Components of causal inference with focus on assumptions and confounding control
Doctoral School in Medicine, Biomedical Science and Technology at Aalborg University
You already know that establishing a causal relationship is distinct from observing an association. While individuals who receive the flu vaccine tend to have a lower mortality rate compared to those who do not, we must consider whether this lower mortality is directly attributable to the vaccine or if it arises from other distinctions between the vaccinated and unvaccinated groups. The concept of confounding introduces a pervasive bias when we compare groups that are not fundamentally similar. It represents a substantial challenge to drawing accurate causal conclusions from observational data. Consequently, the course's primary focus revolves around the essential task of mitigating bias and confounding in epidemiological research using various techniques.