PhD opportunities

ViCBiostat undertakes research in public health and clinical epidemiology marrying statistical methodology with real-world health-related decision making. Within a wide ranging program of research on statistical methodology applied to health and medical research at ViCBiostat, we have various topics that would be suitable for PhD study. 

PhD SCHOLARSHIP IN BIOSTATISTICS - Accelerated failure time models in clinical trials

We welcome applications for a scholarship at Australian Postgraduate Award (APA) award rates to undertake a PhD at Monash University to work on a project that will investigate the methodological aspects of accelerated failure time models as applied in specific clinical trial design and analysis contexts. It is anticipated that the work will encourage clinical acceptance of these models as an alternative to the commonly used Cox proportional hazards model. The project will be collaborative with supervisors Prof Rory Wolfe and A/Prof Stephane Heritier in the Biostatistics Unit and clinical colleagues at Monash University and affiliated hospitals. You will be closely linked in with other biostatistical researchers and biostatistics PhD students at ViCBiostat (http://www.vicbiostat.org.au/).

Students should have a Bachelors or Masters degree in a field of statistics. Applicants will be asked to apply for an Australian Postgraduate Award through Monash University and those successful will be offered a generous top-up to that scholarship.

For further information or to express interest please contact Rory Wolfe.

To lodge an application please email an up-to-date CV and your academic transcripts to: Rory.Wolfe@monash.edu


The following is an example of a PhD currently being undertaken:

Use of the parametric g-formula for causal inference from observational data

Supervisor – Associate Professor Lyle Gurrin, University of Melbourne

In observational studies, where participants have not been randomized to the treatments of interest, the “parametric g-formula” has been proposed as a simplified, computationally practical alternative to G-computation for comparison of marginal mean outcome risks between treatments regimens allowing adjustment for time-varying confounders that may be intermediate on the causal pathway from earlier exposures to later outcomes. This PhD project will investigate the parametric g-formula method, focusing on the extent to which results depend on the parametric modelling assumptions, and whether similar conclusions hold when the method is extended from binary outcomes to time-to-event data.