Yuetian Luo
Rutgers University
Room: M3 3127
Algorithmic Stability in Statistical Inference
This presentation will focus on two different designs that are currently being used to test interventions in children. Each design uses an alternative approach to improve the feasibility of these clinical trials. Firstly, we will discuss a novel method to rank interventions during the trial to determine the best intervention. This method is particularly useful when it is challenging to specify a “control” intervention as there is a great variation in current practice, which is common in paediatrics. We will then discuss a design that uses Bayesian hierarchical modelling to evaluate interventions when sample size is severely limited, e.g., for rare diseases, by combining information across rare and common diagnoses.