|
Interim Analysis
and End-Points
Before beginning the
trial, the timing of any interim analyses should be planned so
the trial is not stopped early with an erroneous positive result.
The probability of achieving a false positive result by chance
alone can exceed 20% if interim analyses are performed every 6
months in a 4-year study. Therefore, it is vitally important that
a p-value smaller than 0.05 be observed prior to a trial's
early termination. (Note: The p-value represents the statistical
likelihood that the finding is not an anomaly or fluke occurrence,
e.g. p < 0.05 means that there is less than a 5 % chance that
that occurred.)
Properly designed antiemetic
trials with a broad eligibility criteria typically accrue quickly.
We seldom use interim analyses since the 2/3 to 3/4 of total patients
are typically entered and on trial before there are mature outcome
data for analyses from the first 1/2 of the patients (assuming
an interim analysis at the mid-way point of accrual).
To ensure the quality
of the trial, end-points must be clearly defined before the trial
begins. Nausea and vomiting should be separately evaluated, not
only because they each depend on different physiological triggers,
but also to distinguish any differing benefits of treatment therapies.
While the occurrence and severity of nausea and vomiting are of
primary clinical concern, the duration of each may provide important
additional clinical information.
|