Title: Disease risk estimation under clustered and outcome-dependent sampling
Authors: Marianne Jonker - Radboudumc (Netherlands) [presenting]
Mar Rodriguez-Girondo - Leiden University Medical Center (Netherlands)
Abstract: Family-based cohort studies are often used sources of data for the estimation of survival and risk prediction in populations with a high genetic predisposition to a disease. Despite the abundance of family-based data available and the increasing need in the medical world for statistical expertise to analyze these data, the development of statistical methodology is slow. Some of the challenges that are encountered when analyzing these types of data are the way the data are collected (families are ascertainment based on the presence of the disease of interest), clustering due latent factors, missing data structure (family structure) and right or interval censoring of the event of interest. An overview of existing methods is given with a focus on more recent developments, illustrated with some medical applications.