网络Meta分析研究进展系列(二十):网络Meta分析的样本量计算及精确性评估

Advances in methodology of network meta-analysis(20):sample size calculation and precision assessment in network meta-analysis

本文着重介绍网络Meta分析所需样本量的估算方法,包括有效合并研究数量法、有效样本量法和等效样本量法,并辅以实例演示计算步骤。在此基础上,我们进一步把样本量和效应估计值的精确性联系起来,介绍网络Meta分析的GRADE证据分级系统中对效应估计值精确性的评估。研究者在进行网络Meta分析后应评估当前样本量是否满足检验效能的要求,并据此对证据的精确性的升降级给出判断意见。

This review introduces the methods of calculating sample size required in network meta-analysis,including the approaches for the effective number of trials,the effective sample size and the equivalent sample size.Given the intrinsic link between sample size and estimate precision,we further introduce the precision assessment in GRADE system for network meta-analysis.It is recommended that whether sample size available in the network meta-analysis satisfies the requirements for statistical power should be usually assessed,which would be helpful for the grading assessment of evidence precision.