Abstract:
|
Adaptive treatment strategies (ATSs) recommend treatments based on evolving patient-specific characteristics, and therefore the identification of an optimal ATS is an important component of the personalized medicine paradigm. Estimation of such treatment strategies using semi-parametric approaches which are doubly robust has gained popularity, as these methods afford some degree of protection in settings where it may not be possible to specify all models correctly. In this talk, I will demonstrate some model diagnostic and selection tools that have been especially adapted for use in doubly robust estimation of optimal DTRs.
|