|
Abstract:
|
Adaptive clinical trial designs allow for changes in the study design, sample size, or analysis methods based on information observed during a trial. Risks of adaptive designs include statistical, operational and misclassification biases, both real and perceived. Benefits include possibilities for more efficient and ethical trial designs. We explore three types of adaptive sample size designs from the published literature and compare them in both superiority and non-inferiority settings to traditional, fixed sample size designs. We also provide some results based on simulations to address the effects of various misspecifications of assumptions used in the initial sample size calculation. Finally, we summarize the risks and benefits of implementing each of the four approaches. Examples pertaining to clinical trial designs for anti-infective drug products are also discussed.
|
