利用真实世界数据的上市后药品安全监测的统计学方法综述  

From Real-World Data to Safety Monitoring:A Overview of Statistical Methods in Post-Marketing Safety Surveillance

作　　者：徐圣 王锴 廖珊妹 XU Sheng;WANG Kai;LIAO Shan-mei(BeiGene(Shanghai)Co.,Ltd.;Peking University First Hospital;Peking University Clinical Research Institute)

机构地区：[1]百济神州(上海)生物科技有限公司 [2]北京大学第一医院,北京大学临床研究所

出　　处：《中国食品药品监管》2021年第11期71-82,共12页China Food & Drug Administration Magazine

摘　　要：药品不良反应,特别是危及生命的药品不良反应,是患者、临床医生和监管机构的重要关注点。严重药品不良反应的发现常伴随强制性四期临床试验、说明书中的黑框警告或者最终被迫退出市场。监管方、支付方和供应方一直致力于利用真实世界数据为包括上市后药品安全监测在内的药品全生命周期管理提供高质高效的决策支持。本文按照不同类型的真实世界数据,对上市后药品安全监测相关统计方法进行回顾和概述,包括基于自发报告系统、电子健康记录、保险索赔数据以及混合数据源的方法。不仅涵盖了开发成熟、得到广泛应用的方法(如比值失衡分析法、疫苗监测中的最大序贯统计检验方法等),还包括仍处于开发验证阶段的方法(如肿瘤药物监测中的最大序贯统计检验方法、药物警戒专用数据处理管道等)。本文主要介绍了各方法的适用场景及优劣,并结合实例进行阐述。此外,还就真实世界数据二次使用过程中的注意事项、安全信号再评估的必要性、统计学方法的底层假设、新兴方法在未来的使用和推广进行了讨论。Adverse drug reactions(ADRs),especially life-threatening ADRs,have been a major concern for patients,clinicians,and regulatory agencies.The identification of serious ADRs usually results in mandatory phase IV clinical trials,black box warnings on the label,or even drug withdrawal from the market.Regulators,payers,and health providers have been leveraging real-world data for effective and efficient decision-making throughout the product life cycle.We reviewed several statistical methods for post-marketing safety surveillance,including methods based on spontaneous reporting databases,electronic health records and insurance claim data,as well as methods based on mixed sources.Some of these approaches are welldeveloped and widely applied(e.g.,disproportionality analysis,MaxSPRT in vaccine surveillance),while others are still emerging or experimental(e.g.,MaxSPRT in oncology drug surveillance,pharmacovigilance data processing pipeline).This paper evaluates the applicable scenarios,strengths and weaknesses of these methods,and provides examples.In addition,we discuss the issues of re-using real-world data,the necessity of re-evaluating safety signals,the underlying assumption of those statistical methods,as well as the challenges and opportunities of emerging methods.

关 键 词：真实世界数据 上市后安全监测 统计方法 自发报告系统 纵向观察性数据 

分 类 号：R95[医药卫生—药学]