Large-Sample Confidence Intervals for the Treatment Difference in a Two-Period Crossover Trial, Utilizing Prior Information

TL;DR

This paper develops a large-sample confidence interval for treatment differences in two-period crossover trials that incorporates prior information suggesting no differential carryover effect, aiming to improve inference accuracy.

Contribution

It introduces a novel large-sample confidence interval method that leverages uncertain prior information to enhance estimation in crossover trials.

Findings

The proposed interval improves coverage probability when prior information is correct.

Simulation studies demonstrate increased efficiency over traditional methods.

Method provides robust inference even with uncertain prior assumptions.

Abstract

Consider a two-treatment, two-period crossover trial, with responses that are continuous random variables. We find a large-sample frequentist 1-alpha confidence interval for the treatment difference that utilizes the uncertain prior information that there is no differential carryover effect.