With almost two decades of experience at the Erasmus MC University and a research background in anesthesiology, Professor Dr. Diederik Gommers has seen healthcare from many corners. He is the department head of operation rooms, day treatment and the Central Sterilization Department (Dutch: CSA). Diederik is also a professor of Intensive Care Medicine, chairman of the Department of Adult Intensive Care, chairman for the topic of Emergency, Perioperative & Intensive (Dutch: SPI) and theme lead for Human-Centered Technology & AI for Health. He was also chairman of the Dutch Society of Intensive Care.
The Dutch healthcare system faces shortages on many fronts, including personnel. Meanwhile, the demand for care is rising and will continue to rise. Intensive Care Units are the frontlines that face this challenge every day, 24/7. IG&H speaks with Professor Dr. Diederik Gommers about his journey with data and AI, and what it could mean for healthcare.
As an Intensive Care doctor, you are naturally very involved in monitoring patient data. How did your interest in data and AI arise?
I come from an entrepreneurial family. My father was a small business owner. I used to run a small pharmaceutical company myself. I view the Intensive Care Unit (ICU) as a shop within a very large hospital. As an ICU, we are of course always busy and everything is difficult; I discovered that the ICU staff was unfriendly to other departments that wanted to transfer a patient to the ICU. To help solve that, we introduced a drastic measure: No more saying no.
Now, if a hospital calls and asks if we have a place for a patient who needs open-heart surgery, they ask which other patient they can take over from us. We started shifting towards collaboration within our region.
How do you make that shift to collaboration?
I have always had a passion for logistics and constantly think about how we can optimize these things. I use data to find that out and to make discussions objective.
What do you still want to achieve and what do you need for that change?
After the Covid period, I asked myself what I was going to do. A shortage of nurses had been going wrong in intensive care for a long time, even before Covid. We have been providing care in the same way for 25 years. That became very clear during the pandemic. We, as doctors, and I, as chairman, were incredibly proud of what we were doing. But on the work floor, I was losing the nurses. They were getting stressed and were not in control. This was a moment of reflection, asking what I wanted to change and could change.
I decided to stay in Erasmus MC and put the learnings I experienced during Covid into practice. Data and AI were a major part of that experience. I realized I should stay in my role but needed to organize the care process to help the care workers, so their job satisfaction remains intact. It quickly became clear that we needed to work with data.
How does data help?
We have 130,000 data points per day. If you talk to an ICU nurse, they currently have little use for that data. I had no overview of my Covid patients during the pandemic. As professionals, we can never have a proper conversation based on a complete overview of past patient data. After Covid, we started the DataHub within our department. This was a necessary step because there is a lot of data, but it must be brought together.
What issues do you run into now?
Two major issues. First, you quickly notice that everyone’s interests are different. As a doctor, I am in it from the patient’s perspective and my daily work. An IT professional is in their world. The data scientist sees a number. The administrator thinks, "nice, but where do I get the money from?" You have to bring those worlds together.
Second, we bypass the quality of the data too quickly. Data quality is incredibly important. Large Language Models are not even that difficult to make but the data you train them on is crucial. We need to arrange that properly together.
What is your ambition with data and AI?
I want to have the capacity to meet the future healthcare demand. The number of patients will increase and the available healthcare workers will decrease. I will have to make my care process more efficient. My strategic goal is innovative care with the help of AI. The minister is also calling for this.
Where do you think that will lead?
Here, every patient with open-heart surgery goes to a monitored bed. Sometimes very briefly, sometimes a bit longer. With that data, you can make much better decisions based on the outcome of care. You have to provide customized care for patients, appropriate to their social circumstances. Where does someone come from? What do they do? What can they do? So, really empathic care. Essentially, I want to buy time. Time so that the care worker can sit by the bed and have contact with the patient again. Thus more empathic care with less time lost to administration.
And what else?
Improve your logistics so that the operating rooms, ICU and clinic are better aligned. A hospital is a whole chain with cross-connections, but the organization is set up in silos. I want less administration; simply talk and then use speech-to-text. The literature is almost impossible to keep up with. In med school, we used to think that the professor was a walking encyclopedia who knew everything. But those times are over.
What is your expectation of this initiative?
We have many smart people here; it’s unbelievable what they do. I always thought building an AI model was difficult. But no, asking a good question is the challenge! We are going to transform once more, only now without a virus. The demand for care will be high, with fewer people. In five years, I want to have a prediction model for the ICU so that we can give the right treatment to the right patient but also refuse a patient because the treatment does not add value.