Time-to-event endpoints are prevalently used for efficacy analyses in clinical trials of oncology and many other diseases. This chapter reviews commonly used statistical methods of time-to-event endpoints for oncology clinical trials under the estimand framework. We focus on the considerations for properly aligning the statistical analysis to each intercurrent event strategy. We explore the complex role of censoring within the estimand framework, given its significant influence on the interpretation of treatment effects. Other aspects of the estimand, including variables, population-level summaries, and sensitivity analyses, are also discussed. A case study of a hematology clinical trial is provided at the end to illustrate the considerations to align statistical analyses to corresponding estimands.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Statistical Analysis Under the Estimand Framework: Time-to-Event Endpoints

  • Xin Zhang,
  • Wentian Guo,
  • Chen Lu,
  • Dong Xi,
  • Longshen Xie,
  • Jianfei Zheng

摘要

Time-to-event endpoints are prevalently used for efficacy analyses in clinical trials of oncology and many other diseases. This chapter reviews commonly used statistical methods of time-to-event endpoints for oncology clinical trials under the estimand framework. We focus on the considerations for properly aligning the statistical analysis to each intercurrent event strategy. We explore the complex role of censoring within the estimand framework, given its significant influence on the interpretation of treatment effects. Other aspects of the estimand, including variables, population-level summaries, and sensitivity analyses, are also discussed. A case study of a hematology clinical trial is provided at the end to illustrate the considerations to align statistical analyses to corresponding estimands.