|
16.1
Study design: This study consists of two sequential trials:
(a) a randomized two-arm, double-blind, placebo-controlled,
crossover phase III trial to assess the effectiveness and
toxicity of SOYEST as a treatment for hot flashes, and (b)
an observational continuation trial to obtain longer term
information about the effects of SOYEST for the management
of hot flashes.
As
we have done in three previous studies (Loprinzi
et al., 1994; Barton
et al., 1998), we will use a one-week period to get
baseline hot-flash data. We will then use a classical two-treatment,
two-period crossover design wherein we will follow patients
for a four-week treatment period on placebo/SOYEST with
a crossover to the alternative substance for the subsequent
four-week treatment period. We will not use a washout period,
understanding that there may be a delayed washout effect
during the second randomized study period. In our previous
three studies, this worked well for us and the presence
of a carryover effect has not posed a problem to the analysis,
power, or interpretation of the results.
16.2
Randomization: Following stratification, the treatment assignment
will be calculated using a dynamic allocation procedure
(Pocock
et al., 1975) which balances the marginal distributions
of the stratification factors between the two treatment
sequences. The factors defined in Section
5.0, together with institution, will be used as stratification
factors. Patients will be allocated equally between the
two treatment sequences. This stratification process should
prevent a "lead time" bias in either of the two treatment
arms.
16.3
Analysis plans (double-blind trial): The efficacy of SOYEST
during each treatment period will be compared between the
two treatment sequences using the following measures: (1)
mean daily hot flash frequency during the last week of each
treatment period; (2) mean hot flash severity during the
last week of each treatment period; and (3) mean daily hot
flash score calculated for each patient during each treatment
period by adding up all the hot flash severity scores recorded
by the patient during each treatment period and dividing
that sum by the total number of days on which the patient
recorded hot flash data during that period. The toxicity
of SOYEST during each treatment period will be compared
using the following measures: (4) the incidence of each
toxicity, (5) the maximum severity of each toxicity, and
(6) the distribution of the overall toxicity score calculated
for each patient during each treatment period by adding
up the maximum toxicity grade for each of the toxicities
recorded by the patient during the period.
Since
this is a crossover study, the therapeutic effects of SOYEST
and placebo can be compared in two different ways for each
efficacy and toxicity variable: (1) subtract the value in
the first treatment period from the value in the second
treatment period, and compare these differences between
the two treatment sequences; or (2) ignore the values recorded
in the second period, and compare between the two treatment
sequences only the values recorded in the first period.
The first analysis method (i.e. the classic crossover analysis)
is, of course, more powerful than the second, but both types
of analysis may be required in case one or both of the following
two events occur: (1) evaluation data are missing for many
patients at one or more of the three-key evaluation times,
i.e. at baseline and at the end of each four-week treatment
period, or (2) there is evidence of unequal carryover effects.
16.31
Treatment effectiveness will be assessed in phases as described
below in order to deal with several issues observed in previous
studies or expected in this population, i.e. (1) the average
number of hot flashes per day may range from 2 to 15 or
more; (2) the frequency and severity of hot flashes are
highly correlated; (3) the side effects from other medication
or diseases may be "blamed" on the SOYEST/placebo; (4) the
length of time before any SOYEST effect is detectable is
unknown; and (5) the length of the time before hot flashing
intensity and frequency return to baseline levels following
discontinuation of therapy (i.e. washout period) is unknown.
16.311
The data for the seven-day baseline period will be used
to assess:
-
the initial comparability of the two treatment sequence
groups, and
-
the accuracy of the estimated flashing frequency and severity
rates supplied by the patient for use as stratification
factors at the time of study entry.
16.312
Plots of the daily incidence counts and severity scores
for flashes over the seven days of baseline measurement
and 56 days of double-blinded medication will be generated
for all randomized patients (including those who terminated
therapy early for any reason) and will be examined for information
regarding:
-
length of time to detectable treatment effect, and
-
length of washout period
|