Patient-centered statistics in clinical prediction, trials and meta-analysis

Abstract

To support the path towards patient-centeredness in clinical research and practice, statistical methodology must also evolve. In this thesis, statistical methods are developed and applied with the aim of better placing patients and their outcomes at the center of clinical research. Part I focuses on the estimation of treatment effects in (meta-)analyses of trials from a patient-centered perspective. Part II brings this perspective to prediction research. The developed methods are applied to clinical studies in the fields of hepatology, oncology, wound healing and cardiology.