|
Question 6.1.3
The
drug company follows your advice in the previous question
to do an extended double-blind pilot study comparing Drug
P to placebo over a 6 week period. Patients rate "pain
over the past 24 hours" on a 0-10 numerical scale,
and "baseline pain" is defined as the mean of
the pain scores for the 7 days before randomization. You
arbitrarily define the "time of onset" as the
earliest time point at which pain has been reduced to 50%
or less of the baseline pain score for three consecutive
days.
What
is the most appropriate way to compute the average time
of onset of pain relief and statistically compare the drug
group to the placebo group?
 Calculate
the mean times to onset and compare the means for the drug
and placebo groups using an unpaired t test.
Calculate
the median time to onset and compare the overall distributions
of the onset times using survival analysis techniques.
|