14 Nov 2017 09:00am to 05:00pm

Propensity score methods for estimating causal effects in non-experimental studies

Event Location
Room 102
Redmond Barry Building, Latham Theatre Tin Alley
Carlton VIC 3052
Prof. Elizabeth Stuart
Johns Hopkins Bloomberg School of Public Health

ViCBiostat is hosting respected speaker Professor Elizabeth Stuart from Johns Hopkins University who will present this one-day workshop.

Propensity scores are an increasingly common tool for estimating the effects of interventions in observational (“non-experimental”) settings and for answering complex questions in randomized controlled trials.  They can be of great use in medicine, public health, and the social sciences, for example examining the effects of long-term use of pharmaceuticals, or of depression treatment for adolescents, or of potentially adverse exposures such as childhood maltreatment.   This short course will discuss the importance of the careful design of observational studies, and the role of propensity scores in that design, with the main goal of providing practical guidance on the use of propensity scores to estimate causal effects.  The course will cover the primary ways of using propensity scores to adjust for confounders when estimating the effect of a particular “cause” or “intervention,” including weighting, subclassification, and matching. 

Topics covered will include:

- How to specify and estimate the propensity score model
- Selecting covariates to include in the modes
- Diagnostics
- Common challenges and solutions. 

The methods will be illustrated using a case study using large-scale administrative data from Denmark to estimate the effects of a suicide prevention program on suicide attempts.    Software for implementing analyses using propensity scores will be briefly discussed, including resources for Stata and R.   The course will also highlight recent advances in the propensity score literature,  including prognostic scores, covariate balancing propensity scores, methods for non-binary treatments (such as dosage levels of a drug or when comparing multiple programs simultaneously), and approaches to be used when there are large numbers of covariates available (as in claims or other “big” data). 

Who should attend?

Attendees will either be biostatisticians or will have completed “applied multivariable regression” subjects in an MPH/M.Epi/MSc.


There will be several demonstrations of implementing the methods discussed in both Stata and R, and example computing code will be provided to participants. There will, however, be no computing practical classes on the day, so there is no need for participants to bring a laptop computer.


Registration via Trybooking now open.