New LLM framework PaReGTA enhances patient representation from EHR data

Researchers have developed PaReGTA, a novel framework that leverages Large Language Models (LLMs) to create patient representations from electronic health records (EHRs). This method converts longitudinal EHR data into text with temporal cues, fine-tunes a sentence-embedding model for domain-specific embeddings, and aggregates these into a fixed-dimensional patient representation. PaReGTA aims to preserve temporal information lost in traditional methods and has shown promising results in a retrospective classification task for migraine patients from the All of Us Research Program, outperforming other approaches in AUC, accuracy, and F1 scores. AI

IMPACT This framework could improve the accuracy and efficiency of analyzing patient data in healthcare, potentially leading to better diagnostic tools and personalized treatment plans.

RANK_REASON The cluster contains an academic paper detailing a new framework for EHR analytics. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• All of Us Research Program
• Bert
• electronic health records
• Kihyuk Yoon
• LightGBM
• PaReGTA
• PaReGTA-RSS

• paper
• model release

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →