During the three days on Artificial Intelligence applied to Intensive Care event “Milan Critical Care Datathon and ESICM’s Big Data Talk“, – organized by Humanitas, in collaboration with Politecnico di Milano, MIT and ESICM (European Society of Intensive Medicine) from 1 to 3 February – there was also a challenge in which participants (doctors, data experts, engineers and statisticians) collaborated in teams to devise new ways to respond to real clinical questions using large data sets of electronic health records.

The best projects were selected and will present their research work at the next annual ESICM event in Berlin.



Winning projects

“The hope and purpose of the challenge and of the multidisciplinary team work among the different figures who participated in the conference is that the projects considered best and most promising do not end on this occasion, but are carried out by the teams in the future,” clarified the organizers.

Among the ideas for the application of AI to the ICU selected that will be illustrated in Berlin:

  • The prediction within 24 hours of arrival in the intensive care unit of the need for hemotransfusion in patients with gastrointestinal bleeding;
  • The prediction of the clinical usefulness of a new arterial sample for haemogasanalysis at a distance of 2 hours from the last sample taken;
  • The analysis of the outcome after 72 hours of post-resuscitation patients for septic shock according to the management of the water balance after the first 24 hours.

The three projects were chosen by the scientific committee of the event on the basis of different criteria – explained Dr. Stefano Falini, doctor in the Department of Anesthesia and Intensive Care and member of the scientific committee of the event -, including the clinical relevance of the demand, the mode of investigation, the ability to detect critical issues in its resolution and in the results obtained, and not least in the presentation itself.

“At Humanitas we have been accustomed for some time to measuring ourselves with the data, and there is so much attention to the application of Artificial Intelligence in Intensive Care: certainly we will use these new technologies more and more in our hospitals”, explained Professor Maurizio Cecconi, Head of Anesthesia and Intensive Care Unit at Humanitas and lecturer at Humanitas University. The hope, in addition to improving the treatment and diagnosis of patients, “is that technologies and computers can make us better interpret data that our brain struggles to process and will free useful time we can devote to the doctor-patient relationship”.