基于真实世界数据构建严重药品不良反应信号数据挖掘模式——以胆木制剂为例进行实证  

A Data Mining Model for Serious Adverse Drug Reaction Signals Based on Real-World Data:Taking Nauclea officinalis Pierre ex Pitard Preparations as an Example for Empirical Analysis

作　　者：黄凌 苗会青 苏小洁 杨婧斐 贺梦娇 柳鹏程[2] 林凯 宋海波 HUANG Ling;MIAO Hui-qing;SU Xiao-jie;YANG Jing-fei;HE Meng-jiao;LIU Peng-cheng;LIN Kai;SONG Hai-bo(Center for Pharmacovigilance of Hainan Province;School of International Pharmaceutical Business,China Pharmaceutical University;Hainan Center for Drug and Medical Device Evaluation and Service;Center for Drug Reevaluation,NMPA)

机构地区：[1]海南省药物警戒中心 [2]中国药科大学国际医药商学院 [3]海南省药品和医疗器械审评服务中心 [4]国家药品监督管理局药品评价中心

出　　处：《中国食品药品监管》2024年第10期102-111,共10页China Food & Drug Administration Magazine

摘　　要：在当前我国药品监管机构信号检测工作的基础上,本文利用真实世界数据以及数据挖掘技术,探索构建一种可以及时发现、准确识别严重药品不良反应的循证证据新模式。该模式基于自发呈报系统数据库和医疗电子病历系统数据库,将药品不良反应信号检测分为信号初步发现、信号筛选、信号验证与信号评价四个环节,通过数据挖掘技术快速高通量地挖掘药品不良反应信号,逐步提升信号的证据等级,提高信号检测效率。与此同时,本文使用该模式进行了胆木制剂不良反应信号数据挖掘的实证研究,初步验证了该模式的可操作性,为药品监管部门顺利开展严重药品不良反应信号数据挖掘工作提供了范式经验。Based on the current signal detection work of Chinese drug regulatory agencies,this paper explores the construction of a new evidence-based model for the timely detection and accurate identification of severe adverse drug reactions(ADRs)utilizing real-world data(RWD)and data mining techniques.This model,based on the spontaneous reporting system database and the electronic medical records system database,divides ADR signal data mining into four stages:initial signal detection,signal screening,signal verification,and signal evaluation.By applying data mining techniques,the model allows for rapid,highthroughput mining of ADR signals,which gradually improves the level of evidence of the signals and improves the efficiency of signal detection.In this paper,an empirical study using this model for the data mining of adverse reactions of Nauclea officinalis Pierre ex Pitard preparations is presented,initially verifying the model's operability and providing a paradigmatic experience for regulatory authorities to carry out data mining on severe ADR signals.

关 键 词：真实世界数据 严重药品不良反应 数据挖掘 信号检测 循证证据 

分 类 号：R95[医药卫生—药学]