Published: 2019-05-10 16:00 | Updated: 2019-05-16 16:15

Dissertation 24 May 2019: Elisabeth Dahlqwist

On Friday May 24 Elisabeth Dahlqwist will defend her thesis "Method developments for the attributable fraction in causal inference". Her opponent will be Professor Xavier de Luna from Handelshögskolan, Umeå universitet. Elisabeth's supervisors are Arvid Sjölander and Yudi Pawitan.

In public health and policy making, understanding the overall impact of an intervention is of essential importance. A way to quantify the disease burden due to some risk factor is by the attributable fraction (AF). The AF is a measure of the proportion of some disease that could be prevented if all would have been unexposed to the risk factor of interest. From the definition of the AF, it is a causal parameter and in order to achieve a causal interpretation of the AF estimate, we have to tackle the challenges of estimating causal effects in observational data. One of them is the problem of confounding, which may cause the researcher to confuse a spurious correlation with a causal effect. In Elisabeth's work, she stresses the importance of using model-based adjustment to estimate the AF and develop novel methods for AF estimation using family data and instrumental variable analysis.

Time and location: 09:00, May 24 2019 in the lecture hall Petrén, Nobels väg 12B, Karolinska Institutet, Solna