Hi there!
I am a second-year PhD candidate in artificial intelligence (AI) at the Intensive Care department, Erasmus Medical Center, Rotterdam, the Netherlands. Our lab, the Datahub, works towards data-driven decision support in the intensive care unit. Supervised by Diederik Gommers, Jasper van Bommel and Michel van Genderen, our studies vary from live-dashboarding for health quality monitoring to prediction modeling and ‘actionable AI’ through Causal Inference. We work together with the Pattern Recognition and Bioinformatics laboratory (Computer Science department Delft University of Technology), where I am supervised by Jesse Krijthe and Marcel Reinders.
I finished my BSc (Clincal Technology) and MSc (Biomedical Engineering) at Delft University of Technology.
New preprint on medRxiv!
Recently, we have been systematically reviewing Causal Inference research in the ICU, discuss quality of these studies and give recommendations for future work towards ‘actionable’ AI in the ICU. Check out the preprint here!
New paper out! Dynamic model updating for Early Warning in COVID-19 patients.
We present an early warning model specifically for COVID-19 patients, showing a better predictive performance compared to traditional Early Warning Scores, ie the National early warning score aka NEWS and the Modified early warning score aka MEWS.
Dynamic model updating: We performed a temporal validation by splitting all included patients into groups admitted before and after August 1, 2020. We simulated the situation in which the model would have been developed after the first and implemented during the second COVID-19 ‘wave’ in the Netherlands. We experimented with different model updating strategies, combining model refitting and hospital specific re-calibration (figure).
Read all about it in our paper published in ICMx!
Dynamic mortality prediction in the ICU
Traditional risk scores like APACHE IV provide mortality risk stratification for ICU patients upon admission. Our new study presents a model for dynamic risk stratification throughout the whole ICU stay, validated in the Dutch Data Warehouse dataset, with more than 3000 COVID-19 patients admitted to 25 different ICUs in the Netherlands.
Check out our paper in the special edition of Intelligence-Based Critical Care and Anesthesia!
