|
Abstract:
|
Stratified medicine has tremendous potential to deliver more effective therapeutic intervention to improve public health. For practical implementation, reliable prediction models and clinically meaningful categorization of predicted individual treatment effect scores are vital elements to identify subpopulations warranting different intervention strategies. A systematic approach integrating 1) prediction model building and variable selection, 2) stratification via constrained optimization, and 3) reproducibility assessment to optimally categorize subjects into different groups with a pre-specified clinically meaningful successive average treatment effect will be presented. We illustrate how the proposed methodology can help identify subgroups with the most beneficial prospect and futile subgroups receiving treatment to inform future trial design or treatment selection strategy.
|
