There are a number of factors involved, such as how
many interim analyses will be run; how many subjects
will be enrolled when each of the interim analyses
are conducted; whether they will assess futility,
efficacy, both, or safety; and what statistical approach
(boundaries) will be used to determine whether
to continue the trial or terminate early. Generally,
multiple designs are examined in order to compare
their relative advantages/disadvantages prior to
making a final selection and incorporating it into
the trial design.
It is critical that interim analyses be conducted in a
pre-planned manner that preserves the trial’s type I
error, maintains the treatment allocation blind, and
limits the number of people involved in the decisions
regarding trial conduct that have access to the results.4
Failure to control these aspects