In face of increasing antibiotic resistance, tools assisting with the prescription of specific antibiotics with few side effects are of great importance. The KINBIOTICS project focuses on particularly serious bacterial infections such as sepsis.
KINBIOTICS will generate a comprehensive dataset for training AI-based algorithms to predict antibiotic efficacy and train advanced AI methods for predicting effective therapies and potential side effects for individual patients* on this dataset. The dataset is composed of thousands of historical cases from the three participating hospitals.
In addition, an open pathogen- and resistance-observatory is to be developed, which can also be used by outpatient physicians beyond the boundaries of the clinics to support the prescription of effective antibiotics.
The KINBIOTICS project aims to provide AI-based decision support to prescribe antibiotics within hours of detecting an infection. Since knowledge of the specific pathogen is essential for identifying a specific antibiotic, a new rapid test for sequencing the antibiotic genome will be developed and prototyped in clinics.
The project is a model project and will provide a blueprint for the clinical integration of such decision support systems. The developed software will be made available as open source and in the form of virtual containers for easy reuse.