The failure rate of drugs in Phase 3 is around 50%. A major factor in this failure rate is making poor dose decisions from Phase 2 data. We will present results from an alternative approach to drug development: a seamless, adaptive Phase 2/3 design. The study employed a Bayesian Analytical approach to allocate patients to 7 doses of the experimental drug and to select doses for the Phase 3 program. The randomization scheme and decision rules were based on a Clinical Utility Index, which balanced safety and efficacy measures.
The study began screening patients in August 2008; adaptive randomization began in December 2008; and the algorithm selected 2 doses in April 2009. Patients continued on the selected doses and the 2 comparator arms and additional patients were randomized to the 4 arms until a targeted sample size was reached, based on a predictive probability calculation, to complete the two year study. Both doses have proven to be safe and efficacious and the drug is now in the market. The study design and operating characteristics will be compared and contrasted to a typical, fixed trial design. The interim results of the dose-finding portion of the study will be presented
|
ASA Meetings Department
732 North Washington Street, Alexandria, VA 22314
(703) 684-1221 • meetings@amstat.org
Copyright © American Statistical Association.