An evaluation of composite endpoints using data from the EXCITE trial

在许多评估干预效果的临床试验中，一个单一的结果 没有充分捕捉到干预的效果。在这些情况下，一个复合体 将来自几个单独结果的信息结合起来的结果可能更可取。 虽然已经提出了许多针对复合材料的统计检验 结果，这些方法倾向于集中在维护类型I的技术问题上 误码率及其对统计功率的影响。这些都是重要的考虑因素 设计试验，但不涉及其他关键方面，如 将构成临床相关综合衡量标准的个体结果。这 该提案是对在#年使用NICHD数据资源的特别关注通知的回应 二次数据分析。该提案将对来自 在NICHD数据和样本中心(DASH)上提供的EXPICE试验，以进行调查 建议的统计方法与实际的临床试验数据，并深入了解 合并成相关综合结果的各个结果的特征。

In many clinical trials that assess the effect of an intervention, a single outcome does not adequately capture the effect of the intervention. In these cases, a composite outcome that combines information from several individual outcomes may be preferable. Although there have been a number of statistical tests proposed for composite outcomes, these methods tend to focus on the technical issues of maintain the type I error rate and the effects on statistical power. These are important considerations for designing trials, but do not address other key aspects, such as the characteristics of individual outcomes that would make up a clinical relevant composite measure. This proposal is in response to a Notice of Special Interest for using NICHD data resources in secondary data analyses. The proposal will do secondary analyses of the data from the EXCITE trial, available on the NICHD Data and Specimen Hub (DASH), to investigate the proposed statistical methods with actual clinical trial data, and to gain insight into the characteristics of individual outcomes that combine into a relevant composite outcome.