Properties of the Random Block Design for Clinical Trials
To avoid deterministic treatment allocations in the permuted block design (PBD), many clinical trialists prefer randomizing the block sizes as well. While such a procedure is rarely formalized, it is generally assumed that the design will be less predictable. In this paper, we formalize the random block design by assuming a discrete uniform distribution for block size. The aim of this study is to provide a statistical understanding of the RBD, by investigating its distributional properties, including the degree of predictability and variability of treatment imbalance.
KeywordsBlock Size Random Block Design Discrete Uniform Distribution Correct Guess Prior Allocation
Professor Rosenberger was supported by a scholarship from the German-American Fulbright Kommission, and also by the Department of Medical Statistics, RWTH Aachen University.
- 4.Heussen, N.: Der Einfluss der Randomisierung in Blöcken zufälliger Länge auf die Auswertung klinischer Studien mittels Randomisationstest. RWTH Aachen University, Aachen (2004)Google Scholar
- 7.Shah, B., Berger, J.S., Amoroso, N.S., Mai, X., Lorin, J.D., Danoff, A., Schwartzbard, A.Z., Lobach, I., Guo, Y., Feit, F., Slater, J., Attubato, M.J., Sedlis, S.P.: Periprocedural glycemic control in patients with diabetes mellitus undergoing coronary angiography with possible percutaneous coronary intervention. Am. J. Cardiol. 113, 1474–1480 (2014)CrossRefGoogle Scholar
- 8.Shao, H.: Exact properties of restricted randomization procedures. George Mason University, Fairfax (2015)Google Scholar