Acceptability of the Dapivirine Vaginal Ring for HIV-1 Prevention and Association with Adherence in a Phase III Trial

More than half of people living with HIV worldwide are women and those living in sub-Saharan Africa bear a disproportionate burden of new HIV infections each year [1]. They require and desire options for HIV prevention that are safe, effective, and acceptable [2, 3]. The dapivirine vaginal ring (DVR) is one potential option. Two randomized, placebo controlled, phase III trials in healthy sexually-active women, MTN-020/ASPIRE and IPM-027/The Ring Study, demonstrated the monthly DVR was well-tolerated and reduced HIV incidence [4, 5]. In ASPIRE, higher risk reduction was estimated with measures indicating higher adherence [6]. In earlier phase I and II trials the DVR was found to be highly acceptable [7‐10].

Assessment of product acceptability is not well standardized and is too often measured as a singular concept. As described by Mensch et al. [11] a framework that evaluates components of acceptability—including dosing regimen, product use attributes, effect of the product on sex, and partner’s attitudes—may help better define the influence of product acceptability on adherence of new HIV prevention methods.

In previous trials, acceptability of vaginal rings was high across several measures, including overall willingness to use in the future (if found effective for HIV prevention) [7, 9], ease of insertion/removal [7, 9, 10], feeling comfortable [7, 9, 10], and being unnoticeable during daily activities or once inserted [7, 9, 10]. Furthermore, early trials indicate most do not feel the ring during sex [7] or that it does not interfere with sex [9]. Per participant report, their male partners were more likely to have felt the ring during sex, however, only a minority reported this to be a problem for continued ring use [7]. Qualitative analysis from ASPIRE indicated that participants generally liked the ring better with experience, that adherence challenges could usually be overcome with staff or peer support, and that male partners were a commonly cited influence on ring acceptability and adherence [12]. The ring also impacted sexual experience in many ways, for some positively and for others negatively [13].

Here we present a quantitative analysis of DVR acceptability in the ASPIRE trial, one of the first cohorts to use the ring for an extended period (up to 3 years). We evaluated dimensions of acceptability, changes throughout the trial, and their independent influence on an objective measure of adherence. Understanding the acceptability of the DVR, the importance of each component, and the potential influence of acceptability on product adherence, is critical for supporting the ring’s successful roll-out and use in sub-Saharan Africa.

Among the 2629 participants enrolled in the ASPIRE trial, 2562 (97%) completed the acceptability measures on at least one follow-up ACASI and were included in the analysis. There were no significant differences between participants included in the analytic sample and those without ACASI who were excluded (all P ≥ 0.09). Descriptive characteristics of the analytic sample are presented in Table 2. Participants across countries differed on almost every characteristic evaluated (Supplemental Table 1). Across all sites, the average age of participants was 27 with 20% considered young (between the ages of 18–21); 46% had completed secondary school, 41% were married, and 45% earned an income of their own. Nearly all (98%) had a primary partner and were parous (92%). At enrollment, 58% had more than one male sex partner in the past 3 months. On average, participants were followed for 21.5 months (median 21, interquartile range 14–29 months).

Across multiple dimensions of acceptability, the dapivirine vaginal ring was highly acceptable to participants and, from their perspective, was also acceptable to their partners. The majority of participants expressed future likelihood of use. However, components of acceptability varied greatly by country, and certain components of acceptability, like effects on sex, had a significant impact on adherence.

Despite vaginal rings being novel in this population, overall acceptability was high and increased over time as participants gained comfort and familiarity with the ring. These findings are similar to early phase dapivirine ring studies [7‐9], and supported by qualitative analysis of ASPIRE data, where participants reported liking the ring more with experience [12]. Without a standardized measurement of acceptability, however, it can be challenging to compare findings across studies. The Mensch model [11] provides a framework to evaluate components of acceptability in the context of clinical trials, which, allowed us to better understand factors that may contribute to overall acceptability and observe that some, but not all, elements of acceptability changed over time. For example, while ease of insertion and general comfort with the ring increased, physically feeling the ring during sex for participants and partners remained relatively stable. By assessing both overall and separate components of acceptability, we gained a deeper understanding of ring acceptability, and demonstrate that even though certain elements of ring use may be less acceptable or may change over time, general acceptability may still be high.

We observed notable differences by country both overall and by individual components of acceptability. Generally, participants in Uganda and Zimbabwe reported higher likelihood of future ring use. Individual elements of acceptability also varied by country. For example, those in Malawi and Uganda minded ring use during sex and during menses, whereas this was not a reported concern among participants from South Africa or Zimbabwe. Furthermore, partner acceptability was much lower in Uganda compared to other countries. It is worth noting that some country-specific acceptability measures may seem incongruent with each other. For example, participants in Malawi and Uganda reported minding wearing the ring during sex more often but were also more likely to report increased sexual pleasure. In Uganda, future likelihood of use was among the highest, but minding ring use during menses or sex was also high. These findings speak to several key points. First, the various components of acceptability are not necessarily mutually exclusive. For example, one may mind wearing the ring during sex with certain partners (who may not know about or approve of ring use) but feel increased pleasure with other partners. Similarly, one may mind wearing the ring during sex or menses, but desire or need for an effective HIV prevention method—and therefore their intention for future use of an effective product—may outweigh this dislike. Alternatively, some participants may have disregarded the consistent use aspect of the willingness of future use variable (“keep [the ring] inserted in your vagina every day”), which may also explain some of these apparent disparate results—lack of cognitive testing of this question is a limitation of our analysis. Ultimately, participants in the study were likely balancing pleasure, protection, inconvenience, and partnerships simultaneously in their decision-making, as they probably do similarly in many aspects of their lives every day, including but not limited to contraception.

Our findings of variability in acceptability by country are supported by other work, which demonstrate that preferences around sex, product use, vaginal practices, and hygiene, differ across and within countries [17, 18]. Relationship dynamics, previously noted in ASPIRE qualitative research to have varied by location [19], may be one factor that contributes to the observed differences in acceptability by country. Our findings speak to the importance of conducting research across a range of settings and populations, as well as emphasize the need for offering choices in HIV prevention methods. Individuals from different countries, cultures and backgrounds have different preferences for HIV prevention products, which has potential to influence their uptake and use.

Given the differences observed in age-stratified efficacy of the dapivirine vaginal ring in ASPIRE [4], we expected to find similar variation in acceptability in younger versus older participants. However, most acceptability measures were similar across ages, apart from younger participants being more likely to report the ring was not acceptable to their primary partner. Although reported partner acceptability varied by age, there was no impact observed on adherence related to this component among younger participants. Younger individuals may be more likely involved in shorter term or more casual relationships, or generally be less comfortable talking to their partners about sexual and reproductive health issues, which could explain lower perceived partner acceptability and/or lack of any observed effect of partner acceptability on adherence in this subpopulation. Our observations may also be limited in that only participants who had access to a ring in the 3 months prior to PUEV were administered acceptability questions—younger participants were less likely to receive a ring during this time, so our findings may underestimate potential effect of partner acceptability on adherence among this group. Nonetheless, these findings highlight that while acceptability may have some influence on adherence, it does not explain age-related differences in observed effectiveness in phase three trials, suggesting that other social and structural factors also contribute to ability to use the ring consistently.

Importantly, although perhaps not unexpectedly, participants with low overall acceptability were less likely to use the ring consistently based on objective adherence measurements. Some attributes of acceptability, including effects on sex, perceived negative change to the vaginal environment, and minding wearing the ring during menses, influenced adherence more strongly, whereas others—comfort and ease of insertion—had no effect. Feeling or minding the ring during sex was the strongest influencer of nonadherence observed at both month 12 and in the last year of the study. This is further supported by our findings of increasing nonadherence among participants who reported no change or decreased sexual pleasure, compared to those who reported increased sexual pleasure with the ring. Our observations are supported by previous qualitative research, which demonstrated that the impact of the ring on sexual experiences was a specific concern that participants felt they had to navigate and that had the potential to impact their adherence [13] and that for some, the ring improved their sexual pleasure for reasons ranging from changes in lubrication to reduced fear of HIV acquisition [12]. Participant report of partner acceptability or their partner feeling the ring during sex had a modest influence on nonadherence, although these results were not statistically significant. A stronger relationship between partner acceptability and adherence was observed among older participants. This trend is supported by other qualitative findings from ASPIRE, where partner perceptions were often cited as a prominent influencer of acceptability and adherence [12] as well as work that looked at the relationship between reported partner support and adherence in ASPIRE [20]. Importantly, perceived partner acceptability does not necessarily equate to partner support, highlighting the importance of exploring these relationships both quantitatively and qualitatively.

Recognizing that acceptability may have some influence on adherence or overall willingness for future use, it is important to acknowledge that actual use (adherence) and intention or willingness to use in the future may not always align. This is especially true when adherence is in the context of a randomized controlled trial (RCT) with a placebo and unknown efficacy of the active product. Future willingness to use a known effective product may be high, even among those who struggled with adherence during the RCT due to challenges with use during sex or other barriers; conversely, a participant with high adherence may not ultimately find a product acceptable. Individuals may have many motivations for joining clinical trials outside of product use, and that these motivations may confound observations between components of acceptability and adherence. Furthermore, other important social and structural factors may also play a role, such as desire to have a child, experience of social harms or intimate partner violence [21], initial worries about the ring [22], perception of HIV risk, or acceptability attributes not measured here, such as product characteristics (size, consistency, smell), dosing regimen, or stigma. That said, understanding what components of acceptability most strongly influence adherence may help identify individuals in need of more support and allow for more targeted counseling during initiation of the ring.

There are a few limitations worth noting in our analysis. First, only participants who received a ring in the past 3 months were asked acceptability questions—meaning that, if those who were lost to follow-up or otherwise not receiving rings (e.g., on product hold) were also more likely to find the ring unacceptable, we may be overestimating acceptability of the ring. However, more than 90% of participants did complete acceptability questions so we anticipate that acceptability was still high despite this. Furthermore, PUEV ACASI data were not available from two sites in South Africa, where overall acceptability was somewhat lower (available data from South African participants reported only 59% ‘very likely’ future use) potentially leading to an underestimate of the strength of association between acceptability and nonadherence. Additionally, we did not evaluate all potential components of acceptability, meaning we could have overlooked other key factors that may or may not influence adherence, or contribute to overall acceptability. As already mentioned, we also only measured some acceptability components at PUEV, meaning our analysis of these factors was retrospective and limited our ability to say what effect those factors had on future ring adherence. Also, collection and testing of used rings for residual drug started approximately one year into the trial. Therefore, we are missing some initial residual drug measurements from participants enrolled earlier which could potentially misclassify their level of adherence. Finally, because the error around measures of residual dapivirine with current methods can be up to 0.5 mg, it is challenging to define high adherence by a certain release rate (provided that also the amount of dapivirine released per month is small, about 4 mg over 28-days of use) [15]. Hence, we focused on nonadherence as the definition is more definitive. When more granular quantitative measures of adherence for the vaginal ring are developed, we may find a different relationship between acceptability components and high adherence.

In conclusion, the DVR was highly acceptable to participants across several dimensions of acceptability, although with country variation. Acceptability of the ring increased over time, demonstrating the DVR’s potential as a prevention product for long-term use. Addressing perceived interference with sex, menses, or problematic changes to the vaginal environment, in future interventions through counseling and skill building could help improve adherence, as could embracing sex-positive messaging related to ring use and increased pleasure. Ultimately, the variability we observed in acceptability of the ring highlight the importance of increased choice in HIV prevention options, as no one method will satisfy the needs of all individuals.