A Text Messaging–Based Support Intervention to Enhance Pre-exposure Prophylaxis for HIV Prevention Adherence During Pregnancy and Breastfeeding: Protocol for a Randomized Controlled Trial

Background: Cisgender women in Kenya are at elevated risk of HIV acquisition during pregnancy and post partum. Acute HIV infection during pregnancy and breastfeeding accounts for approximately one-third of all vertical HIV transmissions. The World Health Organization recommends offering oral tenofovir-based pre-exposure prophylaxis (PrEP) to pregnant and postpartum women who are HIV negative but at substantial and ongoing risk for HIV acquisition. PrEP delivery for pregnant and postpartum women is expanding within routine maternal child health clinics in Kenya. However, approximately half of pregnant women discontinue PrEP within 30 days of initiation. Therefore, it is crucial to develop PrEP adherence strategies that enhance support for adherence when peripartum events and health issues pose challenges to sustaining PrEP adherence. Objective: We are conducting a randomized controlled trial to determine the effect of a bidirectional communication platform named Mobile Solutions for Women’s and Children’s Health (mWACh), which utilizes two-way SMS text messaging between patients and remote nurses to support PrEP adherence and address maternal health concerns in real time during the peripartum period. Methods: The mWACh-PrEP study is a randomized trial designed to support PrEP adherence during the peripartum period by comparing mWACh-PrEP to the standard of care (ie, in-clinic adherence counseling) among women who are HIV negative and initiating PrEP. Purposive sampling was used to select 5 facilities offering PrEP in antenatal clinics in Kisumu and Siaya Counties, and block randomization will be used to divide participants into groups. Participants in the intervention arm will receive a customized messaging curriculum via SMS text messages targeted toward their particular perinatal stage. The primary outcome, PrEP adherence at 6 months post partum, will be evaluated using a log-binomial regression model, adjusting for imbalanced baseline characteristics. Based on a previous study of directly observed dosing conditions, we will use a hair tenofovir concentration cutoff of 0.038 ng/mg (corresponding to 7 doses/week) as the primary adherence outcome measured at 6 months post partum (binary outcome). Qualitative interviews and cost-effective analyses will be conducted to understand the feasibility, acceptability, and economic impact of the intervention. Results: Enrollment began in March 2022 and is projected to continue until July 2023, with follow-up through March 2024. The study results are expected to be reported in 2025. Conclusions: This trial will provide insights into using mobile health to enhance PrEP adherence among pregnant and postpartum mothers. Additionally, the findings will have implications for the use of mobile health technology to improve adherence to other daily medications during the peripartum period. We are conducting a nonblinded individual-level RCT comparing 2 arms of PrEP adherence to determine the effect of the mWACh-PrEP tool on PrEP adherence among pregnant women at risk for HIV acquisition initiating PrEP within ANC clinics. The study is being conducted in 5 public health facilities offering routine MCH services within Kisumu and Siaya Counties in Kenya. The facilities were selected based on high HIV prevalence areas, existing infrastructure, established rapport, and discussions with the Kenyan Ministry of Health and the local government of Kisumu County. Participating facilities are enrolling 600 pregnant women who are HIV negative (300 in each arm), between 24 and 32 weeks of gestational age (to allow for adequate follow-up during pregnancy), and receiving ANC at MCH clinics. Gestational age is ascertained using the last menstrual period and fundal height measurement, which is part of the routine care for women in this region. Eligibility criteria include HIV and tuberculosis negative, HIV acquisition risk score ≥6 (translating to HIV incidence 7.3 per 100 person-years) [40], initiated PrEP during routine ANC clinic in the participating clinics, ≥18 years, has access to a cell phone, plans to reside in the area for at least 1 year post partum, and plans to receive postpartum and infant care at the study clinic. Two PrEP adherence models are compared: standard of care (SOC) comprising in-clinic adherence counseling only versus SOC plus the mWACh-PrEP tool. All women receive routine MCH care and PrEP services that include health education, clinical assessments, safety monitoring, screening for sexually transmitted infections (STIs), and treatment per national guidelines (including expedited partner treatment for those with STI diagnoses), HIV risk reduction counseling, assessment of psychosocial barriers, and standard in-clinic adherence counseling per national guidelines. Participants randomized to SOC receive in-clinic counseling on PrEP. Participants randomized to the intervention arm are registered in the mWACh-PrEP system, a bidirectional SMS text message communication platform that sends PrEP-tailored, theory-based, preprogrammed texts on PrEP adherence weekly during pregnancy and the postpartum period, in addition to receiving SOC services (Figure 1). Screening, enrollment, and follow-up procedure for this study. ANC: antenatal care; MCH: maternal and child health; mWACh: Mobile Solutions for Women’s and Children’s Health; PrEP: pre-exposure prophylaxis; SOC: standard of care; TB: tuberculosis; TFV: tenofovir. Block randomization using a 1:1 allocation in random-sized blocks stratified by recruitment site is being used to ensure a balanced allocation of each arm within sites. Before randomization, study nurses ensure all routine PrEP initiation services and study enrollment procedures are complete. Treatment allocation (mWACh-PrEP or SOC) for each consented participant is done by the study nurses within the individual’s health facility using Research Electronic Data Capture (REDCap; Vanderbilt University). The randomization code is stored in an electronic format, and once assigned, it is unblinded. The randomization key for the database is restricted to the data manager and statistician. Pregnant women initiating PrEP as part of routine care are approached by the study staff while waiting for ANC and invited to participate in the study. Participants undergo screening, and all eligible participants who express interest undergo an informed consent process at the ANC clinic before the initiation of study procedures. Information on facility volume, clients approached, clients screened, and participants enrolled are reported daily by the study staff. Reasons for nonparticipation are captured from those who decline. At enrollment, electronic questionnaires are administered to capture demographic, pregnancy, and medical history; mental health status; HIV results; risk perception; risk assessment; partner characteristics; PrEP attitudes; expected maternity care; health literacy; and COVID-19 experiences. Socioeconomic status is measured using the validated standard Demographic and Health Survey (DHS) Wealth Index [41]. Health literacy is assessed using the standard DHS panel of 8 items across 4 domains (capacity to interpret, obtain, understand, and make appropriate health decisions) [41]. Intimate partner violence (IPV) is assessed using the Hit, Insult, Threaten, Scream scale [42]. Symptoms of depression are assessed using the Edinburgh Postpartum Depression Scale and the Patient Health Questionnaire 9, which has been validated in Kenya [43]. An HIV acquisition risk score validated among pregnant and postpartum Kenyan women is used to assess the risk for HIV [40]. Additional data abstracted from the MCH clinics and PrEP records include details on PrEP initiation, confirmation of syphilis, and HIV test results. Per national guidelines, all women initiated on PrEP in both arms receive in-clinic counseling on the benefits of sustained PrEP adherence at enrollment and subsequent refill clinic visits, HIV risk reduction, and assessment of psychosocial barriers. Participants randomized to the intervention arm are registered in the mWACh-PrEP SMS text messaging system with their preferred name for messaging, messaging language (ie, English, Kiswahili, or Dholuo), and day of the week and time for SMS text message delivery. Once registered, the participants receive weekly automated “push” messages on the day/time of their choice from enrollment until they are 6 months post partum. All automated “push” messages include the participant’s nickname, clinic, and nurse name; an educational message or actionable advice targeting PrEP adherence and continuation and/or MCH topics; and a question related to the content. Through multiple messaging tracks, participants receive messages targeted to their perinatal status as well as supportive messages if they decide to discontinue or restart PrEP. SMS text message topics include adherence encouragement (IMB domain: motivation), PrEP efficacy and safety (IMB domain: information), self-efficacy for prevention of HIV, support for potential PrEP side effects, behavioral skills (eg, tips for remembering PrEP medications; IMB domain: behavioral skills [44,45]), and visit reminders (Figure 2). Participants have the option to respond to automated messages or extemporaneously contact the study nurses at any time. Study nurses encourage participants to respond with concerns and questions (Figure 3). Participants have access to a two-way dialogue with remote nurses until their 6-month postpartum visit. Participants are free to voluntarily exit autonomously by texting “STOP,” thereby ending all communication. All messaging is free to the participant using a reverse-billed short code. The mWACh system also includes multiple features to support study management and personalized counseling based on the participant’s psychosocial status (Table 1). Conceptual model adapted from the Information-Motivation-Behavioral Skills (IMB) Model. mWACh: Mobile Solutions for Women’s and Children’s Health; PrEP: pre-exposure prophylaxis. Screenshot of the SMS messaging system. PrEP: pre-exposure prophylaxis. Example mWACha-PrEPb SMS text messages based on the IMBc model. amWACh: Mobile Solutions for Women’s and Children’s Health. bPrEP: pre-exposure prophylaxis. cIMB: Information-Motivation-Behavioral Skills. Study participants are followed up until 9 months post partum to ascertain sustained PrEP adherence at 3 months after cessation of SMS text messages. All study visits for both arms will be aligned with Kenyan national guidelines for standard ANC and infant immunization visit schedules: monthly visits during pregnancy until delivery and then after delivery at 6 weeks, 14 weeks, 6 months, and 9 months post partum. At each visit, nurses administer questionnaires to assess sexual behavior; infant outcomes; IPV; symptoms of depression; and PrEP attitudes, use, and adherence. PrEP use disclosure is assessed at each follow-up visit by asking participants whether they have disclosed to anyone that they use PrEP and, if so, to whom they disclosed this (Table S1 in Multimedia Appendix 1). Additional data on infant growth, birth weight, estimated gestation at birth, diagnosed preterm birth, and serial growth are abstracted from medical records. Additionally, z scores are calculated using WHO AnthroPlus software. Data on birth outcomes (stillbirth, pregnancy loss, mortality, or congenital anomalies) and birth length are abstracted from the MCH medical record or interview. A combination of red blood cells from DBS and hair samples is used to assess PrEP adherence. A strong correlation between dosing frequency and TFV levels in hair has been demonstrated [46], and significant correlations were also seen between hair, plasma, peripheral blood mononuclear cells, and red blood cell levels (in DBS) of PrEP drugs or metabolites [47,48]. Plasma levels reflect exposure over the past 1 to 3 days [49] and are highly susceptible to intraindividual variation; therefore, measuring hair drug levels is a unique approach to determining drug exposure over a longer period [50,51]. DBS are collected from a finger or heel prick from all mothers and infants at all study visits for future assessment of TFV-DP and emtricitabine triphosphate in red blood cells. Hair samples will be collected from participants at pregnancy follow-up visits and from both the participants and their babies at postpartum follow-up visits. These hair samples will be analyzed for TFV using validated liquid chromatography-tandem mass spectrometry methods at the University of California San Francisco (UCSF) Hair Analytical Lab (HAL). The UCSF HAL found that TFV concentrations in the hair under directly observed dosing conditions are similar by sex, and the same adherence benchmarks can be used for both women and men [50]. Additionally, hair TFV levels have not been shown to alter in pregnancy [51]. The study staff is trained on data collection using REDCap and communication with participants in the intervention arm using the two-way mWACh-PrEP platform to send, receive, and respond to SMS text messages. Baseline and follow-up data are collected and stored electronically via REDCap, transported via a secure socket layer, and only accessible by authenticated users. Designated study staff generate weekly reports on study progress, performance indicators, adverse events, and troubleshooting problems. All study laptops used to collect data and study databases are encrypted and password protected. All data collected in this study will be available free of charge after registration to access or download files on a study-related website (URL to be determined) after the completion of primary analyses. The sample size was calculated to detect a difference between trial arms dependent on the frequency of PrEP adherence in the SOC arm of the cohort. Prior data from our PrEP in pregnancy studies indicate that the frequency of detectable PrEP drug levels is approximately 50% among pregnant women who initiate PrEP in MCH clinics within fully programmatic settings [12]. Since this study will exclusively enroll women who initiate PrEP in MCH, we anticipate that adherence will be similar or slightly higher. Assuming a PrEP adherence estimate of 50% in the SOC arm at 6 months post partum and a sample size of 600 (randomized 1:1), we will have 80% power to detect a minimum detectable difference of 12% between arms in PrEP adherence in the mWACh arm, assuming 15% attrition and a 2-sided test with a Cronbach α of .05 (Table S2 in Multimedia Appendix 1). The primary outcome, PrEP adherence at 6 months post partum, will be assessed in an intention-to-treat analysis comparing randomization arms mWACh-PrEP versus SOC. Enrollment characteristics will be compared between randomization arms to assess the balance between arms and control for any imbalanced factors. Based on a previous study of directly observed dosing conditions, we will use a hair TFV concentration cutoff of 0.038 ng/mg (corresponding to 7 doses/week) as the threshold of detection for the primary adherence outcome measured at 6 months post partum (detectable/nondetectable binary outcome) [52]. Protective PrEP drug level thresholds for cisgender women have not been established; however, current WHO and Kenyan national guidelines recommend daily (7 doses/week) oral PrEP for cisgender women. A log-binomial regression model will be used to estimate the relative risk, adjusting for any imbalances in baseline demographic, clinical, or behavioral characteristics between arms. Missing primary outcomes or loss to follow-up will be assumed nonadherent. To evaluate the sustained effect of mWACh-PrEP on PrEP-taking habits after the SMS text messaging intervention ends, the same model will be used to analyze our secondary outcome of PrEP adherence at 9 months post partum (3 months after cessation of mWACh-PrEP messaging). Focus group discussions and semistructured in-depth interviews will be conducted to evaluate potential barriers and facilitators to the acceptability and feasibility of mWACh-PrEP implementation within routine ANC delivery settings (Figure 4). To evaluate personal experiences with the mWACh-PrEP intervention, a purposively selected sample of study participants will be recruited to participate in in-depth interviews. In-depth interviews will explore community, dyadic, and individual factors influencing decision-making and how these influence individual perceptions of and experiences with mWACh-PrEP use. To gather perspectives on the mWACh-PrEP scale-up, we will invite the study’s community advisory board (CAB) and county-level health officials overseeing ANC-PrEP programs to participate in focus group discussions. These focus group discussions will elicit perspectives on foreseeable barriers and facilitators of larger-scale implementation of mWACh-PrEP based on individual, community, and health system factors. Schematic of qualitative purposive stratified sampling approach (N=70). PrEP: pre-exposure prophylaxis. Discussions will be conducted in English, Kiswahili, or Luo depending on participants’ preferences and be guided by prepiloted topic guides with the flexibility to explore additional probes and newly emerging content as necessary. All in-depth interviews and focus group discussions will be audio recorded, transcribed, and translated to English for analysis as necessary. Thematic analysis will be used to identify key elements of the patient/provider experience, including convenience, comfort level, and satisfaction of the mWACh-PrEP model, that influence intervention acceptability, feasibility, and potential scale-up. We will project the cost-effectiveness of implementing mWACh-PrEP intervention within ANC compared to the SOC per averted disability-adjusted life year (DALY) using a payer perspective and collect intervention costs from expense reports, staff, and expert interviews. Costs will be divided into mutually exclusive categories, including personnel, equipment, supplies, buildings and overhead, and start-up. We will convert the local currency, Kenyan shillings, to US dollars. Sensitivity analyses will be conducted to identify influential assumptions. Time and motion observation of intervention activities will be conducted to calculate staff time and resources necessary for the mWACh-PrEP intervention. A research assistant (RA) will collect data on the time and motion of patient/provider interactions to inform costs and productivity assumptions. The RA will also conduct a survey with health care workers to assess daily responsibilities associated with the mWACh-PrEP intervention, including time to send and respond to client SMS text messages and routine service provision. Research time (eg, administering study questionnaires) and other research costs will be separated from programmatic costs. We will parameterize a previously developed mathematical model with cost and outcome data from the mWACh-PrEP trial and project the averted HIV infections, HIV-related deaths, and DALYs associated with scaling up the mWACh-PrEP intervention in Kenya. We will calculate the incremental cost-effectiveness ratio as the difference in costs divided by the difference in effects (DALYs) for the intervention compared to SOC over a 20-year period. Costs and benefits will be discounted annually at 3%. Secondary outcome analyses will include sustained PrEP adherence by determining levels of detectable TFV in the hair at 9 months post partum, adherence cofactors, STI incidence, and prevention-effective adherence (time-varying alignment of adherence with risk behaviors). Exploratory analyses will include HIV incidence, perinatal outcomes, and expedited partner treatment outcomes by randomization arms (Table S3 in Multimedia Appendix 1). Ethical approval was obtained from the institutional review board of the University of Washington (STUDY00010797) and the Kenyatta National Hospital Ethics and Research Committee (P319.05/2021). This study was registered at ClinicalTrials.gov. ({“type”:”clinical-trial”,”attrs”:{“text”:”NCT04472884″,”term_id”:”NCT04472884″}}NCT04472884). Any critical protocol deviations, adverse events, and future modifications will be communicated promptly to these organizations. Additional permissions will be sought from all participating clinical sites. A written informed consent before recruitment is administered to all study participants, with each receiving a copy of the signed form. Additional permission will be obtained for audio recording and note-taking during the qualitative interviews. All participants are informed that their participation is voluntary, and they can stop receiving the SMS text messages at any time by contacting the study staff or texting “STOP,” which will cease communication between the study staff and the participant. All study staff is trained in protecting human participants and the clinical and emotional needs of pregnant and postpartum women on PrEP. The MWACh system is point-to-point encrypted; therefore, only study team members can access it through a password-protected laptop stored in a locked office. All data are deidentified and kept in password-protected databases in a locked study office that is accessible only to study personnel. A CAB has been established in Kisumu County to review the study objectives, methodology, and other study activities. The team comprises local community leaders, women from the community, the Kenyan Ministry of Health, and county health representatives from MCH departments. The CAB is anticipated to meet 2 times per year to review this study’s progress. An external advisory panel (EAP) has been established to oversee the data and safety of study participants by reviewing study aims, statistical analysis plan, and progress. EAP meetings will be routinely convened to review study enrollment, retention, data missingness, and pooled study outcomes. Study results will be disseminated through peer-reviewed journals, technical reports, policy briefings, and oral presentations at relevant local and international conferences. Additionally, we will present results to clinical experts and policy advisors at the local county level, the Kenyan Ministry of Health, and the WHO. Routine hybrid meetings in the form of progress updates with CABs engaged in the study activities, facility staff, and study collaborators will be conducted to share study progress and findings.