Data-driven precision medicine for improved antibiotic stewardship and infection control

Objective: To enhance infection control and antibiotic stewardship through data-driven precision medicine, utilizing advanced data science and machine learning for individualized risk profiling and prevention of healthcare-associated infections (HAI).

The aim is to develop automated algorithms for HAI surveillance and predictive models to improve antibiotic use.Methods: The project employs state-of-the-art machine learning, natural language processing, multimodal methods and large language models to analyze electronic health record data where we integrate structured and unstructured data for improved decision-making.

We use a large corpus of electronic health record data from >19 million admissions at Karolinska University Hospital and validate results on patient data from Region Västerbotten.Scientific novelty: The project leverages a unique clinical data warehouse and IT software deployment, enabling the development and validation of decision support systems from bench-to-bedside.

It stands at the forefront of automated prediction of adverse events, contributing to trustworthy machine learning with explainable and uncertainty-quantified predictions.Significance: This research represents a paradigm shift in patient safety, potentially reducing unnecessary antibiotic usage and combating antimicrobial resistance and HAI.

The implementation of automated surveillance systems is expected to lead to more proactive and efficient healthcare management.