Design and Analysis of Group Sequential Trials for Repeated Measurements When Pipeline Data Occurs: A Tutorial

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Statistical Inference

Background:

• Group sequential trials (GST) enable early stopping for efficacy or futility.
• Established GST methods are less developed for delayed or repeated outcome measurements.
• Pipeline subjects with early data present analytical challenges in interim analyses.

Purpose of the Study:

• To provide guidance on planning and analyzing GST with repeated measurements and delayed endpoints.
• To enhance statistical power by incorporating early outcome data from pipeline subjects.
• To offer practical tools, including an R package, for implementing these advanced GST methods.

Main Methods:

• Discussing and expanding upon existing statistical methods for GST with delayed endpoints.
• Developing computational details for valid p-values and confidence intervals.
• Incorporating nonbinding stopping rules for futility in interim analyses.

Main Results:

• Methods are presented to formally incorporate all available data, including from pipeline subjects.
• The approach increases statistical power by utilizing early measurements.
• Valid statistical inference (p-values, confidence intervals) is achievable with the proposed methods.

Conclusions:

• This paper offers a comprehensive guide for designing and analyzing GST with delayed endpoints and repeated measures.
• The proposed methods effectively leverage data from pipeline patients to improve trial efficiency and power.
• Accessible R code and a package are provided to facilitate the application of these advanced statistical techniques.