贝叶斯统计在儿童用药有效性外推中的应用  

Bayesian Statistics for Extrapolation of Efficacy in Pediatric Medicines

作　　者：李若冰 言方荣 王骏 LI Ruobing;YAN Fangrong;WANG Jun(Center for Drug Evaluation,National Medical Products Administration,Beijing 100022;Research Center of Biostatistics and Computational Pharmacy,School of Science,China Pharmaceutical University,Nanjing 211198)

机构地区：[1]国家药品监督管理局药品审评中心,北京100022 [2]中国药科大学理学院,生物统计与计算药学研究中心,江苏南京211198

出　　处：《中国医药工业杂志》2022年第11期1606-1611,共6页Chinese Journal of Pharmaceuticals

摘　　要：由于儿童人群的生理特殊性,使用外推技术将成人人群数据外推至儿童人群是当前儿童用药的研发热点。贝叶斯统计作为一种可更充分利用已获得临床数据的灵活方法,正成为儿童用药有效性外推中的有效工具。文章在介绍儿童临床研究特点的基础上,阐述了贝叶斯统计的基本概念和基础模型框架,并介绍了幂先验、混合先验和meta分析预测先验,以及模拟技术在贝叶斯统计中的重要作用。然后介绍了一项使用稳健meta分析预测先验借用成人人群历史数据的儿童人群多发性硬化症研究案例。最后讨论了贝叶斯统计在儿童用药外推中其他领域的应用场景,以及实际使用中的重点关注内容。Due to the physiological characteristic of the pediatric population,extrapolation of clinical data from the adult population to the pediatric population is a development focus on pediatric medication.Bayesian statistics,as a flexible method to efficiently utilize the occupied early clinical data,is becoming a powerful tool in the extrapolation of efficacy in pediatric medicine.Based on the introducing of pediatric research features,this paper illustrates the basic concepts and framework of Bayesian statistics,introduces power priors,mixed priors and meta-analysis predictive priors,as well as the important role of simulation technology in Bayesian statistics.Then,a case study of multiple sclerosis in a pediatric population using robust meta-analysis predictive priors to borrow historical data from adult population is present.Finally,this paper discusses other fields of Bayesian statistics in pediatric medication extrapolation and the key consideration in clinical development.

关 键 词：儿童用药 有效性外推 贝叶斯统计 

分 类 号：R945[医药卫生—微生物与生化药学]