Skip to main content
. 2012 Dec 12;25(2):716–727. doi: 10.1177/0962280212465498

Figure 1.

Figure 1.

Maximum sample size and maximum cost (arbitrary units) of treatment as allocation ratio changes. Designs are chosen using triangular stopping boundaries such that they give 5% type-I error and 90% power. Maximum cost assumes that the cost of allocating a patient to the control group is c, and the cost of allocating a patient to an experimental treatment is 1 where c ∈ {1, 0.5, 0.25, 0.1}.