Two-stage group-sequential designs with delayed responses - what is the point of applying corresponding methods?

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Methodology

Background:

• Group-sequential designs typically assume no lag between patient enrollment and outcome measurement.
• In practice, response delays can impact statistical analysis, particularly in trials with interim analyses.
• Existing methods, like Hampson and Jennison's error-spending boundaries, address delayed responses.

Purpose of the Study:

• To compare the performance of group-sequential designs that account for delayed responses versus those that do not.
• To evaluate designs using the R package rpact and provide insights into global and conditional performance.
• To determine if performance gains justify the complexity of designs incorporating delayed responses.

Main Methods:

• Investigated extensions of the delayed response group sequential test (DR-GSD) design.
• Incorporated nonbinding lower recruitment stopping boundaries into the existing framework.
• Evaluated design performance using the R package rpact and defined suitable performance measures.

Main Results:

• The DR-GSD framework can accommodate both binding and nonbinding rules with additional constraints.
• Performance enhancements from delayed response methods depend significantly on interim sample size and data volume.
• Overall performance gains from incorporating delayed responses were found to be limited.

Conclusions:

• This research provides insights into performance differences between group-sequential designs with and without delayed response considerations.
• Given the marginal performance differences, operational feasibility should be a key factor in trial design discussions.
• The findings extend existing literature on group-sequential designs and their practical application.