In dose-response studies with censored time-to-event outcomes, D-optimal designs
depend on the true model parameters and the number of censored outcomes. In order
to implement such a design in practice, an adaptive design that incorporates updated
knowledge about the dose-response curve at interim analyses can be used [1]. Further,
treatment allocation should involve randomization, which is essential to mitigate various
experimental biases and perform valid statistical inference at the end of the trial. Here,
we compare several randomization procedures and their impact on model estimation.
Full Article:
