Summary of the UiO:RealArt convergence environment. Illustration: Emilie Willoch Olstad.

Real world – artificial worlds: improving causal inference in perinatal pharmaco-epidemiology using machine learning approaches on real-world and artificial data.

Perinatal pharmacoepidemiological studies are essential for providing society with knowledge about medication safety in pregnancy. Existing studies are most commonly based on observational data, due to the ethical challenges of using randomized-controlled trials in this field. Observational studies often have critical methodological flaws due to bias and confounding, making causal interpretation very difficult.

The UiO:RealArt convergence environment aims to combat the methodological stagnation in perinatal epidemiology, by uniting machine learning, pharmacological and biological sciences, and social and education sciences. We explore artificial worlds and the potential of embedding machine learning techniques within the causal inference framework. UiO:RealArt is a novel endeavor, capitalizing on the advantages Norway has through access to linked health and education registries in combination with observational and biological data from a national birth cohort. 

Image may contain: Glasses, Hair, Nose, Head, Photograph.
The UiO:RealArt leader group. Photos: private.

To unite the study of real-world and artificial data, we have expertise within a range of different disciplines, including pharmacology, machine learning, statistics, genetics, epigenetics, psychology, language development and educational attainment. The leader group reflects the interdisciplinary approach, representing six different departments (Dpt. of Informatics, Dpt. of Pharmacy, Dpt. of Psychology, Dpt. of Research and Innovation [Oslo University Hospital] and Dpt. of Special Needs Education).


Published Aug. 2, 2022 3:29 PM - Last modified Oct. 8, 2022 8:48 PM