Heidelberger Kolloquium

Schedule of event:

27/01/2020, 16:15

More event dates:

For this event there are no more dates.


Moritz Danzer, Institut f. Biometrie u. Klinische Forschung, Westfälische Wilhelms-Universität Münster






Adaptive Designs for Multivariate Single-Arm Survival Trials

Confirmatory sequential and adaptive designs comprise a pool of methods that allow the investigator to conclude an ongoing trial or to modify its design based on interim data without compromising type I error rate control. Current sequential and adaptive survival tests were developed for confirmatory survival trials with only one survival endpoint. Those methods allow design modifications to be based on interim data of the primary endpoints, whereas the use of interim data from further time-to-event endpoints is possible only with disadvantages. For survival trials with two or more primary time-to-event endpoints (multivariate survival data), however, no adaptive sequential design methodology has been provided so far.

By applying techniques similar to those described in (Schmidt et al. 2017} we develop a sequential multivariate one-sample log-rank test. Using it as an adaptive design, provision is made for design modifications based on interim data from all involved survival endpoints simultaneously. These modifications can be made at initially planned interim analyses of the original sequential design as well as at initially unplanned dates.

Our methodology applies in confirmatory single-arm Phase IIa survival trials, especially in pediatric oncology, where reliable reference curves are obtainable. Slight modifications of the general procedure enable an application to settings with chronologically ordered endpoints (e.g. progression and overall survival) or settings in which one endpoint is associated to an absorbing state.

We also conduct a simulation study to check properties of the test statistics in trials with small sample sizes.


K13 Marsilius-Arkaden, Turm Süd

 back to overview
to top