Enhancing Clinical Trial Design Through Automated Eligibility Criteria Mapping to Real-World Data (ECLIPSE)
Clinical trial eligibility criteria are critical for identifying patients and evaluating trial feasibility, yet they are often expressed in narrow, inconsistent, and non-computable formats distributed across protocols, registries, and unstructured electronic health record (EHR) notes. This creates significant operational delays and results in up to 50% of promising trials never opening due to uncertain feasibility. Our project, ECLIPSE (Eligibility Criteria Linking for Integrated Patient Selection and Emulation), develops a protocol-first, AI-enabled framework to transform eligibility criteria into machine-readable, standards-based specifications that can be instantiated against both structured and unstructured EHR data. By integrating large language models (LLMs) for note extraction, real-time protocol amendment tracking, and mapping to common data models, ECLIPSE enables rapid trial feasibility assessments, accurate cohort identification, and real-world target trial emulation.