Consumer-Centered Selection of Relevant Clinical Trials Using Agent AI
摘要
Current clinical trial matching solutions lack practical, consumer-accessible tools that enable independent trial discovery. This paper presents a solution that makes use of HL7 FHIR standard, and several AI Agents that empower consumers to independently identify relevant clinical trials using their personal health information. The system employs four specialized agents using HL7 FHIR data: (1) extracting medical conditions from consumer records; (2) pre-filtering trials using vector embeddings; (3) structuring unstructured eligibility criteria; and (4) evaluating patient eligibility against trial requirements. The framework generates relevance scores and provides ranked trial recommendations within 5-10 min, which would typically require days of manual effort. Integrated into a web-based platform, this AI-driven approach enables consumers or patients to directly discover pertinent clinical trials without clinical staff intervention, significantly improving access to potentially life-saving research opportunities.