《使用非结构化电子健康数据开展真实世界比较效果和安全性研究的报告规范》要点解读及思考  

Transparent Reporting on Research Using Unstructured Electronic Health Record Data to Generate‘Real World’Evidence of Comparative Effectiveness and Safety:Interpretation and Reflections

作　　者：聂晓璐[1,2] 雷毅[3] 尉耘翠 孙子墨 王青[3] 彭晓霞[1] NIE Xiao-lu;LEI Yi;YU Yun-cui;SUN Zi-mo;WANG Qing;PENG Xiao-xia(Center for Clinical Epidemiology and Evidence-based Medicine,Beijing Children’s Hospital,Capital Medical University,National Center for Children’s Health;Department of Epidemiology and Biostatistics,School of Public Health,Peking University;E-Health Engineering Research Center,Department of Automation,Tsinghua University;Department of Pharmacy,Beijing Children’s Hospital,Capital Medical University,National Center for Children’s Health)

机构地区：[1]国家儿童医学中心/首都医科大学附属北京儿童医院临床流行病与循证医学中心 [2]北京大学公共卫生学院流行病与卫生统计学系 [3]清华大学自动化系数字医疗健康工程研究中心 [4]国家儿童医学中心/首都医科大学附属北京儿童医院临床研究中心 [5]国家儿童医学中心/首都医科大学附属北京儿童医院药学部

出　　处：《中国食品药品监管》2021年第11期47-55,共9页China Food & Drug Administration Magazine

基　　金：国家自然科学基金(No.72174128);首都医科大学附属北京儿童医院国家自然科学基金培育基金(GPQN202005)。

摘　　要：近些年来,利用行政管理和临床保健数据库等常规收集的卫生数据开展真实世界比较效果与安全性的研究越来越多地影响药品监管、报销和其他医疗保健决策。电子健康记录(Electronic Health Records,EHR),尤其是电子病历数据中的非结构化数据蕴含大量症状、体征、诊断相关数据,结合高效可行的临床真实世界数据采集模式,将其整理为可供分析的结构化数据,可以更好地利用这些信息开展研究。目前已发表的多个报告规范详细说明了关于如何规范报告使用常规收集卫生数据开展观察性研究。然而,现有报告规范未对电子医疗记录、登记数据或其他医疗保健数据源中所包含的结构化和非结构化信息加以区分。如何更加透明、规范地报告,即将非结构化文本提取,整理成为可以开展比较效果研究和安全性研究分析的结构化字段,对于此类因果推断研究、结果解释有重要意义。鉴于此,哈佛医学院Shirley V.Wang教授带领的研究团队提出并制定《使用非结构化电子健康数据开展真实世界比较效果和安全性研究的报告规范》。本文对基于非结构化EHR开展真实世界比较效果和安全性研究过程中涉及的专业术语和相关技术进行简单归纳,着重介绍现已发表的报告规范中对于非结构化文本处理,如使用自然语言处理或机器学习方法时需重点报告的核心要点,以期为研究人员今后更好地开展和报告此类研究提供参考。Research that makes secondary use of administrative and clinical healthcare databases is increasingly influential for regulatory,reimbursement,and other healthcare decision-making.Electronic health records(EHRs),especially electronic medical records,contain unstructured data that record information on symptoms,signs and diagnoses.With the help of efficient and practical clinical real-world data collection models,unstructured data can be cleaned and put to better use.Several guidance documents have been published to improve transparency,reproducibility of observational studies using routinely collected health data.However,existing guidance does not differentiate between structured and unstructured information contained in EHRs,registries,or other healthcare data sources.More transparent and standardized reporting on the extraction and organization of unstructured text into structured fields that can be used for comparative effectiveness and safety studies is of great significance to such causal inference research and interpretation.To this end,a research team led by Professor Shirley V.Wang of Harvard Medical School proposed and developed the Transparent Reporting on Research Using Unstructured Electronic Health Record Data to Generate Real World Evidence of Comparative Effectiveness and Safety.This paper summarizes the terminology and technologies involved in real-world comparative effectiveness and safety research based on unstructured EHRs,with a focus on the core points for the transparent reporting of unstructured text processing involving use of natural language processing-or machine learning-derived data fields,to provide reference for future research.

关 键 词：药品监管 电子健康数据 非结构化文本 真实世界研究 报告规范 

分 类 号：R95[医药卫生—药学]