Comparative Effectiveness of Individual Versus Group-Level Interventions to Increase PrEP Uptake and Reduce HIV/STI Incidence

Background and Significance

The HIV diagnosis rate among African-born (immigrant, foreign-born) Black women is the highest of all Black individuals living in the US. Nationally, the 2014 HIV diagnosis rate among African-born Black women was 100.5 per 100,000-a rate 400% (5.3 times) higher than US-born Black women (19.1), 1.4 times the rate of US-born Black men (72.8), and 2.0 times the rate of African-born men (51.3). While the rate of new HIV diagnoses among all Black women has decreased significantly overall, the disparity between African-born and US-born Black women has increased. More than half of newly diagnosed African-born Black individuals are women (60.7%), compared to 27.0% among the US-born Black population.

In Massachusetts (MA), a proposed multi-site study location, 70% of new HIV diagnoses among Black women are among non-US born women (60% African-born). In New York City (NYC), 34% of recently diagnosed women were non-US born, with 23% from Africa. Crucially, a significant proportion of HIV transmission among African-born women occurs post-immigration within the US. Phylogenetic and qualitative analyses indicate that between 30% and 62% of HIV-infected African-born individuals most likely contracted the virus in the US.

HIV risk among African-born Black women is driven by complex cultural, psychosocial, and structural factors along a defined causal pathway. These include culturally based gender inequalities, high levels of HIV-related stigma, knowledge deficits, acculturative stress, intimate partner violence (IPV), low socioeconomic and undocumented immigration status, and culturally specific sexual practices like "dry sex". Despite their vulnerability, correct condom use and the uptake of Pre-Exposure Prophylaxis (PrEP) are suboptimal. The key evidence gap is that no culturally-appropriate, evidence-based interventions have been developed or adapted to increase condom use and PrEP uptake specifically for diverse African-born women.

Significance This study is designed to address a critical gap and inform an important decisional dilemma for stakeholders and community-based organizations (CBOs) by comparing the effectiveness of two distinct HIV/STI prevention interventions. Sister-to-Sister (S2S) is a brief (20-minute), individual-level, health professional-led intervention focused on skill-building and condom usage. SISTA (Sisters Informing Sisters about Topics on AIDS) is a peer-led, group-level intervention utilizing the Theory of Gender and Power to overcome gender inequality and promote ethnic pride.

By culturally adapting these interventions, the study aligns with national efforts to "End the HIV Epidemic." The project builds upon the infrastructure of Tulumbe! ("Let's Engage"), a PCORI-funded partnership of over 70 African-born patients, clinicians, and stakeholders. The community identified three primary objectives that this study addresses: (1) increase safer sex/testing among youth/young adults, (2) increase self-efficacy for HIV prevention to overcome gender inequality, and (3) reduce HIV stigma. We hypothesize that the culturally adapted SISTA (renamed DADA, meaning "Sister" in Swahili) will be more effective than the culturally adapted S2S (renamed Dada Kwa Dada or DKD) because it targets gender inequality, is peer-led, and fosters community support.

Specific Aims and Objectives The central research question is: What effect does participation in a culturally-adapted group-level versus an individual-level intervention have on the use of HIV prevention strategies among at-risk African-born women living in the US?

• Specific Aim 1 (Cultural Adaptation and Pilot Test): To culturally adapt two widely utilized, evidence-based HIV prevention interventions originally designed for US-born Black women (S2S and SISTA) utilizing a revised 7-step ADAPT-ITT framework. This aim includes conducting focus groups and a pilot test to determine feasibility and acceptability prior to the comparative effectiveness trial.
• Specific Aim 2 (Comparative Effectiveness RCT): To conduct a randomized controlled comparative effectiveness trial (RCT) of the culturally adapted versions of S2S (DKD) versus SISTA (DADA) to determine their effectiveness in increasing condom use and PrEP uptake (two patient-centered primary outcomes) among African-born women.

Subject Selection Participants will be African-born cisgender women, ages 18 to 45, who are fluent in English or French. Eligibility criteria dictate that participants must self-identify as Black or mixed race, report condomless vaginal or anal sex with one or more males in the last 3 months prior to enrollment, and have an HIV-negative or unknown status (offered testing at baseline). Pregnant women are explicitly excluded as there is no data regarding the selected interventions in pregnancy.

The study spans sites in MA and NYC. The clinical trial components (Aim 2) will take place through the Brigham and Women's Hospital (BWH) Clinic and Whittier Street Health Center (WSHC) in MA, and the Muslim Women's Institute for Research and Development (MWIRD) in NYC. WSHC serves roughly 8,793 African immigrant women annually (47% of clientele), and MWIRD serves 1,755 (with >40% French-speaking). The adaptation phase (Aim 1) involves the Multicultural AIDS Coalition (MAC) in MA and the African Services Committee (ASC) in NYC.

Recruitment Strategy

Recruitment will follow the MGB IRB-approved protocol #2023P001336. A multi-pronged strategy will be employed by study coordinators and African-born Patient Investigators:

1. Study Advertisements: Flyers will be posted in clinical waiting areas, CBOs (e.g., MAC, Harvard Street Neighborhood Health Center), African grocery stores, and restaurants. Online social media groups, radio, and mass transit ads will also be utilized.
2. Study Outreach: Research Assistants (RAs) will attend 1-2 social gatherings/events monthly sponsored by African community organizations to describe the study.
3. Participant Referrals (Snowballing): The study leverages community networks by offering a $10 referral incentive for any current participant (or non-participant) who refers an eligible candidate (up to $30 maximum for three or more referrals).
4. Provider Referrals: Health care providers at BWH, WSHC, and MWIRD will inform eligible patients and provide contact information.

Randomization Process and Treatment Assignment The study utilizes a blocked 1:1 randomization design with stratification by site, African region of origin (West vs. East Africa, the most common regions of origin), and language (English vs. French). The biostatistician will use a random number generator to devise a randomization log via REDCap. To prevent anticipation of the condition, alternating blocks of 2 and 4 will be used.

Once randomized via REDCap, DADA participants will wait to be scheduled into a session until at least 4 additional participants within their site and language group have been randomized to form a cohort of 5-8 women. This group size is purposefully kept small to enhance opportunities for pro-social, peer feedback regarding sexual health behavior. We anticipate that both interventions will be delivered approximately two weeks post-enrollment.

Study Procedures Due to the COVID-19 pandemic, intervention delivery will occur entirely virtually via Zoom. Upon enrollment and consent, participants complete an online REDCap survey.

• DKD (Individual-Level): A brief (approximately 1 to 1.5 hours) skill-building session delivered by study team staff in English or French.
• DADA (Group-Level): A comprehensive peer-led intervention delivered across two 3-hour sessions, designed to reflect African cultural norms, poetry, and storytelling.

Participants are followed at 3 and 6 months. To increase the validity of self-reported condom use, participants will be provided a daily calendar to document sexual activity (focusing on 30-day data). HIV testing (4th generation assay) and STI testing (urine nucleic acid amplification for Chlamydia and Gonorrhea) are required at baseline and 6 months. Participants may choose in-person lab visits (covering transport or a $25 reimbursement), use mailed self-test kits, or provide official lab results from the past 3 months.

Compensation includes stipends for data plans. To encourage completion of the lengthy group sessions, DADA participants receive $120 at baseline, while DKD participants receive $50. Both groups receive $25 at the 3- and 6-month follow-ups, along with intervention completion certificates.

Sample Size and Power The RCT targets a total sample size of 120 participants (60 per condition), accounting for an estimated 10% dropout rate. Based on prior data, baseline condom use is estimated at 22%.

• Condom Use: Assuming a 26% increase in the DKD arm and a 53% increase in the DADA arm at 6 months, 120 participants provides 80% power to detect a statistically significant difference (alpha = 0.05).
• PrEP Uptake: The sample provides 80% power to detect a difference of 10% increased uptake in the DKD arm versus 33% in the DADA arm.
• STI Incidence: With 60 per arm, differences of 20% STI incidence in the DKD arm versus 2% in the DADA arm can be detected with 80% power.

These calculations account for site-level clustering using an intraclass correlation coefficient (ICC) of 0.005 and a design effect of 1.175. The team projects screening 928 individuals, with 474 (50%) eligible, to achieve the target 120.

Internal and External Validity Internal validity is safeguarded through multiple mechanisms: selection bias is minimized by the block randomization design; attrition bias is controlled by adhering strictly to an intention-to-treat (ITT) approach; and performance bias is mitigated by recording interventions and conducting ongoing supervision. Contamination is reduced by requesting participants not to share information.

External validity is highly robust. The study recruits from massive target populations (>100,000 African-born in MA, >145,000 in NYC). The interventions are based in real-world CBO and clinical settings, and the eligibility criteria capture a diverse cross-section of women with varying degrees of HIV/STI risk. By heavily involving community stakeholders and Patient Investigators throughout the adaptation and implementation phases, the study ensures the findings are highly generalizable and ready for national scale-up.

Heterogeneity of Treatment Effects (HTE) The study will estimate intervention effects across clinically relevant subgroups, particularly regional origin (West vs. East African) and time since immigration (≥10 years vs <10 years). Subgroup analysis will not be conducted based on immigration status due to the high likelihood of missing data surrounding this sensitive variable.

The study acknowledges that a sample size of 120 is statistically underpowered for definitive HTE interaction testing. Assuming two strata of equal size (n=60 per stratum), achieving 80% power to detect statistically significant HTE for condom use would require detecting extreme odds ratio differences of 11.67 versus 1.16. Therefore, while formal interactions will be tested, any clinically meaningful but statistically non-significant stratum-specific differences will be treated strictly as hypothesis-generating data for future larger-scale studies.

Analytic Plan Data will be analyzed using an intention-to-treat (ITT) protocol, analyzing participants in their original assigned conditions regardless of session attendance. Initial unadjusted analyses will use t-tests for continuous variables and chi-square testing for categorical outcomes.

The primary statistical analyses will employ generalized estimating equation (GEE) regression models (linear and logistic) for repeated measures to obtain adjusted mean differences or odds ratios over the 6-month period. Models will adjust for sociodemographic covariates, study site, and age. Because the DADA arm is a group-level intervention (causing clustering) while DKD is individual-level, heteroscedastic partially nested mixed-effects models will be used. Additionally, an "as treated" analysis will assess the DADA arm by dividing subgroups based on their level of dose/compliance to the two 3-hour sessions. All statistical testing will employ a 2-sided type I error rate of 5%.

Monitoring and Quality Assurance Given the behavioral nature of the interventions and the minimal risk profile, a multiple-person Data Safety and Monitoring Board (DSMB) is deemed unnecessary. Instead, an independent, trained mental health provider-Dr. Christina Psaros, a licensed clinical social worker and Associate Professor with research experience-will serve as the independent monitor.

Dr. Psaros will receive periodic reports every 3 months detailing subject enrollment, retention, dropouts (with reasons), and adverse events. She will make formal recommendations regarding the continuation, modification, or termination of the study. Any unexpected, serious adverse events will be reported by telephone by the PI to the IRB and the independent monitor within one business day, followed by a comprehensive written narrative report within 3 business days.

Process (Acceptability and Feasibility) Evaluation Aim 1 incorporates an extensive adaptation and evaluation process using the ADAPT-ITT framework. Virtual focus groups will be held with 68 potential participants (11 groups) and 39 organizational/clinic staff (4 groups) across MA and NYC to assess culturally specific prevention needs. A "theatre test" will feature live mock interventions with trained facilitators and CEAB members to refine the manuals.

Following adaptation, a non-randomized pilot test involving 28 women (14 NYC, 14 MA) will assess feasibility and acceptability. Acceptability will be measured using an adapted Information Systems Success Model (ISSM), evaluating four domains via Likert scales: information quality, delivery quality, perceived usefulness, and overall satisfaction. Feasibility will be evaluated through post-intervention facilitator feedback on session flow and content adequacy.

Fidelity Assessment Intervention fidelity is a critical component and will be evaluated based on three dimensions: (1) Content (was each component implemented as planned?); (2) Coverage (what proportion of the target group participated?); and (3) Frequency/Duration (were components implemented as often and for as long as planned?).

During the pilot and RCT, post-intervention interviews will be conducted with facilitators and participants to capture these metrics. Furthermore, a random sample of 10 recorded Zoom intervention deliveries will be formally reviewed to assess strict adherence to the intervention manual. The team will also evaluate moderators of fidelity, such as intervention complexity, facilitation strategies, and patient responsiveness.

Data Management The Principal Investigator (Bisola Ojikutu) holds ultimate responsibility for data management and safeguarding. All study data will be collected, stored, and managed using REDCap (Research Electronic Data Capture), a secure, HIPAA-compliant web-based application hosted by Partners HealthCare Research Computing. Data will be identified by a unique identification (ID) code and kept entirely separate from study samples, lab results, informed consent documents, and locator information. To prevent bias, the analytical study team will remain blinded to the data and results until the conclusion of the trial.

Preventing and Handling Missing Data Trial design strategies to limit missing data include offering substantial financial incentives, frequent follow-up periods, non-HIV-related contact (e.g., birthday/holiday emails), and leveraging the deep community trust of participating CBOs.

If item nonresponse is substantial, it will be addressed using multiple imputation techniques via PROC MI in SAS, assuming data are missing completely at random (MCAR) or at random (MAR). The team will generate 20 imputed datasets and conduct ITT analyses, combining results using PROC MIANALYZE. If there are statistically significant differences between respondents and non-respondents at follow-up, inverse predicted probability nonresponse weights will be constructed. The team will also randomly impute data in sensitivity analyses to account for non-ignorable missing data scenarios.

Reporting of Subject Enrollment and Follow-up The study is committed to rigorous reporting aligned with the 2010 CONSORT guidelines. The research team will track and report the precise number of patients approached, screened, deemed ineligible, and those refusing participation. Subject attrition and all adverse events will be explicitly noted. For any participant who drops out, the team will thoroughly document the specific reason, who initiated the dropout decision, and whether the dropout pertained to intervention participation, data collection, or both. All included and excluded participants will be accounted for in the final CONSORT diagram.

Clinical monitoring structure/site monitoring plan The study features a highly integrated clinical and site monitoring structure. Site Principal Investigators (e.g., Federica Williams at WSHC, Dr. Nurah Amatullah at MWIRD, Chioma Nnaji at MAC, and Mulusew Bekele at ASC) are responsible for overall site oversight. They supervise local staff, manage recruitment, oversee intervention implementation, facilitate lab testing pathways, and report regularly to the BWH study team.

The Community Engagement Advisory Board (CEAB), comprising community experts, stakeholders, and Tulumbe! members, will meet weekly during Years 1 and 2, and monthly in Year 3. Patient Investigators (African-born women, some living with HIV) will act as peer facilitators, directly managing recruitment, consent, and delivery of the DADA interventions.

Data Sensitivity Because the study collects highly sensitive data (including HIV status, behavioral risks, and undocumented immigration status), stringent protections are in place. Participant names will only appear on written consent forms and a master tracking list. This master list links the unique ID numbers to contact information. It will be stored securely in a locked file cabinet and on a single password-protected, encrypted computer inside a locked office, accessible only to the site PI and coordinating project staff. Furthermore, a federal Certificate of Confidentiality will be obtained. Participants experiencing distress or disclosing Intimate Partner Violence (IPV) will be immediately referred to established counseling and clinical care pathways at WSHC, BWH, or the Institute for Family Health.

Disposition of Data After the Study Hard and soft copies of raw data, the master linking list, and all identifiable participant information will be securely and permanently destroyed within one year following the completion of the study. Audio recordings from the focus groups and qualitative interviews will be destroyed immediately after their transcription accuracy has been verified. De-identified quantitative data sets will be maintained for a reasonable period to allow for the publication of final reports and journal articles, after which they too will be securely destroyed.

Data Integrity Data integrity is foundational to the study's validity. The research team will undergo extensive training to ensure standardized, objective collection and recording of participant information. Site coordinators will rigorously edit and audit all REDCap assessments for legibility, consistency, and completeness. All electronic files will be heavily password-protected, encrypted, and transmitted securely (via encrypted flash drives or MGB Dropbox for business). Paper files will be kept exclusively in locked file cabinets. Mass General Brigham Institutional Review Board will serve as the single IRB, utilizing reliance agreements with the community sites to maintain strict ethical and data integrity oversight throughout the project lifecycle.