Allocation in platform trials to maintain comparability across time and eligibility.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Medical Research Methodology

Background:

• Platform trials involve dynamic changes, with arms entering and leaving over time.
• Patient eligibility restrictions within certain arms can further complicate trial design and analysis.
• Existing methods often fail to address biases arising from time-dependent factors and patient eligibility in platform trials.

Purpose of the Study:

• To identify and address biases in platform trial comparisons caused by time-varying factors and patient eligibility.
• To develop a robust algorithm for ensuring comparability between active and control arms in platform trials.
• To provide a flexible and implementable solution for complex clinical trial designs.

Main Methods:

• The study builds upon existing biostatistical methods to develop a novel algorithm.
• The algorithm addresses biases related to non-concurrent controls and ineligible patients.
• It incorporates mechanisms for re-randomization and two-stage randomization procedures.

Main Results:

• Even with concurrent and eligible controls, biases can arise if allocation ratios are not maintained.
• The proposed algorithm guarantees comparability between active and control groups, accounting for time and eligibility.
• The method is shown to be flexible and easily implemented in practice.

Conclusions:

• Proper allocation ratios are crucial in platform trials to avoid bias, even with concurrent, eligible controls.
• The developed algorithm offers a reliable solution for unbiased comparisons in complex platform trials.
• This approach enhances the integrity of clinical research by ensuring robust comparisons under challenging trial conditions.