Weighing up the risks and benefits of trial participation: understanding non-adherence in a PrEP trial

Participants' explanations for non-adherence in the FEM-PrEP clinical trial.

Corneli A, Perry B, McKenna K, Agot K, Ahmed K, Taylor J, Malamatsho F, Odhiambo J, Skhosana J, Van Damme L. J Acquir Immune Defic Syndr. 2015 Nov 3. [Epub ahead of print]

Background: FEM-PrEP - a clinical trial of daily, oral emtricitabine/tenofovir disoproxil fumarate for HIV prevention among women in sub-Saharan Africa - did not show a reduction in HIV acquisition because of low adherence to the study pill. We conducted a follow-up study to identify reasons for non-adherence.

Methods: Qualitative, semi-structured interviews (n=88) and quantitative, audio computer-assisted self-interviews (n=224) were conducted with former FEM-PrEP participants in Bondo, Kenya, and Pretoria, South Africa. Thematic analysis was used to analyze the qualitative data, and descriptive statistics were used to describe ACASI responses. Data are presented within the five categories of Ickovics' and Meisler's conceptual framework on adherence: 1) the individual, 2) trial characteristics and study pill regimen, 3) patient-provider relationship, 4) clinical setting, and 5) the disease.

Results: Participants' explanations for non-adherence were primarily situated within three of the framework's five categories: 1) the individual, 2) trial characteristics and study pill regimen, and 3) the disease. Concerns about the investigational nature of the drug being tested and side effects were the prominent reasons reported for non-adherence. Participants also described being discouraged from taking the study pill by members of the community, their sexual partners, and other participants, primarily because of these same concerns. Limited acceptability of the pill's attributes influenced non-adherence for some participants as did concerns about HIV-related stigma. Additionally, many participants reported that others continued in FEM-PrEP while not taking the study pill because of the trial's ancillary benefits and visit reimbursement - factors related to the clinical setting. Negative patient-provider relationships were infrequently reported as a factor that influenced non-adherence.

Conclusion: Despite substantial study staff engagement with participants and communities, concerns about the study pill and discouragement from others appeared to have influenced non-adherence considerably. Alternative study designs or procedures and enhanced community engagement paradigms may be needed in future studies.

Abstract access  [1]

Editor’s notes: The authors of this important paper on a PrEP trial, end with a note of caution. They note that when interpreting the findings we should remember that the women in this study were taking a ‘study product’. The women were not taking a product of proven efficacy. Therefore, as the authors state, it would be wrong to assume that ‘African women cannot and will not be adherent if provided with PrEP outside of a clinical trial setting’. If they had been told that the product was efficacious, they may have behaved differently. This is important because a key message of the paper is that trial participants managed their participation so they felt comfortable in the trial. Many wanted to ensure they received benefits from their participation, including good health care, but they also wanted to manage risk. Risk associated with fears about the trial drug and risk from the disapproval of sexual partners about their participation. It is also very clear in these findings that the participants could manage the expectations of the trial team, by telling them what they wanted to hear during the trial. This suggests the limited value of ‘adherence questionnaires’ in some settings. The authors provide a powerful illustration of the value of mixed methods in trials of this sort. Drug concentration data told the researchers that many women were not adhering to the drug. Qualitative semi-structured interviews using this drug concentration data with the individual women helped the team to understand why. The authors also discuss the influence of community and family members in undermining participant faith in the trial. They explain the lengths that the trial team went to, to inform community members about the trial. Considerable time was given to sharing information. Doubts remained; concerns that were enough to discourage participation. This too is an important finding underlining the value of investing in community engagement in research. But it also highlights the need to find ways to enhance not just engagement, but also understanding and trust. 

Africa [9]
Kenya [10], South Africa [11]
  • [12]