Simultaneous Confidence Intervals for Odds Ratios in Bilateral Correlated Data Under Equal Correlation Coefficient Model

Bilateral data arising from paired organs or limbs within the same subject frequently occur in clinical trials and biomedical research. Neglecting the intraclass correlation inherent in such data can lead to inaccurate variance estimation and invalid statistical inference. Donner’s equal correlation coefficient model addresses this issue by assuming a single, common intraclass correlation coefficient across all individuals to characterize the dependence between paired observations from the same subject. Although prior research under this model has developed asymptotic testing procedures and confidence interval approaches for two-group comparisons of proportions, simultaneous confidence intervals for the odds ratio have not been thoroughly investigated. This article derives nine asymptotic simultaneous confidence intervals for the odds ratio and evaluates their performance through extensive Monte Carlo simulations, focusing on the empirical coverage probability and mean interval width. A real-world example is presented to illustrate the practical application of the proposed methods.