最小临床意义差值的确定和应用  

Determination of the Minimal Clinically Important Difference and Its Application

作　　者：杨丽虹[1] 肖晶旻 刘少南[1] 吴大嵘[1] 葛龙 郭新峰[1] YANG Li-Hong;XIAO Jing-Min;LIU Shao-Nan;WU Da-Rong;GE Long;GUO Xin-Feng(The Second Affiliated Hospital of Guangzhou University of Chinese Medicine,Guangdong Provincial Hospital of Traditional Chinese Medicine,Guangzhou 510120 Guangdong,China;Evidence-based Social Science Research Center,School of Public Health,Lanzhou University,Lanzhou 730000 Gansu,China)

机构地区：[1]广州中医药大学第二附属医院/广东省中医院,广东广州510120 [2]兰州大学公共卫生学院循证社会科学研究中心,甘肃兰州730000

出　　处：《广州中医药大学学报》2022年第4期959-965,共7页Journal of Guangzhou University of Traditional Chinese Medicine

基　　金：国家重点研发计划项目(编号:2019YFC1709805);广东省中医药局科研项目(编号:20201151)。

摘　　要：最小临床意义差值(minimal clinically important difference,MCID)的常用估算方法包括效标法、分布法和文献分析法等。目前MCID的估算方法众多且不统一,给MCID的确定、结果解释和应用带来了困难。鉴于目前各种估算方法均存在一定局限,建议同时使用多种估算方法获得多个MCID初步估算值,并以效标法为主、其他方法为辅或将多种方法通过统计整合的估算值来确定最终的MCID。MCID可协助进行临床研究结果的临床意义判断、样本量估算以及临床决策等,在具体应用之前,应充分了解该MCID的估算方法和样本特征等相关细节以判断是否适用于所开展的研究或临床场景。Common methods for ascertaining minimal clinically important difference(MCID)usually include anchor-based method, distribution-based method and literature analysis method. At present, the estimation methods of MCID are numerous and inconsistent,which brings difficulties to the determination,interpretation and application of MCID. In view of the limitations of all current estimation methods, it is recommended that multiple estimation methods can be used simultaneously to obtain multiple preliminary MCID estimates,and that the final MCID can be determined by using the anchor-based method as the primary method and the remaining methods as the secondary method or by statistically integrating the estimates from multiple methods. MCID can assist to interpret the clinical significance of the clinical research results,estimate the sample size,and make the clinical decision. For the application of MCID, fully understanding of the characteristics of the methods for ascertaining MCID and the sample size should be made firstly,so as to ensure the application of MCID meeting the research and clinical situation.

关 键 词：最小临床意义差值(MCID) 效标法 分布法 文献分析法 临床意义解释 样本量估算 临床决策 

分 类 号：R195.1[医药卫生—卫生统计学]