Optimizing Patient Enrollment in Global Clinical Trials by Metaheuristics

• Activity Number: 217 - Theory and Algorithms for Adaptive Clinical Trial Design with Multiple Objectives
• Type: Invited
• Date/Time: Tuesday, August 9, 2022 : 8:30 AM to 10:20 AM
• Sponsor: WNAR
• Author(s): Weng Kee Wong* and Mitchell Schepps and Austin Matt and Vlad Anisimov
• Companies: UCLA and UCLA and Amgen and Amgen
• Keywords: Optimal Enrollment Design; Constrained Optimization; Genetic Algorithm; Particle Swarm Optimization; Poisson-Gamma Enrollment Model; Tuning Paramters

Abstract

Multi-center clinical trials enroll a large number of participants from multiple centers and frequently they do not meet the target number due to a myriad of administrative, cost and unanticipated issues. For multi-center trials conducted globally, the problem to create an optimal enrollment design is very complex, and may include multiple objectives and nonlinear constraints that arise from physical, regulatory or budgetary considerations, and additional practical limits imposed by the various centers in different countries. Standard optimization techniques and mathematical programming methods do not usually work well for solving such large high-dimensional optimization problems. Nature-inspired metaheuristic algorithms have emerged as a powerful and dominating optimization tools in computer science and engineering to solve complex and high-dimensional optimization problems. We review this class of flexible and assumptions free algorithms and demonstrate, for the first time, their capability and utility in solving the problems of optimizing enrollment design under different constraints for multi-center clinical trials.