Matching patients to clinical trials is a major hurdle in medical research. Current methods, relying on manual screening of medical records, are slow and expensive. This talk explores the potential of Large Language Models (LLMs) to automate this process. We present a two-stage approach using LLMs trained on medical data. First, a retrieval stage efficiently narrows down potential trials for a patient. Then, a zero-shot LLM analyzes the patient's medical history (unstructured text) to determine their eligibility for the shortlisted trials. Advanced prompting techniques guide the LLM for faster and more accurate assessments. This method offers significant promise for streamlining clinical trial recruitment, accelerating medical progress, and potentially leading to new breakthroughs.
