Evaluation of possible designs for a three-arm clinical trial: Comparing the CLL11 design to potential alternative designs


To optimize resources, randomized clinical trials with multiple arms can be considered an attractive option to simultaneously testing various treatment regimens. Motivated by the success of a three-arm randomized clinical trial whose inference strategy was based on a closed testing procedure (CLL11), we compare four different potential strategies to run a three arm clinical trial and quantify the differences in power and the time it takes until approval of the new drug for these designs. We discuss statistical and operational aspects of implementing a design using a closed testing procedure, and situations where it is of limited use. In conclusion, using a closed testing procedure is an innovative method for a multiple-arm clinical trial, may bring the new drug fastest to patients, the power loss due to the global test in a closed testing procedure is minimal, and – according to our assessment – was the ideal method to achieve the goals in CLL11 in a time- and resource-efficient way.

Roche Global Biostatistics Conference (contributed talk)
San Francisco, United States