ABSTRACT
Clinical trials are often multicentric, that is, patients are recruited in several
centers (medical practices or hospitals). In so doing, a larger sample size is
possible, among other things. A disadvantage, however, is the possibility of
differences between the centers. Therefore, the randomization of patients is
usually done separately for every center in order to have in each center a
similar allocation of patients to the different treatments. A consequence for
the statistical analysis is that the stratification by center must be considered
(EMEA, 2003).