Smart application of AI keeps people working in care
What will the hospital of the future look like?
Only two percent of all AI models developed reach production, and even fewer are eventually used in the clinic. The rest get bogged down in the concept stage. Even though the smart use of data is essential to safeguard care in the future. Michel van Genderen is a specialist in internal medicine and intensive care. He wants to harness data in such a way that it can be used to improve patient care, in the ICU and beyond. From bytes to bedside.
The power of using data in smart ways was demonstrated during the corona crisis, when every nurse and care professional was short of hands and time. The rapid upscaling of ICU capacity, which more than doubled, brought challenges. In order to learn quickly and maintain an overview, it was essential to have a real-time insight into the data, particularly because of our lack, in the early stages, of a clear understanding of the disease. “Working with the Researchsuite, an IT program at Erasmus MC, we were able to link different data sources and establish a real-time overview of all patients. As a result, we had a clear picture of all patients in intensive care at every moment of the day. This understanding of the data improved the quality of care because it meant that we were one of the first hospitals to know that the ventilator problems we had with COVID patients were partly due to blood clots in the lungs.
“After the COVID crisis, we pressed on. Our aim is to use advanced analytics (aka artificial intelligence) at the patient’s bedside: from bytes to bedside, in other words. To bring all the stakeholders together, we set up the Datahub. This is a physical space at the Erasmus MC where we want to doctors, nurses, IT staff, data engineers and data scientists to meet in order to optimise co-creation. In this way, we can learn from each other and have a clearer idea of what is needed in the clinic/in practice, and what is technologically possible. And particularly to see how this technology can be used to improve care and make it sustainable for the future,” says Van Genderen, who is one of the initiators of the Datahub. “In the meantime, we have access to more than 1 million data points for ICU patients from the past five years and, with our analytics partner SAS, TU Delft and the Researchsuite, we are looking for trends with potential predictive value in order to treat patients better in the future,” says Van Genderen.
We have access to more than 1 million data points for IC patients from the past five years and, with our analytics partner SAS, TU Delft and the Researchsuite, we are looking for trends with potential predictive value in order to treat patients better in the future.
Shaping the ICU of the future, ICU 2.0, is the largest Datahub project. “With the ICUs at the Sophia Children’s Hospital, we want to design and implement data – and therefore artificial intelligence models – in such a way that they can be implemented in the care process. Most of the IC nurses are enthusiastic about working on data-driven lines,” Van Genderen says. “For instance, they want a traffic light with red, amber and green on a live dashboard that shows whether a patient is stable or needs immediate care. But they also want to use that dashboard to report on the workload so that others can help. We are working on the remote monitoring of patients, for example on nursing wards or at home, and better monitoring for patients after they have been discharged from intensive care.” The Datahub is not just for IC: ultimately, we want to roll out the knowledge we acquire and the AI-related decision support we develop to the entire hospital. So, we now working with the Adult Intensive Care Units, the Intensive Care Units at the Sophia Hospital and the Surgery department, but also with the Cardiology and Thoracic Surgery. With TU Delft, we therefore want to ensure that conceptual AI knowledge – but also knowledge relating to ethical and governance issues – is implemented in the clinic, while guaranteeing safety. Here, we are collaborating with the Faculty of Electrical Engineering, Mathematics and Computer Science and with the Faculty of Engineering, Governance and Management. “This is just the beginning as far as we are concerned,” concludes Van Genderen.
For more info on Datahub please visit https://datahub.health/