Clinical trial randomization schemes to control patient characteristics that can influence clinical response are not sufficient to control for their effects on operational efficiency and cost within the provider institution. This variation can result in biased economic findings when patients are allowed to select a provider themselves.