ADIT: Liberal and flexible statistical methods for the evaluation of individualized therapies

The overall objective of the joint project ADIT is to combine the flexibility of adaptive designs for clinical trials with the new requirements and dynamic developments of individualized therapy approaches. A good example is pediatric oncology where highly stratified therapies are becoming increasingly important, however, the underlying population is such limited that a formal proof of efficacy is hardly possible under the common statistical principles. Hence, there is a need for new statistical principles and methods that foster the difficult development process by reducing the burden of formal statistical requirements.

The subproject in Bremen will address this goal by two general approaches. The first is the development of statistical methods that follow a new, more liberal but still acceptable approach for multiple type I error rate control. The new principle is less conservative than the common family wise error rate control but still bounds the future patients’ risks caused by the potential erroneous trial decisions. In a second approach we aim to increase the flexibility of the new approach by using less rigorous statistical methods in the early stages of an adaptive trial while in the later stages the level of rigor is increased in order to extend and/or confirm the earlier findings.

Project start: Okt 2015
Funding: BMBF
Subproject members: K. Rohmeyer and W. Brannath (former member: Sarah Nanninga)
Project partner: Dr. Robert Kwiecien (IBKF, WWU Münster) and Prof. Olaf Witt (DKFZ Heidelberg)