|
|
Enriched Enrollment Designs
A variant of the crossover design, the "enriched
enrollment" design will be presented in the following example.
Problem
7.1
Based on clinical experience
and a small placebo-controlled trial, a pain researcher
inferred that approximately one-quarter of patients with
diabetic neuropathy will consistently respond to chronic
treatment with trandermal clonidine far more than they respond
to placebo. To prove this, he designed a two-stage clinical
trial. In the first state, 40 patients were treated with
3 weeks of both clonidine, 0.3 mg/day, and placebo in a
crossover trial. Patients with moderate or better relief
with clonidine but a lower grade of relief with placebo
were invited to enter a second clinical trial, "enriched"
for apparent clonidine responders, and were randomized to
receive two pairs of clonidine and placebo treatment, each
treatment one week in duration. With each pair, treatment
order was randomized.
In the first trial, there was
no significant difference between clonidine and placebo
over the entire 40 patients. The figure below illustrates
results of the second study. In the 12 patients who entered
this repeated crossover study, the mean level of pain was
lower during clonidine than placebo treatment (p = 0.015).
|
Figure
7.2
|
 |
Pain
intensity during treatment of 12 diabetic neuropathy
patients with either transdermal clonidine or
placebo. This group was "enriched" in
clonidine responders based upon the results of
an initial larger trial. While the order of clonidine
or placebo in each treatment pair was randomly
selected, the clonidine treatment in each pair
is displayed on the left for clarity. |
|
|