Demo: A Framework for Ranking The Clinical Trials with Large Language Model Prompting

In this study, we propose a framework for clinical trial ranking with the integration of Large Language Models (LLMs). Leveraging the strength of LLMs, we expect to accelerate and enhance the trial ranking process with the development of a vector database generated by encoding some crucial parts of MIMIC-IV discharge notes, such as medication of admission, history of illness, and brief hospital course, along with the list of clinical trials and their associated PubMed literature. Based on the filtered list of clinical trials extracted from the vector database, we prompt LLMs to determine their relevance to the PubMed literature. We aim to investigate the efficacy and efficiency of LLMs integration to accelerate clinical trial ranking and intelligent recruitment of clinical trials.