CMStatistics 2021: Start Registration
View Submission - CMStatistics
Title: Optimal allocation strategies for blocked clinical trial designs with heterogeneous outcomes Authors:  Lida Mavrogonatou - University of Cambridge (United Kingdom) [presenting]
David Robertson - University of Cambridge (United Kingdom)
Sofia Villar - University of Cambridge (United Kingdom)
Abstract: Clinical trials are multi-objective experiments, incorporating criteria such as power, type I error control, efficiency and patient benefit within and outside the trial. Depending on the trial at hand, optimal performance in terms of one or more of these criteria may take priority over others. The interrelation of the often-conflicting objectives is investigated in such cases. A blocked design setup is considered in which the variance of the treatment response error is allowed to vary both among treatments and blocks. It is shown that the balanced allocation no longer achieves optimal power or efficiency as in the homogeneous case and so alternative allocation strategies are examined. The D-optimal design is shown to be a promising allocation strategy, offering substantial gains in terms of power and estimation accuracy under variance heterogeneity.