Multiplicity in clinical trials may appear under several different guises:
multiple endpoints, multiple treatment arm comparisons, and multiple looks
at the data during interim monitoring, to name a few. It is well recognized
by statisticians and nonstatisticians alike that multiplicity inflates the
type I error rate of the experiment, and this has prompted the development
of many multiple comparison adjustment procedures. What has remained one o
f the thornier and more controversial points of contention among trialists
today is the philosophy surrounding the need for multiplicity adjustment in
clinical trials. This paper provides guidelines on how to deal with this c
omplex issue in a practical manner. Through a series of scenarios and examp
les, we illustrate the fundamental issues surrounding the concept of multip
licity and point to some key questions one should ask when deliberating on
the necessity and extent of adjustment for multiple comparisons. (C) Elsevi
er Science Inc. 2000.