The Center for Data Innovation spoke with Dr. Mathias Goyen, chief medical officer EMEA at GE Healthcare, a branch of General Electric which strives to create an ecosystem for precision healthcare. Dr. Goyen discussed the opportunities and challenges of using AI in healthcare and the changes digital transformation requires the industry to make. He also spoke about how his team had ramped up its efforts to support healthcare delivery during the pandemic.
Eline Chivot: You joined GE Healthcare almost nine years ago. What is your mission in the organization?
Mathias Goyen: I am currently the chief medical officer EMEA for GE Healthcare, a 16 billion dollar division of General Electric. I am responsible for leading GE Healthcare’s medical, clinical, and evidence generation strategies for product modalities in Europe, the Middle East and Africa. Together with my team I provide leadership in healthcare economics and outcomes research to develop customer value propositions for new and existing products. Being a radiologist myself, I am engaging in many conversations with our customers ranging from precision health to artificial intelligence and the future of radiology.
Chivot: Which challenges to healthcare delivery could AI address?
Goyen: When we talk about AI we talk about big data. The amount of data that is being developed in healthcare is mind-blowing. In 2010 it took three and a half years for medical data to double. Only 10 years later, in 2020, so this year, it will take 0.2 years—just 73 days. This is from now (today, November 2) till mid-January.
There are close to 6,000 journals being published putting out 900,000 articles a year. A radiologist looks at an average of 50,000 images in a 12-hour shift, 15 years ago it was 500.
One promising big-data approach to disease prevention and detection is precision health, whereby we don’t only look at traditional radiology data coming from magnetic resonance imaging (MRI), computerized tomography (CT), ultrasound, etc. (so called in-vivo data). We also look at in-vitro data or -omics-data, which is data coming from lab, pathology, electronic medical records (EMRs), but also from wearables.
Why do we do this? Because we would like to do the right thing, at the right time, for every patient. This is what personalized medicine or precision health, enabled by data, is about.
This data surge can be overwhelming for clinicians; on the other hand, they might be able to see patterns by looking at all these data that might allow them to predict a course of disease. AI has the potential to process all this data to enable precision health and augment the work of the clinician to see more and diagnose disease more accurately and faster.
Chivot: When talking about AI in healthcare, we often hear how the technology is used to interpret medical images, accelerate the detection of diseases, develop new drugs, or enhance hospital management. How is that deployed throughout the different levels of our healthcare systems? Can you also explain what are the challenges of applying AI in healthcare?
Goyen: Let’s maybe go one step back and ask ourselves: Where are AI and advanced analytics helping to deliver better patient outcomes?
I see three levels in the health system.
The first level is the individual level. That means we are embedding AI-capabilities right into our scanners, our ultrasound-devices, our X-ray equipment, or into our PACS-systems.
The second level where we apply AI is the departmental level. We can use deep learning capabilities to streamline workflows in radiology departments or private imaging centers.
And finally, there is the hospital or network level, where AI and predictive analytics can be used to better manage patient safety, patient experience, and patient volume in an entire hospital or even a hospital system.
To make this more tangible let me give you just two quick examples.
On our mobile X-ray, we’ve incorporated the first-ever FDA-cleared AI functionality deployed on an imaging device to help radiologists prioritize scans for review – so if there is a potential abnormality in a chest X-ray, for example, the radiologist is alerted and can review this first.
It is not a question about whether the radiologist is better than the AI or the AI outperforms the radiologist. If there is a patient with a pneumothorax in the middle of the night when there are only a few radiologists on call, I think having a smart computer that can flag that a patient needs acute care could be the difference of life and death.
The second example I would like to share is our “command center” approach. Clinical command centers are using AI and advanced analytics to coordinate patient care and resources across an entire health system. They streamline care delivery, they can find you an empty bed, they can direct resources to areas of high need, and give healthcare providers the critical information they need to best care for their patients. It is a beautiful marriage I’d say between artificial and human intelligence coming together 24 hours a day, seven days a week, to detect risk and coordinate complex care. We have over 10 command centers up and running in the United States, and we have a first command center live up in Bradford in the United Kingdom.
To your question about the challenges: Interestingly, hospital CEOs are reporting that they have progressed more in the last six months in their thinking about mobile and virtual health than they had expected to for the next couple of years.
Patients’ attitudes are changing as well. Patients don’t necessarily want to go into a provider for a physical visit if a virtual visit could work just as well. I think COVID-19 has in many ways made clinicians more open to embracing new technologies that can help them work better and smarter.
Now, this is important: It is strategy, not technology that drives digital transformation. We are talking about a change management process, a real paradigm shift that you need the entire organization to buy into that includes the clinicians, the IT department, and technical staff. People must be really committed; you need ambassadors or change agents within the organization who are not only ready to get used to the new environment but actively promote it internally.
From a clinician’s and radiologist’s perspective, all these efforts are only accepted if it doesn’t get more complicated. The irony is that while this technology is designed to help productivity, it actually can add more work and complexity.
The best AI application is the one that is invisible, that is seamlessly integrated into the workflow unfolding its magic in the background.
One example is our Edison Open AI Orchestrator. It is a workflow management system that simplifies the selection, deployment, and usage of AI. You could say we are bringing the power of AI into the radiologist workflow. Radiologists don’t even know that AI is working. They just realize that their worklist is prioritized and that their tasks are completed more efficiently.
As my colleague and friend Karley Yoder, Vice President and General Manager AI at GE Healthcare, likes to say: “It’s like an invisible friend helping you run faster.” Patients are the winner in this game as AI will lead to more precise diagnosis and a faster time to treatment.
And finally, you need of course some budget and a cooperation with a vendor to make digital transformation happen.
Chivot: How has COVID-19 impacted the healthcare industry in Europe? How have you and your team been using AI during the pandemic, for instance to help hospitals cope?
Goyen: The entire medtech and healthcare industry has been heavily impacted as healthcare systems were under stress. We probably didn’t see the kinds of overflow situations that we saw in China, but many elective and more urgent cases were postponed.
What is striking by the way is that many of the emergencies have disappeared, too. Heart attacks and strokes of course continue to happen, but doctors don’t see them. There are scientific reports indicating that there is an up to 60 percent reduction for those emergencies. The most likely explanation is that people prefer to stay home and suffer rather than to come to the hospital and get infected with coronavirus.
The U.S. National Cancer Institute “warned that mammogram and colonoscopy screening delays…could lead to 10,000 additional deaths over the next 10 years.”
COVID-19 makes people fear face-to-face medical care. As a result, many people with urgent health problems remain at home rather than call for help. (And when they do finally seek medical attention, it is often only after their condition has worsened.)
For my team and I, during the pandemic, providing a rapid response has been and continues to be key. For example: We not only think outside the box, but also inside the box.
CT in a box is a response for providing faster access to CT in situations requiring increased CT scans and a temporary set up to manage patients in a pandemic situation.
We have ramped up our ventilator, monitoring, and ultrasound production.
Another important role we play is in installing and maintaining essential medical equipment—my healthcare heroes are our field engineers, they have been key workers throughout the crisis—putting customers first, improving lives in moments that really matter.
Finally, another aspect that comes to my mind is: Managing many COVID-19 patients in a traditional intensive care unit (ICU). This would be nearly impossible, without technology like a virtual ICU. We are providing digital solutions to increase capacity in the ICU and to help hospitals work better by setting up command centers to manage patient experience and patient flow in hospitals.
Chivot: To what extent has the pandemic accelerated the application of AI in healthcare? How is it reshaping medical-tech priorities and research?
Goyen: For me the COVID-19 pandemic is a clear wake-up call that the entire healthcare industry needs to build and invest in a modern digitalized healthcare infrastructure.
Moving forward, I believe that digital solutions powered by AI will become even more important because they are revealing their great benefits: What are these? Greater productivity, faster availability, less administrative work, and the spatial separation of physician and patient. We are undergoing a paradigm shift from “Why digital?” to “Why still analog?”
Our customers ask us how we can make their machines more productive and how they can interface with patients using innovative technology. At one level, it’s clear that the technology must be simple—patients and doctors need technology to be intuitive.
At the same time, building digital health systems requires a complete modernization of the current infrastructure to enable virtual hospitals or smart hospitals, to enable greater access to care, and ultimately to lower the cost of care delivery.
COVID-19 has shown how we can quickly come together as an industry together with our customers to accelerate innovation, making digital solutions and AI not something to fear, but rather a helping tool to enhance the work of healthcare professionals.