We developed a human-centered, publicly accessible, and regularly updated dashboard that summarizes and synthesizes domestic HIV implementation research conducted within EHE priority jurisdictions and aligned with EHE goals and objectives. Translating evidence into action beyond the individual study level depends on broad access to study results, as well as on transferability and generalizability. Achieving the goals of the EHE initiative requires understanding outcomes at a national scale, which, in turn, relies on the ability to rapidly assess data across studies and geographies. Therefore, to conceptualize and operationalize the product, we employed an iterative, reflexive development process grounded in the Design Council’s Double Diamond framework.

The core dashboard development team (hereafter, “the team”) comprised faculty members of the Johns Hopkins University IS Hub (formerly the Mid-Atlantic Centers for AIDS Research [CFAR] Consortium IS Hub), full-time research associate staff, PhD/DrPH students, postdoctoral fellows, and a professional web developer. The Johns Hopkins University IS Hub is one of several NIH-funded consultation entities that provide technical assistance and support capacity-building to strengthen domestic HIV implementation research funded by the NIH. NIH funding supported the design and development of this dashboard as part of its mandate to advance IS tools and resources for the EHE initiative.

Discover-phase consultations were conducted between November 2022 and January 2023. Define-phase publication identification and extraction work were conducted between March and July 2023. Develop-phase wireframing and prototyping occurred between September 2023 and June 2024. Deliver-phase user-testing sessions were conducted between June 2024 and April 2025.

We used the Design Council’s Double Diamond framework to guide the dashboard development process across 4 iterative phases: Discover, Define, Develop, and Deliver [17]. The Discover phase sought to understand the landscape of potential end users and identify the types of data and visualizations most valuable for decision-making, with the objective of defining the dashboard’s overall scope and purpose. The Define phase translated user priorities into actionable design requirements and data specifications, clarifying what the dashboard needed to do and how it could best communicate findings to distinct user groups. In the Develop phase, we translated conceptual requirements from the Discover and Define phases into tangible design elements, which were then synthesized into a prototype of the dashboard’s structure, navigation, and content. During the Deliver phase, we evaluated the dashboard’s usability and relevance through structured user testing and iterative feedback.

Faculty members conducted initial consultations with representatives from the NIH and an NIH-funded national HIV IS coordination team [18]. These stakeholders were uniquely positioned to support coordination across EHE IS studies and to encourage the use of cross-study resources to advance the field of IS while supporting the rapid achievement of EHE goals. Meetings focused on presenting the initial dashboard concept for feedback and discussing proposed implementation research characteristics and variables for inclusion. The team synthesized insights from the initial consultations to produce a core dashboard concept, a set of defined user groups, and preliminary use cases.

We identified a working corpus of IS-focused EHE publications (n=15) to guide activities throughout the Define phase. A colleague from the NIH-funded HIV IS Coordination Initiative provided the team with an initial list of 22 EHE publications that EHE investigators had self-reported as grant-related outputs. To support the linkage of publications to specific EHE projects, the Coordination Initiative also provided project descriptions that included the project title, CFAR affiliation, principal investigator, project director, EHE jurisdiction(s), and community or implementing partner. Of the 22 self-reported publications, 15 met the eligibility criteria for inclusion: original research articles that could be linked to a specific EHE project. This purposeful working corpus was intended to support tool development rather than serve as a systematic review of the published EHE IS literature.

We identified variables for data extraction using a hybrid deductive-inductive approach. Deductively, we drew on Hickey et al [19] specification of HIV implementation interventions, the Consolidated Framework for Implementation Research for determinants [20], and Proctor et al [21] taxonomy of implementation outcomes. Inductively, we reviewed the 15 publications to identify emergent variables and data elements that captured how implementation was described in practice. We iteratively refined the variables through team consensus meetings during extraction, drawing on patterns observed across publications and the team’s collective expertise in HIV IS. We then organized these variables into a structured online database (Airtable [22]) that served as the data extraction tool.

Five team members (TSJ, ASK, RD, LL, and LKB) conducted extractions from the initial set of publications. Each record’s data were first extracted by 1 team member and then independently reviewed by a second for accuracy and completeness. Any discrepancies identified during the second review were resolved through group consensus. The online extraction tool was iteratively refined throughout the process and ultimately grouped data into the categories of study design, study population, outcomes, and community engagement.

Following the initial extractions, the team held internal consultations to map available data to the use cases identified in the Discover phase. We aligned each use case with specific user queries and priority variables to predict how distinct user groups might interact with the dashboard and answer questions most relevant to their roles and positionality within the broader EHE initiative. The consultations also clarified how the data architecture might inform visualization design and filtering options to answer user-specific questions.

We created low-fidelity wireframes to demonstrate how the extracted data might be visualized to both leverage their opportunities and manage their limitations in meeting the dashboard’s objectives and use cases. We conducted a “storyboarding” exercise to visualize a hypothetical user’s journey through the website, considering how the data might be grouped to be accessible and informative for all users while still distinguishing among user types. We used PowerPoint and physical drawings to imagine which data would be available on each page and how they might be visualized. We shared the wireframes with an external collaborator, who produced a low-fidelity Python prototype. Both the wireframes and the prototype were used to facilitate the initial round of user testing.

We conducted 14 user-testing sessions with EHE stakeholders across 3 rounds. Sessions were facilitated by members of the research team (LKB and SRS: professors, PhD; RD: postdoctoral researcher, PhD; TSJ: DrPH candidate, MPH; ASK: research manager, MPH), all trained in qualitative methods. Participants were selected primarily by user group affiliation; candidate recruitment discussions also considered proximity to EHE implementation, geographic experience across EHE priority jurisdictions, and familiarity with IS. Participants were known professional contacts of the research team before recruitment; invitation materials described the project as developing a dashboard to synthesize NIH-funded EHE IS research and framed the sessions as consultations rather than formal research. Formal demographic data were not collected, given the nonhuman subjects institutional review board determination; all participants were selected based on at least a decade of relevant experience within their respective user-group domain. No invited participants declined or withdrew.

One team member facilitated each 60-minute session; a second was present to take notes and provide supplementary facilitation as needed. No additional nonparticipants were present. Sessions were conducted virtually via Zoom, recorded, and transcribed. Sessions began with a guided walkthrough of the prototype, accompanied by live feedback from the participant. Participants then responded to embedded semistructured questions on both function (practical interaction/logistics) and form (content), including how they might use the dashboard in daily work and how accessibility and usability might be improved. Each session was intentionally structured to mirror the stages of the Double Diamond framework, guiding participants through questions that sequentially reflected the Discover, Define, Develop, and Deliver phases, and allowing each session to function as its own nested HCD cycle (Multimedia Appendix 1).

We used a rapid qualitative analysis approach to synthesize session data [23,24]. During each session, one team member captured real-time notes to support in-session probing; session transcripts were later reviewed to inform postsession synthesis. Notes and transcript observations were organized into a structured spreadsheet with columns for dashboard sections and elements, user observations, and suggested adaptations. Three team members (ASK, LKB, and TSJ) led the analysis; findings from each session were synthesized across all sessions within a given round during team consensus meetings, where interpretive discrepancies were discussed and resolved. This formative study was not designed to achieve thematic saturation; the sample size was determined a priori to ensure representation of all planned user groups in each testing round [23]. Themes were derived inductively from the structured templates through team consensus; no prespecified coding framework or formal coding tree was used, consistent with the rapid qualitative approach [23]. Synthesized findings were converted into actionable recommendations for the external web developer, covering both microlevel interface adjustments (such as terminology, navigation, and layout) and macrolevel conceptual shifts, including user tailoring, data transparency, and dissemination features.

This research project was reviewed by the Johns Hopkins Bloomberg School of Public Health Institutional Review Board and received a nonhuman subjects research determination. Verbal consent for recording and the use of deidentified feedback was obtained at the start of each session. Community partner participants received a US $50 gift card in recognition of their time.

Consultations and team synthesis produced 3 primary outputs: a core dashboard concept, a defined set of user groups, and an initial set of use cases.

The envisioned product was a publicly accessible, interactive dashboard for reviewing and summarizing published EHE implementation research studies. The dashboard was intended to tailor content to distinct user priorities, offer visualizations of implementation research characteristics, and represent qualitative work, which was a significant part of the published EHE research. Importantly, this new dashboard was conceived as research outcomes–focused, in contrast to existing dashboards that primarily provide programmatic updates.

We identified 5 primary user groups: (1) the project’s funders, the NIH; (2) academics; (3) community partners, such as nongovernmental organizations and community-based organizations; (4) implementing partners, including local and state health departments; and (5) experts in HIV-related IS in the United States. For each stakeholder group, we used data from consultative meetings to answer three questions: (1) what might interest a representative of that group about the dashboard; (2) what questions might that individual ask about published EHE research results; and (3) how can the dashboard meet that individual’s needs? (Table 1).

We then used insights from Table 1 to identify 5 primary use cases that would guide the dashboard’s design: (1) collate published EHE IS data into a single repository; (2) support rapid review of the current “state of the (implementation) science”; (3) provide accessible data to answer implementation questions; (4) facilitate the generation of research questions and the development of sampling frames for EHE-related systematic reviews and other evidence syntheses; and (5) facilitate partnership and network development by identifying academic and implementation researchers.

The Define phase translated the 5 use cases identified in the Discover phase into concrete data requirements and a structured information architecture for the dashboard. The team mapped each use case to the types of user queries it would need to support and the specific variables that would need to be captured to answer those queries (Table 2). This process clarified both what data would need to be extracted from the published literature and how the resulting database would need to be structured to support filtering and visualization across user groups. Extraction activities ultimately grouped variables into 4 primary domains—study design, study population, outcomes, and community engagement—forming the backbone of the dashboard’s data architecture.

The Develop phase produced 3 outputs: low-fidelity wireframes and storyboarded user journeys; a Python-based prototype that integrated priority variables and views; and an initial information architecture specifying page types, navigation, and filters aligned with the Discover and Define insights. The wireframes included mock-ups of a landing page, a table summarizing research to date, maps of EHE project and community partner locations, and pages for barriers and facilitators, IS frameworks, implementation outcomes, client outcomes, service outcomes, HIV cascade targets, and study designs (Figure 1).