An order restricted multi-arm multi-stage clinical trial design

Area of Science:

• Clinical Trial Design
• Pharmaceutical Statistics
• Drug Development Optimization

Background:

• Multi-arm multi-stage (MAMS) designs enhance drug development efficiency by allowing concurrent study of multiple arms and early stopping.
• Conventional MAMS designs assume independent treatment arms, which can be suboptimal when treatment effects have a predictable order (e.g., dose-response).

Purpose of the Study:

• To extend the MAMS framework to incorporate ordering information of treatment effects without assuming a specific dose-response or duration-response model.
• To develop a design that maintains high probability of identifying all truly effective treatments while controlling the family-wise error rate.

Main Methods:

• Developed an extended MAMS framework that integrates the order of treatment effects.
• Validated the design's performance in a symptomatic asthma study.
• Conducted simulations to compare the proposed design against fixed-sample and conventional MAMS designs.

Main Results:

• The proposed design effectively incorporates ordering information, leading to improved decision-making.
• Demonstrated strong control of the family-wise error rate.
• Simulations showed approximately 15% reduction in sample size compared to conventional MAMS designs in considered settings.

Conclusions:

• The extended MAMS design offers enhanced efficiency and decision-making power in drug development by leveraging treatment effect ordering.
• This approach provides a valuable alternative to conventional MAMS designs, particularly when treatment effects are expected to follow a trend.