Journal of Clinical Trials
Open Access
ISSN: 2167-0870
+44 1478 350008
ISSN: 2167-0870
+44 1478 350008
Vishal Ahuja
Southern Methodist University, USA
Posters & Accepted Abstracts: J Clin Trials
Clinical trials have traditionally followed a fixed randomized design, where patients are typically allocated once, at random, and usually equally to various treatments. Such designs provide a clean way of separating out the effects of alternate treatments. Response-adaptive designs, where assignment to treatments evolves dynamically as patient outcomes are observed, are gaining in popularity due to potential for improvements in both cost and efficiency over traditional designs. An ideal adaptive design is one where patients are treated as effectively as possible without sacrificing the potential learning or compromising the integrity of the trial. We propose such a design, termed Jointly Adaptive, which uses forward-looking algorithms to fully exploit learning from multiple patients simultaneously. Compared to the best existing implementable adaptive design that employs a multi-armed bandit framework in a setting where multiple patients arrive sequentially, we show that our proposed design improves health outcomes of patients in the trial, in expectation, by 8.6% under a set of considered scenarios. We also demonstrate our design��?s effectiveness using data from a recently conducted stent trial, where we demonstrate an improvement of over 37%, in expectation. A consequence of using forward-looking algorithms in the above approach is that the problem size grows exponentially with the number of patients and time periods, making it computationally challenging to solve. To address this, we propose grid-based approximation methods that reduce problem dimensionality and allow for the implementation of adaptive designs to large clinical trials. We use numerical examples to demonstrate the effectiveness of our approach.
Email: vahuja@mail.smu.edu