Tierney Jayne F, Burdett Sarah, Fisher David J
MRC Clinical Trials Unit, Medical Research Council Clinical Trials Unit, University College London, London, UK.
Syst Rev. 2025 Apr 10;14(1):84. doi: 10.1186/s13643-025-02752-z.
Our previous guide to estimating hazard ratios (HRs) from published summary (aggregate) data has become very widely used, but many still have difficulties knowing when and how to apply the methods. Informed by our increased experience of applying them across a range of settings, the queries we have received and results of a survey of Cochrane editors on the methods, we have updated the guidance comprehensively. Previously, we described a range of scenarios for deriving a HR and logrank variance (V) from published time-to-event analyses. They are incorporated in this update, together with clarification of ambiguities and additional scenarios. We also provide further guidance on extracting and using data from publications and Kaplan-Meier (KM) curves, raise some of the challenges, and discuss recent alternatives to the "Parmar" KM methods. A new calculations spreadsheet will perform all possible calculations given the data that are entered and includes new features to enhance the user experience. This updated guidance and associated spreadsheet represent valuable additional tools for those conducting meta-analyses based on published, summary, time-to-event data.
我们之前关于从已发表的汇总数据中估计风险比（HRs）的指南已被广泛使用，但许多人在了解何时以及如何应用这些方法时仍有困难。基于我们在一系列环境中应用这些方法的更多经验、收到的疑问以及对Cochrane编辑关于这些方法的调查结果，我们对该指南进行了全面更新。之前，我们描述了一系列从已发表的事件发生时间分析中得出HR和对数秩方差（V）的情景。它们被纳入本次更新中，同时对模糊之处进行了澄清并增加了其他情景。我们还提供了关于从出版物和Kaplan-Meier（KM）曲线中提取和使用数据的进一步指导，提出了一些挑战，并讨论了“Parmar”KM方法的最新替代方法。一个新的计算电子表格将根据输入的数据执行所有可能的计算，并包含增强用户体验的新功能。这份更新后的指南及相关电子表格为那些基于已发表的汇总事件发生时间数据进行荟萃分析的人提供了有价值的额外工具。
