|
Sample Size
Sample sizes should
be calculated to achieve adequate power (at least 80%; 90% is
preferable) to allow for the detection of clinically meaningful
differences. Although a large improvement may be expected (or
hoped for), the study should be designed with the more modest,
but still clinically relevant, difference in mind.
With the recent advancements
in the control of acute emesis, an improvement of 15% or smaller
may be clinically important and may add to the improvement of
patient care. Conversely, this small measure of improvement would
serve to increase the sample size.
As Table 4 shows, the
number of patients needed in a trial's sample size increases as
the size of the improvement that the trial is attempting to detect
decreases. Improvement in this table refers to the percent of
all patients who achieve complete control, e.g. a 10% improvement
starting from a baseline of 40% with complete control would mean
50% with complete control.
| Table
4: Sample Size
|
|---|
|
% Complete control with "standard"
| 5%
Improvement
| 10%
Improvement
| 15%
Improvement
|
|---|
| 40% | 3146 | 814 | 372 |
| 60% | 3020 | 752 | 330 |
| 80% | 1890 | 438 | 176 |
|
|