This research demonstrates how fine-tuned LLMs can help healthcare providers make better medication decisions, reducing the risks of overprescription.

Key insights:

• Large Language Models can be effectively fine-tuned to process Electronic Health Records and recommend appropriate medications
• The approach addresses challenges in adapting to diverse EHR systems by leveraging unstructured data
• Fine-tuned models show improved generalization capabilities over traditional medication recommendation systems
• The solution directly addresses clinical safety risks associated with medication errors and overprescription

For healthcare organizations, this research offers a practical path to implement AI-assisted medication review that works across different clinical systems while prioritizing patient safety.

Addressing Overprescribing Challenges: Fine-Tuning Large Language Models for Medication Recommendation Tasks