Traditionally, clinical trials are designed using a frequentist paradigm and aim to establish if a treatment is superior (non-inferior) to control on average.
Precision medicine trials goes further and seeks to establish whether a treatment works for a particular patient and to what extent.
Basket trials are one particular type of clinical trial with the goals of precision medicine in mind.
They simultaneously evaluate a new targeted therapy in different patient subgroups (e.g., defined by disease subtype) that share a commonality (e.g., genetic aberration) that the new treatment targets – all under one overarching protocol.
Efficient analysis of basket trials often features ‘borrowing of information’ across the subgroups and a number of highly efficient, robust methods have been proposed to this end.
In this course, through lectures and practical sessions, we will formally introduce Basket trials, review several efficient approaches for the analysis of Basket trials and discuss their merits and demerits. Particular topics discussed include:
Robust Bayesian hierarchical models and their variations
Sample size determination for basket trials
Inclusion of interim analyses in trial with information borrowing