面向电子病历的医学术语规范化系统设计及实现

Design and Implementation of the Medical Term Standardization System for Electronic Medical Record

电子病历中蕴含着丰富的真实世界医学知识,其所用术语的规范化是该医学知识得以应用的前提,但是由于医生自由书写的原因,术语往往出现不规范的问题。当前的研究往往侧重于术语规范化的算法,没有形成算法推荐与人工校审的闭环。本研究构建的是面向电子病历的术语规范化系统。首先,基于命名实体识别算法,从电子病历自由文本中抽取医学术语;然后,通过术语相似度匹配算法,映射到自主研发的术语集;最后,通过Java技术实现了该系统,支持算法匹配和人工校审。以某医院的电子病历作为案例,进行术语规范化的操作,验证了该系统的可用性和有效性。

Electronic medical record contain a lot of medical knowledge in the real world.The standardization of the terms used in them is the premise for the application of the medical knowledge.However,due to the free writing of doctors,the terms are often non-standard.The current studies often focus on the algorithm of term standardization and the closed loop of algorithm recommendation and manual proofreading are not formed.In this study,the term standardization system for electronic medical records is constructed.First,medical terms are extracted from the free text of electronic medical records based on named entity recognition algorithm;then,the term similarity matching algorithm is used to map to the self-developed term set;and finally,the system is implemented with Java technology and algorithm matching and manual proofreading are supported.Take the electronic medical records of a hospital as an example.The term standardization operation is made and the availability and effectiveness of the system are verified.