News

Get your tickets now for the 2024 ViCBiostat Summer School!

We are pleased to be delivering three courses led by local members Thao Le, Julie Simpson, Rory Wolfe, Margarita Moreno-Betancur and Ghazaleh Dashti; as well as special guest Ronald Geskus (OUCRU, Vietnam).

All courses will be available in-person and online via Zoom.

 

Prediction modelling

12&13 February, 9am – 5pm

Presented by: Thao Le, Ronald Geskus, Julie Simpson, Megha Rajasekhar and Rory Wolfe

This course provides an introduction to prognostic modelling in the context of health care research. It covers important principles for model development, internal and external validation, and the use of machine learning methods in model building. Computer practicals will be in both R and Stata. Upon completion of the course, participants will be familiar with the main approaches to developing and validating new prediction models in health and will have an understanding of how to implement the methods in widely used statistical software. This course is suitable for participants who are already familiar with commonly-seen statistical concepts and methods including multivariable regression models such as gained from statistics units in a Masters of Public Health.

Prerequisites: Participants should be familiar with one of R or Stata, e.g. able to create new variables, run regression commands and obtain basic plots.

 

Advanced survival analysis

14&15 February, 9am – 5pm

Presented by: Ronald Geskus, Julie Simpson, Thao Le, Rory Wolfe, Anais Charles-Nelson and Stephane Heritier

 

This course provides a deeper understanding of the main concepts, quantities and models in the analysis of time-to-event data. We explain the different types of censored and truncated data and the assumptions behind the basic estimation methods. The basic Cox proportional hazards model is extended by covering nonproportional hazards, the equivalent Poisson regression approach, time-varying covariates and immortal time bias. We explain “the hazard of hazard ratios” and introduce models based on restricted mean survival time.  A specific focus will be on key concepts in competing risks analysis. Computer practicals will be provided in both R and Stata.  Participants should have some understanding of the basics of survival analysis (familiarity with right censored data, Kaplan-Meier and the Cox proportional hazards regression model).

Prerequisites: Participants should have basic familiarity with one of R or Stata and ability to conduct basic survival analyses in the package, e.g. obtain a Kaplan-Meier plot.

 

Causal mediation analysis

16 February, 9am – 5pm

Presented by: Margarita Moreno-Betancur and Ghazaleh Dashti

 

Many health research questions concern the multiple pathways that are presumed to mediate a relationship between an exposure and an outcome. Very often, the translational intent of such research questions is to inform potential intervention targets. However, the usual causal mediation approaches do not consider this interventional intent and/or rely on assumptions that are either too stringent or not assessable in practice. Recently an alternative approach has emerged based on “interventional effects” that assess the impact of relevant interventions on one or multiple mediators and are identifiable under relaxed assumptions. This approach is gaining popularity in applications, making it timely to present this topic.

This course provides an overview of the conceptual issues surrounding causal mediation analysis. It then presents the interventional effects approach, using a recently proposed framework that defines these effects by mapping them to a “target trial” that evaluates interventions on one or several mediators. We describe how to define and emulate a target trial for mediation analysis and introduce an extended g-computation approach for estimating these effects.

Lectures and tutorials will ground understanding of the methods, whilst a hands-on computer practical (in R and Stata) will cover their practical implementation. Illustrations from real-world epidemiological studies are included throughout.

Prerequisites:  It is strongly recommended that participants have previously taken an introductory course on causal inference, such as a prior ViCBiostat causal inference workshop, and have familiarity with concepts and methods such as the target trial and g-computation. To do the computer practical, students must also have a sound working familiarity with Stata or R and have the corresponding software installed on their computer or laptop. A list of R packages required will be circulated with electronic copies of the course material beforehand.

 

 

Venue:

Ground Floor Conference Rooms

Monash University SPHPM

553 St Kilda Road, Melbourne

and online via Zoom meeting

 

Register here now: https://shop.monash.edu/vicbiostat-2024-summer-school.html 

Registration prices