Human Factors in Open-Source Medical Software: Compliance Barriers, AI/LLM Adoption, and Community Dynamics

Abstract

Open-source medical applications (OMA) play an increasingly central role in the digital health ecosystem, supporting electronic health records, medical imaging, clinical datasets, and emerging artificial intelligence (AI) and large language models (LLM). While these systems promise rapid innovation and reduced development costs, they operate within highly regulated and safety-critical environments that pose significant human factors challenges for developers. This study examines how regulatory compliance, AI/LLM adoption and community dynamics intersect in contemporary OMAs. Using a structured GitHub search strategy, we identified 12 representative and actively maintained repositories spanning electronic health records, medical imaging, datasets, AI/LLM utilities, and development frameworks. We conducted a metadata-based analysis capturing project activity, programming languages, adopted compliance (e.g., GDPR, HIPAA, FHIR, DICOM), and AI/LLM integration. The results indicate that compliance adoption remains uneven, with roughly half of repositories lacking explicit regulatory integration and GDPR emerging as the most consistently addressed standard. AI/LLM use is limited but growing, primarily concentrated in medical imaging and exploratory text-processing applications. Across projects, active community support emerged as a critical factor in sustainability, compliance adoption, and developer onboarding. These findings position developers as an under-studied stakeholder whose challenges directly shape downstream clinical safety and system reliability. The study highlights opportunities for human factors research to improve the usability of compliance frameworks, guide safe human-AI integration, and strengthen sociotechnical support structures in open-source health IT.