2349. Integrated Machine Learning-based patient clustering and prediction as a tool to improve decision-making in ICU resource utilization
Invited abstract in session WE-8: Risk Management in Healthcare, stream Health Care Management.
Wednesday, 16:30-18:00Room: H8
Authors (first author is the speaker)
| 1. | Christina Bartenschlager
|
| Nürnberg School of Health, Technical University of Applied Sciences Nürnberg |
Abstract
Maintaining high-quality patient care while managing scarce resources is a constant challenge for the management of intensive care units. Thanks to its data-rich nature, on the one hand, critical care is in a prime position to receive innovative decision support by means of Machine Learning. On the other hand, in literature either traditional supervised or unsupervised Machine Learning techniques are trained based on the American MIMIC intensive care unit data set. In this work, we research the application of an integrated supervised and unsupervised Machine Learning approach, called Cluster-then-predict, to a MIMIC-equivalent real-world intensive care unit dataset of a German University Hospital. The ultimate goal is to improve decision-making in intensive care unit resource utilization based on integrated clustering and prediction of a common denominator for resource use, the length of stay of a patient.
Keywords
- Machine Learning
- Health Care
- Decision Support Systems
Status: accepted
Back to the list of papers