面向科研的临床结构化数据预处理方法研究  

Study on the Preprocessing Method of Clinical Structured Data for Scientific Research

作　　者：车贺宾 何昆仑[1] 吴欢 陈媛媛[1] 王万玲 王飞 刘立永 尚亚飞 CHE Hebin;HE Kunlun;WU Huan;CHEN Yuanyuan;WANG Wanling;WANG Fei;LIU Liyong;SHANG Yafei(Medical Innovation Research Department of PLA General Hospital,Beijing 100853,China;Beijing Goodwill Hessian Health Technology Co.Ltd.,Beijing 100085,China)

机构地区：[1]中国人民解放军总医院医学创新研究部,北京100853 [2]北京嘉和海森健康科技有限公司,北京100085

出　　处：《医学信息学杂志》2025年第3期55-60,共6页Journal of Medical Informatics

基　　金：新一代人工智能国家科技重大专项(项目编号:2021ZD0140406)。

摘　　要：目的/意义分析临床回顾性研究中结构化数据的主要问题和预处理技术,为临床数据预处理研究提供新思路。方法/过程基于中国人民解放军总医院多个医学中心10余年历史数据治理实践,梳理一套以数据融合、简约、清洗为主的预处理方法,并总结其效果。结果/结论数据预处理解决了临床科研中历史数据的常见问题,减轻了数据二次加工工作量。基于此建立的临床科研检索系统,有效提高了数据的可用性和价值。Purpose/Significance To analyze the main problems and preprocessing technical methods of clinical structured data in clinical retrospective studies,so as to provide new perspectives for clinical data preprocessing research.Method/Process On the basis of more than 10 years historical data governance in several medical centers of the PLA General Hospital,a set of preprocessing methods focusing on data fusion,simplification and cleaning are sorted out,and the effects are summarized.Result/Conclusion Data preprocessing solves the common problems of historical data in clinical research and reduces the workload of data secondary processing.The clinical research retrieval system is established on this basis to effectively improve the availability and value of data.

关 键 词：临床科研 数据预处理 数据清洗 

分 类 号：R-058[医药卫生]