Improving care for COVID-19 patients on the intensive care using machine learning

ENVISION is a H2020 project, started in December 2020, aiming to improve treatment of COVID-19 patients in the ICU. In a number of European countries, COVID-19 patients hospitalized in the ICU will be monitored by means of the so-called Sandman system. Data from the monitoring system and guidelines are then combined, via artificial intelligence and other techniques, to support the treating physicians in their clinical decisions.

Dr. Thea van Asselt and prof. dr. Maarten Postma from the Global Health Unit (GHU) at the department of Health Sciences are in the lead of the work package concerned with the health economics of the Sandman system – if the system were to be implemented, what would the benefits be for the patient and the health system, and what would the associated costs and potential savings be? Mathematician and psychologist Leslie Zwerwer, PhD student at the GHU for ENVISION, aims to find out. As the project is still in the startup stages and data is yet to be collected, current activities are centered around literature study and other preparations.
The Sandman system is hypothesized to have an effect via a number of predefined ‘use cases’. One of these cases is the early prediction of sepsis, a very frequent complication in COVID-19 ICU patients. During the project, the data will reveal whether this is indeed possible. We will estimate the health gains and cost savings to be achieved when sepsis can indeed be predicted earlier. We do this by means of a mathematical model, a technique often used in health economics. 

ENVISION is led by the University Hospital in Frankfurt, and participating centres are located in Germany, Portugal, Bulgaria, Romania, Spain, Slovenia, Lithuania, Italy, and the United Kingdom. Read more on ENVISION here.

Project team: Maarten Postma, Thea van Asselt, Simon van der Pol, Leslie Zwerwer, Hanneke Vervoort.