中国药物警戒 ›› 2023, Vol. 20 ›› Issue (9): 1011-1016.
DOI: 10.19803/j.1672-8629.20220736

• 安全与合理用药 • 上一篇    下一篇

基于自然语言处理技术的儿童病例注射用尖吻蝮蛇血凝酶不良反应主动监测方法构建与验证研究

邰隽1, 于淼2, 张琳3, 雷毅3, 刘军廷4, 李涛4, 黄贵民4, 程毅菁4, 聂晓璐5,6, 王青3,7,*   

  1. 1首都儿科研究所附属儿童医院耳鼻咽喉头颈外科,北京 100020;
    2大连理工大学系统工程所,辽宁 大连 116024;
    3清华大学自动化系,药物警戒信息技术与数据科学创新中心,北京 100084;
    4首都儿科研究所,儿童健康大数据研究中心,北京 100020;
    5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,北京 100045;
    6北京大学公共卫生学院流行病与卫生统计学系,北京 100091;
    7清华海峡研究院,药物警戒信息技术与数据科学研究中心,福建 厦门 36100
  • 收稿日期:2022-12-26 出版日期:2023-09-15 发布日期:2023-09-14
  • 通讯作者: *王青,男,博士,研究员,人工智能辅助诊断技术。E-mail: qing.wang@tsinghua.edu.cn
  • 作者简介:邰隽,男,博士,教授,儿童咽喉头颈常见及重大疾病。
  • 基金资助:
    国家自然科学基金资助项目（82204149）

Active monitoring of adverse reactions related to haemocoagulase agkistrodon for injection in children

TAI Jun1, YU Miao2, ZHANG Lin3, LEI Yi3, LIU Junting4, LI Tao4, HUANG Guimin4, CHENG Yijing4, NIE Xiaolu5,6, WANG Qing3,7,*   

  1. 1Otorhinolaryngology Head and Neck Surgery, Children's Hospital, Capital Institute of Pediatrics, Beijing 100020, China;
    2Institute of Systems Engineering, Dalian University of Technology, Dalian Liaoning 116024, China;
    3Research Center for Pharmacovigilance IT and Data Science(PVID), Department of Automation, Tsinghua University, Beijing 100084, China;
    4Child Health Big Data Research Center, Capital Institute of Pediatrics, Beijing 100020, China;
    5Center for Clinical Epidemiology and Evidence-based Medicine, Beijing Children's Hospital, Capital Medical University, National Center for Children's Health, Beijing 100045, China;
    6Department of Epidemiology and Biostatistics, School of Public Health, Peking University, Beijing 100091, China;
    7Pharmacovigilance Research Center for Information Technology and Data Science, Cross-strait Tsinghua Research Institute, Xiamen Fujian 36100, China
  • Received:2022-12-26 Online:2023-09-15 Published:2023-09-14

PDF (PC)

129

摘要： 目的 探索真实世界大样本儿童病例注射用尖吻蝮蛇不良反应主动监测方法,以期为儿童病例注射用尖吻蝮蛇血凝酶不良反应的揭示和挖掘提供支持。方法 利用实体识别技术对2014年2月25日至2021年1月8日首都儿科研究所住院并使用注射用尖吻蝮蛇血凝酶的35 720儿童病例进行回顾性不良反应主动监测,采用查准率（precision,P）、查全率（recall,R）和F1值3个指标来评估实验结果。结果 经过3轮分析,查准率、查全率和F1值3个指标均达到了100%。结论 应用自然语言处理技术可以用于真实世界大样本儿童病例注射用尖吻蝮蛇血凝酶不良反应监测。

关键词: 尖吻蝮蛇血凝酶, 注射剂, 儿童, 药品不良反应, 主动监测, 自然语言处理

Abstract: Objective To explore methods of active monitoring of adverse reactions related to haemocoagulase agkistrodon for injection in children so as to contribute to revelation and mining of the adverse reactions. Methods Using entity recognition technology, adverse reactions in children who used haemocoagulase agkistrodo and were hospitalized at the Capital Institute of Pediatrics between February 25, 2014 and January 8, 2021 were monitored actively and retrospectively. The results of experiments were assessed using three indicators: Precision (P), Recall (R) and F1 value. Results After three rounds of analysis, all the three indicators achieved 100%. Conclusion The application of natural language processing technology can be used to monitor adverse reactions related to haemocoagulase agkistrodon for injection in children.

Key words: haemocoagulase agkistrodon, injection, children, adverse drug reaction, active monitoring, natural language processing

中图分类号: