Implementation of maximin efficient designs in dose-finding studies.

Area of Science:

• Clinical trial design
• Biostatistics
• Pharmacometrics

Background:

• Designing clinical studies requires robust methods to ensure efficiency across various parameter scenarios.
• Traditional designs may not optimally balance efficiency when parameters vary.
• Maximin efficient designs offer a strategy to maximize the minimum efficiency.

Purpose of the Study:

• To propose a feasible maximin approach for clinical study design, particularly for dose-finding studies.
• To introduce a practical four-step expert elicitation procedure for defining parameter subsets.
• To present a novel algorithm simplifying the construction of maximin efficient designs.

Main Methods:

• A four-step expert opinion elicitation procedure to define parameter subsets.
• Extension of parameter subsets for broader applicability.
• Development and application of the H-algorithm for constructing maximin efficient designs.
• Utilizing a sigmoid Emax model for dose-response relationship analysis.

Main Results:

• The proposed procedure makes the maximin approach feasible for dose-finding studies.
• The H-algorithm significantly simplifies the construction of maximin efficient designs.
• Maximin efficient designs demonstrated at least 15% greater inferential precision compared to uniform designs in a sigmoid Emax model.

Conclusions:

• The maximin approach, facilitated by expert elicitation and the H-algorithm, provides a powerful tool for clinical study design.
• This method enhances inferential precision in dose-response studies.
• Maximin efficient designs represent a significant improvement over standard uniform designs.