What drives adolescent girls and young women’s decisions to persist on PrEP? Results of a comparative qualitative study from a DREAMS program in western Kenya

Abstract

Background:

Oral pre-exposure prophylaxis (PrEP) is a valuable tool to help end HIV, but persistent PrEP use is low amongst adolescent girls and young women (AGYW) in sub-Saharan Africa (SSA), who are at high risk for HIV. To improve persistent PrEP use in AGYW, more research is needed to understand what drives their decisions to continue with PrEP.

Methods:

We conducted a thematic, comparative qualitative analysis of in-depth interviews with 32 AGYW who participated in PEPFAR’s DREAMS program in western Kenya: 16 were randomly sampled among those who persisted with PrEP and 16 among those who discontinued. We compared results between these groups to explore drivers for the decision to persist with PrEP. We interpreted findings using the Integrated Behavior Model, which posits that the decision to persist is influenced by attitudes, social norms, and personal agency.

Results:

Three themes emerged that illuminated AGYW’s decision to persist with PrEP. First, having positive attitudes towards PrEP was insufficient to ensure persistence in absence of correct knowledge. All AGYW were positive about PrEP, but those who discontinued often had insufficient/misinformation about PrEP, specifically on its appropriate use during pregnancy/breastfeeding. Second, support from family and friends was critical to overcome societal stigma and could counter lack of support from romantic partners. Third, strong mentors supported PrEP persistence by increasing self-agency, but this was secondary in importance to the need for robust knowledge and social support.

Conclusion:

To improve PrEP persistence amongst AGYW in SSA, PrEP programs may benefit from increasing education delivered to AGYWs and their communities, specifically family and friends. In particular, messages that PrEP may be used when pregnant/breastfeeding should be reinforced. Other health programs for AGYWs may also benefit from an increased focus on educational messaging to program recipients and people in their existing social networks.

INTRODUCTION

Adolescent girls and young women (AGYW) (i.e., cis-gender females aged 15–24 years old) remain at a disproportionately high risk of acquiring HIV, despite the availability of highly effective HIV prevention methods (1,2). In 2023, an estimated 44% of all new HIV infections globally occurred among AGYW, with more than three-quarters of these infections occurring in sub-Saharan Africa (SSA) (3). Oral pre-exposure prophylaxis (PrEP) can significantly decrease the likelihood of women acquiring HIV when taken daily (4–6), but uptake and persistence amongst AGYW has remained low, including in high prevalence countries in SSA (7–10).

In 2014, the United States President’s Emergency Plan for AIDS Relief (PEPFAR) launched a multifaceted program comprised of biomedical, behavioral, economical, and social interventions to reduce HIV incidence in AGYW in SSA: the Determined, Resilient, Empowered, AIDS-free, Mentored, and Safe (DREAMS) Initiative (11,12). DREAMS has been implemented across the highest incidence HIV communities in SSA, providing HIV testing and screening, education subsidies, violence prevention programs, accessible contraception, comprehensive economic strengthening, health education sessions, and access to a comprehensive PrEP program, including a dedicated PrEP mentor, for AGYW (12). While PrEP has been a cornerstone of DREAMS, and PrEP across SSA has largely been provided through PEPFAR (12,13), AGYW persistence with PrEP through DREAMS has been low (8,14–16).

The challenge of retaining AGYW on PrEP, even in the well-financed DREAMS program, has fueled implementation science research to explore barriers and facilitators for PrEP use amongst AGYW in SSA. Research has identified many programmatic factors that can help or hinder PrEP implementation (e.g., sufficient supply of PrEP, decentralization of PrEP delivery, use of peer mentors, and community engagement) (17–26). Recent research has also increasingly identified the critical role of relationships in PrEP persistence (23,26–30). Despite this research, our understanding on how to improve PrEP persistence among AGYW in SSA remains incomplete. Applying behavior change models to PrEP research can further elucidate ways to improve PrEP persistence amongst AGYWs. The Integrated Behavioral Model (IBM) is one of the most comprehensive models to understand uptake of health behaviors, positing that the decision (or intention) to engage in a health behavior is the most important driver for behavior change (31); it further suggests that attitude, perceived norms, and personal agency directly drive this decision. We used the IBM to conduct a qualitative analysis to understand drivers of the decision to persist with PrEP amongst AGYW in a SSA setting. The results of this analysis can be used to help improve the design of PrEP programs and other health promotion programs for AGYWs in SSA.

MATERIALS AND METHODS

Study Design

We conducted a comparative, qualitative study to explore drivers of the decision to persist with PrEP in a SSA setting with a high burden of HIV amongst AGYW. Semi-structured in-depth interviews were conducted in 2019 amongst 32 AGYW participating in PEPFAR’s DREAMS program in Kisumu County and Homa Bay County, Kenya as part of a broader evaluation of PrEP (8,9). Sixteen AGYW who persisted with PrEP and 16 who discontinued PrEP were randomly sampled for interviews from 336 participants who were part of the primary study’s prospective cohort (9); none of the participants selected for interviews declined to participate or dropped out. An AGYW was defined as persisting with PrEP if she reported to be continuing PrEP and returned for refills at least once prior to returning for an interview three months post-enrollment. Persistence was determined through surveys, the results of which have been previously published (8,9).

Data Collection

All interviews were conducted by female Kenyan research staff using a semi-structured interview guide. Interviews were conducted in English or Kiswahili, based on the participant’s preference. Interviewers were recent university graduates who were trained and supervised by public health researchers who lead the study (CO and HG); they had interacted with participants previously in the broader evaluation. Interviews were conducted in person at a location selected by the participant where auditory privacy could be ensured, which was most often the DREAMS Safe Space. Each interview lasted no more than 60 minutes and was recorded (audio only) then transcribed and translated into English for analysis. In addition to interview data, basic sociodemographic and reproductive health data was pulled from surveys conducted in the larger, primary study (9). Additional details are documented in the completed COREQ checklist (Supplement A).

Setting

The PEPFAR DREAMS program in Kenya’s Kisumu County and Homa Bay County ran from 2015 to 2021, and included PrEP distribution, amongst other educational and economic programs (12). Kisumu and Homa Bay counties have the highest incidence of HIV in Kenya, accounting for approximately one-in-six HIV cases in the country (32). The program and location of this intervention have been previously described (8,9). Importantly, all participants in this study were engaged in the same PrEP program implemented through DREAMS.

Data Analysis

We conducted a rapid, inductive analysis of interview data using the framework method (33–35). Based on a preliminary reading of the transcripts, three researchers (AM, DT, and JC) identified recurring codes related to topics AGYWs discussed regarding their PrEP use (e.g., PrEP knowledge, attitudes towards PrEP, individual economic situation, living arrangements, ease of PrEP access, PrEP stigma, mentor relationships, partner/family involvement). We then created an Excel-based summary matrix to extract data, including quotes, related to these codes (Supplement B). Two researchers summarized key concepts from each participant using this matrix. Additional codes were added to the matrix as they emerged. This analytic process entailed one researcher (AM) first reviewing all transcripts to complete the summary matrix for each participant; a second researcher (DT) then reviewed all transcripts and the completed summary matrix, making notes on their agreement/disagreement with the information included in the summary matrix. Both analysts came to agreement on the completed summary matrix before using it to identify themes.

Researchers reviewed the summary matrix to compare findings between AGYWs who persisted and those who discontinued, identifying core themes based on these findings. We enumerated positive and negative derivations of core themes for each participant and compared these counts between groups (36). We interpreted themes using the IBM, specifically the three areas that the model posits contribute to a decision to engage in a health behavior: attitude, perceived norms, and personal agency (Figure 1) (31).

Figure 1. Overview of the Integrated Behavior Model as applied to PrEP persistence.

This diagram is adapted from published versions of the Integrated Behavior Model. Definitions for these components have been taken from the companion materials for Glanz, Rimer, and Viswanath’s Health Behavior and Health Education: Theory, Research, and Practice, 4^th Edition (Accessible via: https://www.med.upenn.edu/hbhe4/part2-ch4-integrated-behavior-model.shtml)

Ethics

Ethics approvals for this observational human subjects research study were obtained from the Institutional Review Boards (IRB) at the Kenya Medical Research Institute (Kisumu, Kenya), PATH (Seattle, WA), and Fred Hutchinson Cancer Research Center (Seattle, WA). Written informed consent was obtained from all participants interviewed as well as those enrolled in the broader PrEP adherence study. All data used in this analysis were de-identified.

RESULTS

On average, the 32 AGYW interviewed in this study were 22–23 years old, were equally likely to reside with relatives or a romantic partner, and generally had children (Table 1). There were few differences in sociodemographic or reproductive health measures between AGYW interviewees who persisted with PrEP and those who did not. The most prominent differences were related to education and the number of partners, both of which were greater in the group that persisted with PrEP.

Table 1.

Overview of study participants, based on persistence with PrEP during study period

	Persisted (N=16)	Discontinued (N=16)	Total (N=32)
Sociodemographic Variables
Age
Average (SD)	23.0 (1.6)	22.4 (1.4)	22.7 (1.5)
Range	19–25	19–25	19–25
Living Situation
Resides with relatives	9 (56%)	7 (44%)	16 (50%)
Resides with partner	7 (44%)	9 (56%)	16 (50%)
Has children	12 (75%)	13 (81%)	25 (78%)
Currently in school	7 (44%)	3 (19%)	10 (31%)
Highest education
Primary	2 (13%)	5 (31%)	7 (22%)
Secondary	10 (63%)	11 (69%)	21 (66%)
Secondary +	4 (25%)	0 (0%)	4 (12%)
Reproductive Health Variables
Uses contraception	11 (69%)	10 (63%)	21 (66%)
Has multiple partners	5 (31%)	1 (6%)	6 (19%)
Partner engages in high-risk behaviors	8 (50%)	8 (50%)	16 (50%)
Has a partner who is HIV+	2 (12.5%)	0 (0%)	2 (6%)

Three themes emerged in our analysis that illuminate how attitudes, perceived norms, and personal agency contribute to an AGYW’s decision to persist with PrEP (Figure 2). We provide an overview of these themes below. A comparison of these themes by PrEP persistence can be seen in Table 2. An additional case-level breakdown can be found in Supplement C.

Figure 2. Summary of key themes as interpreted through the Integrated Behavior Model.

Drivers of decision to persist are indicated by the ‘+’ symbol. Drivers of discontinuation are indicated by the ‘—’ symbol.

Table 2.

Comparison of frequency and direction of topics raised in interviews with AGYW, by PrEP persistence.

Theme	IBM Domain	Topic	Persisted (N=16)			Discontinued (N=16)
			Positive	Negative	N/A	Positive	Negative	N/A
1: Necessity of positive attitudes & correct knowledge	Attitudes & Knowledge	Attitude	16	0	0	16	0	0
		Knowledge	11	5	0	2	14	0
2: Support from family and friends is critical	Perceived Norms	Family/ Friends	15	0	1	9	3	6
		Partner(s)	11	5	1	4	4	8
3: Strong mentors provide auxiliary support	Personal agency	N/A	11	7	0	9	10	1

‘Positive’ direction of theme refers to having positive attitudes, correct knowledge, support of family/friends, positive reflections on mentorship. Theme 1: ‘Positive’ denotes positive sentiments towards PrEP and correct knowledge about PrEP; ‘negative’ denotes negative sentiments towards PrEP and incorrect knowledge or professed incomplete knowledge about PrEP. Theme 2: An ‘N/A’ response indicates that they did not disclose to this individual that they were on PrEP and were uncertain how someone would respond. Theme 3: An ‘N/A’ response indicates this topic was not discussed in the interview. NOTE: An interviewee could reflect both positive and negative sentiments for Themes 2 and 3.

Theme 1: Positive attitudes are insufficient without correct knowledge

The most prominent theme that emerged revealed an important connection between two concepts in the IBM: attitudes and knowledge. Attitude, defined here as an AGYW’s overall perception of, and experience with, taking PrEP, is one of the three domains of decision (intention) in the IBM model (Figure 1). However, our findings suggest that an AGYW’s attitude may be less influential than the knowledge she has about PrEP when making decisions about PrEP persistence.

All AGYW interviewed had positive attitudes about PrEP, describing PrEP to offer them protection, security, and the opportunity to lead a healthy life. AGYW who ceased PrEP did so despite being overtly positive about how PrEP could help them; these AGYW even expressed concern about their safety after ceasing PrEP.

“I always want to protect myself from contracting HIV. That’s what motivates me… I feel I'm protected.”

(Persisted, age 24: Interview 6)

“The reason as to why I loved it [PrEP] most was just that fact that I was taking it to protect myself… it became my protector….. [After stopping PrEP] I have to be afraid. I have even been so worried because [I] am not so pleased with the way my husband walks around.”

(Discontinued, age 23: Interview 25)

While AGYW in both groups shared positive attitudes towards PrEP, there was a stark difference in the knowledge they had about PrEP. Most interviewees who discontinued PrEP expressed misinformation or incomplete information about PrEP (14/16), both of which were less common amongst AGYW who persisted (5/16). AGYW in both groups generally had an accurate description of the purpose of PrEP and how PrEP should be consumed (e.g., the need to take PrEP daily to prevent HIV). However, those who discontinued expressed uncertainty or believed erroneous information about long-lasting effects of PrEP, concurrent use of contraception with PrEP, and dangers of pregnancy and breastfeeding when on PrEP. Specifically, 11 of 16 AGYW who discontinued PrEP gave reasons that could be categorized as misinformation, with six of these AGYW stating they discontinued PrEP because they believed it was contraindicated during pregnancy and breastfeeding.

“The reason I stopped is that am hearing rumors that if you are pregnant and taking the drugs it is not good, or the time the baby is suckling it’s not good, so I decided to stop. It’s not good, so I decided to stop, so that when I have given birth and the baby stopped breastfeeding then I will continue.”

(Discontinued, age 23: Interview 27)

Finally, it was common for interviewees in both groups to express a desire for more knowledge about PrEP. AGYW who remained on PrEP described receiving continued education from mentors and healthcare providers during refill appointments as critical for persistence with PrEP. AGYW who had stopped PrEP often expressed a desire to re-start but did not have the knowledge on if, or how, this could occur.

“[At PrEP pick-ups] you find people who have knowledge about PrEP, they can educate you if in any case you still have challenges in PrEP. When you have any question, they can always respond.”

(Persisted, age 19: Interview 9)

“I had stopped taking prep, can I start taking prep again?….Maybe they can say that if I had stopped, I can’t go back. Yet as for me, I still want to start again.”

(Discontinued, age 22: Interview 28)

Theme 2: Support from family and friends is critical to overcome societal stigma

The second major theme to emerge was that a decision to persist with PrEP was driven by the emotional support an AGYW received from across her social circle. This theme relates to “perceived norms” in the IBM, as these close relationships were seen to be instrumental in negating, or further promoting, community stigma surrounding PrEP use.

AGYWs described stigma around PrEP to be pervasive in their communities largely due to people conflating PrEP with ARVs, and thus HIV infection.

“…PrEP drugs look like ARVs…if you carry the container of PrEP…he can think that you are taking ARVs.”

(Persisted, age 23: Interview 12)

“There are some girls who come to your house. When they see that drug container, they go around saying that you are taking the ARVs.”

(Discontinued, age 22: Interview 18)

Support from family, friends, and neighbors appeared critical in overcoming this stigma to persist with PrEP. Almost all AGYW who persisted reported receiving emotional support from family and friends (15/16); such support was not as common among AGYW who discontinued (9/16). Interestingly, it was common for both AGYW who persisted and AGYW who discontinued to have partners be dismissive, skeptical, or critical of their PrEP use (5/16 and 4/16, respectively). However, those who persisted despite having unsupportive partners described receiving support from family and friends; those with unsupportive partners who discontinued often expressed receiving no support, or mixed support, from family and friends.

“My boyfriend is always like ‘Why are you taking that stuff? Don’t you trust me?”…My mum always tells [me] that it is good”….”[My friends] do remind me. They say, ‘For us, we are going back today. What about you?”

(Persisted, age 21: Interview 4)

“The only person I told [about PrEP] was just my husband that I stay with here, and I did not receive any support from him.”

(Discontinued, age 23: Interview 25)

Overall, AGYW who discontinued PrEP described receiving less support from all relationships. Among AGYW who persisted, 14/16 disclosed their PrEP use to family, friends, or partners, but just 4/16 AGYW who discontinued disclosed their PrEP use to those in their social circle. Secrecy made persistence challenging.

“I knew I could not sustain it because I was doing it secretly.”

(Discontinued, age 19: Interview 26)

Theme 3: Strong mentors promote self-agency but their support is ancillary

The final theme that emerged was about the role of professional mentors with PrEP persistence. The positive experiences with mentors may help increase persistence by influencing an AGYW’s sense of personal agency, self-efficacy (i.e., how well they believe they can stick with PrEP), and perceived control (i.e., how much control they have over environmental factors that make it harder or easier to stick with PrEP).

Strong mentorship appeared to promote self-efficacy by providing AGYW opportunities to troubleshoot barriers to persistence, receive encouragement, learn more about PrEP, and be reminded to pick up refills. Together, this appeared to positively influence an AGYW’s belief in her ability, and her resolve, to stick with PrEP.

“They [the mentors] teach you, talk to you and encourage you. They even ask you about the difficulties that you are experiencing. They remind you not to miss to take the drugs when the time comes.”

(Persisted, age 24, Interview 2)

Strong mentorship also appeared to affect an AGYW’s perceived control by fostering an environment that made it easier to persist with PrEP. Specifically, mentors shared information on how AGYW could overcome side effects and personally worked to remove environmental constraints related to PrEP refills.

“[My mentor] used to bring [PrEP] for us because she lived near us….she could call us, if she sees a challenge, she calls [us] to go and pick them from her place, or she used to come after picking them and she follows [me to] my place.”

(Discontinued, age 25, Interview 17)

While strong mentorship played a positive role in many AGYW’s journey with PrEP, the role of mentorship appeared less critical to the decision to stick with PrEP than the factual knowledge AGYW had on PrEP (Theme 1) and the support they received from their existing social structure, namely family and friends (Theme 2). Overall, misinformation about PrEP and stigma were by far the most common reasons discontinuing PrEP, with the role of mentors being secondary in these discussions.

DISCUSSION

This comparative qualitative study adds to the literature on barriers and facilitators for AGYW PrEP use by exploring drivers specific to an AGYW’s decision to persist on PrEP. Within the context of a well-resourced DREAMS program, we found that AGYW were enthusiastic about PrEP but faced hurdles in their decision to continue due to community stigma and misinformation, particularly related to PrEP use while pregnant and breastfeeding. Support from family and friends was critical to overcoming stigma to foster persistence and was seemingly more influential than support received from romantic partners and mentors.

Perhaps most importantly, our study highlights the critical role of knowledge about PrEP in an AGYW’s decision to persist on PrEP, particularly as reproductive needs changed. While AGYW attitudes about PrEP were extremely positive, knowledge about PrEP appeared insufficient to ensure these positive attitudes translated into decisions to stay on PrEP. In particular, our study found misinformation about PrEP use during pregnancy and breastfeeding led to discontinuing PrEP. Our finding mirrors results of other studies which highlight AGYW hesitancy regarding PrEP during pregnancy or breastfeeding (37,38). This is unfortunate, as PrEP is considered safe during periconception, pregnancy, and breastfeeding (39) and has been prioritized for pregnant and lactating women at high-risk for HIV in national PrEP guidelines across SSA (40). While we acknowledge the need to strengthen evidence around PrEP safety while pregnant/breastfeeding (41), our study highlights the need for AGYWs to have access to comprehensive education about how PrEP can be used while pregnant and breastfeeding, based on current knowledge and national guidelines. It may be beneficial to review curriculum used in AGYW PrEP programs to ensure messaging about PrEP use during periconception, pregnancy, and breastfeeding is clearly communicated. The push to integrate PrEP with antenatal care clinics can help with this messaging (42,43), but providing comprehensive education around PrEP and reproduction prior to pregnancy may help maximize PrEP use during periconception and beyond.

Secondly, our study suggests that it may behoove PrEP programs to expand PrEP education to the broader community. Like many studies in SSA, our study identified pervasive community stigma around PrEP due to misinformation, which has been found to hinder PrEP use (20,21,27,30,44). Our study echoes past research that highlights challenges AGYW face if they take PrEP without support from their social circles, with family acceptance being particularly important (16,26,30). While the need to engage AGYW partners in PrEP sensitization is recognized (18–20,23,26–28,45), our findings suggest that bringing education programs to an AGYW’s family and broader community could be equally, if not more, important for PrEP persistence. We propose that community-wide PrEP education could broadly increase knowledge about PrEP, thereby decreasing stigma surrounding PrEP, and thus lead to increased support for AGYW who initiate PrEP. In particular, engaging mothers in PrEP education and outreach could be one way to normalize PrEP and strengthen the support families can provide AGYWs to persist, as mothers have been found to be particularly influential in decisions surrounding PrEP and reproductive health (30,46). Finally, as our findings suggest mentor support is secondary in importance to family or community support for PrEP persistence, program resources may be more impactful if they are focused more on community education and engagement than on intensive mentoring; this is an important finding, given the massive funding reductions that have occurred since this study was conducted. We hypothesize that intensive mentoring may be less necessary if the community is better educated about PrEP, but more research on this topic is needed.

Limitations

Our study is limited to its cross-sectional, qualitative design. Firstly, the study design limits the generalizability of our results, but the congruence of our results with other studies conducted across SSA strengthens our confidence in our findings and their relevance outside this study setting. Secondly, we recognize that social desirability bias may have been present in interviews. We attempted to minimize this bias by ensuring interviews occurred with study staff who had no association with DREAMS and with whom study participants had already built rapport through engagement in previous studies. Further, we believe that the interviewers being young Kenyan women, like the study participants, reduced the extent of social desirability bias and made participants more likely to feel at-ease in discussions but recognize there remained a power differential due to their educational backgrounds that could have contributed to some bias. Thirdly, we know that persistence on PrEP did not equate to adherence to PrEP in our study population. We have previously reported that blood tests for PrEP metabolites found PrEP adherence among AGYW participants to be very low in our study population (9). As a result, our study can only speak to an AGYW’s decision to persist on PrEP, but not on whether she actually adhered to PrEP. Finally, the analysis reported on in this manuscript occurred a few years after data collection and thus it was not possible to share findings with study participants for their comments. We believe interpretation of these interviews is strong, as co-authors on this paper include researchers with a deep understanding of the local research context and findings are largely congruent with published research from across SSA; however, we recognize that we have viewed the qualitative data with our academic lenses and that this interpretation would be stronger with additional reflection from the study participants/community.

CONCLUSION

Results of this qualitative study conducted in western Kenya suggest that PrEP programs for AGYWs may benefit by increasing education for both AGYWs and their communities, specifically the families of AGYWs. In particular, messages about appropriate PrEP use during pregnancy, breastfeeding, or while contracepting should be reinforced. As HIV prevention work faces unprecedented reductions in funding, our findings point to a need for additional research to disentangle the utility of PrEP mentors, as well as to explore cost-effective best practices for implementing community-based PrEP education. Finally, the results of our study may be applicable to help strengthen other health programs for AGYWs in SSA (e.g., family planning, antenatal care, domestic violence prevention).

Appendices

Supplement A – COREQ checklist and reflexivity statement (see accompanying document)

Supplement B - Framework analysis matrix (see accompanying document)

Supplement C – Participant-level summary table (see accompanying document)