|
Activity Number:
|
193
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 10, 2004 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Biopharmaceutical Section
|
| Abstract - #300796 |
|
Title:
|
Maximum Imbalance in Treatment Assignments in a Clinical Trial that Uses the Minimization Allocation Procedure
|
|
Author(s):
|
Olga M. Kuznetsova*+ and John K. Troxell
|
|
Companies:
|
Merck & Co., Inc. and Merck & Co., Inc.
|
|
Address:
|
PO Box 2000, RY34-316, Rahway, NJ, 07065-0900,
|
|
Keywords:
|
minimization ; allocation in clinical trials ; randomization ; balance in treatment assignments
|
|
Abstract:
|
Minimization is a covariate-adaptive allocation procedure designed to balance the treatment groups in a clinical trial over a large number of factors. Although the balancing properties of minimization have been studied extensively through computer simulations, little attention has been paid to the fact that in Taves' minimization the maximum possible imbalance in treatment assignments can be established by a complete enumeration of combinations of marginal imbalances. The maximum imbalances in treatment assignments will be established for Taves' minimization with the Friedman-White overall imbalance function. For example, for minimization that balances over four factors with two levels each, the imbalance in treatment assignments across a single level of any factor is shown never to exceed 11; the overall imbalance in treatment assignments will never exceed nine. Since minimization is defined through an algorithm that sets no limit to maximum allowed imbalance, eliminating a concern that an allocation sequence with any given large imbalance in treatment assignments might occur as a result of the minimization algorithm is important.
|
