Simple sequential boundaries for treatment selection in multi-armed randomized clinical trials with a control.

Area of Science:

• Clinical Trials Methodology
• Biostatistics
• Medical Decision Making

Background:

• Phase III clinical trials often face resource limitations when evaluating multiple treatment regimens against a standard.
• Multi-armed randomized selection trials offer a strategy to eliminate inferior treatments efficiently.
• Quick endpoints, like imaging in acute stroke, make frequent interim monitoring desirable.

Purpose of the Study:

• To propose a class of sequential selection boundaries for multi-armed clinical trials.
• To enable the selection of treatments with clinically significant improvement over the control or declare futility.
• To facilitate flexible, blinded monitoring and simplify design calibration.

Main Methods:

• Development of sequential selection boundaries for multi-armed trials.
• Application of methods to a selection trial with an imaging endpoint in acute stroke patients.
• Simulation-based comparison with an optimal two-stage design.

Main Results:

• The proposed boundaries are easy to implement in a blinded manner.
• Flexible monitoring schedules based on calendar time are supported.
• Design calibration is simplified, even with interval-known control group response rates.
• The proposed method demonstrated a smaller average sample size compared to the two-stage design in simulations.

Conclusions:

• Sequential selection boundaries provide an efficient strategy for multi-armed clinical trials.
• The proposed method reduces sample size, particularly when a superior treatment exists.
• This approach is advantageous for trials with quick endpoints and limited resources, such as in acute stroke research.