- Programme: BMBF, AI-based assistance systems for in-process health applications
- Funding amount: around EUR 1 million for Bavaria
- Funding period: 08/2022–08/2025
- Coordinator: Prof. Dr Jens Brunner, University of Augsburg/Technical University of Denmark
KISIK project: AI-based forecasting and optimisation methods in assistance systems for efficient control of intensive care capacity in German hospitals
Along with surgical departments, intensive care units (ICU) are the most important and expensive resource in a hospital. From an economic point of view, the ICU should ideally always be at full capacity. At the same time, the capacity of the unit is limited. The KISIK project aims to develop software to support rational decisions about optimal ICU occupancy.
Capacity management in intensive care units in Germany: Challenges and opportunities
An intensive care unit in Germany costs around 1,500 euros per day and patient. The capacity of these units is limited due to the high demand for equipment and staff. In principle, both planned (elective) and unplanned (emergency) patients are admitted to the hospital and, depending on the severity of their condition, require intensive care. The capacity of an ICU therefore depends on a number of random factors, such as the arrival of patients or the length of stay of patients in the unit. From an economic point of view, utilization of the ICU at full capacity makes sense. Today, capacity planning is mainly based on the experience of the doctors in charge. Data-driven forecasting and optimisation methods are not yet widely used although mismanagement of ICU occupancy can lead not only to high costs, but also life-threatening situations.
Development of AI-based algorithms to support decision-making in the management of intensive care capacities
The aim of the KISIK project is to develop AI-based decision support algorithms for managing intensive care capacity. The algorithms are designed to automatically classify new patients and categorise existing patients, e.g. by predicting length of stay. Based on the occupancy of the unit and the type of patient reported, the system makes a decision recommendation to the doctor. Such a system is currently not available for those involved in ICU capacity planning. Yet there is a great need to optimise the use of the ICU as well as reduce the number of operations postponed at short notice or premature transfers.
Three partners from Bavaria are working together to realize this project: the University of Augsburg (Prof. Dr Jens O. Brunner), the University Hospital Augsburg (Prof. Dr Axel R. Heller, Prof. Dr. Christina Bartenschlager) and XITASO GmbH (Dr. Thomas Geislinger).
"The BayFOR team supported us intensively during the application process. The professional and well-founded feedback significantly improved the structure and precision of the application. The efficient and fast processing of requests was very impressive and the responses were always very friendly. Great cooperation!”
Prof. Dr Jens Brunner, University of Augsburg/Technical University of Denmark
Prof. Dr. Jens O. Brunner
University of Augsburg/Technical University of Denmark
Eichleitnerstraße 30, 86159 Augsburg
Phone: +49 821 598 6440