A general family of dynamic treatment allocations is defined, and it is shown that the permuted block procedure (Zelen 1974) and Begg and Iglewicz method (1980) are extreme choices in this family. A compromise method is suggested. The framework of this general family allows the relationships between these methods to be examined. By means of a simulation study these three methods plus the complete randomization method are compared in terms of efficiency and balance. The compromise method is shown to have good overall properties. In addition, an illustrative example is given.
|Number of pages||11|
|Journal||Communications in Statistics Part B: Simulation and Computation|
|State||Published - 1 Jan 1991|
Bibliographical noteFunding Information:
This work was begun while the first author was visiting the Department of Biostatistics, Harvard School of Public Health and was partially supported by a grant from the National Cancer Institute, National Institute of Health, CA-23415. The strocd author was on leave from the University of Haifa. We would like to thank Professor Iglewicz and a referee for their very helpful comments .
- clinical trials
- dynamic treatment allocation
- permuted block
- variance minimization
ASJC Scopus subject areas
- Statistics and Probability
- Modeling and Simulation