NIASRA diagram

Causal Diagrams: A tool to better understand results from epidemiological and machine learning analyses

This session will introduce participants to the field of causal inference. We will review the distinction between causal and predictive research questions, and where machine learning and traditional epidemiological techniques may be misinterpreted. We will review the fundamentals of directed acyclic graphs (DAGs), and how they can be used to determine appropriate sets of variables for estimating total causal effects of exposure (treatment). The session will highlight when automated “sophisticated analyses” will likely fail, and why clinician expertise is required to make sense of any data that will be used in the development of interventions.


Dr Ian Shrier is an Associate Professor, Department of Family Medicine, McGill University, Canada and Senior Investigator, Centre for Clinical Epidemiology, Lady Davis Institute.

Dr. Shrier has been practicing sport medicine for over 30 years and is a past President of the Canadian Academy of Sport Medicine. He has a PhD in Physiology, post-doctoral training in Epidemiology and over 290 peer-reviewed publications on topics related to the effects of exercise and sport medicine injuries, return-to-play decision making, injury epidemiology, causal inference, and meta-analyses. He was the co-Editor in Chief of Review Synthesis Methods, and is an editorial board member of three international sport medicine journals.