The Design of Large-Scale Plasma Donation Trials.

Area of Science:

• Biostatistics
• Clinical Trial Design
• Blood Donation Safety

Background:

• Evaluating new blood donation devices requires robust statistical methods due to challenges like non-independent donations, rare adverse events, and donor heterogeneity.
• Existing methods may not adequately address the complexities of analyzing safety outcomes in repeated donation scenarios.

Purpose of the Study:

• To develop a statistical framework for designing superiority and non-inferiority trials to assess the safety of new blood donation devices.
• To provide a method for planning accrual and follow-up periods based on historical donation data.

Main Methods:

• A statistical framework using generalized estimating equations with a working independence assumption was developed.
• A Poisson approximation with a log link and identity variance function was used to model binary donation-specific outcomes.
• Historical data on donation intensity, heterogeneity, adverse event rates, and serial dependence informed the planning and variance estimation.

Main Results:

• The derived statistical formulae demonstrated good control of type I error rates and statistical power.
• The framework enables effective planning of clinical trials for blood donation devices.
• Application to a plasma donation trial showed the utility for investigating serious hypotensive adverse events.

Conclusions:

• The developed statistical framework provides a reliable approach for designing and analyzing clinical trials comparing blood donation devices.
• This methodology addresses key challenges in donation safety research, including rare events and donor-specific factors.
• The study offers a practical tool for ensuring donor safety in the development of new blood collection technologies.