Adjusting for baseline measurements in the two-period crossover study: a cautionary note

Area of Science:

• Biostatistics
• Clinical Trial Design

Background:

• Two-period crossover studies are common in clinical research.
• Patient response is often measured as the change from the start to the end of each period.

Purpose of the Study:

• To investigate the potential for bias when using period change as a response measure in two-period crossover designs.
• To analyze the impact on estimating direct treatment effects and residual effects.

Main Methods:

• The study examines a standard two-period crossover design.
• It analyzes the consequences of using the change in response within each period as the primary outcome measure.

Main Results:

• Using period change as the response measure can lead to biased estimates.
• This bias affects the assessment of both direct treatment effects and residual effects.

Conclusions:

• The intuitive method of using period change in two-period crossover studies can be misleading.
• Researchers should exercise caution and consider alternative analysis methods to avoid bias.