A confidence interval approach for comparative studies involving binary outcomes in paired organs

Bilateral binary data are very common in medical comparative studies and clinical trials (e.g., comparison of two treatments in ophthalmologic studies) in which information involving paired organs (e.g., eyes) is available from each subject. Previous studies reported that ignoring the dependence feature of bilateral data could lead to severely incorrect coverage probabilities. In this article, we study various confidence interval estimators for the difference of response rates under the dependence nature of bilateral data. Their performances are evaluated with respect to their expected coverage probabilities, mesial/distal non-coverage probabilities and interval widths via simulation studies. A real example from an otolaryngologic study is used to demonstrate our proposed methodologies.