血脂四项检验项目及结果临床描述与国际规范术语集映射
Names and Clinical Result Descriptions of Four-items Blood Lipid Tests and Mapping to the International Standard Terminology
赵云松 1杨鹏 1张林 2谭志军 1梁英 1张海悦 1徐勇勇1
作者信息
- 1. 第四军医大学卫生统计学教研室(卫生信息研究所),西安市,710032
- 2. 石家庄解放军白求恩国际和平医院核医学科,石家庄市,050082
- 折叠
摘要
目的 分析体检报告血脂四项检验项目及结果临床描述常用的中文词条,尝试建立与中文规范术语及与LOINC、SNOMED CT英文术语集的映射关系.方法 以在8家医院进行健康体检的11953名员工体检报告的"血脂四项检验结论"为研究对象,利用R语言文本挖掘tm包及中文分词Rwordseg包,将每一份体检报告的"血脂四项检验结论"的临床描述文档转化为词条矩阵.血脂四项的检验项目名称与LOINC建立映射关系,异常结果中文描述词条与全国科学技术名词审定委员会建议的规范医学名词、《诊断学》、ICD-10的中文规范术语建立映射关系,与SNOMEDCT英文描述建立映射关系.结果 ①中文分词词库包含39个词条,其中17个词条为规范术语,占43.6％;②血脂四项检验项目名称可与LOINC代码完全映射;③17个规范术语有15个可以与SNOMED CT术语集建立映射关系.结论 ①体检医院的"血脂四项检验结论"临床描述术语还需要进一步规范;②血脂四项的检验项目名称可通过《居民健康档案医学检验项目常用代码》与LOINC术语集建立映射关系;③检验结果的临床描述与SNOMED CT术语映射可建立"多对一"的映射关系.
Abstract
Objective This study aims at mining and analyzing the terms of names, clinical result descriptions of four-items blood lipid tests, attempting to establish the mapping relations between the terms of names, clinical descriptions and the Chinese standard terminology, or the international standard terminology systems, such as LOINC, SNOMED CT. Methods Regarding the four-items blood lipid test results of 11953 staffs as research object which come from 8 health check-up institutions, every document of clinical descriptions of four-items blood lipid test results were transformed to term matrix by program package of tm for text mining and program package of Rwordseg for Chinese segmentation. The mapping relations were established between the names of four-items blood lipid tests and LOINC, and other mapping relations were established between the terms of clinical descriptions of abnormal test results and the Chinese standard terminology from China National Committee for Terms in Sciences and Technologies, Diagnostics or ICD-10, and the international standard terminology systems of SNOMED CT. Results ①Among the 39 terms of lexicon database, there are 17 standard terminologies, accounting for 43.6％ of all terms; ② The names of four-items blood lipid tests were mapping the LOINC by "one to one"; ③Among the 17 standard terminologies, there are 15 that could map the terminology systems of SNOMED CT. Conclusions ① It is essential to unify and standardize the clinical descriptions of four-items blood lipid test results which come from health check-up institutions; ② The mapping relations between the names of four-items blood lipid tests and the terminology systems of LOIINC were established by Code for common laboratory tests in health records; ③ The terminology mapping relations between the clinical description of test results and the terminology systems of SNOMED CT were established by "many to one".
关键词
文本数据挖掘/R语言/临床信息管理/术语标准/血脂检验Key words
Text mining/R programming language/Clinical information management/Terminology standard/Blood lipid test引用本文复制引用
基金项目
国家自然科学基金资助项目(81573250)
出版年
2017
NETL
NSTL
维普
万方数据