目标最大似然估计在药品安全性主动监测中的应用

中国药物警戒 ›› 2017, Vol. 14 ›› Issue (6): 341-345.

• 统计学在药物警戒中的应用专栏 • 上一篇下一篇

目标最大似然估计在药品安全性主动监测中的应用

韩贺东¹,郭威¹,叶小飞¹,许金芳¹,郭晓晶¹,朱田田¹,史文涛¹,王蒙¹,侯永芳²,贺佳^1,*

¹ 第二军医大学卫勤系卫生统计学教研室，上海 200433；
² 国家食品药品监督管理总局药品评价中心，北京 100045

收稿日期:2017-05-31 修回日期:2017-08-17 出版日期:2017-06-20 发布日期:2017-08-17
通讯作者: 贺佳，女，博导，教授，药物流行病学及新药评价的统计分析。E-mail：hejia@smmu.edu.cn
作者简介:韩贺东，男，在读硕士，流行病与卫生统计学。
基金资助:
国家自然科学基金（81373105）：贝叶斯倾向性评分方法研究及其在药品不良反应信号检测中的应用；上海市公共卫生重点学科（815GWZK0901）：循证公共卫生与卫生经济学；上海市卫计委科研课题（201440379）：药品不良反应信号检测中混杂因素控制的倾向性评分方法研究。

Application of Targeted Maximum Likelihood Estimation in Active Drug Safety Surveillance

HAN He-Dong¹, GUO Wei¹, YE Xiao-Fei¹, XU Jin-Fang¹, GUO Xiao-Jing¹, ZHU Tian-Tian¹, SHI Wen-Tao¹, WANG-Meng¹, HOU Yong-Fang², HE Jia^1,*

¹ Department of Health Statistics, Faculty of Health Service, Second Military Medical University, Shanghai 200433, China;
² Center for Drug Reevaluation,CFDA,Bei jing 100045,China

Received:2017-05-31 Revised:2017-08-17 Online:2017-06-20 Published:2017-08-17

摘要/Abstract

摘要： 目的利用观察性数据进行因果效应的推断一直是研究者关注的焦点。在药物流行病学领域，探索药物疗效或不良反应也需要因果推断。由van deer laan提出的目标最大似然估计（targeted maximum likelihood estimation，TMLE）被证明具有良好的特性，本文对该方法的原理及应用现状做一介绍。方法因果推断常用的方法包括逆概率加权法，G-算法（G-formulation）及一些双稳健的估计方法。通过文献检索，对TMLE法的原理、模型构建、效应估计、统计推断、性质及应用等方面进行回顾与综述，同时比较TMLE法与其他几种因果推断方法的异同。结果与其他因果推断方法相比，TMLE法具有一定的优势，然而其实现过程更复杂。结论 TMLE法是一种双稳健的因果推断方法，能产生目标参数的有效无偏估计，可为TMLE法在药品不良反应主动监测因果推断中的应用提供参考，以期进一步加强药品风险管理。

关键词: 目标最大似然估计, 因果推断, 药物流行病学, 药品不良反应主动监测

Abstract: Objective Causal inference using observational data has long been the focus of epidemiologic researchers. In pharmacoepidemiologic studies, exploring effect of a drug or adverse drug reaction caused by a drug also fits into causal inference. Targeted maximum likelihood estimation (TMLE) proposed by van deer laan is proved of desirable properties, we introduce its principle and application in the study. Method The common methods for causal inference include inverse probability of treatment weighting (IPTW), G-formulation and some double robust estimators. We provide an overview of the theory, model, estimation, statistical inference and properties. Meanwhile, we also compare the performances of several causal effects estimators. Result Compared with other methods, TMLE has certain advantages. However, its implementation is difficult. Conclusion TMLE is a method with double robustness and could also produce efficient unbiased estimator of targeted parameter. There is potential value of TMLE enhancing drug risk management in active drug safety surveillance.

Key words: targeted maximum likelihood estimation, causal inference, pharmacoepidemiolog, active surve illance of adverse drug reactions

中图分类号:

R994.11

韩贺东,郭威,叶小飞,许金芳,郭晓晶,朱田田,史文涛,王蒙,侯永芳,贺佳. 目标最大似然估计在药品安全性主动监测中的应用[J]. 中国药物警戒, 2017, 14(6): 341-345.

HAN He-Dong, GUO Wei, YE Xiao-Fei, XU Jin-Fang, GUO Xiao-Jing, ZHU Tian-Tian, SHI Wen-Tao, WANG-Meng, HOU Yong-Fang, HE Jia. Application of Targeted Maximum Likelihood Estimation in Active Drug Safety Surveillance[J]. Chinese Journal of Pharmacovigilance, 2017, 14(6): 341-345.

参考文献

[1] 王玲.基于医院信息系统开展药品不良反应监测研究[J].中国药物警戒,2015, 12(4):229-231.
[2] 陈超,郭代红,薛万国,等.住院患者药品不良事件主动监测与评估警示系统的研发[J].中国药物警戒,2013,10(7):411-414.
[3] Rubin D B. Estimating causal effects of treatments in randomized and nonrandomized studies[J]. Journal of Educational Psychology, 2014, 66(1):688-701.
[4] Paul W H. Satistics and Causal Inference[J]. Journal of the American Statistical Association, 1986, 81(396): 945-960.
[5] 张颖,杨国彦,王禹毅,等.因果关系模型在中医药观察性研究中应用的探讨[J].现代中医临床,2014,21(5):19-22.
[6] Austin P C. An introduction to propensity-score methods for reducing confounding in observational studies[M]. Multivariate Behavioral Research, 2011, 46(3): 399-424.
[7] Snowden J M, Rose S, Mortimer K M. Implementation of G-computation on a simulated data set: demonstration of a causal inference technique[J]. American Journal of Epidemiology, 2011, 173(7):731-738.
[8] M. J. van der Laan, Rubin D B. Targeted Maximum Likelihood Learning[J]. International Journal of Biostatistics, 2006, 2(1):1043-1043.
[9] Rubin D B. Statistics and Causal Inference: Comment: Which Ifs Have Causal Answers[J]. Journal of the American Statistical Association, 1986, 81(396):961-962.
[10] Greenland S, Robins J M. Identifiability, exchangeability, and epidemiological confounding[J]. International Journal of Epidemiology, 1986, 15(3):413-419.
[11] Petersen M L, Porter K, Gruber S, et al. Diagnosing and responding to violations in the positivity assumption[M].Statistical Methods in Medical Research. 2010:31-54.
[12] Porter K E, Gruber S, M. J. van der Laan, et al. The relative performance of targeted maximum likelihood estimators.[J]. International Journal of Biostatistics, 2010, 7(1):1-34.
[13] M. J. van der Laan, Gruber S. Collaborative double robust targeted maximum likelihood estimation[J]. International Journal of Biostatistics, 2010, 6(1):391-408.
[14] K. L. Moore, M. J. van der Laan. Covariate adjustment in randomized trials with binary outcomes: Targeted maximum likelihood estimation[J]. Statistics in Medicine, 2009, 28(1):39-64.
[15] Rose S, M. J. van der Laan. A double robust approach to causal effects in case-control studies[J]. American Journal of Epidemiology, 2014, 179(6):663-669.
[16] Stitelman O M, De G V, M. J. van der Laan. A general implementation of TMLE for longitudinal data applied to causal inference in survival analysis[J]. International Journal of Biostatistics, 2012, 8(1):207-217.
[17] Rosenblum M, Mark J. van der Laan. Targeted Maximum Likelihood Estimation of the Parameter of a Marginal Structural Model[J]. International Journal of Biostatistics, 2009, 6(2):Article 19.
[18] Wenjing Z, M. J. van der Laan. Targeted Maximum Likelihood Estimation of Natural Direct Effects[M]. U.C. Berkeley Division of Biostatistics Working Paper Series. 2012:1-40.
[19] Schuler M S, Rose S. Targeted Maximum Likelihood Estimation for Causal Inference in Observational Studies[J]. American Journal of Epidemiology, 2017, 185(1): 65-73.
[20] Pang M, Schuster T, Filion K B, et al. Targeted Maximum Likelihood Estimation for Pharmacoepidemiologic Research[J]. Epidemiology, 2016, 27(4):570-577.
[21] Pirracchio R, Yue J K, Manley G T, et al. Collaborative targeted maximum likelihood estimation for variable importance measure: Illustration for functional outcome prediction in mild traumatic brain injuries[J]. Statistical Methods in Medical Research, 2016. DOI: 10.1177/0962280215627335.
[22] Lendle S D, Fireman B, M. J. van der Laan. Targeted maximum likelihood estimation in safety analysis[J]. Journal of Clinical Epidemiology, 2013, 66(8):91-98.
[23] Gruber S, M. J. van der Laan. An application of targeted maximum likelihood estimation to the meta-analysis of safety data[J]. Biometrics, 2013, 69(1):254-262.
[24] M. J. van der Laan, Rose S. Targeted Learning: Causal Inference for Observational and Experimental Data[J]. Targeted Learning Causal Inference for Observational & Experimental Data, 2011.
[25] Gruber S, M. J. van der Laan. tmle : An R Package for Targeted Maximum Likelihood Estimation[J]. Journal of Causal Inference, 2012, 51(13):1161-1189.
[26] Rassen J A, Schneeweiss S. Using high-dimensional propensity scores to automate confounding control in a distributed medical product safety surveillance system[J]. Pharmacoepidemiology & Drug Safety, 2012, 21 S1):41-49.
[27] Gruber S, Mark J. van der Laan. A Targeted Maximum Likelihood Estimator of a Causal Effect on a Bounded Continuous Outcome[J]. International Journal of Biostatistics, 2010, 6(1):Article 26.
[28] Lumley T. Network meta-analysis for indirect treatment comparisons[J]. Statistics in Medicine, 2002, 21(16):2313-2324.
[29] Arbogast P G, Ray W A. Use of disease risk scores in pharmacoepidemiologic studies[J]. Statistical Methods in Medical Research, 2009, 18(1):67-80.

目标最大似然估计在药品安全性主动监测中的应用

Application of Targeted Maximum Likelihood Estimation in Active Drug Safety Surveillance

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

参考文献

相关文章 7

编辑推荐

Metrics

[1]	王丽, 王晓墨, 杨晨露. 基于医疗大数据开展药品上市后监测与评价研究的偏倚及其控制[J]. 中国药物警戒, 2022, 19(3): 239-243.
[2]	黄丽红, 陈峰. 用于支持药械注册的现实世界研究统计学思考[J]. 中国药物警戒, 2022, 19(1): 27-31.
[3]	黎国威, 李立康, 刘颖欣, 王若婷, 吴大嵘. 观察性药物流行病学研究的P值相关问题探讨[J]. 中国药物警戒, 2022, 19(1): 23-26.
[4]	胡方圆, 叶小飞, 许金芳, 郭晓晶, 贺佳. 使用常规收集的卫生数据开展药物流行病学研究报告规范（RECORD-PE）及解读[J]. 中国药物警戒, 2019, 16(3): 144-148.
[5]	朱田田, 叶小飞, 郭晓晶, 张天一, 汪晶, 史文涛, 贺佳. 基于电子医疗数据库的药物流行病学安全性研究的特点及要求[J]. 中国药物警戒, 2014, 11(7): 413-315.
[6]	王大猷. 药品不良反应监测的定义和范畴研究[J]. 中国药物警戒, 2014, 11(12): 732-738.
[7]	刘建平, 邢建民. 循证的药品不良反应评价方法[J]. 中国药物警戒, 2010, 7(1): 12-15.