The investigators will conduct a randomized controlled trial to test the efficacy of Shine, an mHealth HIV behavioral intervention for African American transgender women (TGW). Participants will be 215 African American TGW recruited by Trans Women of Color Collective (TWOCC) via in-person and online recruitment strategies. First, TWOCC's leadership and administrative teams will distribute palm cards and display posters describing the study during all community-based events that occur during the recruitment period. Second, ISA and TWOCC will develop an online strategy to disseminate a study flyer via social media (e.g., Facebook, Twitter). The flyer will be posted on both social media channels weekly during the study recruitment period. Interested women recruited via either method (in-person or online) will be instructed to contact the ISA field test manager for more information and to complete eligibility screening.
Interested participants who contact the field test manager will complete an eligibility screener. Specifically, participants must : 1) identify as a Black transgender woman using the recommended 2-step method (Step 1: assigned male birth sex, Step 2: current gender identity is female); 2) be aged 18 or older; 3) report risk of HIV transmission in the past 3 months (i.e., CAS with a serodiscordant or risky male partner AND either \<90% ART adherence or \<5 daily doses of PrEP per week); 4) own a smartphone; and 5) be able to read and speak English.
Ineligible women will be thanked, dismissed, and offered alternatives to the type of information being offered in this research study. Eligible women will provide contact information receive a detailed overview of the main points of the study. Once any participant questions are answered, the field test manager will trigger a text to be sent to the potential participant's study phone number. The text will include a link to an online version of the informed consent form presented on a secure web page (i.e., communication between the phone's browser and web page are encrypted). Women who agree to participate will acknowledge consent by clicking a button at the bottom of the online consent form that reads "I understand and CONSENT to participate." Women who choose not to participate after reading the informed consent form will press a button that reads "I understand and choose NOT to participate." These women will be thanked and all contact will cease.
Participants will be assigned to conditions using a permuted-blocks randomization scheme. Blocks of size 4 and 6 will be used. Assignments within blocks will be random but balanced among the two conditions and the order of the blocks (4 vs 6) also will be random. This procedure will ensure that the number of participants randomized to each condition will be equal, that any imbalance among the conditions at any point during the recruitment and randomization will be modest, and that it will be very difficult to guess the assignment of the next participant. After randomization has occurred, participants will complete the baseline research survey via telephone.
The primary outcome will be a dichotomous composite HIV risk transmission score, calculated by measuring whether a participant or the participant's partner consistently used at least one effective method for reducing HIV transmission risk (i.e., adherence to biomedical \[ART or PrEP\] and behavioral \[condoms, serosorting\] strategies) during each instance of anal sex over the past month. Every anal sex act must be protected by one of these methods to receive an overall composite score of 0. The individual behavioral components of the composite score (i.e., percentage of condomless anal sex, serosorting, PrEP or ART adherence) will serve as secondary measures. In addition, specific outcomes by HIV status will include HIV medical care utilization and adherence, HIV testing, and PrEP knowledge, interest, and uptake. Finally, further secondary measures will assess gender affirmation, well-being, social support, and sexual communication.
Once the baseline research survey has been completed, participants will receive instructions for accessing the assigned materials. Those in the experimental group will be instructed to text the word "join" to the Shine study phone number. After joining, experimental group participants will be enrolled in a mobile intervention that aims to reduce HIV health disparities among African American TGW along the entire care continuum. Participants randomly assigned to the control condition will be texted a link to a set of videos on healthy HIV-related behaviors for transgender women. These videos cover a variety of topics, including recommended testing frequencies, PrEP, and ART. All control participants will have access to Shine at the end of the research study (i.e., approximately six months after recruitment).
Six months after the baseline research survey, all participants will complete a follow-up research survey assessing the same outcome measures in the baseline research survey.
Given a sample size of 215 and a predicted retention rate of 80%, the investigators anticipate having complete data for at least 172 participants at the close of the study. Because randomization carries the expectation of creating treatment groups equivalent with respect to known and unknown prognostic factors, removing randomized participants from the analysis runs the risk of tampering with this balance and introducing bias into the treatment comparisons. As a result, an intention-to-treat analytic approach will be followed such that all participants randomized into the study will be included in all analyses irrespective of protocol violations post randomization. Multiple imputation is now widely regarded as an effective method for replacing missing data, and the investigators will use this approach.
Binomial logistic regression will be the main analytic technique for the primary outcome (composite HIV transmission risk score) and other non-continuous measures. For continuous measures, hierarchical linear modeling (HLM) will be used. All models will control for any demographic covariate (e.g., age) that varies at the .2 significance level due to randomization failure at baseline.
To ensure this study will be fully powered, the investigators estimated power for the binomial logistic regression analysis on the binary primary outcome measure (composite HIV transmission risk score). To conduct this calculation, the investigators examined power from Garofalo and colleagues' RCT of LifeSkills, a 6-session group-based, in person intervention for young trans women of any HIV status. At the 12-month follow-up, TGW in the empowerment-based HIV intervention showed a 39.8% greater reduction in condomless sex compared to participants in a standard care control condition (risk ratio = .60). This effect is described as "robust" by the authors but is typically interpreted as a small-medium effect size (Cohen's d=.35). This study provides an adequate power comparison for two reasons. First, this intervention (empowerment-based) and population (specifically tailored for TGW) closely align with the proposed work. Second, this effect size is consistent with the average effect size of SMS-based interventions on health behaviors (d = .33). As the effect size for the binomial logistic regression is an odds ratio, the investigators translated the Cohen's d of .35 to an odds ratio of 1.89. The G\*Power 3.1 program was then used to determine that an odds ratio of 1.89 and an α error probability of 0.05 requires a sample size of 172 to be sufficiently powered at 1-β error probability = 0.95. Therefore, the 172 participants the investigators expect to retain through the follow-up, given a sample size of 215 and a 20% attrition rate, should be sufficient to detect significant differences in composite risk for HIV transmission at follow-up.
