[2510.21084] MediRec: Enhancing Chinese language Treatment Suggestion with Explainable Medical Reasoning

View a PDF of the paper titled MediRec: Enhancing Chinese language Treatment Suggestion with Explainable Medical Reasoning, by Juntao Li and 5 different authors

View PDF
HTML (experimental)

Summary:Massive language fashions (LLMs) have proven sturdy potential for medical choice assist via their superior language understanding and reasoning capabilities. Nevertheless, their software to Chinese language medical remedy suggestion stays largely unexplored. Current approaches are primarily developed on English digital well being report datasets and concentrate on coarse-grained remedy code prediction, providing restricted assist for interpretable medical decision-making. On this work, we suggest MediRec, an explainable LLM-based framework for Chinese language remedy suggestion from digital well being data. MediRec combines clinically grounded reasoning-chain distillation with reinforcement studying to enhance each suggestion accuracy and interpretability. Complete experiments on a Chinese language remedy suggestion benchmark present that MediRec achieves sturdy efficiency, with an F1 rating of 0.5813 and a Jaccard rating of 0.4626. Additional analyses point out that MediRec generates clinically believable suggestions with clear reasoning, demonstrating its effectiveness for explainable remedy choice assist in Chinese language healthcare settings.