Retrieval-Reasoning Large Language Model-based Synthetic Clinical Trial Generation

We introduce a novel Retrieval-Reasoning few-shot setting framework that employs LLMs to generate clinical trials with binary labels of success/failure.