药物流行病学研究混杂因素的控制  

Guide on Confounders Control in Pharmacoepidemiology

作　　者：王树月 陈明珠 张兴会 贾敏 沈智文 王俊彦 张灵健 宗晓郁 李慧[1] 周锦 吴明智[1] 原哲 赵振营[3] 杨长青 赵立波 邱晓燕[6] 方世平[7] 詹思延 宫建[1] Wang Shuyue;Chen Mingzhu;Zhang Xinghui;Jia Min;Shen Zhiwen;Wang Junyan;Zhang Lingjian;Zong Xiaoyu;Li Hui;Zhou Jin;Wu Mingzhi;Yuan Zhe;Zhao Zhenying;Yang Changqing;Zhao Libo;Qiu Xiaoyan;Fang Shiping;Zhan Siyan;Gong Jian(Research Group of Jian Gong on Pharmacoepidemiology and Clinical Drug Evaluation,Shenyang Pharmaceutical University,Shenyang 110016,China;Department of Pharmacy,School of Heping Hospital,General Hospital of Northern Theater Command of China Medical University;Department of Pharmacy,Tianjin People’s Hospital;Department of Clinical Pharmacy,School of Basic Medicine and Clinical Pharmacy,China Pharmaceutical University;Beijing Children’s Hospital Clinical Research Center,Capital Medical University;Huashan Hospital Affiliated to Fudan University;Department of Pharmacy,Zhongnan Hospital of Wuhan University;Department of Epidemiology and Biostatistics,Peking University School of Public Health)

机构地区：[1]沈阳药科大学药物流行病与临床药物评价课题组,沈阳110016 [2]中国医科大学北部战区总医院和平院区学院药学部 [3]天津市人民医院药学部 [4]中国药科大学基础医学与临床药学学院临床药学教研室 [5]首都医科大学附属北京儿童医院临床研究中心 [6]复旦大学附属华山医院 [7]武汉大学中南医院药学部 [8]北京大学医学部公共卫生学院

出　　处：《药物流行病学杂志》2021年第8期503-507,共5页Chinese Journal of Pharmacoepidemiology

摘　　要：中国药学会于2019年9月发布《中国药物流行病学研究方法学指南》(以下简称"《指南》")。本文检索2007~2021年Web of Science、CNKI等数据库中有关药物流行病学研究中混杂偏倚控制的相关文献,依据《指南》相关内容,进一步描述混杂偏倚控制,为药物流行病学研究中的混杂偏倚控制提出新思路。本文发现,混杂因素包括单因素、多因素和未知混杂因素,应用较广泛的混杂偏倚控制方法有分层分析、Logistic回归、倾向性评分、工具变量分析和敏感性分析等。《指南》提示,药物流行病学的研究质量需结合健康中国战略,有效控制研究中不可避免的混杂偏倚,提高药物流行病研究结果的安全、有效性,激励公共医疗健康的创新,促进全民健康发展。Chinese Pharmaceutical Association(CPA)issued Guide on Methodological Standards in Pharmacoepidemiology On September,2019.Web of Science and CNKI were searched to collect studies about the control of confounding bias in pharmacoepidemiology from 2007 to 2021,and further describes the control of confounding bias based on the relevant content of Guide on Methodological Standards in Pharmacoepidemiology to propose new ideas for the control of confounding bias in pharmacoepidemiology research.This paper found that confounding factors include single factor,multiple factors and unmeasured confounding factors.The more widely used confounding bias control methods include stratified analysis,logistic regression,propensity score,instrumental variable analysis and sensitivity analysis.The Guide on Methodological Standards in Pharmacoepidemiology suggests that the quality of pharmacoepidemiology research needs to be combined with the strategy of healthy China to effectively control the inevitable confounding bias in research.Improve the safety and effectiveness of research results on pharmacoepidemiology,encourage innovation in public medical and health care,and then promote the healthy development of the whole people.

关 键 词：混杂偏倚控制 分层分析 多因素调整分析 倾向性评分分析 工具变量分析 敏感性分析 

分 类 号：R181.3[医药卫生—流行病学]