CHI/CANSSI Prairies Workshop: Causal Inference: Insights and Applications

Date: Monday, Wednesday, Friday, May 12, 14, and 16, 2025
Time: 9:00 a.m.–3:30 p.m.
Place: Hybrid (in person and on Zoom); University of Manitoba, Bannatyne Campus, Chown Building, Room 207 A&B

We encourage participants within Manitoba to attend in person.

Workshop Description

This three-day workshop on “Causal Inference: Insights and Applications” is organized by the George & Fay Yee Centre for Healthcare Innovation (CHI) at the University of Manitoba and is the fourth workshop in the CANSSI Prairies Workshop Series in Data Science. It will equip researchers with the essential tools needed to effectively analyze and interpret observational data. Participants will explore foundational concepts, causal diagrams, and statistical methods for adjusting confounding variables. Topics will include non-parametric techniques, propensity score methods, doubly robust approaches, and machine learning strategies.

The workshop will be presented by Sumeet Kalia (Department of Statistics, University of Manitoba), Brenden Dufault (George & Fay Yee Centre for Healthcare Innovation, University of Manitoba), and Amani Hamad (Department of Community Health Sciences, University of Manitoba) and features hands-on sessions using R, real-world case studies, and interactive discussions to enhance understanding of study design and data analysis. This workshop is ideal for professionals, academics, and graduate students in statistics, data science, and health sciences who wish to improve their expertise in observational research.

Refine your research skills and gain confidence in tackling complex observational studies—register today!

Cost

• Academic trainees and students: $100
• Staff of non-profit organizations (including postdocs and early career researchers): $300
• Industry professionals: $600

In line with CANSSI Prairies’ commitment to enhancing knowledge and skills in various areas of data science, we are pleased to announce that support is available to provide a 50% discount on the registration fees for the first 15 students and five postdocs or early career researchers (held their first independent academic position within the past five years) who register.

For more information, please contact Olawale Ayilara at olawale.ayilara@umanitoba.ca.

Registration

REGISTER ON EVENTBRITE

Workshop Outline

Day 1: Foundations and Core Concepts

1. Welcome and Introduction
   • Brief overview of workshop objectives/topics
   • Introduction of instructors and participants
2. Key Concepts in Observational Studies – Dr. Amani Hamad
   • Role of observational studies in causal inference
   • Main observational study designs and their key features, strengths and limitations
   • Common biases in observational studies (selection bias, information bias, confounding)
3. Counterfactuals and Causal Diagrams – Brenden Dufault
   • Non-parametric adjustment
   • Encoding our causal assumptions with directed acyclic graphs (DAGs)
   • How to understand and diagnose common biases using DAGs (selection bias, information bias, confounding)
   • Software for DAGs
   • Practical applications of DAGs for observational studies and imperfect RCTs
4. Statistical Methods for Confounding Adjustment Part I – Brenden Dufault
   • Stratification
   • Multivariable regression
   • G-computation
   • Front door adjustment

Day 2: Statistical Methods for Addressing Bias – Brenden Dufault

5. Propensity Scores
   • Theory of balancing scores for confounder adjustment
   • Estimands beyond the average treatment effect
   • Propensity score methods: matching, stratification, and weighting
   • Guided exercises using statistical software (RStudio)
   • Visualization and diagnostics
6. Specialized Methods
   • Parametric G-computation for mediation
   • IPTW for handling censoring/dropout
   • Causal forests for confounder adjustment
7. Group Discussion and Q&A
   • Discussion on the challenges and limitations of the discussed methods
   • Q&A session to address participant questions

Day 3: Advanced Topics in Observational Studies – Dr. Sumeet Kalia

8. Advanced Topics
   • Treatment-confounder feedback
   • Parametric and non-parametric G-estimation
   • Doubly robust estimation
   • Targeted maximum likelihood estimation (TMLE)
   • Machine learning methods (regression trees; super learner)
   • Instrumental variable with binary and continuous treatments
   • Sensitivity analysis for unmeasured confounding (negative control exposure and outcome; E-value)
9. Case Studies in Observational Research
   • Analysis of real-world observational studies
   • Group discussions on methodology and interpretation
10. Conclusion and Next Steps
    • Summary of the workshop, highlights, and key takeaways

About the Speakers

Dr. Sumeet Kalia is an Assistant Professor in the Department of Statistics at the University of Manitoba. He earned his PhD in Biostatistics from the University of Toronto with a dissertation titled Causal Inference Using Electronic Health Records in Primary Care. Dr. Kalia also holds an MSc in Biostatistics from Western University. Previously, he worked as a Research Analyst (Biostatistician) in the Department of Family and Community Medicine at the University of Toronto, conducting applied and methodological research on causal inference using primary care electronic health records.

Brenden Dufault is a Biostatistical Consultant with the George & Fay Yee Centre for Healthcare Innovation, University of Manitoba, with over 15 years working in clinical trials, medicine, and epidemiology. He specializes in the analysis of observational data using causal methods, and teaches workshops on statistical programming and applied statistics.

Dr. Amani Hamad is an Assistant Professor in the Department of Community Health Sciences at the University of Manitoba. She is the Canada Research Chair in Population Data Science and Data Curation (Tier II) and is a Research Scientist at the Manitoba Centre for Health Policy (MCHP). Dr. Hamad earned her PhD in Pharmacy from the University of Manitoba and completed a postdoctoral fellowship at the George & Fay Yee Centre for Healthcare Innovation Data Science platform. Her research expertise includes population data science, pharmacoepidemiology, maternal and child health, and mental health.