This research leverages machine learning and large language models to predict the necessity of blood cultures in emergency departments, reducing unnecessary testing and antibiotic use.

• ML models achieved up to 0.79 AUC for bacteremia risk prediction using structured EHR data
• LLMs analyzed provider notes to enhance prediction capabilities
• The approach helps address antibiotic shortages and optimize healthcare resource allocation
• Implementation could significantly reduce unnecessary blood cultures and improve clinical decision-making

Why it matters: Unnecessary blood cultures strain healthcare resources and contribute to inappropriate antibiotic use during global shortages. This ML-driven approach provides a practical solution for more efficient medical resource utilization.

A Multi-Phase Analysis of Blood Culture Stewardship: Machine Learning Prediction, Expert Recommendation Assessment, and LLM Automation