The 2 x 2 matched-pairs trial: exact unconditional design and analysis

Area of Science:

• Biostatistics
• Clinical Trial Design
• Statistical Methodology

Background:

• Traditional analysis of 2x2 matched-pairs trials relies on the exact conditional method.
• This method's sample size calculations may not be optimal, potentially leading to less powerful studies.

Purpose of the Study:

• To introduce and evaluate an exact unconditional method for designing and analyzing 2x2 matched-pairs trials.
• To compare the efficiency and power of the unconditional method against the conditional method.

Main Methods:

• Development of an exact unconditional statistical test for 2x2 matched-pairs data.
• Utilizing the total number of sampled pairs (N) in the analysis, unlike the conditional method's focus on discordant pairs.
• A simple Z statistic forms the basis of the unconditional test.

Main Results:

• The exact unconditional method generally yields smaller required sample sizes compared to the exact conditional method.
• The unconditional test demonstrates uniform power superiority across various alpha levels (.01, .025, .05) and sample sizes (N=10-200).

Conclusions:

• The exact unconditional method provides a more powerful and efficient approach for the design and analysis of 2x2 matched-pairs trials.
• This method is applicable to various research settings, including in vitro studies evaluating antimicrobial agent effectiveness.