three months, with as few as six to eight percent using PrEP beyond six months after initiation [2, 12, 13] and rapidly dropping adherence within the first six to twelve months [2, 12, 14], thus affecting PrEP persistence [2, 12, 13, 15]. Various studies indicate that AGYW pause PrEP because of fears of trying out something new due to concerns about PrEP side effects, not liking or being able to take daily pills, unsupportive partners, reduced HIV risk, change in relationship status, and PrEP access barriers [14–17]. However, some of these studies do not indicate whether those who discontinued PrEP later restarted.

Evidence from some of the PrEP studies indicate that a small proportion of the AGYW who discontinue PrEP later restart [18, 19]. Mayanja et al. (2022) found low PrEP adherence and continuation among AGYW in Kampala, with 55 out of 81 PrEP starters discontinuing PrEP, and only four of them restarting [20]. Similarly, Baron et al. (2020) reported three participants out of six participants willing to restart PrEP in South Africa, but, apart from the influence of club meetings, no other motivating factors were discussed that could have influenced PrEP restart [21]. Rao et al. (2023), in a larger South African study of 40,000 volunteers, found that 9% of female sex workers (FSWs) and 3% of AGYW re-initiated PrEP within a year, although reasons for re-initiation were not covered, it was assumed that FSWs’ knowledge of PrEP played a role [22]. Ndimande-Khoza et al. (2023) further found familial influences to have impacted PrEP use and restart decisions, particularly among AGYW [23]. Whilst several studies have explored PrEP uptake and adherence [24–29], there is limited data on PrEP use among AGYW at high risk who initiate and discontinue PrEP at least once but restart it, and what motivates them to restart PrEP. Therefore, in this study we explored reasons for pausing and restarting among AGYW who initiated PrEP in an HIV prevention program in Kampala, Uganda. Understanding the motivations for restarting PrEP after discontinuation can inform health education practices, strengthen HIV prevention programs and improve PrEP adherence.

Methods

Study design, sampling and site

A descriptive qualitative study was carried out between Nov 2023 and Mar 2024 using in-depth interviews (IDIs) among 14–24-year-old AGYW in Kampala, Uganda, who are at high risk of contracting HIV, to investigate factors influencing pausing of PrEP and restarting. The study was nested within a randomised controlled trial (RCT) aiming to assess the effect of peer support on oral PrEP uptake and adherence. The 394 RCT participants were recruited from urban slums, enrolled at the AIDS Information Centre (AIC), Kampala and randomly assigned to health worker counselling (control) vs. health worker counselling plus peer support (intervention). The criteria for selecting AGYW to take part in the RCT were: being HIV and Hepatitis B-negative, not pregnant, having never used PrEP and willingness to take part in the study [30]. In this study, the findings we use are from follow-up IDIs with participants purposively selected from age range (14–19 and 20–24), study arm (control and intervention groups) and having initiated PrEP. Participants were contacted by phone to come for the interviews.

Ethical considerations

The study received ethical clearance from the Uganda Virus Research Institute ethics committee (ref: GC/127/918), the Uganda National Council for Science and Technology (ref: HS2490ES) and the London School of Hygiene and Tropical Medicine (ref: 28193). Participants took part in in-depth interviews after providing written informed consent. The information forms were read to the participants in Luganda, the most common language used in Kampala. For those who could not read and write, both the participants and an impartial witness identified by the participant had the consent forms read to them and if the participant agreed to participate, the witness wrote the participant’s name, and the participant put a thumbprint on two copies, one for the study site and another kept by the participant. Interviews were conducted in private places comfortable for both the researcher and the participant to guarantee privacy. Pseudonyms and unique numbers were used instead of their names for anonymity.

Data collection

IDIs were conducted on 17 AGYW after six months in the RCT when the peer support intervention had ended. A semi-structured topic guide exploring various topics, including PrEP use and experiences, was used by two trained female research assistants. Interviews were conducted in Luganda, the local language participants were familiar with, and they lasted averagely 45 min ranging between 30 and 60 min. Interviews were audio-recorded using a voice recorder after obtaining written informed consent. Interviews were conducted from the AIC clinic facility in private rooms.

Researcher positionality and reflexivity

The study team comprised four researchers (three females YM, IK and ZN, and one male LM) with backgrounds in social and behavioural science (LM, IK, ZN) and YM with a background in medicine. The RCT team included one female counsellor and two nurses. LM, IK, ZN and YM had prior experience in qualitative research, and everyone on the team had prior knowledge in HIV prevention, and familiarity with the social context of Kampala were the study was conducted. These aided relating with the study participants and attending to them based on the prior experience of the team in conducting HIV prevention programs in Kampala. YM, IK, ZN and the study counsellor and nurses have done similar work in the same population before which helped in participant mobilisation, and handling due to prior interaction with participants of similar characteristics. LM who had not worked in this kind of population before relied on the research team to understand the different dynamics of the study population.

Interviews were conducted by two younger female research assistants (IK and ZN), who were selected partly because their gender and relative social proximity to participants were considered important for facilitating open discussion on sensitive issues such as PrEP use, relationships, and sexual health. The male team member (LM) and principal investigator (YM) did not conduct interviews in order to minimise potential power and gender-related barriers to disclosure from LM. In addition, the interviewers were asked not to take the role of counsellors for emotional stories from participants, since we had a counsellor on the team, the counsellor (a female with extensive experience working with AGYW in Kampala) handled these issues that needed counselling. The interviewers also referred participants who had medical concerns such as pregnancy related issues, wanted more HIV prevention services or treatment of STI’s to the nurses for medical attention.

Data coding and analysis involved all four team members. We recognise that researchers’ disciplinary training, gender, and social positions may shape interpretation. To enhance reflexivity, emerging codes and themes were discussed collectively, including consideration of how gendered and contextual perspectives influenced interpretation. Interdisciplinary discussions also helped examine biomedical issues raised by participants, such as the participant perceived effects of PrEP on pregnancy, LM a social behavioural scientist asked YM a medical doctor for clarification on the effects of PrEP on pregnancy.

YM was the most senior person on the team, a medical doctor who was doing her PhD, and the project PI, LM and IK were educated to Masters and ZN to Bachelors. However, despite the differences in academic profile, everyone’s contribution based on their area of strength, YM was the PI of the study, and we always consulted her on biomedical explanations during the analysis and interpretation and the general trends in HIV. LM, ZN and IK were social scientists with experience in qualitative research and supported YM with the social science approaches. Thus, interdisciplinary teams and having people who are locally situated support interpretation of different findings and contextualising them.

Data analysis and management

Data were transcribed verbatim and translated into English by three trained social scientists. All the transcripts were checked and cleaned by IK before uploading them to a secure server at the MRC/UVRI&LSHTM with access to only the study staff.

Data were coded both inductively and deductively [31] in NVivo 14 to allow the new codes to emerge during the coding process. Coding was done by IK, ZN, LM and YM. The coded data were analyzed thematically using iterative categorization [32]. An initial codebook was developed using four transcripts representing two transcripts each from control and intervention. All four transcripts were reviewed by the four-study staff to identify the initial codes; the coded transcripts were then compared across the different researchers to refine and agree on the codes, which were used to develop the draft codebook. This draft codebook was tried on four other transcripts, still using two scripts from control and intervention to test if the draft codebook suited other transcripts. This formed the final codebook that was imported to NVivo 14 and used by the team to code all the transcripts. To ensure intercoder reliability between coders, we started by coding one similar transcript amongst the four team members, and after we had a discussion to compare if we were all interpreting the codebook in the same way and coding consistently. Any differences in coding were discussed among the coders until we reached a mutual agreement on how best to code the text. We applied this same approach with another transcript and then proceeded to code independently with regular consultations and check-ins to discuss any challenges that arose in the coding process. After coding, the coding report for each code were exported to Word documents to do line-by-line analysis using iterative categorization and summaries were created for each code [32]. Data were analyzed with a focus on PrEP knowledge, understanding, perceptions and experiences.

We followed the steps described by Neale (2021) [32] to conduct iterative categorization and line-by-line analysis. LM, who had previous experience conducting line-by-line analysis using iterative categorization [33], led this process and trained the research team.

After exporting the coded data from NVivo into Word documents, we created a folder called analysis in which each code was assigned three separate Word files. Using the sub-code risk perception as an example, the files were named; (1) Risk perception codes, (2) Risk perception analysis, and (3) Risk perception summary. We then opened the risk perception codes and risk perception analysis documents side by side on the computer screen. The coded data were read line by line, and the meaning conveyed in each line was analysed and interpreted. Each identified meaning was recorded in the risk perception analysis document together with its source.

Where similar meanings were identified across different lines, these were grouped together and the corresponding sources were recorded under the same category. This enabled us to identify how many participants expressed similar views across the coded data while also capturing different meanings emerging from the same line. The process facilitated the identification of patterns and variations in participants’ accounts, including perspectives reported by the majority and those expressed by only a few participants. Following this stage, a one-page summary was prepared for each code, highlighting the key interpretations generated through the line-by-line analysis.

Health belief model

The findings were interpreted using the Health Belief Model (HBM). HBM is a psychological framework for explaining and predicting health behaviors based on people’s beliefs and attitudes. It was created in the 1950s by social psychologists Hochbaum, Rosenstock, and Kegels of the United States Public Health Service [34–36]. The paradigm was originally designed to explain why people did not engage in health prevention behaviors such as tuberculosis (TB) screening [35], but it has since been extended to a variety of health conditions such as HIV, cancer and vaccine uptake [37–40]. We used this model to categorize factors that influence pausing of PrEP and what motivates restarting it.

The Health Belief Model is built around 5 constructs that influence health behavior, which we applied to interpret the findings. The first construct is perceived susceptibility, referring to individuals’ beliefs about their risk of developing a health condition, because the more vulnerable the person feels, the more likely they are to use preventive measures. In this study, we considered the individual perceived risk of AGYW to get HIV. The second construct is perceived severity, which relates to a personal belief about the seriousness of the condition and its potential consequences, which may result in taking an action, in this study we looked at participant perceived severity of HIV. The third construct, perceived benefits, refers to the effectiveness of the health action taken to avert the risk or severity of the disease; with this model, the belief that health action will be of benefit influences its use. In the context of this study, we considered how the perceived benefits of PrEP influenced restart. Fourth is the perceived barriers, which focus on the perceived obstacles in adopting new health behavior such as financial, physical or emotional factors. For this study, we considered the factors that hinder PrEP use leading to pausing of PrEP. The fifth construct is the cues to action, which are the triggers that prompt an individual to take part in healthy behavior, and we considered the triggers that led AGYW to restart PrEP and the process involved in pausing it.

Findings

We start by providing the participant characteristics listed below in Table 1.

Table 1.

Demographic characteristics of AGYW invited for Follow-Up In-depth interviews (N = 17)

Characteristics	Categories	Intervention	Control	Total
Age range	14–19	4	6	10
	20–24	4	3	7
Relationship status	Married	1	0	1
	Single (Never married)	5	8	13
	Single (separated)	2	1	3
Educational level	Primary	4	5	9
	Secondary	1	2	3
	Tertiary	1	1	2
	None	2	1	3
Occupation	Student	0	1	1
	Sex work	6	7	13
	Entertainment	2	1	3
PrEP use after enrolment	PrEP initiation	8	9	17
	Paused PrEP at least once	6	9	15
	Never paused PrEP	2	0	2
	Restarted PrEP	5	1	6
	Hoped to restart	0	8	8
	No hope to restart	1	0	1

We present the findings starting with pausing of PrEP and then restarting PrEP. Of the 17 AGYW who initiated PrEP, 15 participants (six in intervention and nine in control) paused PrEP at least once during the study. Generally, from the qualitative accounts, PrEP pause varied with some pausing it within the first few weeks due to side effects although other factors also contributed, while others paused after a month of use. Pausing later was mainly due to re-evaluation of risk, pregnancy, and mobility. Participants discussed various reasons for pausing PrEP as discussed below in line with the constructs of the Health Belief Model.

Perceived susceptibility: The perceived susceptibility that led to pausing of PrEP was the reduced HIV risk perception due to changes in sexual behavior. Those who were in a stable relationship or staying with their partner and not active in sex work during that period perceived themselves to be at a lower risk of getting HIV and thus paused/stopped PrEP.

“I (Interviewer): How long did you take while taking PrEP?

R (Respondent): Two months.

I: Then you stopped?

R: Yes.

I: Why did you stop?

R: Because I saw I was out of risk of getting HIV.

I: You say you were out of risk of getting the virus, what were the indicators that you were out of risk?

R: Because I was in relationship with one man and he’s the one they knew officially. IDI01-14DEC2023-FU.

Perceived severity: The perceived severity that made AGYW pause PrEP was linked to fears about the consequences of taking (and not taking) PrEP. The prevailing health conditions and concurrent use of other oral medication made some AGYW pause PrEP during periods of illness because they were already on prolonged oral medications and felt that adding the daily PrEP pill alongside these would mean taking too many tablets.

“I have been using it but now I have decided to pause it a bit because I am taking a lot of drugs. The last time I was here, I was given drugs because I had an infection. When I went to the hospital for a scan, they found out that I had [condition x] and they also gave me a lot of drugs because I did not want to be injected. I realized that the drugs were for 2 months, and the infection had not yet cleared. When I thought of adding PrEP, I realized that the drugs were too many for my body to handle” IDI03-27NOV2023 –FU.

“I first paused taking them [PrEP] because I had fallen sick with a certain disease, but I thought…, I was wondering whether it was caused by the contraception I was using, or it was caused by PrEP. So, I was there (confused). I even thought I had cervical cancer”. IDI11-13DEC2023-FU.

In addition, some of the AGYW paused PrEP when they got pregnant due to fears that PrEP would affect their unborn baby. “I stopped taking PrEP from the time when I realized that I was pregnant. Then I stopped using it” IDI22-29NOV2023-FU. However, they had plans to resume after giving birth. “In my opinion after giving birth, I may start [PrEP] again” IDI01-14DEC2023-FU.

Perceived barriers: The presence of side effects was the main reason for pausing PrEP. Participants reported getting side effects at the start such as headache, nausea, vomiting, body weakness and others, and decided to first pause or stop PrEP.

“I swallowed at the inception of the study, the side effects were too much, and I first paused it, however, I restarted it”. IDI02 -−27NOV2023-FU.

“I: Did you ever start using oral PrEP while in this study?

R: Yes.

I: How long did you take while swallowing these pills?

R: Three months

I: Three months and you said, “you later discontinued them”?

R: Yes.

I: What influenced you to stop taking the pills?

R: Side effects

I: Side effects like what?

R: I used to vomit; I experienced constant headache.

I: You have said vomiting?

R: And headache; when I couldn’t do my duties well, you are ever vomiting

I: Okay, how long did the side effects take?

R: They took long because whenever I would swallow them… I used to swallow them every morning and I would spend the whole day in bed”. IDI21-07DEC2023.

The second barrier was mobility that caused pausing of PrEP. Some AGYW reported going to distant places either for formal work or sex work and visiting relatives/friends; for some, their PrEP pills got finished before they returned to Kampala, and they missed their refill appointments as they had moved far from the study clinic.

“I got a trip, I had a research whatever I was attending in [district X, 500 kilometers away from the study clinic]. I travelled with the pills though I didn’t realize that it would not be enough because I was unable to estimate the period, we were to spend there. So, the pills were used up, and I was very far from Kampala, and I had no access, so I decided to pause”. IDI03-27NOV2023 -FU.

“I: I want to know why you stopped. Because you said you started, so why did you stop using PrEP?

R: I will not say that I got a problem with PrEP but it is just that we would have to go to some places, since we normally go and dance in those bars. So, when we went there, it was far and yet I had taken it [PrEP dose] and it got finished. I stayed there for long; I spent about 2 months there. By the time I got back here, my scheduled appointment date had passed" IDI05-27NOV2023-FU.

In addition to the above, the AGYW reported being unable to access PrEP in the areas they moved to for formal work or sex work, and this resulted into stopping/pausing of PrEP.

“I: So, when you started moving around, you stopped using it?

R: Mm (yes). Because when I am that side (up country and islands), I cannot get the drugs you understand? It is difficult to get them…” IDI07-30JAN2024-FU.

This indicates PrEP access challenges in hard-to-reach areas of the country which lead to pausing of PrEP hence affecting PrEP persistence.

Financial constraints also caused AGYW to pause PrEP due to the lack of money for transport to pick up PrEP refills and for meals required to take daily PrEP. The financial constraints made participants prioritize other competing life needs such as paying rent, staying at their workplace opposed to spending money on picking up PrEP.

“R: We had gone to Kanyanya.

I: Kanyanya. Okay, let me ask, like how much is the transport from here to Kanyanya?

R: Eh musawo (health worker), it would be much!

I: Like how much? Try and think about it

R: It can cost maybe 10,000/=.

I: So, you saw that that was a lot of money and you could not come back.

R: I could not go and then come back. Remember you may not get much [income], some men can use you, he doesn’t give you but tells you that he will send you yet we had to pay rent because we had got a small house. So I couldn’t manage it”. IDI05–2-IK-27NOV2023-FU.

Despite 15 participants pausing PrEP at least once, six participants restarted, and eight participants had hopes to restart. Most participants who paused and restarted PrEP were mainly in the intervention arm (five out of one) and majority aged 14–19 years (four out of six). However, there were no differences observed in factors for restarting amongst those who paused earlier or later. The motivations for PrEP restart and/hope to restart were as below in line with the HBM constructs.

Perceived susceptibility: High HIV risk perception was the major motivator for PrEP restart as most participants felt they were still at high risk of acquiring HIV due to the nature of their job (sex work), having many sexual partners of unknown HIV status, some of whom refused HIV testing and sometimes had sex while drunk. This high self-perceived risk motivated them to restart PrEP to prevent HIV infection.

“What motivated me to restart PrEP is because I know my job that I do, it has issues with the sickness (HIV); so, I want to get that protection when every time I am safe so that I don’t get the virus (HIV) because I am still doing the job (sex work)”. IDI24-28NOV2023-FU.

“Because I have more than one person who I told you I’m in a relationship with and it is not like all of them allow to use [HIV prevention] methods. Some even refuse to go for blood testing”. IDI06–06DEC2023-FU.

Perceived benefits: The perceived benefits of restarting PrEP included some of the participants considering PrEP to be a better method of controlling HIV because it not only offered protection but was also more reliable, as compared to other methods such as condoms and HIV testing. Oral PrEP was reported to have reduced anxiety while having sex.

“I: Let me ask, in the future, would you want to restart using PrEP?

R: Mm (yes)! It is not bad by the way. Even though condoms, okay all of them are helpful, but PrEP helps so much like 100%. Because a condom can burst along with other things which is not the case with PrEP”. IDI07-30JAN2024-FU.

In addition, participants reported uncertainties around other HIV prevention methods such as the option of condom use they had used during the period they paused PrEP. They reported that some men did not like to use condoms, yet such clients offered a lot of money for condom-less sex. This motivated them to have hopes of restarting PrEP to earn more money from condom-less sex. It was reported that some clients accepted using condoms but removed them during sexual intercourse. Some AGYW had fears that condoms could burst during sexual intercourse and also argued that condoms not being available at their workplaces sometimes would put them at risk.

R: I will restart [PrEP].

I: Why do you expect to restart it?

R: “Because I also fear HIV, I fear it and you should know that with the other method [condom], it is not that every time the condom can be available. Do you see it? So, you would be at risk if it is not available or if it bursts in the process, you see that you would have gotten HIV”. IDI11-13DEC2023-FU.

Another perceived benefit was the future marriage prospects that motivated AGYW to restart PrEP because they wanted to keep themselves healthy so that in the future, when they got a marriage partner, they would navigate HIV testing without anxiety.

“I: So, would you like to restart PrEP in the future?

R: Yes.

I: What would influence you to use PrEP in the future?

R: Right now, I do not expect that I am going to leave this job because it gives me a lot of money. So that means that, since I mind about my health, I don’t want to get sick even though I am sex worker, I have to swallow PrEP because it will help me in the future, in case I am to leave sex work I can get a man, but that is when I am healthy” IDI07-30JAN2024-FU.

Cues to action: AGYW made some actions before and during pausing of PrEP. Firstly, a few of them sought medical advice from the healthcare workers before they paused. Some of them received health worker counselling that enabled them to continue taking it after being informed that the side effects were temporary.

“I have been with the health worker and asked her for a talk or counseling session before I make a final decision, though I had thought of pausing PrEP first, I wanted to get guidance from the health worker…. I wanted medical advice from the health worker who understands this”. IDI03-27NOV2023 -FU.

“I had ever…I used them but the reason I quit them was just as I told you [side effects], but the health worker explained to me that it is temporary and that’s what will influence me to get back on them”. IDI08-01DEC-2023-FU.

It is important to note that most of these girls, even after pausing PrEP, substituted it with other HIV prevention methods such as condoms or regular testing, indicating continued risk mitigation measures while not on PrEP.

“I: So, when you started moving around, you stopped using it [PrEP}?

R: Because when I am that side, I cannot get the drugs you understand? It is difficult to get them that is why I was using condoms, because they are everywhere. You can find them everywhere” IDI07-30JAN2024-FU.

Some girls were motivated to think about restarting PrEP due to peer/group influence because they were the only ones not using PrEP among their peers. One girl who had paused taking PrEP felt inconvenienced carrying boxes of condoms yet her friends who continued taking PrEP were not carrying them.

I: Okay. So now, they take PrEP?

R: They take; I am the only one who is not taking.

I: Okay. Is there a way it makes you feel when you see your friends taking PrEP?

R: Yes, because you be moving around with a box of condoms yet for them, they go free, when they don’t have anything. It somehow makes me feel bad.

I: Okay. Now you feel okay, you want to come back and get PrEP?

R: Yes. I am about to come back and get it again. IDI05-27NOV2023-FU

Discussion

In this study, we used qualitative interviews to explore pausing and restarting of PrEP amongst AGYW who were mainly engaged in sex work, who initiated PrEP in an HIV prevention program in Kampala, Uganda. We mainly found that AGYW who paused PrEP restarted at some point and others had hopes to restart later. The main reasons for pausing PrEP were the perceived barriers, such as the presence of side effects, or perceived susceptibility, such as reduced perceptions of HIV risk, while PrEP restart or intentions to restart were mainly motivated by the perceived severity, such as high HIV risk perception. However, most importantly, during the period they paused PrEP, they used other HIV risk mitigation measures such as condom use.

The study’s findings are interpreted in light of the Health Belief Model [37, 38], which helps to explain AGYW’s decisions to pause or restart PrEP use depending on their perceived health risks and benefits. Perceived barriers, such as side effects, mobility, PrEP access during travel far away from the study clinic, and financial constraints, were barriers to continuing PrEP use. The perceived susceptibility was also evident when some AGYW in stable relationships or with less engagement in sex work had lower perceived risk of getting HIV, resulting in a pause on PrEP. On the other hand, the high perceived HIV risk among others who continued to engage in sex work and had multiple partners of unknown HIV status motivated PrEP restart. These decisions are consistent with the model’s premise that people act based on their beliefs about the likelihood of developing an illness [34], making it important to understand people’s beliefs about an intervention such as taking oral PrEP for it to succeed.

In this study, we found that side effects were the main reason for pausing of PrEP. This is similar to various PrEP adherence studies conducted in Uganda [20, 24, 25, 41] and elsewhere in East Africa [16] and South Africa [15, 42]. What differentiates our study is the observation that some AGYW consulted healthcare workers or counselors before deciding to pause PrEP. During these discussions, some AGYW received explanations about the temporary nature of the side effects, leading them to continue PrEP, while others opted to pause but expressed intentions to restart later. Our study’s contribution is supported by the recommendation of Reddy et al., (2023) where they stress that challenges experienced during PrEP use could be managed by the provision of effective counselling and management of side effects [42]. Indeed, those in our study who sought counselling planned to continue or continued with PrEP after pausing. The few individuals hoping to restart PrEP even after side effects implies that AGYW should be encouraged to always consult with the healthcare workers before choosing to pause or stop. This is more likely to lead to safe pauses and to increase PrEP restart and continuation, hence improving adherence.

The need to travel far from their residences for sex work or other work also resulted in AWYG pausing of PrEP. This concurs with other studies that found mobility of participants to be affecting PrEP adherence in Uganda [24], in Namibia [43] and elsewhere in South Africa [14]. This necessitates the need for long-acting PrEP, like lenacapavir, which can be offered twice a year and can alleviate challenges of AGYW missing PrEP refills or failing to find PrEP providers due to mobility. Rossue et al. (2021) in South Africa implemented a successful community based mobile clinic to provide PrEP to AGYW; however, mobility of AGYW was a challenge as the mobile clinic van would be in their communities when they were away [44]. Like in our study, this indicates that mobility is a big challenge to adherence and points to the need for making PrEP options, including long-acting injectable PrEP, widely available and accessible to overcome some of these challenges, and also the need to have more accessible HIV prevention options such as the dapivirine vaginal ring that has been found acceptable in clinical trials in Uganda by both men and women [45] and other sub Saharan African countries [46], when accessible, it can be an option alongside oral and injectable PrEP. Related to mobility, PrEP access has been identified as a challenge in achieving PrEP adherence [44]. Some of the AGYW in our study had PrEP access challenges in the areas they moved to for sex work. Therefore, in addition to providing long-term injectable PrEP, provision of PrEP in hard-to-reach areas that are frequented by young women selling sex can also help to solve access challenges to mobile AGYW. Butler et al., [47] offered remote PrEP delivery services during COVID-19 through placing PrEP in local private pharmacies, community based organizations and university halls of residence. Similar models could be tested and adopted or tailored to different contexts to ensure availability of PrEP in local pharmacies/clinics in the hard-to-reach areas to improve access challenges.

The World Health Organization (WHO) recommends offering PrEP to pregnant and breastfeeding women who are at high risk of acquiring HIV [8]. In our study, some participants paused PrEP upon conceiving due to concerns over possible harm to their unborn babies. Despite participants receiving information that oral PrEP could be used during pregnancy, this perception among participants could be attributed to the common knowledge that some drugs are dangerous to the fetus during pregnancy [41]. This finding is similar to other studies that found anxiety among participants in taking PrEP while pregnant [41]. This necessitates the need for robust health education by the healthcare workers, counsellors and peer educators addressing the fears around the unborn baby for pregnant mothers and lactating mothers as well. In other studies, such as a study conducted amongst pregnant women in South Africa, a large number of women were taking PrEP while pregnant and continued with sex work for up to 5 months and paused PrEP and sex work due to perceived dangers of having sex with multiple partners during pregnancy [41]. Our participants reported fear of taking PrEP while pregnant; however, even after pausing PrEP, some may have continued with sex work while pregnant, leading to a risk of mother-baby transmission in case they acquired HIV infection during that period. A woman’s risk of HIV acquisition doubles during pregnancy and breastfeeding [48, 49]. Therefore, even though it is widely perceived and known that women do not take some medications in early pregnancy, information about the proven safety of oral PrEP during pregnancy must be incorporated into PrEP messages to encourage PrEP continuation among those who become pregnant while taking PrEP to prevent HIV acquisition and onward mother-to-baby transmission.

The AGYW in our study paused PrEP while having other health conditions or taking other medications, as it was considered burdensome to take PrEP alongside other medications. This is similar to other studies in Australia where participants paused PrEP while taking other medications due to failure to manage schedules for taking both medications [50]. This necessitates the need to include the safety of taking PrEP alongside other medications in health education programs and counselling to foster PrEP persistence during such situations.

In our study, financial constraints led to prioritizing other competing needs such as rent and work resulting in PrEP pause. This highlights how health behaviors are largely shaped by other social and structural determinants of health such poverty. This is common is sub–Saharan African settings where poverty affects uptake of HIV prevention programs. Our findings are similar to a study in Central Uganda that found poverty and attending to work affecting PrEP continuation amongst fisherfolks [51]. In the same study, women reported more financial challenges and food insecurity compared to men highlighting the economic disadvantage and vulnerability of AGYW [51]. Financial constraints have hindered PrEP consistence elsewhere such as Zimbabwe and South Africa [52], Eastern, southern, and western Africa [53] and in other continents [54].

Our study also found reduced HIV risk perception as a reason for pausing PrEP. This is similar to other studies that found AGYW and FSWs pausing PrEP due to perceived low HIV risk [15, 16, 55–57]. This indicates changes in risk perception or risk re-evaluation. The baseline interviews of our study indicated all participants perceived themselves to be at high risk, but over time the risk perception changed based on sexual behavior. This is not only unique to our study, but similar with other findings where risk re-evaluation occurred due to factors such as changes in relationships and getting pregnant, which resulted in pausing sex work [14, 16, 41]. Risk perception is very crucial in PrEP uptake and continuation. PrEP uptake in this population is influenced by high HIV perceived risk [25, 58, 59]. However, what is unique with our study is that we found risk mitigation amongst AGYW who pause PrEP because of reduced risk perception. They replaced PrEP with other HIV prevention methods such as condom use or HIV testing, especially those who mentioned reducing sexual partners. This implies PrEP education and HIV prevention programs should also emphasize and promote other HIV prevention methods alongside PrEP, this would give AGYW other options in case they paused PrEP due to intentional or unintentional factors.

Whereas to some AGYW, reduced HIV risk perception resulted in pausing of PrEP, high HIV risk perception was the major factor that motivated PrEP restart, consistent with findings from other studies in East and Southern Africa [14, 42]. This indicates the importance of HIV risk re-evaluation on PrEP uptake, thus necessitating the need to have programs that make AGYW aware of their HIV risk. A study in South Africa recommends the use of HIV risk assessment tools to make AGYW recognize their HIV risk [42] because poor assessment of risk may lead to poor use of PrEP [23]. These risk assessment tools could be adopted at community levels as well as clinical levels or incorporated into sexual and reproductive health services to make girls aware of their HIV risk. Additionally, peer support groups should integrate discussions on HIV risk so that AGYW can learn from peers’ experiences and better assess their own vulnerability. The self-awareness of HIV risk can result in PrEP initiation and continuation after pausing.

Opposed to our study findings, a study in South Africa [57] found that the fear of being scolded by health care workers for missing refill resulted into fear to return for refill despite South Africa having good adolescent friendly services, however, it’s important to note that in this South African study, the fear reported is just perceived but not experienced. In our study, however, we did not receive cases of AGYW fearing to approach healthcare workers, this could have been the reason why AGYW consulted them before pausing of PrEP. The study team were well trained and had experience working with AGYW in previous HIV prevention studies, hence provided a friendly environment. Thus, AGYW should be encouraged to always make consultations without fear.

Study strength and limitations

We used in-depth interviews among purposively selected participants (PrEP starters) which offered rich and deeper insights about PrEP experiences, pause and restart. However, as with qualitative interviews, these findings may not be generalized to all populations but based on the depths of the discussion with the AGYW interviewed on pausing of PrEP and restart, the findings can be transferable to AGYW in similar settings and social economic demographics such as AGYW who practice sex work and dealing in hospitality and entertainment.

Data reported in this study is based on the individuals who were interviewed after six months in the study. They provided reasons for pausing PrEP and restarting within this time frame, but there is a need for understanding pausing PrEP and restarting in longitudinal studies, such as after one or two years and beyond to examine the changes in PrEP journeys over a long period. This would help to capture PrEP persistence amongst those who restarted it and compare factors based on duration such as six months, 12 months and above. This study does not report data on the persistence of those who restarted which would be picked from following up those who restarted over a certain period in a longitudinal study.

Conclusion

The findings suggest that some of the AGYW who pause PrEP typically do not stop entirely but rather interrupt their usage due to specific challenges encountered during their PrEP journey while maintaining hopes of restarting in the future. The study also highlights that AGYW often engage in risk mitigation strategies after pausing or discontinuing PrEP. Therefore, comprehensive tailored health education messages should be given, emphasizing the importance of resuming PrEP when challenges arise and promoting the use of alternative HIV prevention methods during periods of interruption e.g., pill burden due to taking oral medications for other illnesses and improving accessibility to long-acting injectable PrEP for highly mobile AGYW. Our findings further suggest that effective health worker support, and targeted education, particularly around creating risk awareness, managing side effects and addressing concerns during pregnancy, can significantly enhance oral PrEP continuation and restart rates among AGYW whose preferred HIV prevention method is the oral pill.

Abbreviations

PrEP

Oral pre-exposure prophylaxis

AIC

Aids information centre

I

Interviewer

R

Respondent

WHO

World Health Organization

AGYW

Adolescent girls and young women

HIV

Human immunodeficiency virus

SSA

Sub-Saharan Africa

TDF

Tenofovir disoproxil fumarate

RCT

Randomised controlled trial

HBM

Health belief model

MRC/UVRI & LSHTM

Medical Research Council/ Uganda Virus Research Institute and London School of Hygiene and Tropical Medicine Uganda Research Unit

IDI

In-depth interview

FSWs

Female sex workers

Author contributions

YM conceptualized and designed the study, developed the data collection forms, and was the principal investigator of the project. IK and ZN collected data. IK, ZN, LM did data transcription and translation, LM planned the analysis process (Iterative Categorization) and led the draft of the manuscript. IK led the analysis process, filling all transcripts on a secure drive and anonymizing them. IK, YM, ZN, and LM participated in developing the codebook, coding and iterative categorization. All the authors iteratively provided feedback and reviewed the manuscript.

Funding

This project was part of the second European & Developing Countries Clinical Trials Partnership (EDCTP2) program supported by the European Union (CSA2020NoE-3102). The contents of this manuscript are the responsibility of the authors and do not necessarily state or reflect those of EDCTP.

Data availability

The data used for the analysis of this project can be accessed through an email request to [datasharing@mrcuganda.org].

Declarations

Ethics approval and consent to participate

The study received ethical approvals, and the participants also provided written informed consent.

Consent for publication

The participants consented for anonymized information to be published.

Competing interests

The authors declare no competing interests.