Automatic Posology Structuration : What role for LLMs?

2025-06-24

AI Summary

French prescription instructions suffer from ambiguity, nonstandard phrasing, and colloquialism, hindering reliable structured extraction and thereby compromising medication safety and clinical decision support. Method: We propose a hybrid pipeline integrating a lightweight Named Entity Recognition and Linking (NERL) system with a fine-tuned large language model (LLM), employing confidence-driven dynamic routing: high-confidence samples are processed by NERL, while low-confidence ones are delegated to the LLM for fine-grained parsing. The approach synergistically combines prompt engineering and parameter-efficient fine-tuning, augmented by a result fusion mechanism. Contribution/Results: Our method achieves 91% structured extraction accuracy under low-latency constraints—substantially outperforming pure prompt-based approaches and matching the performance of conventional rule- or model-based systems. It demonstrates strong clinical deployability and scalability, marking the first application of a confidence-driven LLM–NERL collaborative paradigm to French prescription structuring.

Technology Category

• Natural Language Processing: Information Extraction
• Machine Learning: Large Multimodal Models (LMMs)
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Multilingual and cross-lingual Web search
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation

Abstract

Automatically structuring posology instructions is essential for improving medication safety and enabling clinical decision support. In French prescriptions, these instructions are often ambiguous, irregular, or colloquial, limiting the effectiveness of classic ML pipelines. We explore the use of Large Language Models (LLMs) to convert free-text posologies into structured formats, comparing prompt-based methods and fine-tuning against a "pre-LLM" system based on Named Entity Recognition and Linking (NERL). Our results show that while prompting improves performance, only fine-tuned LLMs match the accuracy of the baseline. Through error analysis, we observe complementary strengths: NERL offers structural precision, while LLMs better handle semantic nuances. Based on this, we propose a hybrid pipeline that routes low-confidence cases from NERL (<0.8) to the LLM, selecting outputs based on confidence scores. This strategy achieves 91% structuration accuracy while minimizing latency and compute. Our results show that this hybrid approach improves structuration accuracy while limiting computational cost, offering a scalable solution for real-world clinical use.

Problem

• Convert free-text posologies into structured formats
• Improve accuracy of medication instruction structuration
• Combine NERL and LLMs for hybrid clinical solution

Innovation

• LLMs convert free-text posologies into structured formats
• Hybrid pipeline combines NERL and LLMs for accuracy
• Confidence-based routing optimizes performance and cost