Summary
Objectives:
The aim of this paper is to show that even in a highly regulated area such as clinical
research and development in pharmaceutical industry, there are needs and ample opportunities
for statisticians and other medical informatics professionals to further creatively
develop and implement methods in order to support the collection, analysis and interpretation
of clinical data.
Methods:
The recently published “Critical Path” initiative of the US Food and Drug Administration
discusses the decline in new drug submissions in the last decade and illustrates potential
causes in the present clinical development process. Areas where statisticians can
and have begun to look for new innovative ways to overcome these shortcomings are
presented and examples of such novel approaches that have been developed by statistical
methodologists in the pharmaceutical industry together with statisticians in academia
are given.
Results:
In Early Development, i.e., in the first studies in man with a new compound, a combination
of Bayesian methods and modeling approaches is particularly promising to increase
the efficiency of decision making whereas in later phases (IIb and III) a marriage
of modeling and classical frequentist approaches together with novel adaptive designs
is expected to help to chose the right dose regimen and to perform the trials more
efficiently in reduced time.
Conclusions:
The combination of known statistical methods and thinking and the development of
new approaches are in line with the present paradigm of “learning and confirming”
in regulated clinical development while increasing the efficiency of both.
Keywords
Drug evaluation - drug approval - clinical trials - biometry - statistical models
