Integrating pharmacoepidemiology and AI/ML for maternal–fetal health research

Our research

We combine advanced pharmacoepidemiological methods with machine learning to address key questions in maternal–fetal health. We work with multiple data modalities, including tabular register data, ultrasound images, text from patient records, and time series data such as cardiotocography (CTG). We apply large language models to identify adverse maternal and fetal outcomes in clinical text.

Our studies range from hypothesis-generating research to target trial emulations. We use unsupervised learning for descriptive analyses, and causal machine learning and causal inference to explore associations. We also develop predictive models and decision support systems to improve clinical decision-making.

Projects​

Funding​

Collaborations ​

• Adam Hulman, Associate Professor & Senior Data Scientist, Steno Diabetes Center Aarhus, Denmark