ABSTRACT
The rapidly escalating costs and increasing failure rate in drug development
decrease willingness and ability to bring many candidates forward into the
clinic. There are multiple reasons: (1) Those “low-hanging fruit” in the chem-
ical compound space have already been picked, (2) due to ethical considera-
tions, a pivotal clinical trial cannot be a placebo-controlled trial when a drug
for the disease is available; however, an active-controlled, in which the test
drug is compared against the best drug in the market, makes the efficacy mar-
gin much smaller than in a placebo-controlled trial. However, the regulatory
agencies in general still require type-I error rates of at most 5%, although
the meaning of the type-I error is completely different: one is compared to
placebo and the other is compared to active drug. Thus a diminished margin
for improvement escalates the level of difficulty in statistically proving drug
benefits or the sample size must increase, as must the cost and time. To stay in
competition, many pharmaceutical companies adopt different strategies, such
as company mergers or acquisitions, late phase outsourcing, reducing work-
force in the early research but buying promising compounds that are ready for
clinical trials. In addition to implementations of strategies in business, there
is also a significant change in the technological side: a paradigm shift from
classical statistic clinical trial designs to adaptive designs, which can lead to
a reduction in cost and time, and an increase in success rate in drug devel-
opment. In this chapter, we will review commonly used adaptive designs and
questions and answers that a newcomer may have.