多水平模型及其在医药政策评价研究领域的应用  

Multilevel model and its application in evaluation of medicine policy intervention

作　　者：任燕[1,2,3] 黄云香 仉元瑾 姚明宏 贾玉龙 杨珉 孙鑫[1,2,3] REN Yan;HUANG Yunxiang;ZHANG Yuanjin;YAO Minghong;JIA Yulong;YANG Min;SUN Xin(Chinese Evidence-based Medicine Center,West China Hospital,Sichuan University,Chengdu 610041,P.R.China;NMPA Key Laboratory for Real World Data Research and Evaluation in Hainan,Chengdu 610041,P.R.China;Sichuan Real World Data Research and Innovation Center,Chengdu 610041,P.R.China;West China School of Public Health,Sichuan University,Chengdu 610041,P.R.China)

机构地区：[1]四川大学华西医院中国循证医学中心,成都610041 [2]国家药品监督管理局海南真实世界数据研究与评价重点实验室,成都610041 [3]四川省真实世界数据技术创新中心,成都610041 [4]四川大学华西公共卫生学院,成都610041

出　　处：《中国循证医学杂志》2021年第12期1474-1479,共6页Chinese Journal of Evidence-based Medicine

基　　金：国家自然科学基金项目(编号:72004149);美国中华医学会基金会项目(编号:CMB19-324);四川省青年科技创新研究团队项目(编号:2020JDTD0015);四川大学华西医院学科卓越发展1·3·5工程项目(编号:ZYYC08003)。

摘　　要：随着区域医疗大数据平台的建立和不断完善,区域医疗大数据在医药政策研究评价中越来越凸显出重要的作用。区域医疗大数据通常具有层次结构,传统的统计分析模型因不能处理层次结构的数据而受到限制,多水平模型是处理层次结构数据的有力统计分析工具,近年来受到国外政策评价研究者的广泛关注,而国内缺乏其在医药政策评价中的应用。本文介绍了多水平模型构建的总体思路及其在医药政策评价领域中的几种常见应用场景,以期为未来利用区域医疗大数据或层次结构数据评价医疗政策实施效果提供方法学参考。With the establishment and development of regional healthcare big data platforms,regional healthcare big data is playing an increasingly important role in health policy program evaluations.Regional healthcare big data is usually structured hierarchically.Traditional statistical models have limitations in analyzing hierarchical data,and multilevel models are powerful statistical analysis tools for processing hierarchical data.This method has frequently been used by healthcare researchers overseas,however,it lacks application in China.This paper aimed to introduce the multilevel model and several common application scenarios in medicine policy evaluations.We expected to provide a methodological framework for medicine policy evaluation using regional healthcare big data or hierarchical data.

关 键 词：多水平模型 医药政策 效果评价 区域医疗大数据 

分 类 号：R19-0[医药卫生—卫生事业管理]