Background Adaptation plays a critical role in implementation science (IS) by optimizing the fit of evidence-based
interventions (EBIs) to local contexts and populations. Adaptation is especially relevant in digital health, where tech
nologies, user preferences, and needs evolve rapidly. Yet, methodological approaches to guide and document adapta
tion processes are seldom included. Grounded in an IS framework, we describe a multi-phase process and results
for the adaptation of a digital health intervention (DHI) for HIV prevention and treatment for sexual and gender minor
ity (SGM) youth in Malawi, Kenya, Nigeria, and Zambia.
Methods Guided by the Assessment, Decision, Administration, Production, Topical Experts, Integration, Training,
and Testing (ADAPT-ITT) model, we adapted the EBI HealthMpowerment (HMP) DHI for SGM youth. We conducted
eight focus group discussions with SGM youth (N=80) living with or at risk for HIV and four with other key experts
(N=29) to inform app adaptations. Rapid qualitative analyses were conducted, and the findings were summarized
using the Mobile App Rating Scale. Subsequently, we conducted beta testing with 40 SGM youth to assess the app’s
usability with the validated System Usability Scale. The data were collected from June 2024 to April 2025.
Results Most HMP adaptations were within two domains of the Mobile App Rating Scale: Engagement and Infor
mation Quality. Engagement adaptations focused on increasing support and reducing provider response
times through the app, along with added gamification. Information Quality adaptations focused on expanding
resources (e.g., mental health, economic strengthening) and ensuring that healthcare referrals made via the app were
to SGM-competent providers. Functionality enhancements included adding content in local languages and integrat
ing WhatsApp. Data safety was prioritized and addressed by creating an emergency lockout code. For Subjective
Quality, participants perceived the app favorably for its support for medication adherence. Results from the beta test
ing indicated a score of 70.0, indicating above-average usability. Modifications were made to the ADAPT-ITT model
to embed consultations with key experts throughout the different phases