Statistics and Its Interface

Volume 11 (2018)

Number 1

Optimal responses-adaptive designs based on efficiency, ethic, and cost

Pages: 99 – 107

DOI: http://dx.doi.org/10.4310/SII.2018.v11.n1.a9

Authors

Chen Feng (School of Industrial & Systems Engineering, Georgia Institute of Technology, Atlanta, Georgia, U.S.A.)

Feifang Hu (Department of Statistics, George Washington University, Washington, D.C., U.S.A.)

Abstract

The trade-off between power and ethical concerns has been well discussed by researchers. The total costs, however, has hardly ever been considered in the adaptive design of clinical trials. In this article, we derive the compromised optimal allocations based on costs, ethical concerns, and efficiency for clinical trials with binary and normal responses. The compromised optimal allocations are implemented with a doubly biased coin design (DBCD) based on Hu and Zhang’s allocation function. The properties of the proposed designs are studied both theoretically and numerically. In many cases, the proposed designs are more efficient, economical and ethical than complete randomization (equal allocation) under both binary and normal responses.

Keywords

asymptotical normality, binary response, clinical trial, doubly adaptive biased coin design (DBCD), normal response, sequential method

Full Text (PDF format)

The research was partially supported by NSF Awards DMS-1442192 and DMS-1612970, the National Natural Science Foundation of China (No. 11371366).

Paper received on 10 November 2016.