[论文] DIASENTINEL: An Auditable Multi-Agent System for Guideline-Grounded Di...

论文概要

研究领域: NLP 作者: Yung Wei Shueh, Zhi-Jie Chen, Chia-Hsuan Hsu 发布时间: 2025-09-01 arXiv: 2509.00147

中文摘要

大型语言模型（LLM）为临床决策支持提供了前景，但仍易受幻觉事实、无支持建议和引用错误的影响。我们提出DIASENTINEL，一个完全本地部署的多智能体系统，用于从电子健康记录（EHR）进行一年期2型糖尿病（T2DM）风险筛查和基于指南的报告生成。该系统集成了校准风险预测、确定性临床信号提取、基于美国糖尿病协会（ADA）指南的互惠排名融合，以及结合基于规则检查与LLM蕴涵的混合验证层。该演示提供实时批量筛查仪表板和交互式患者报告界面，包含引用建议、验证结果和原始EHR比较。DIASENTINEL展示了一个用于可靠、可审计和隐私保护的基于LLM的临床决策支持的实用框架。

原文摘要

Large language models (LLMs) offer promising clinical decision support but remain vulnerable to hallucinated facts, unsupported recommendations, and citation errors. We present DIASENTINEL, a fully on-premise multi-agent system for one-year type 2 diabetes mellitus (T2DM) risk screening and guideline-grounded report generation from electronic health records (EHRs). The system integrates calibrated risk prediction, deterministic clinical signal extraction, Reciprocal Rank Fusion over American Diabetes Association (ADA) guidelines, and a hybrid verification layer combining rule-based checks with LLM蕴涵. The demonstration provides a real-time batch-screening dashboard and an interactive patient报告界面 with cited recommendations, verification results, and raw EHR comparison. DIASENTINEL demonstrat...

#论文 #arXiv #NLP #小凯