Skip to the content
 

Causal approaches to the design and analysis of epidemiological studies

This course has reached maximum capacity - no exceptions.

Course

Causal approaches to the design and analysis of epidemiological studies

Convenor

Professor Andrew Forbes
Presenter

Professor Jonathan Sterne
Department of Social Medicine, University of Bristol, UK

Date

August 9, 2010

Time

8:45am – 5:15pm

Venue

Seminar Room 1, Level 5,
The Alfred Centre, 99 Commercial Road. Melbourne, Vic, 3004

Suitable for

Epidemiologists, Biostatisticians, PhD students, and anyone interested in learning about causal concepts in epidemiological research.
Prerequisite knowledge for this workshop is a good understanding of epidemiological and statistical concepts including multivariable regression models and survival analysis.

Aims
  1. To explain causal terminology and how causal graphs can inform approaches to statistical analyses.
  2. To provide an in-depth understanding of confounding and its control.
  3. To explain the role and potential of different methodological approaches to overcome confounding and selection bias, including inverse probability weighting and marginal structural models.
  4. To provide an opportunity for participants to cast their own substantive problems in causal terms and diagrams.
  5. To provide an opportunity for practical data analysis experience using Stata software (for participants who are Stata users).

Overview

In the past two decades causal diagrams and statistical methods for causal inferences have increasingly informed the design and been used for the analysis of epidemiological studies.  Causal diagrams offer an intuitive approach to the documentation and communication of assumptions about the relationships between variables in observational and randomised studies.  They provide a unified approach to the study of confounding and selection biases and they present an excellent opportunity for integration into epidemiological teaching at introductory and advanced levels.  These diagrams have also informed the development of new analytic methods to deal with confounding and selection biases, particularly for the analysis of longitudinal studies with time-varying exposures and confounders. 
This workshop will cover the above areas using a mixture of presentations, small-group discussions and data analysis practical examples.

Workshop timetable  (PDF, 2pg, 59kb)
Materials Course notes will be included.
Note on computing:  Familiarity with the Stata software package is advantageous but not required.  Participants familiar with Stata are encouraged to bring a laptop computer with Stata pre-loaded to enable hands-on practical data analysis.  Participants not familiar with Stata will receive a guided interactive demonstration of data analysis with detailed discussion.
Catering Morning tea, lunch and afternoon tea will be provided
Further information Contact Andrew.Forbes@monash.edu

Fee

$300 (incl GST)
$250 Early Bird (registration by July 18)
Discounts for students/staff as per the registration form.

NOTE:  Places are strictly limited so early registration is encouraged
While the information contained herein was correct at the date of publication, Monash reserves the right to alter procedures, fees and regulations should the need arise.