This research demonstrates how Large Language Models can analyze patient medical records to predict drug overdose risk with superior accuracy compared to traditional machine learning approaches.

• Uses LLMs' natural language processing capabilities to interpret complex longitudinal medical data
• Enables earlier and more accurate identification of at-risk patients
• Provides healthcare providers with actionable insights for timely intervention
• Integrates medical knowledge embedded within LLMs to enhance prediction performance

This innovation represents a significant advancement for clinical decision support systems, potentially saving lives through earlier intervention in high-risk cases and optimizing resource allocation in healthcare settings.

Large Language Models for Drug Overdose Prediction from Longitudinal Medical Records