Using the CFIR 2.0 Framework to Assess the Implementation of GARDE: A Population Health Management Tool for...

Background: Identifying individuals at increased hereditary cancer risk remains challenging in primary care settings due to limited time with providers, low self‑efficacy in family history collection, and evolving referral guidelines. Genetic Cancer Risk Detector (GARDE), a population‑level open-source software platform, uses electronic health record (EHR)-based family history data and automated patient outreach tools to identify and engage individuals who meet criteria for genetic testing. To better understand how such tools can be integrated into diverse health systems, this study applied the Consolidated Framework for Implementation Research (CFIR) to characterize facilitators, barriers, and key processes involved in implementing GARDE at two large academic medical centers. Methods: Between September 2023 and June 2025, a total of 36 biweekly meetings with principal investigators and key stakeholders involved in GARDE implementations were recorded, date‑stamped, transcribed, verified, and coded using CFIR 2.0 domains and subdomains. Facilitators, barriers, and implementation processes were categorized by each site (n =2) and meeting date. Coding captured site‑specific contextual factors, meeting‑level variation, and co‑occurring themes across the implementation timeline. A secondary coder reviewed the transcripts to further contextualize themes and ensure accuracy in coding. Results: Principal investigators and stakeholders discussed 22 facilitators and 32 barriers to implementation in 14 of 36 total meetings. Facilitators predominantly reflected individual characteristics, including stakeholder buy-in and leadership support while the barriers most frequently reflected inner setting constraints (e.g., limited IT or staffing resources), implementation climate challenges (e.g., competing institutional priorities), and innovation characteristics (e.g., installation or integration difficulties), which often co-occurred. Implementation processes were similar across the two sites, with each site engaging in teaming, planning, tailoring strategies, and adapting workflows. Key cross‑cutting themes included data privacy considerations, patient outreach strategies (e.g., messaging modality and frequency), and differing institutional motivations for implementation. Conclusion: Findings highlight the importance of coordinated planning, stakeholder engagement, and attention to organizational context when implementing GARDE. These insights will support the development of an implementation toolkit and guide future efforts to integrate population‑level genomic decision support tools into routine care.