Big Data: early warning system for the ICU

Patient monitoring systems in the intensive care unit sound up to 700 alarms on average per patient per day, which boils down to one alarm every two minutes. An excessive number of them are false alarms. This generates vast amounts of data, which can make it difficult for doctors and nurses to identify the most critical alarms to manage. It also has a negative effect on the treatment of intensive care patients. The ICU Cockpit project aims to address this problem.

In this MEDICA-tradefair.com interview, Professor Emanuela Keller talks about how big data can be used more effectively, explains how digitalization can improve conditions for humanity, and reveals what might be the hallmarks of an intensive care unit of the future.

Professor Keller, what is the goal of the ICU cockpit project?

Prof. Emanuela Keller: Our goal is to use big data to identify and predict critical complications in patients and to avoid false alarms.

In intensive care medicine and especially in neurocritical care there are a multitude of medical devices that measure biosignals. This generates a vast amount of data that is incredibly varied, making it difficult to integrate and analyze it effectively. That means we have several gigabytes of data for every patient that are not shown in his/her electronic medical record. This data is challenging to integrate, and we simply don't use it. We now plan to use artificial intelligence to analyze these diverse datasets to facilitate the early detection of critical events in real-time. Our research group develops algorithms for this and aims to then present these events in a user-friendly way in an IT system.

What are your project successes so far?

Keller: We have used video recordings and electroencephalograms and have already developed our first algorithms. This enables the early detection of epileptic seizures, cerebrovascular disorders, and secondary brain injury damage in neurocritical care patients.

How does this improve or change the treatment of ICU patients?

Keller: We can improve patient safety by automatically identifying and eliminating artifacts – inaccurate measurements – in the biosignals. The early detection of risk constellations and critical complications also supports an earlier therapeutic intervention, which can improve patient care in the future.

I also believe that digitalization in the intensive care unit can improve conditions for humanity, because it can help eliminate some fatiguing tasks for nurses and doctors. This allows us to focus more on the individual patients, freeing up time to engage with them and their family members.

What will the intensive care unit of the future look like in your opinion?

Keller: Big data and AI will play a big role in the future of ICU since they can help improve patient safety and the quality of care. Ultimately, this reduces cost and saves money. Digital technologies such as telemedicine also make it possible to transfer patients out of the ICU into general wards much sooner. Another benefit is that critical care patients could also be monitored remotely. From my perspective, AI and telemedicine will significantly change the intensive care unit in the future.