Artificial Intelligence in the ICU: The Challenge of Integrating Clinical Data
The intensive care unit generates data continuously: from devices, clinical records and each patient’s evolution. According to Otavio Ranzani, the main challenge is to integrate all of it and be able to read it without errors. He explains it in this Connecting Voices conversation, recorded during the 39th Annual Congress ESICM LIVES 2026.
The challenge in the ICU: lots of data, from many devices
Ranzani describes an UCI where data from different devices and records coexist. To make it usable, the challenge is to integrate it and read it without errors. His proposal is to do it automatically, with processing pipelines that guarantee data quality.
What he expects from artificial intelligence in the ICU
In his view, the role of AI will be to integrate the many signals in the UCI (from patients, from devices, from their evolution and trajectories) to help identify alarms that are not perceived today and to anticipate events, in line with precision medicine.
Two motivations: less workload and precision medicine
What excites him most are two lines of progress. The first is reducing workload in the UCI, which he believes AI and better data will help achieve. The second is moving towards precision medicine, when it becomes possible to integrate all these signals and predict which patients will respond better to treatments.
Connected data, the starting point
At Better Care we connect data, people and clinical decisions with BC Link®. If this approach interests you, we also recommend:
- Why Connected Clinical Data is the Foundation of Artificial Intelligence in the ICU
- How Predictive Models Are Transforming the ICU
Sobre Otavio Ranzani
Otavio Ranzani is Group Leader at the DataHealth Lab of the Sant Pau Research Institute in Barcelona. In this video he takes part from ESICM LIVES 2026, in Lisbon.
Frequently asked questions
What is the main data challenge in the ICU according to Otavio Ranzani?
IIntegrating data from different devices and records, and being able to read it without errors.
How does he propose to integrate ICU clinical data?
Automatically, with processing pipelines that guarantee data quality.
What excites him most?
Que ayude a integrar las señales de pacientes, dispositivos, evolución y trayectorias, a identificar alarmas que hoy no se perciben y a anticipar eventos, en la línea de la medicina de precisión.
¿Qué es lo que más le entusiasma?
Two things: reducing workload in the ICU, and moving towards precision medicine, when it becomes possible to integrate all these signals and predict which patients will respond better to treatments.
¿Would you link to drive the continuity of care in your hospital? ¡Request a demo!