| |
Randomization
Randomization is a
premiere consideration for it distributes known and unknown prognostic
factors among the arms of the trial in an unbiased manner, so
that the effects of these prognostic factors are averaged out
in tests of statistical significance. Randomization also avoids
the conscious or subconscious assignment of certain patients to
particular arms of a study, thereby maintaining the integrity
of the data collected. Without randomization, any data collected
might be suspect, for it could not be assured the results were
obtained in a non-biased manner. For obvious reasons, therefore,
nonrandomized designs are unacceptable for comparative antiemetic
trials. They may, however, have a role in early open-label trials.
|