Fisherian Evidential Approach for Adaptive Doubly Randomized Withdrawal Designs

• Activity Number: 631
• Type: Topic Contributed
• Date/Time: Thursday, August 8, 2013 : 8:30 AM to 10:20 AM
• Sponsor: Biopharmaceutical Section
• Author(s): Qing Liu*+
• Companies: Johnson & Johnson
• Keywords: Doubly Randomized Withdrawal Design ; Fisherian Evidential Approach ; Multiple Type 1 Error Rates ; Sample Size Adjustment ; Adaptive Design ; Enrichment

Abstract

Many diseases require initial (induction) and maintenance treatments. Traditionally, such treatment regimen has often been developed through separate clinical trials in different patient cohorts. We propose to develop the treatment regimen in a single clinical trial with a doubly randomized withdrawal design. Within this trial, it is still necessary to answer the two basic research questions regarding the acute and maintenance effects of the new treatment regimen. With the traditional approach of separate trials, each of the research questions is tested at the trial significance level, and the multiple type 1 error rates are not adjusted between the trials. However, it is not clear if the multiple type 1 error rates need to be adjusted for a single trial with the doubly randomized withdrawal design. We develop a Fisherian evidential approach by which no adjustment to the p-values or significance levels are made. For the single trial, the approach also controls strongly the multiple type 1 error rates for false clinical claims of the acute and maintenance effects. We also discuss features of sample size calculation and adjustment.