Mendelian Randomisation [GE10]

Detailed information about this course:

Description

This 3-day course aims to give an overview of recent developments in drawing causal inferences from epidemiological data using Mendelian randomization and is aimed at individuals with a background in statistics or epidemiology with a strong statistics component. Theoretical discussion will be accompanied by introduction to available software and lab practicals.

The first day will comprise a brief introduction to graphical models since these provide a natural framework for expressing and manipulating many of the concepts involved. We will then go on to causal modelling and the need for a formal causal framework before explicitly considering instrumental variable methods with Mendelian randomisation applications as the primary example.

On the second day, a practical session in analysing Mendelian randomization studies will put into focus what can and cannot be done. We then continue with the discussion of various theoretical issues, including the introduction of Bayesian approaches to Mendelian randomization.

The final day will start with a practical session in analysing Mendelian randomization studies using a Bayesian approach. We conclude by introducing current topics of interest including checking for violations of assumptions, the use of multiple instruments and implications for case-control data.

N.B.: This course is organised every other year.

Objectives

• Gain insight in new developments in the field of genetic epidemiology.
• To gain insight into disentangling the concepts of causation and association.
  

Participant profile

Epidemiologists, geneticists, biostatisticians, investigators/physicians involved in clinical and pharmacological research.

Assessment

Assignment(s)