Comparing the response rates for superiority, noninferiority and equivalence testing with multiple-to-one matched...

Area of Science:

• Biostatistics
• Clinical Trials
• Epidemiological Safety Studies

Background:

• Paired and multiple-to-one matched data are common in clinical trials and safety studies.
• Sample size and power for binary responses depend on discordant pairs or matched sets.
• Existing methods focus on fixed sample size determination for paired binary data.

Purpose of the Study:

• To extend fixed sample size determination methods to multiple-to-one matched binary data.
• To examine simultaneous and two-stage testing for superiority, noninferiority, and equivalence hypotheses.
• To propose an efficient two-stage adaptive design for sample size reestimation.

Main Methods:

• Extending existing fixed sample size determination methodologies for paired binary data to multiple-to-one matched data.
• Investigating the validity of simultaneous testing and hypothesis switching for superiority and noninferiority/equivalence.
• Developing a two-stage adaptive design for sample size reestimation using interim data.

Main Results:

• The monotonic property of superiority and noninferiority/equivalence tests allows for valid simultaneous testing and switching.
• A two-stage adaptive design improves sample size determination efficiency by utilizing interim data.
• The proposed method addresses the inefficiency of sample size calculations based solely on control group data or unrealistic alternative hypotheses.

Conclusions:

• The study provides a robust framework for sample size determination in complex matched binary data settings.
• The proposed two-stage adaptive design enhances the efficiency and practicality of clinical trial sample size planning.
• These advancements are crucial for optimizing resource allocation and statistical power in epidemiological and clinical research.