Exploring changes in treatment effects across design stages in adaptive trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Pharmaceutical Medicine

Background:

• Recent guidelines from the Committee for Medicinal Products for Human Use (CMHPU) on flexible designs introduce controversy regarding adaptive trial interpretation.
• The guideline proposes pre-planned heterogeneity testing, potentially preventing data combination across design stages if interim analysis reveals significant treatment effect differences.

Purpose of the Study:

• To investigate the error rates of the CMHPU guideline's proposed procedure in the presence of calendar-time effects.
• To introduce and evaluate an alternative testing strategy using change point methods for adaptive clinical trials.

Main Methods:

• Error rate investigation of the CMHPU guideline's heterogeneity testing procedure under calendar-time effects.
• Development and simulation study of a novel change point-based testing strategy.

Main Results:

• The CMHPU guideline's approach may be susceptible to errors when calendar-time effects are present.
• The proposed change point method demonstrates favorable performance in simulation studies compared to the guideline's suggested procedure.

Conclusions:

• The CMHPU guideline's interpretation of adaptive trials may require refinement to account for time-dependent covariates like calendar-time effects.
• Change point methods offer a promising alternative for robust analysis of adaptive clinical trials, potentially improving data integrity and reducing analytical errors.