SBRI19012U Big Data Analytics and Machine Learning II – Biostatistics and Epidemiology in Translational Medicine

Volume 2022/2023

BRIDGE - Translational Excellence Programme



The main topic of the course is the introduction of causal learning methods for observational data such as encountered in the Danish national registries. The general aim is to frame better scientific questions with tools from causal inference and to design better statistical analyses that answer our questions. Register-based studies are, by nature, observational and may as such be impacted by confounding, i.e., both drug use and disease outcome may be affected by other factors. We put much emphasis on the methodological issues in the analysis to understand challenges, limitations and possibilities and what it takes to move beyond associations. In order to make valid inference in such studies adjustment for confounding is crucial. We will discuss under which conditions causal conclusions are warranted and methods by which the analysis may be performed.

Learning Outcome

On completion of the course, the participants should be able to:


  • List and distinguish common biases and pitfalls in the analysis of observational studies.
  • Identify and discuss the potential problems with drawing causal conclusions from observational studies, and describe possible approaches to circumvent these problems.
  • Explain the different properties including their advantages and limitations of a number of statistical and machine learning methods for causal inference.



  • Employ tools from causal inference to clearly translate a scientific question to a well-defined causal and statistical parameter.
  • Utilize causal diagrams to identify and vizualize confounding and selection biases.
  • Exercise critical thinking generally in scientific collaboration.
  • Apply inverse probability weighting, standardization and targeted learning in simple statistical analyses to draw causal inference from observational studies.



  • Discuss the general problem of confounding and its critical impact on inference.
  • Discuss and address the effect of time-dependent variables and confounders, and their influence on analysis and inference.
  • Understand the central aspects of observational data analysis, and be able to discuss and communicate these.

Course literature is published on Absalon.

Participants must meet the admission criteria in BRIDGE - Translational Excellence Programme
5 full days with forum lectures and computer exercises.
  • Category
  • Hours
  • Lectures
  • 15
  • Theory exercises
  • 15
  • Total
  • 30
Continuous feedback during the course
Type of assessment
Continuous assessment
Course participation
Type of assessment details
Attendance and active participation
Exam registration requirements

Participants are automatically registered for the Examination upon course registration.

All aids allowed
Marking scale
passed/not passed
Censorship form
No external censorship
Criteria for exam assesment

Active contribution and course participation according to the BRIDGE Guidelines.