The most common method in Meta-analysis involving survival data is pooled hazard ratios (HR), and the data base is HR in original studies and 95% confidence interval (95%CI). However, non-reports or incomplete reports of HR in original studies and 95%CI are not uncommon, which limits the conducting of Meta-analysis based on HR. The most of original studies related to survival outcomes can provide survival curves, and the survival outcomes can be completely reviewed through pooling survival curve in the study population. This paper introduces the approach of extracting and pooling survival curve data for generating a pooled survival curve by using R software.%涉及生存资料的Meta分析最常用的方法是对两组比较的风险比(hazard ratio,HR)进行合并,其数据基础是原始研究的HR及其95%可信区间(95%CI).然而,原始研究中HR及其95%CI未报道或不完整报道的情况并不少见,这限制了基于HR的Meta分析的开展.绝大多数涉及生存结局的原始研究均提供了生存曲线,通过对生存曲线进行合并能对人群的生存结局进行全面评估.本文介绍利用R软件实现生存曲线数据提取并对其进行合并以生成汇总生存曲线的方法.
展开▼