Admit One: QEG 2019 Event


Register Now


The Queensland Epidemiology Group (QEG) is the Queensland chapter of the Australasian Epidemiology Association (AEA). Together with the Menzies School of Health Research and the University of Queensland, QEG and the AEA are proud to present:

An Evening With Dr Ellie Murray: 

a epi masterclass on causal inference methodology

5:00 - 8:00 pm (including light refreshments)

Monday 25th November, 2019

University of Queensland, Oral Health Centre Auditorium, Herston

Registration is necessary for catering and seating capacity purposes. 

Eleanor Murray is an assistant professor at the Boston University School of Public Health. Her research is on causal inference methodology for improving evidence-based decision-making by patients, clinicians, and policy makers. She uses novel statistical methods to answer comparative effectiveness questions for complex and time-varying treatments using observational data and randomized trials when available, and individual-level simulation modeling when insufficient data exists in the time frame required for decision-making. She is applying these methods to a variety of medical conditions including HIV progression, cancer, psychiatric conditions, and cardiovascular disease. She was a postdoctoral research fellow at the Harvard T.H. Chan School of Public Health, working on causal inference for comparative effectiveness and real-world evidence in the HSPH Program on Causal Inference. She holds an ScD in Epidemiology and MSc in Biostatistics from Harvard, an MPH in Epidemiology from Columbia Mailman School of Public Health, and a BSc in Biology from McGill University. Ellie is the Associated Editor for Social Media at the American Journal of Epidemiology and is also known as @EpiEllie of #EpiTwitter. 

Improving causal inference from observational data using the target trial framework

Causal inference from observational data is often regarded as unreliable because of the high potential for bias in the design and analysis of observational studies. In contrast, randomized controlled trials are generally expected to be free of confounding at baseline, by definition have well-defined interventions, and are guaranteed to have positivity - the three most important conditions that must hold in order to estimate causal effects. Randomized controlled trials are also typically free from immortal time bias, which is a less commonly discussed but pervasive bias in observational data analyses. We can exploit these properties of randomized trials to design better observational data analyses via the target trial framework. 

In this talk, I will describe the theoretical and practical application of the target trial framework to improving causal inference from observational data with examples from HIV progression and cardiovascular disease.

Causal inference from pragmatic randomized trials requires analytic methods from observational studies

Pragmatic randomized trials are key tools for research on the comparative effectiveness of medical interventions. Unlike other randomized trials, pragmatic trials are specifically designed to address real-world questions about options for care and therefore to guide decisions by patients, clinicians, and other stakeholders. Therefore, characteristics of a pragmatic trial include typical patients and care settings, clinically relevant comparators, unconcealed assignment to treatment, and follow-up time long enough to study long-term clinical outcomes without having to rely on surrogates. 

While pragmatic trials are useful to guide decision making, they are also especially vulnerable to post-randomization confounding, from incomplete adherence and post-randomization selection bias from loss to follow-up. These sources of bias are common in observational epidemiology, and the use of analytic approaches pioneered for observational studies can improve inference from pragmatic trials. 

We propose causal inference guidelines tailored for the analysis of pragmatic randomized trials using methods from observational research. Importantly, conventional methods to adjust for confounding and selection bias do not generally work for post-randomization variables. In fact, conventional methods such s multivariate outcome regression, stratified analyses, propensity score regression and matching, and others may themselves introduce bias. Therefore, our guidelines are based on so-called g-methods, developed by Robins and collaborators since 1986, which can appropriately adjust for post-randomization biases. Because g-methods require data on post-randomization (that is, time-varying) treatments and covariates, embracing these guidelines will require a revised framework for both the design and conduct of both pragmatic trials and other trials with substantial loss to follow-up or non-adherence. 

Live streaming of this event may be available; details for this will become available closer to the event date. If you are able to come along in person, we encourage you to do so. This is a wonderful opportunity to meet with other QEG members and epidemiology aficionados, including our esteemed speaker, Dr Eleanor Murray.

For more information about the AEA, including other events and membership, please visit:

Event Link: Register Now