Motivation:
The growth of individual level data available from national health care registries offers unprecedented opportunities to answer clinically relevant questions in a variety of medical areas. This type of data also comes with unique challenges for study design and analysis. Register data are often of the form of diagnostic codes, procedures, and medications observed repeatedly over time, and outcomes such as cancer and death from different causes. The session will focus on statistical solutions to the issues and methodological challenges that arise in the use of register data for medical research. Although we are focused on register-based research, these challenges are shared by many other biostatistics applications, and methodology developed for register-based situations will be translatable to many others.
Organizer:
Michael C. Sachs, Department of Medical Epidemiology and Biostatistics, Karolinska Institute
Speakers:
Yun-Hee Choi, Department of Epidemiology and Biostatistics, Western University, Ontario
A Competing Risk Model with Time Varying Covariates for Estimating Breast and Ovarian Cancer Risks in BRAC1/2 Families
Michael Crowther, Department of Health Sciences, University of Leicester
Modelling complex disease trajectories using multi-state survival models: Estimation, prediction and software
Erin Gabriel
Discussant:
Paul Lambert, University of Leicester