同一主题的系统综述常见但常未达到关键方法学标准:一项研究之研究
该研究之研究随机抽取2022年发表的107篇系统综述(SR)与Meta分析,提取105个问题,并检索2018至2023年所有相关SR与Meta分析。33个问题(31.4%,95%CI 22.9-41.3)存在重叠SR,中位重叠数为3篇;230篇SR中仅15篇(6.5%)满足六项关键方法学标准;33个重叠问题中7个(21.2%)结果不一致。
要点
- 2022年发表的SR中约三分之一存在2018至2023年发表的重叠SR。
- 230篇SR中仅15篇(6.5%)满足六项关键方法学标准。
- 33个重叠问题中7个(21.2%)的重叠SR结果不一致。
核心结论
- 2022年发表的SR中约三分之一有至少一篇2018至2023年发表的重叠SR,且多数未达到最低方法学标准;约五分之一的重叠问题其SR结果不一致。
研究设计与样本
- 研究之研究:随机抽取2022年发表的107篇SR与Meta分析,提取105个问题,并针对每个问题检索2018至2023年发表的所有SR与Meta分析以评估重叠频率。
- 评估SR是否满足六项关键方法学特征:方案注册、检索主要电子数据库、检索试验注册库、双重筛选与提取、使用Cochrane偏倚风险工具、进行推荐分级的评估、制定与评价(GRADE)评估。
主要发现
- 107篇2022年发表的SR与Meta分析共提取105个不同问题,并识别出2018至2023年发表的123篇其他SR与Meta分析。
- 33个问题(31.4%,95%CI 22.9-41.3)存在重叠SR,每个问题中位重叠SR数为3篇(IQR 2-6;范围2-19)。
- 230篇SR中15篇(6.5%)满足六项关键方法学标准,12个问题(11.4%)至少有一篇SR满足该标准。
- 33个存在重叠SR的问题中,7个(21.2%)的重叠SR结果不一致。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
原文
OBJECTIVES: To 1) assess the frequency of overlapping systematic reviews (SRs) on the same topic including overlap in outcomes, 2) assess whether SRs meet some key methodological characteristics, and 3) describe discrepancies in results. STUDY DESIGN AND SETTING: For this research-on-research study, we gathered a random sample of SRs with meta-analysis (MA) published in 2022, identified the questions they addressed and, for each question, searched all SRs with MA published from 2018 to 2023 to assess the frequency of overlap. We assessed whether SRs met a minimum set of six key methodological characteristics: protocol registration, search of major electronic databases, search of trial registries, double selection and extraction, use of the Cochrane Risk-of-Bias tool, and Grading of Recommendations, Assessment, Development, and Evaluations assessment. RESULTS: From a sample of 107 SRs with MA published in 2022, we extracted 105 different questions and identified 123 other SRs with MA published from 2018 to 2023. There were overlapping SRs for 33 questions (31.4%, 95% CI: 22.9-41.3), with a median of three overlapping SRs per question (IQR 2-6; range 2-19). Of the 230 SRs, 15 (6.5%) met the minimum set of six key methodological characteristics, and 12 (11.4%) questions had at least one SR meeting this criterion. Among the 33 questions with overlapping SRs, for 7 (21.2%), the SRs had discrepant results. CONCLUSION: One-third of the SRs published in 2022 had at least one overlapping SR published from 2018 to 2023, and most did not meet a minimum set of methodological standards. For one-fifth of the questions, overlapping SRs provided discrepant results.
引用
- Kwok W, Dallant T, Martin G, et al. Systematic reviews on the same topic are common but often fail to meet key methodological standards: a research-on-research study. J Clin Epidemiol. 2025. PMID: 41110775. doi:10.1016/j.jclinepi.2025.112018
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。