This page contains a Flash digital edition of a book.
and sample size determination for adequate Expert Team (SET) is addressing QbD, pri- There are initiatives relating to ben-
target dose estimation. The working group marily through developing and promoting efit/risk, helping to combine the two into
plans to release white papers on these topics a better understanding of one of the princi- a single measure and emphasizing that one
this year and to request a new face-to-face pal concepts of QbD: design space. should never make decisions based on only
meeting with FDA to discuss them. one. One consideration regarding this topic
Data Management
is benefit/risk as a patient-specific phenom-
Noninferiority Designs
Data management is a key component of enon. That is, one patient might be willing
When an approved treatment for a disease or clinical research. Statisticians can use the to accept the risk because other agents do
condition exists, placebo-controlled trials to best statistical methods that produce mini- not work for him/her, while another patient
evaluate an experimental treatment may not mal or no bias; however, if the data fed into might find other agents acceptable and so
be considered ethical. In situations such as these methods are inaccurate, incomplete, would not want to take the risk.
this, a common approach is to conduct an or slow to become available, then the infer- There are two biomarker groups. One
active-control noninferiority trial, directly ences are undermined. Therefore, BDMTG is a statistics expert team, and the other
comparing the experimental treatment with supports clinical data management activi- a cross-functional team. Both have been
the approved treatment. ties to achieve efficiency and accuracy in focused recently on biomarker qualifica-
In addition to providing a direct estimate clinical data. Data Management members tion, which is important, as a biomarker
of the relative effects of the approved and of BDMTG have been involved in a range might be a type used to predict ultimate
experimental treatments, when combined of activities selected because of their poten- clinical benefit of a new drug, or it might
with historical information about the effect of tial positive impact on the overall clinical be one used to tailor a new drug to a sub-
the approved treatment, these trials can pro- development process. group of the full patient population, where
vide an indirect estimate of the effect of the the benefit/risk is ideal.
experimental treatment relative to placebo.
Other Clinical Statistics Topics
Finally, there are four working groups
This team, led by statisticians, argues BDMTG members and colleagues are led or co-led by statisticians. These relate
that only one standard of evidence should engaged in many initiatives and working to more statistical or clinical design top-
be applied when deciding whether a phar- groups. Several years ago, BDMTG began ics. The first is a team focused on the mer-
maceutical treatment has demonstrated suf- forming “expert teams” in a variety of its of flexible dose designs. Such designs
ficient efficacy for regulatory approval. areas. In some cases, the topics were being allow various levels of dose flexibility that
This standard of evidence is already addressed only through BDMTG. In other depend on patient tolerability and/or effi-
well established: The new drug must have cases, it was a cross-functional initiative, cacy achieved. The second team focuses on
superior efficacy in comparison to placebo. and, in still other cases, it was both. dichotomizing continuous data for statisti-
Team members hope to influence regula- The first team to highlight is one cal analysis. The recommendation from this
tory policy; in particular, they hope to have where there was both a SET and a cross- group is that the primary statistical testing
an impact on a new FDA guidance docu- functional expert team. Both teams focused should be on the original continuous mea-
ment about noninferiority trials, a draft of on addressing the QT prolongation issue. sure, with a potential secondary analysis
which is expected to be released this year. It has become expected that new molecular of the dichotomized endpoint. The basic
entities be evaluated for effects on QT, which
premise to this argument is that dichoto-
Nonclinical Statistics
is a surrogate for more serious cardiovascu-
mization discards information, resulting in
There are groups and initiatives not directly lar events. Many new drugs must conduct
a loss of statistical power.
connected to BDMTG for nonclinical sta- thorough QT studies that are large, expen-
The third team focuses on co-primary
tistics. However, PhRMA offers support sive, and ethically challenging with an active
endpoints. Having co-primary endpoints
as it can to such initiatives. And PhRMA’s control known to increase the QT interval.
can lead to grossly reduced power to detect
nonclinical members help educate everyone Issues include how much of an increase is
important treatment effects, so the recom-
about important topics, such as Quality-by- too much, especially for the upper limit of
mendation to regulators is to move toward
Design (QbD). the 95% confidence interval. Multiplicity is
accepting a single primary endpoint wher-
Just as clinical trials establish the effi- an issue, as ECG measurements are obtained
ever possible. Finally, missing data is a com-
cacy and safety of a new drug, chemistry, at numerous time points in the study. These
plex topic that has had much attention in
manufacturing, and controls development two QT teams have had success in influenc-
the statistical literature for many years. The
establishes the process of manufacturing the ing regulators and the broad cardiovascular
paper prepared by the PhRMA working
product to meet clinical requirements. QbD, scientific community.
group on this topic recently was accepted
an FDA initiative, is based on designing nec- There are a number of ongoing initia-
for publication in the Drug Information
essary quality into a manufacturing process, tives to evaluate for advocacy purposes opti-
Journal. The primary argument is regard-
instead of “testing” quality into the finished mal methods to assess the safety of a phar-
ing why we should move away from the
product. This requires developing a mecha- maceutical product. One team is focused
three-decade standard of last-observation-
nistic understanding of how formulation and on pre-approval evaluations, including
carried-forward (LOCF) and use mixed
process factors affect product performance. creation of statistical analysis plans for the
models repeated measures (MMRM)
Statistical thought process and methodology evaluation of safety. Another is focused on
analysis instead.
is critical for achieving this objective. observational data post-approval. A third
For more information about BDMTG
Although not formally connected to group is evaluating data mining tools for
and its activities, contact Walt Offen at
BDMTG, the PhRMA CMC Statistical such observational data.
wwo@lilly.com. n
APRIL 2008 AMSTAT NEWS 39
Page 1  |  Page 2  |  Page 3  |  Page 4  |  Page 5  |  Page 6  |  Page 7  |  Page 8  |  Page 9  |  Page 10  |  Page 11  |  Page 12  |  Page 13  |  Page 14  |  Page 15  |  Page 16  |  Page 17  |  Page 18  |  Page 19  |  Page 20  |  Page 21  |  Page 22  |  Page 23  |  Page 24  |  Page 25  |  Page 26  |  Page 27  |  Page 28  |  Page 29  |  Page 30  |  Page 31  |  Page 32  |  Page 33  |  Page 34  |  Page 35  |  Page 36  |  Page 37  |  Page 38  |  Page 39  |  Page 40  |  Page 41  |  Page 42  |  Page 43  |  Page 44  |  Page 45  |  Page 46  |  Page 47  |  Page 48  |  Page 49  |  Page 50  |  Page 51  |  Page 52  |  Page 53  |  Page 54  |  Page 55  |  Page 56  |  Page 57  |  Page 58  |  Page 59  |  Page 60  |  Page 61  |  Page 62  |  Page 63  |  Page 64  |  Page 65  |  Page 66  |  Page 67  |  Page 68  |  Page 69  |  Page 70  |  Page 71  |  Page 72  |  Page 73  |  Page 74  |  Page 75  |  Page 76
Produced with Yudu - www.yudu.com