The study is a pilot randomized trial to test the acceptability to participants, feasibility for providers, and impact on long-term metric of adherence among participants (assessed via tenofovir hair levels) of implementing a POC urine tenofovir test to provide real-time adherence feedback and enhanced adherence counseling among women on PrEP in Thika, Kenya. Eligible women (n=100) will be HIV-negative, ≥18 years old, and on PrEP. Participants will be randomized 1:1 to the intervention of providing real-time feedback via the urine assay versus standard-of-care adherence counseling. Acceptability will be assessed by a quantitative survey of participants at the end of the study, as well as by qualitative data collected via in-depth interviews (n=20) and focus group discussions (n=4 groups of 5-10 women each). Feasibility will be assessed by the proportion of women retained in the study, the mean number of missed visits, the proportion of planned urine assessments completed and messages delivered, while in-depth interviews with providers will explore the ease of administering the urine test. Tenofovir levels in hair serve as the long-term metric of adherence.
Feasibility outcome: The investigators will assess the feasibility of the intervention by interviewing health care providers, who will be administering this test at the clinical point of care in the future. The investigators will examine provider perceptions of the assay using in-depth interviews. These key informant interviews will be performed at the end of this study with the healthcare providers (up to 8) who delivered the counseling messages to intervention arm participants after performing the POC urine TFV test. The semi-structured interview guide will draw from the Unified Theory of Acceptance and Use of Technology (UTAUT) model.200 This model incorporates factors that influence technology acceptance (in this case, of the POC immunoassay): perceived usefulness (performance expectancy), complexity to use (effort expectancy), stigma/social harm (social influence), and benefits (facilitating PrEP adherence among patients). These interviews will also elicit barriers and facilitators to delivering the TFV assay-informed counseling messages.
Acceptability outcome: The investigators will conduct a mixed-methods assessment of the intervention arm participants' experiences with real-time monitoring and feedback at the end of the study. A quantitative survey and a qualitative interview guide for in-depth interviews of participants will draw from the Information-Motivation-Behavioral skills (IMB) model.205-212 Quantitative data collection will occur via standardized interviewer-administered questionnaires. Items to be assessed include the following: 1) Feelings about receiving their PrEP adherence results in real time; 2) Likelihood of participating in other studies using a similar design; 3) Likelihood of wanting to receive results of urine testing outside of a study while they are on PrEP; 4) Concern about the privacy and security of the data regarding their urine results; 5) Grading of the potential impact of knowing their urine TFV results on subsequent medication adherence; 6) Advantages and disadvantages of being told about their adherence in real time; 7) Likelihood of taking PrEP just before later study visits because they knew the urine test was being conducted; 8) Preferences regarding a yes/no assay versus an assay that provides information on "high", "moderate" or "low" adherence. A 5-point Likert item format will be used to assess graded items (such as the likelihood of wanting continued urine testing in the context of PrEP; feelings about the urine testing, ranging from negative to positive; concerns about privacy, ranging from low to high; the potential impact of real-time feedback on subsequent adherence, ranging from low to high). Other items (advantages and disadvantages of being told about adherence results) will provide pre-specified options with one "other" option for open-ended text fields.
Adherence outcome: A linear mixed effects linear regression model will estimate the effect of the intervention versus standard-of-care on logarithmically transformed levels of tenofovir in hair.
