Explore the foundations of causal analysis in epidemiological research. This course offers a thorough introduction to the statistical and methodological principles of causal inference, with a particular emphasis on estimating treatment effects using patient data.
You will become familiar with the Neyman-Rubin causal model, also known as the potential outcomes framework, which formalises causal effects by comparing the outcomes each patient could potentially experience under different treatments. Building on this framework, the course highlights how randomized controlled trials (RCTs) provide unbiased estimates of causal effects thanks to the property of exchangeability, which ensures that mean differences between treatment and control groups correspond to average causal effects. This foundation establishes why RCTs are considered the gold standard in causal inference.
The course then explores the challenges of estimating causal effects in observational studies, where treatment assignment is not randomized and exchangeability cannot be assumed. You will learn about confounding variables and the conditions under which valid causal inference is still possible, focusing on statistical methods that adjust for these challenges. Beginning with regression adjustment, the course critically examines its limitations and then introduces propensity score methods - weighting, stratification, and matching - as robust alternatives that reduce reliance on strict modeling assumptions. You will also gain exposure to double robust estimators and the Target Trial Framework, equipping them not only with technical tools but also with the ability to critically assess assumptions and design rigorous epidemiological studies. By the end, you are prepared to apply these techniques in practice while maintaining a clear understanding of their strengths, limitations, and underlying assumptions.
Key takeaways:
- A clear understanding of the Neyman–Rubin causal model and the distinction between causal and associational effects
- Insight into randomised experiments, the concept of exchangeability, and the central role of RCTs in causal inference
- The ability to identify and evaluate the assumptions required for valid causal inference in observational study designs
- Proficiency in regression-based estimation, alongside an awareness of its limitations for estimating causal effects
- Competence in applying propensity score methods, including estimation, assessment of overlap and covariate balance, weighting, stratification, and matchingPractical experience in implementing causal inference techniques in R, supported by tutorials and collaborative study groups
- Familiarity with the Target Trial Framework for designing and critically assessing observational studies
- Training in advanced methods such as double robust estimation and the integration of matching, regression, and stratification techniques
Start Date: 22 January 2026
Tuition fee: € 1,450
Duration: 4 days
Time Commitment: 40–60 hours
Study Load: 2 ECTS
Location: Amsterdam, Frans Otten Stadium
Registration Deadline: 8 January 2026
Accreditation (if applicable): This course is accredited only for Dutch professionals: general practitioners, geriatric specialists, doctors for the mentally handicapped, medical specialist, social physicians, company doctors, insurance doctors, doctors for society and health.
The course “Causaal Inference and Propensity Score Methods” (WK87) is accredited for 20 hours.
To qualify for these credits, there is an attendance requirement of 100%.