SYMPCODER introduces a novel approach to medical symptom coding by leveraging Large Language Models with task as context prompting.

• Transforms symptom coding from a multi-step process into a single unified workflow
• Creates a specialized dataset from the Vaccine Adverse Event Reporting System
• Achieves significant improvements in identifying and linking symptoms to standardized medical vocabularies
• Demonstrates how LLMs can effectively handle complex medical text extraction

This innovation matters because accurate symptom coding is critical for pharmacovigilance and vaccine safety monitoring, enabling healthcare systems to detect adverse events more efficiently and improve patient safety.

Task as Context Prompting for Accurate Medical Symptom Coding Using Large Language Models