Enrollment Speed, Data Heterogeneity, and Outcome of Adaptive Designs

• Activity Number: 58
• Type: Topic Contributed
• Date/Time: Sunday, July 29, 2012 : 4:00 PM to 5:50 PM
• Sponsor: Biopharmaceutical Section
• Author(s): Yili Pritchett*+
• Companies: Abbott Laboratories
• Keywords: Adaptive Designs ; Enrollment Speed ; Data Heterogeneity

Abstract

With the notion of using accumulated interim data to modify aspects of a study design, it is anticipated that a design with adaptive features will gain both speed and efficiency without compromising data quality and integrity. In this presentation, we illustrate that the outcome of adaptive designs, i.e., the successful execution of the decision rules, can be affected if (1) there is data heterogeneity across study sites (or regions) and enrollment among the sites (regions) is initiated sequentially rather than simultaneously, or (2) if there are outliers (either real or due to erroneous data) in the interim data sets. We also illustrate that study enrollment speed also impacts on the intended outcome of adaptive designs. Adaptability could be greatly reduced if subjects' accrual pattern follows a steep exponential shape. We conclude that to achieve the intended outcome of any adaptive design, it is important to manage sites initiation and enrollment speed, and to place measures in the adaptive algorithms such that design engine can possess certain level of robustness in handling real world data. (Contents are subject to change during the abstract revision season)