输血辅助决策系统构建与应用效果分析

采集医院HIS、LIS、输血管理平台、手术麻醉管理等系统中申请用血、发血、用血相关数据，采用自然语言处理、数据挖掘清洗方式生成用血信息数据库，基于循环神经网络及贝叶斯算法等机器学习构建用血智能评估模型，通过临床红细胞输注前后效果评估、历史用血数据对模型验证修订反馈，不断完善。初步建立了精准、有效的红细胞输注辅助决策系统，该系统可为临床医师提供红细胞输注决策建议，指导临床合理用血，实现精准用血管理。

Construction and Effect Analysis of Blood Transfusion Decision-Making Supporting System

Data related to blood application, blood delivery and blood consumption were collected from the electronic medical record of hospital information system (HIS), laboratory information system (LIS), blood transfusion management platform and surgical anesthesia management system. The blood information database was generated by natural language processing, data mining and cleaning methods. An intelligent evaluation model of blood consumption was built based on machine learning such as circulating neural network and Bayesian algorithm. The model was continuously improved through the effect evaluation before and after clinical red blood cell infusion and the verification and revision feedback of historical blood data. An accurate and effective decision-making system for red blood cell infusion had been established. The system can provide clinicians with red blood cell transfusion decision suggestions, guide clinical rational blood use, and achieve accurate blood use management.