The BeT study is one of 8 sponsored research projects chosen to be a part of a Eunice Kennedy Shriver National Institute of Child Health and Human Development–funded Prevention and Treatment through a Comprehensive Care Continuum for HIV-affected Adolescents in Resource Constrained Settings (PATC³H) consortium [31], which conducts clinical research and evaluation of a variety of combination interventions to improve health outcomes among adolescents at risk for HIV and youths living with HIV in resource-limited settings. The BeT study is the only PATC³H project conducted in Latin America and among transgender youths and builds on decades of research and community engagement with transgender communities. The BeT study is a peer-led, digital-based intervention study aiming to enroll 150 young transgender women in Rio de Janeiro, Brazil.

Inclusion criteria are (1) having been assigned the male sex at birth and self-identifying as travesti, a transgender woman, or having a gender identity different from identities typically associated with the assigned sex at birth; (2) being aged between 18 and 24 years; (3) providing consent for study participation; (4) living in Rio de Janeiro or the greater metropolitan area; and (5) being sexually active. Additional criteria for HIV-negative individuals include (1) no current PrEP use, (2) PrEP indication per Brazilian standards of care, and (3) no PrEP contraindication. For participants living with HIV, additional inclusion criteria are (1) having a confirmed HIV diagnosis, (2) not currently using antiretroviral therapy (ART) or having a detectable viral load and treatment that is not linked to HIV care.

Exclusion criteria include participation in any HIV prevention or care intervention study in the last year, refusal to undergo HIV testing, or having any health condition that, in the opinion of the investigator, precludes participation in the study or may lead to harm to the participant.

The study had a formative phase, which comprised developing and implementing an antistigma campaign, and adaptation and pilot testing of the system navigation intervention [32]. The campaign content included information on basic gender definitions (sex, gender identity, gender expression, and sexual orientation), regulations linked to gender and sexual orientation minorities, the medical transition process, overview of mental health and hormone therapy among transgender people, transphobia and violence against transgender people, and HIV among transgender women (Figure 1). The antistigma campaign, focused on health care settings, clinical providers, and support staff, took place at 4 health clinics serving transgender women in Rio de Janeiro, Brazil. We collected pre- and postcampaign qualitative and quantitative data with providers (N=10) to assess whether antitransgender stigma beliefs and attitudes changed among providers and support staff.

We also conducted formative qualitative interviews with 10 young transgender women (mean age 22.2 years) to inform the development of the pilot intervention. The BeT study uses an effective US Centers for Disease Control and Prevention intervention called Anti-Retroviral Treatment and Access to Services for system navigation and linkage [33,34]. Three components of the intervention required adaptation: (1) content of the Anti-Retroviral Treatment and Access to Services intervention session plan and asynchronous interactions with young transgender women, (2) adaptation to the cultural context of Brazil, and (3) the digital interface. Ten young transgender women were recruited through community outreach by our community engagement specialists (n=2) and through referrals from other studies at Fiocruz (n=8).

The BeT study pilot-assessed the intervention’s preliminary efficacy among 20 participants. Results from the pilot intervention demonstrated preliminary efficacy for improving HIV prevention and care behaviors among young transgender women, as all HIV-negative participants were engaged in PrEP and all but one participant living with HIV became virally suppressed [32]. After the pilot phase, the team refined the intervention, which focused on streamlining the adapted intervention manual to improve peer navigator and participant accessibility. In addition, the team added additional medical monitoring to include important biomarkers of sexual health and HIV as part of the study protocol.

Table 1 describes the study procedures. Potential participants are recruited by a community education team, through word of mouth, and via referrals from enrolled participants. Recruitment also occurs at the HIV prevention services based at our site and at an on-site mobile legal unit that offers weekly access to free legal services for transgender people.

Participants have 5 in-person encounters at the National Institute of Infectious Diseases Evandro Chagas, Fiocruz, including a screening and enrollment visit followed by 4 quarterly visits (at weeks 12, 24, 36, and 48). The BeT intervention takes place between the enrollment visit and the 12-week visit and is completed within that time frame. Trained staff conduct interviewer-administered surveys to participants at all study visits.

At baseline, all participants respond to a screening questionnaire and demographic information, after signing the written informed consent form. Baseline and quarterly survey content build upon questionnaires previously developed and tested with transgender women in Brazil and in the United States, as described in Table 2. During follow-up visits, we will collect information on study outcomes: HIV incidence, PrEP uptake and adherence among HIV-negative participants, and linkage to HIV care and viral load suppression among those living with HIV. In addition, we will also collect data on implementation science outcomes: (1) maintenance (pre- and poststudy) for study coordination, feasibility (poststudy) for peer health navigators, and feasibility, acceptability, and appropriateness (poststudy) for study participants.

The 3-month BeT intervention comprises three components: (1) BeT sessions, (2) digital interactions, and (3) automatic messages (Figure 2). During the first BeT session, the participant works with a peer navigator to build a case management plan aimed at meeting HIV prevention and care goals, and getting other needs met to promote health and wellness (BeT plan). BeT sessions 1 and 3 are conducted in person at enrollment and week 12 visits, respectively. BeT session 2 is conducted remotely via a mobile phone. Participants select a preferred app (eg, WhatsApp or Facebook messenger) for communicating asynchronously with the peer navigator during BeT session 2. Asynchronous digital interactions with peer navigators occurs at least weekly for 3 months and can be more frequent based on participant needs. Digital interactions were not restricted to study participation or HIV prevention or care and can include social support, food basket provisions, and guidance on employment, self-care, and other supports. Finally, participants received weekly automatic messages with health information on HIV testing, prevention, and care.

Statistical analysis will assess recruitment, baseline, and follow-up data and will describe the rapidity of enrollment of young transgender women, their demographic and risk characteristics, and participation as the percent intervention uptake (ie, the number of participants in the BeT intervention divided by the total number of individuals who were assessed for participation and were preliminarily eligible for the study). We will describe PrEP uptake and adherence and HIV care service linkage at intervention initiation and completion. For the primary endpoint, we will compare PrEP uptake and adherence and HIV care linkage between baseline and follow-up data (ie, before and after the intervention). Primary analyses will test the hypothesis that the odds of PrEP uptake and adherence and HIV care linkage will be higher post intervention relative to baseline data. We anticipate behavior change within 3 instead of 6 months because of our ability to more frequently interact with participants (3 months). In fact, we may achieve a greater dose response and intervention effect in a shorter amount of time because of the possibility for a higher number of encounters. At minimum, participants will engage with their digital navigator 4 to 5 times a week in the form of short digital encounters, and over the course of the intervention, participants will have 48 to 60 short digital encounters in total. We will use a pooled logistic model to estimate treatment effects on time to PrEP uptake and adherence and HIV care linkage.

We will also assess outcomes comparing data from the external comparison group to the BeT intervention group. First, we will conduct sensitivity analyses to compare the demographics of our intervention group to those of the external comparison group. In the analysis, we will focus on comparisons between the BeT intervention group and the external comparison group on measures of PrEP and ART adherence and viral suppression. Primary analyses will involve intention-to-treat analysis without considering adherence to the BeT intervention. This analysis will be a comparison of initiators of PrEP and ART, and we shall adjust for baseline confounders related to age or race. We have chosen to compare initiators of PrEP or ART and not assess PrEP uptake because a comparison of initiators parallels intention-to-treat analysis in which assignment and initiation of the treatment strategies occur together at baseline, regardless of whether individuals continue on the strategies after baseline [52]. We will combine outcome estimates with the cost estimates of the intervention component to calculate cost per participant and roughly estimate costs per HIV infection averted or poor health outcomes based on cohort data being collected in Brazil on HIV incidence and care outcomes.

Data from the baseline and follow-up study surveys will be used to assess whether there were improvements among young transgender women in (1) feeling of safety in health clinics; (2) perceptions of provider, staff, and other patient attitudes toward transgender women; (3) anticipated stigma; (4) medical mistrust; (5) health care participation; and (6) avoidance of health care. We will use repeated measures ANOVA to compare changes in mean stigma scores over the course of time in participating in the intervention (ie, before and after the intervention). In intent-to-treat analyses, we will use generalized estimating equations (GEE) Poisson or negative binomial models with robust SEs to compute the ratio of the single arm intervention rates of change in the mean value of stigma over the study period. GEE methods produce estimates that account for within-subject correlation of stigma scores across time [53]. GEE methods will also be used to assess whether changes in stigma are associated with changes in PrEP uptake and adherence and HIV care linkage over time.

We propose to enroll 150 participants and have estimated 80% power to detect effects for our primary outcomes using formulas for longitudinal study designs with attrition. These estimates account for an expected attrition of 15% of the sample by 12 months, as well as adjustment for time-dependent covariates in longitudinal analysis. We will also have 80% power to detect correlations of 0.23 of continuous baseline covariates with baseline responses and changes in study responses from baseline to 6 months.

In the analysis of changes, we will have 80% power in 2-sided tests with a type I error rate of 5% to detect changes from baseline to 12 months of 0.41 to 0.53 SDs in continuous responses, depending on within-subject correlation of the response, and increases of 10-15 percentage points in the frequency of binary responses, depending on baseline frequency as well as within-subject correlation. Thus, this study is powered to detect small to medium effect sizes that are consistent with or smaller than others reported in the literature.

Missing items at any visit should be minimal owing to the electronic data collection protocol described above. However, some loss to follow-up is inevitable. To assess sensitivity to dropout, multiple imputation will be used, under both the standard conditionally missing-at-random assumption, as well as plausible missing not-at-random scenarios. GEE methods can also be weighted to handle missing data.

HIV rapid tests will be performed at baseline for all participants and quarterly for HIV-negative participants. Participants reporting recent anal condomless sex (<30 days) with a negative HIV rapid test will be screened for acute HIV infection through HIV RNA viral load testing. The Limiting Antigen Avidity assay will be used to identify recent HIV seroconversions and may further support the calculation of HIV incidence by providing the timing of seroconversion [54]. Syphilis will be screened using treponemal testing, followed by venereal disease laboratory testing if the former test yields a positive result. Hepatitis B and C will be first screened using rapid tests followed by serology assessments, if needed. All participants will collect rectal swabs at baseline and the 48-week visit for the diagnosis of Chlamydia trachomatis and Neisseria gonorrhea infections. Sexually transmitted infections will be treated in accordance with Brazilian guidelines [55]. Participants who screened positive for hepatitis B and C will undergo complete the diagnosis algorithm (ie, hepatitis C viral load) [56] and will be referred to treatment. Participants with negative hepatitis B test results who have not been vaccinated will be referred for hepatitis B vaccination. PrEP adherence will be evaluated on the basis of tenofovir diphosphate concentrations in dried-blood spots.

Medical and psychiatric evaluation could occur as needed. Study referrals may include endocrinology, mental health evaluation, and proctology, among others. Participants may also access feminizing hormones on site.

The BeT intervention study was reviewed and approved by the institutional review board of the National Institute of Infectious Diseases, Fiocruz (#CAAE 05018818.0.0000.5262) and the University of California, San Francisco Committee on Human Research (#18-26770) in accordance with all applicable regulations. All participants were asked to sign an informed consent form prior to any study procedure.