Closure procedures for monotone bi-factorial dose-response designs

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmacometrics

Background:

• Multiple-dose factorial trials aim to prove combination therapy effectiveness and determine optimal dose ranges.
• Establishing safe and effective dose ranges is crucial for drug development.
• Statistical methods are needed to control error rates in complex trial designs.

Purpose of the Study:

• To present closure procedures for bi-factorial dose-response designs that guarantee strong control of the familywise error rate.
• To investigate two families of null hypotheses for identifying minimum and maximum effective doses.
• To demonstrate the application of these methods in an unbalanced clinical trial.

Main Methods:

• Utilizing closure procedures to ensure strong control of the familywise error rate.
• Applying likelihood ratio tests and multiple contrast tests.
• Investigating bi-factorial dose-response designs with monotone properties.

Main Results:

• The proposed methods effectively address the dual goals of demonstrating combination effectiveness and identifying dose ranges.
• Two distinct families of null hypotheses allow for the identification of minimum and maximum effective doses.
• The methods were successfully applied to a real-world unbalanced clinical trial example.

Conclusions:

• The developed statistical procedures provide a robust framework for dose-finding in factorial trials.
• These methods enhance the precision and reliability of identifying optimal therapeutic doses.
• The availability of R code facilitates the practical implementation of these advanced statistical techniques.