随机对照试验中多重结局处理需进一步指导:方法学文献的范围综述
该范围综述检索Embase、PubMed和Web of Science截至2025年1月16日的文献及6个网站,纳入123篇文章(另从更新检索获25篇、网站检索获8篇),用主题分析总结随机对照试验中多重结局(MOCs)的处理与报告方法学建议。74/123篇(60%)给出6类方法学建议,53/123篇(43%)讨论统计分析方法,多数为对既有方法的改良;现有建议主要关注统计学考量及药品或医疗器械注册试验。
要点
- 74/123篇(60%)文章提出多重结局的6类方法学建议,涉及预设、多重性校正、把握度与样本量等。
- 53/123篇(43%)文章讨论多重结局的统计分析方法,多数为对既有统计方法的改良。
- 现有建议主要聚焦统计学考量及药品或医疗器械注册试验,非注册场景与复杂干预需进一步方法学研究。
核心结论
- 现有关于随机对照试验中使用多重结局的建议主要聚焦统计学考量以及药品或医疗器械注册试验。
- 需进一步开展方法学研究与指导的方面包括:报告在随机对照试验中使用和选择多重结局的理由,以及对非监管场景(包括复杂干预)试验的考量。
研究设计与样本
- 采用范围综述设计,检索Embase、PubMed和Web of Science等文献数据库,检索截至2025年1月16日。
- 纳入标准为:提供关于使用多重结局进行随机对照试验的设计、分析或报告建议的文章,和/或讨论随机对照试验中处理多重结局统计方法的文章。
- 另核查6个特定网站以获取正式指南和报告指南;纳入文章采用主题分析进行总结。
- 检索共获得1716篇文章,其中123篇被纳入,更新检索另获25篇,特定网站检索另识别8篇。
主要发现
- 从123篇中的74篇(60%)识别出关于使用多重结局的方法学建议的6个主要子主题。
- 6个主要子主题为:需要预设结局与分析、多重性校正、把握度与样本量影响、次要结局的多重性、多重结局相关性的考量、多重结局的具体应用。
- 还识别出关于共同主要结局与复合结局的建议,包括其特征、分析方法、报告与挑战。
- 123篇中的53篇(43%)讨论了分析多重结局的统计方法,其中大多数描述对既有统计方法的改良。
局限性
- 原文指出需进一步开展方法学研究与指导的方面包括报告在随机对照试验中使用和选择多重结局的理由,以及对非监管场景(包括复杂干预)试验的考量。
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 review current methodological guidance for handling and reporting of multiple outcomes (MOCs) in randomized controlled trials (RCTs). STUDY DESIGN AND SETTING: A scoping review with bibliographic database searches including Embase, PubMed, and Web of Science up to January 16, 2025 was conducted. Inclusion criteria were articles that: (1) provide advice on the design, analysis, or reporting of RCTs using MOCs; and/or (2) discuss statistical approaches for handling MOCs in RCTs. Six specific websites were also checked for formal and reporting guidelines. Included articles were summarized using thematic analysis. RESULTS: Searches retrieved 1716 articles of which 123 were included with additional 25 articles from updated search. Eight additional articles were identified by the specific website search. Six main subthemes on methodological recommendations for using MOCs were identified from 74 of 123 articles (60%): (1) need to prespecify outcomes and analysis, (2) multiplicity adjustment, (3) power and sample size implications, (4) secondary outcomes multiplicity, (5) considerations of MOCs correlation, and (6) specific applications of MOCs. Recommendations on coprimary and composite outcomes were also identified, including their features, analyses methods, reporting, and challenges. Statistical methods for analyzing MOCs were discussed in 53 of 123 articles (43%), with the majority describing modifications of pre-existing statistical approaches. CONCLUSION: Current recommendations on using MOCs in RCTs focus primarily on statistical considerations and trials of licensing drugs or medical devices. Areas for further methodological research and guidance include reporting of the rationale for the use and selection of MOCs in RCTs and considerations for trials undertaken in nonregulatory setting, including complex interventions.
引用
- Hussein H, Taylor RS, Manyara AM, et al. The need for further guidance on the handling of multiple outcomes in randomized controlled trials: a scoping review of the methodological literature. J Clin Epidemiol. 2025. PMID: 39971166. doi:10.1016/j.jclinepi.2025.111724
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。