药物流行病学杂志2024，Vol.33Issue(9) ：971-977.DOI:10.12173/j.issn.1005-0698.202401024

基于医院信息系统数据的中枢神经系统不良反应自动监测模块构建优化与实践

Establishment,optimization and practice of an automatic central nervous system adverse reactions monitoring module based on hospital information system data

李海艳 郭代红 朱曼 高奥 卢京川 伏安 李超 李鹏 赵安琪
药物流行病学杂志2024，Vol.33Issue(9) ：971-977.DOI:10.12173/j.issn.1005-0698.202401024
✕
  • 维普
  • 万方数据

基于医院信息系统数据的中枢神经系统不良反应自动监测模块构建优化与实践

Establishment,optimization and practice of an automatic central nervous system adverse reactions monitoring module based on hospital information system data

李海艳 1郭代红 2朱曼 2高奥 2卢京川 1伏安 2李超 2李鹏 2赵安琪2
扫码查看

作者信息

  • 1. 中国人民解放军总医院医疗保障中心药剂科(北京 100853);重庆医科大学药学院(重庆 400016)
  • 2. 中国人民解放军总医院医疗保障中心药剂科(北京 100853)
  • 折叠

摘要

目的 构建基于临床药品不良事件主动监测与智能评估警示系统-Ⅱ(ADE-ASAS-Ⅱ)的中枢神经系统不良反应(CNS-ADR)模块,开展亚胺培南西司他丁相关CNS-ADR大样本真实世界主动监测评价研究.方法 以文献、自发报告、电子病历中CNS-ADR相关描述词作为初始词集,利用文本识别技术构建并优化CNS-ADR自动监测模块条件设置.回顾性监测2017-2021年使用注射用亚胺培南西司他丁钠的住院患者,对发生CNS-ADR的阳性病例的人口学特征、中枢神经系统症状、住院科室进行统计描述.结果 基于1 185例人工监测结果进行反复测试优化后,最终确定的模块最佳设置包括62个关键词集,阳性预测值(PPV)为13.63％,召回率为100％.利用该模块拓展监测8 222例用药人群,报警2 366例,PPV为11.88％,关联性阳性281例,发生率为3.42％,其中60岁以上患者占比50.17％,CNS-ADR表现主要为癫痫发作、头痛、躁狂、谵妄等,发生CNS-ADR主要分布在血液科、呼吸科、肿瘤内科等.结论 基于ADE-ASAS-Ⅱ建立的CNS-ADR自动监测模块,为开展CNS-ADR的真实世界研究提供了快捷可靠的文本数据挖掘支持.

Abstract

Objective To construct a module for drug-induced central nervous system adverse reactions(CNS-ADR)within the Clinical Adverse Drug Event Active Monitoring and Intelligent Assessment Alert System-Ⅱ(ADE-ASAS-Ⅱ),and to conduct a large-scale,real-world active monitoring and evaluation of CNS-ADR specifically related to imipenem/cilastatin.Methods Based on literature review,spontaneous report evaluation,and initial word set of CNS-ADR related descriptions in electronic medical records,text recognition technology was used to construct and optimize the condition settings of the CNS-ADR automatic monitoring module.Hospitalized patients using imipenem/cilastatin were retrospectively monitored from 2017 to 2021,and the positive patients which had CNS-ADR were statistically described in terms of the demographic characteristics,CNS symptoms,and hospital departments.Results Based on a repeated testing optimization using 1 185 manually monitored results,the best setting for the determined module includes 62 sets of keywords,with a positive predictive value(PPV)of 13.63％and a recall rate of 100％.Expanding the monitoring to 8 222 medication users using this module,281 cases of positive causality were identified,with an incidence rate of 3.42％.Among them,patients over 60 years old accounted for 50.17％,and the main manifestations of CNS-ADR were epileptic seizures,headaches,mania,and delirium.Conclusion The CNS-ADR automatic monitoring module established based on ADE-ASAS-Ⅱ provides fast and reliable text data mining support for conducting real-world research on CNS-ADR.

关键词

中枢神经系统/药品不良反应/亚胺培南西司他丁/文本分类技术/真实世界研究

Key words

Central nervous system/Adverse drug reaction/Imipenem/cilastatin/Text classification technology/Real world study

引用本文复制引用

基金项目

中国研究型医院学会专项项目(Y2022FH-YWPJ01)

出版年

2024
药物流行病学杂志
中国药学会 武汉医药（集团）股份有限公司

药物流行病学杂志

CSTPCD
影响因子：0.746
ISSN：1005-0698
段落导航相关论文