An ideal randomization procedure would achieve the following goals: Maximize statistical power, especially in subgroup analyses.
However, some trends were apparent in exploratory subgroup analyses.
