An Improved Design for Clinical Trials Evaluating Multiple Agents

• Activity Number: 264
• Type: Topic Contributed
• Date/Time: Tuesday, August 10, 2004 : 2:00 PM to 3:50 PM
• Sponsor: Biometrics Section
• Author(s): Christine E. McLaren*+ and Vernon M. Chinchilli and Jennifer J. Roan and Wen-Pin Chen and Frank L. Meyskens
• Companies: University of California, Irvine and Pennsylvania State University and Minitab Inc. and Chao Family Comprehensive Cancer Center and Chao Family Comprehensive Cancer Center
• Keywords: 2x2 factorial design ; Cochran-Armitage trend test ; simple loop alternative ; contrast statistic ; Jonckheere-Terpstra test ; chemoprevention

Abstract

For clinical trials that intend to answer multiple questions, new designs are needed to achieve results with fewer patients and less expense. Adenomatous polyps are precursors to colorectal cancer, the fourth most common incident cancer in the United States. We propose a 2x2 balanced factorial design to test whether a combination of agents, difluoromethylornithine and sulindac, is better than the individual treatments or placebo with regard to reducing the proportion of recurrent adenomas. To compare proportions of patients developing at least one adenoma between treatment groups, we derive a pair of two-sided tests for the ordered alternative of increasing or decreasing proportions across treatment groups, called the simple-loop alternative, and a method for power and sample size calculations. We compare this test with alternative tests under comparable assumptions. The proposed design to study an intermediate marker for cancer incidence, achieves a reduction in sample size of over 30% and should lead to a significant reduction in the cost of the clinical trial. We suggest the use of this experimental design to achieve economy in terms of numbers of patients and monetary expense.