On April 27th 2022 I presented my masters thesis in statistics, titled:
Decision Making in Early Stage Clinical Trials with Multiple Endpoints
Soon available: - Pre-print version. - Slides for the presentation at the university.
The associated R code will also be published soon.
The thesis was written in collaboration with AstraZeneca; with Karin Nelander and Marcus Millegård as supervisors and José Sánchez as the examiner.
The seminar took place at 10:00am CEST in:
Pascal Lecture Hall
Mathematical Sciences building
Campus Johanneberg (Chalmers)
Abstract: Before a drug is granted regulatory approval, it must be shown to be safe and effective for a specific indication in a large controlled clinical trial. Such studies are costly and expose large numbers of patients to potential risks. Smaller studies are run beforehand to decide whether running such a large study is worthwhile. To obtain sufficient statistical power, these smaller studies often use one or more surrogate variables related to the main indication of interest. In this thesis, an existing framework proposed by Lalonde (2007) for making a decision from a single surrogate variable is generalized to multiple endpoints. Simulations of both synthetic and realistic study designs are used to assess how the choice of endpoints affects the probability of correct and incorrect decisions.