Efficient and flexible simulation-based sample size determination for clinical trials with multiple design...

Area of Science:

• Biostatistics
• Clinical Trial Design
• Computational Statistics

Background:

• Simulation is valuable for clinical trial power estimation when analytic methods fail.
• Computational demands limit simulation to simple sample size problems.
• Current methods often optimize only overall sample size against a single power threshold.

Purpose of the Study:

• To develop a generalizable framework for simulation-based sample size determination.
• To address problems with multiple optimizing design parameters and conflicting minimization criteria.
• To overcome the computational burden of complex clinical trial simulations.

Main Methods:

• Utilizes a global optimization algorithm common in computer experiments.
• Employs a non-parametric regression model to approximate the power function.
• Integrates simulation-based power estimation with optimization for sample size determination.

Main Results:

• The proposed framework efficiently handles complex sample size determination problems.
• It allows optimization across multiple design parameters and conflicting criteria.
• Demonstrates flexibility for various simulation-based power estimation scenarios.

Conclusions:

• The developed framework offers a flexible and computationally feasible approach to complex sample size determination.
• It expands the applicability of simulation methods in clinical trial design.
• The method is adaptable for diverse trial structures, including those with complex clustering and multiple endpoints.