The score-goldilocks design for phase 3 clinical trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Bayesian Statistics

Background:

• Adaptive clinical trial designs allow for modifications during a study.
• Bayesian methods offer a flexible framework for incorporating prior information.
• Efficient trial designs are crucial for reducing costs and time.

Purpose of the Study:

• To propose a novel Bayesian adaptive design, the score-goldilocks design.
• To develop a uniform formula for calculating trial success probability using normal approximation.
• To evaluate the performance and efficiency of the score-goldilocks design compared to existing methods.

Main Methods:

• The score-goldilocks design is based on the algorithmic principles of the goldilocks design.
• Normal approximation is employed to derive a formula for the probability of trial success.
• Simulations are conducted to compare operating characteristics with the goldilocks design.

Main Results:

• The score-goldilocks design yields a uniform formula for success probability across different endpoints.
• Operating characteristics (Type I error, power, sample size, futility/success stopping) are comparable to the goldilocks design.
• Significant improvements in calculation time and operational efficiency are demonstrated.

Conclusions:

• The score-goldilocks design offers a computationally efficient and effective Bayesian adaptive approach for clinical trials.
• It maintains desirable operating characteristics while reducing analytical burden.
• This design enhances the practicality and speed of adaptive clinical trial implementation.