以手法为主要干预措施的随机对照试验在样本量估计方面的常见问题分析  

Common problems in sample size estimation of randomized controlled trials with manipulation as the primary intervention measure

作　　者：冯天笑 肖想玉 秦晓宽 孙传睿 王旭 卜寒梅 朱立国[1] 胡晶[2] 魏戌[1] FENG Tianxiao;XIAO Xiangyu;QIN Xiaokuan;SUN Chuanrui;WANG Xu;BU Hanmei;ZHU Liguo;HU Jing;WEI Xu(Wangjing Hospital of CACMS,Beijing 100102,China;Beijing Hospital of Traditional Chinese Medicine Affiliated to the Capital Medical University,Beijing 100191,China)

机构地区：[1]中国中医科学院望京医院,北京100102 [2]首都医科大学附属北京中医医院,北京100191

出　　处：《中医正骨》2024年第10期39-43,51,共6页The Journal of Traditional Chinese Orthopedics and Traumatology

基　　金：国家重点研发计划项目(2021YFC1712800,2021YFC1712802);北京市科技新星计划交叉合作课题(20220484228);中国中医科学院望京医院高水平中医医院建设项目——中医药临床循证研究专项(WJYY-XZKT-2023-05)。

摘　　要：目的:分析以手法为主要干预措施的随机对照试验(randomized controlled trial,RCT)在样本量估计方面存在的问题。方法:应用计算机检索PubMed、Embase和Cochrane Library中以手法为主要干预措施的RCT,检索时限均为2017年1月至2024年6月。依据文献检索及筛选方案筛选出符合要求的文献后,由2名经过培训的研究人员独立提取是否报告样本量估计、假设检验类型、Ⅰ类错误概率α、把握度(1-β)、预期效应量、单侧或双侧检验、最小临床重要性差值(minimal clinically important difference,MCID)、脱落率、样本量计算工具、主要结局指标及多重检验等信息,以频数和百分比的形式进行描述性分析。结果:共检索到48688篇文献,去除重复文献后随机抽取20%的文献,通过筛选最终纳入274篇文献。在纳入的274项研究中,192项(70.07%)报告了样本量估计。在报告了样本量估计的192项研究中,16项(8.33%)报告了假设检验类型,146项(76.04%)报告了Ⅰ类错误概率α,164项(85.42%)报告了把握度(1-β),98项(51.04%)报告了预期效应量,88项(45.83%)报告了单侧或双侧检验,81项(42.19%)报告了MCID,156项(81.25%)报告了脱落率,167项(86.98%)报告了样本量计算工具,131项(68.23%)报告了主要结局指标。16项报告了假设检验类型的研究均为优效性试验,其中9项采用单侧检验,7项采取双侧检验;37项报告了2个及以上主要结局指标的研究中,8项(21.62%)调整了Ⅰ类错误概率α;23项报告了多个组别比较的研究中,1项(4.35%)调整了Ⅰ类错误概率α。结论:以手法为主要干预措施的RCT在样本量估计方面,存在未报告样本量估计、样本量估计参数缺失或不明、样本量估计方法与假设检验类型及主要结局指标不对应、多重检验时未调整Ⅰ类错误概率α等问题。Objective:To analyze the common problems in sample size estimation of randomized controlled trial(RCT)with manipulation as the primary intervention measure.Methods:All the RCT articles with manipulation as the primary intervention measure included from January 2017 to June 2024 were retrieved from the PubMed,Embase and Cochrane Library through computer.The eligible articles were screened according to the retrieval and screening scheme,and the information,including whether reporting sample size estimation,hypothesis test type,typeⅠerror probabilityα,power of test(1-β),expected effect size,one-sided or two-sided test,minimal clinically important difference(MCID),dropout rate,sample size calculation tool,primary outcome indicators,and multiple tests,was extracted independently by two trained researchers for a descriptive analysis in the form of frequency or percentage.Results:Forty-eight thousand six hundred and eighty-eight RCT articles were searched out,after removing the duplicate articles,20%of the remained ones were randomly selected for further screening,and 274 RCTs were included in the final analysis,among which 192 ones(70.07%)reported the sample size estimation.Furthermore,among the 192 RCTs,16 ones(8.33%),146 ones(76.04%),164 ones(85.42%),98 ones(51.04%),88 ones(45.83%),81 ones(42.19%),156 ones(81.25%),167 ones(86.98%),and 131 ones(68.23%)reported the hypothesis test type,typeⅠerror probabilityα,power of test(1-β),expected effect size,one-sided or two-sided test,MCID,dropout rate,sample size calculation tool,and primary outcome indicators,respectively.In addition,16 RCTs that reported the type of hypothesis test were all superiority trials,with 9 ones using the one-sided test and 7 ones employing the two-sided test.Among the 37 RCTs that reported 2 or more primary outcome indicators,8 ones(21.62%)mentioned the adjustment of typeⅠerror probabilityα.Among the 23 RCTs that reported multiple comparisons,one RCT(4.35%)referred to the adjustment of typeⅠerror probabilityα.Conclusion:Problems,such

关 键 词：按摩(中医) 正骨手法 舒筋整复手法 推拿疗法 随机对照试验专题 样本量 

分 类 号：R274[医药卫生—中医骨伤科学]