Optimal Adaptive Designs in Phase-II Trials

Anindita Banerjee
Statistics
North Carolina State University
*Candidate for the Assistant/Associate Professor Position

Thursday, March 16th
3:30pm
Moos 2-620
Minneapolis Campus

Abstract:

Phase-II trials provide a platform where, on the basis of the efficacy of the drugs, ineffective drugs are screened out and promising drugs move on to the next phase. The efficacy of the drug is often evaluated by a binary response, i.e. success or failure of the drug and is determined by testing the response probability. Two stage designs are widely used for this purpose because they result in correct decisions with the same accuracy as a one-stage trial but with smaller average sample sizes. Simon (1989) proposed optimal fixed two-stage designs which minimize the expected sample size under the null hypothesis. We have derived optimal adaptive designs at the null that perform better than Simon's design, though the gains are modest (BanerjeeandTsiatis,2005). By adaptive we mean that the second stage sample size will be dependent on the results from the first stage. We further explore optimal adaptive designs that minimize the expected sample size at the alternative hypothesis, at a probability midpoint between the null and alternative hypothesis and a weighted combination of the null, alternative and midpoint value. We also construct an envelope function that gives the lowest expected sample size for any possible value of the response probability. The different designs are compared to each other as well as the envelope function. Mostly, the designs that minimize the expected sample size at the midpoint between the null and alternative hypothesis and the design that minimizes a weighted average of the response probabilities perform best across a range of the parameter values, is closest to the envelope function, and generally outperform Simon's design.
This is a joint work with Dr. Anastasios A.Tsiatis, North Carolina State University ,North Carolina.

A social tea will be held at 3:00P.M. in A434 Mayo. All are Welcome.
For more details contact 612-624-4655 or see http://www.biostat.umn.edu/seminar_academic.html