Event

Even though the advantages of randomised trials to assess treatment effectiveness are well known, particularly to address confounding, the limitations are also well established; trials are often restricted to highly selective populations, involve high costs and may be infeasible due to ethic or practical constraints. Analyses of observational data can help to overcome these drawbacks, but confounding remains a major concern. The purpose of jointly analysing trial and observational data to answer a relevant clinical question is precisely to exploit the best of both worlds. In this talk, we discuss the identifiability assumptions required for the joint analysis, considering selection of participants into the trial, assignment into treatment groups and adherence to the assigned treatment. We propose estimators that jointly use data from trial participants and non-participants under different sets of assumptions and illustrate them using data from a major cardiovascular trial nested in the nationwide health registry.

Camila Olarte Parra is a postdoctoral researcher in the CAUSALab of the Unit of Epidemiology at Karolinska Institutet. She was trained as a medical doctor, has a Masters in Epidemiology and a PhD in Statistical Data Analysis. Her research interests include causal inference methodology, causal language, registry-based studies and the estimand framework.

This seminar will be held via Zoom and will be recorded.

Date and time

Thursday 25 September

4-5pm AEST

Zoom details

Please click this URL to start or join. https://monash.zoom.us/j/88387257964?pwd=QbJjJdakGy130gRb6liQxTdPhAFpaU.1     

Or, go to https://monash.zoom.us/join and enter:

Meeting ID: 883 8725 7964

Passcode: 729533