Confidence Interval Construction for the Ratio of Two Treatment Means: An Application to Clinical Trial Data

• Activity Number: 175
• Type: Contributed
• Date/Time: Monday, August 5, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Biometrics Section
• Author(s): Krishna K. Saha*+ and Roger Bilisoly and Darius Dziuda
• Companies: Central Connecticut State University and Central Connecticut State University and Central CT State University
• Keywords: Delta Method ; GEE Approach ; Mean Ratio ; Over-dispersed Poisson data ; Square-and-Add Approach

Abstract

In clinical trials and medical studies, the mean count response in an exposed group is often of interest to compare with that in a control group. The confidence interval is frequently used to evaluate the treatment performance. Previous studies usually reported comparison between two treatments in terms of confidence intervals for the difference between two means in the over-dispersed count data. In this paper, we consider confidence interval construction for ratio of two treatment means in the analysis of over-dispersed count data. We develop various confidence intervals based on the large sample theory and the hybrid methods. Extensive simulations show that all the confidence intervals perform reasonably well in terms of coverage. However, the hybrid GEE confidence interval outperforms other interval procedures. Finally, a real world data set in clinical trials is analyzed by all the methods.