The BIOSTATISTICS Section of The Belgian Statistical Society organizes a short course

INTRODUCTION TO IN BIOMEDICAL APPLICATIONS

May 3-4, 9:00 - 17:00, Louvain-la-Neuve, Belgium

Instructor: (Imperial College, London, UK)

COURSE OUTLINE ------Day 1 Introduction to Bayesian Inference

* Introduction to the basic principles of Bayesian inference, including simple examples of conjugate prior-to-posterior inference; * Discussion of choice of non-informative and informative prior distributions for Bayesian inference; * Introduction to Bayesian hierarchical models and use of Bayesian graphical models as a tool for building complex hierarchical models; * Comparison of Bayesian and classical approaches to statistical inference; * Case-studies providing a practical illustration of the use and interpretation of Bayesian methods in medical, pharmaceutical and epidemiological applications. Examples will be drawn from topics including population pharmacokinetics, meta-analysis, analysis of repeated measures and non-linear models in clinical trials, accounting for multiple comparisons in drug safety trials, disease mapping.

Day 2 Overview of MCMC simulation methods and introduction to WinBUGS

* Overview of Markov chain Monte Carlo (MCMC) simulation methods used to carry out Bayesian calculations; * Discussion of practical issues concerning interpretation of MCMC output; * Demonstration of the WinBUGS software; * Practical exercises using WinBUGS to fit Bayesian hierarchical models to biomedical datasets.

INSTRUCTOR: DR NICKY BEST

Dr Nicky Best is a Reader in Statistics and at the Department of Epidemiology & Public Health at Imperial College, London, UK. She completed her PhD in biostatistics at Cambridge University in 1994 before moving to Imperial in 1996. Her current research interests include spatial modelling of environmental and epidemiological data and other applications of Bayesian hierarchical models in medical research, and she is part of the project team developing the WinBUGS Bayesian software package. She has taught numerous short courses on Bayesian methods and use of the WinBUGS software around the world. In 2004 she was awarded the Royal Statistical Society in Bronze for her pioneering work on the effective use of Bayesian hierarchical models for complex variability in medicine and epidemiology, and for her work in promoting the BUGS software to a wide statistical audience.

WHO SHOULD ATTEND?

The course is aimed at statisticians who will make their first moves into Bayesian statistics.

REGISTRATION AND FURTHER DETAILS

Please, visit the website http://www.stat.ucl.ac.be/~lambert/BiostatSection/