What if the future of healthcare rested on a three-way alliance? Speaking at Nexus Luxembourg, GE Healthcare's global head of AI advocacy, Jan Beger, sketched out the contours of a new medical paradigm: a modern trinity made up of humans, care... and machines.
The American medical industry giant develops solutions and innovations that benefit more than a billion patients every year. Worldwide, the company employs more than 51,000 people and has installed more than 4m devices. It also has 58 AI-enabled medical devices. And to maintain its position as the world leader in precision care, the company innovates to deliver patient-centric technologies and improve hospital efficiency, investing around $1bn each year and employing 9,700 specialist R&D engineers and researchers.
But the leader makes a somewhat alarming observation. Every day, healthcare professionals are confronted with an exponential explosion of data. However, progress is being made: today, it takes just 72 days to double the amount of medical data worldwide, compared with three and a half years just a few years ago. At the same time, the amount of data is exploding. “We are therefore producing an enormous amount of data in this sector. Every 26 seconds, a new medical article is published. If you’re a dermatologist, you should be reading 3,900 scientific publications a year to stay up to date. And if that already sounds like a lot, look at this: if you’re a general practitioner, you’d have to read 21 hours a day to keep up to date with primary care knowledge,” Beger points out. That’s mission impossible.
Faced with this “cognitive overload,” doctors are less available, or even more solicited by screens than by their patients. Burnout, a global shortage of carers that’s estimated to reach 10m by 2030, exploding waiting times: healthcare seems to be running out of steam. But AI could well offer a way out.
Only 3% of data is used
Another striking finding is that, despite a growing volume of data, “only 3% of this data is actually used to produce analyses or improve care. Ninety-seven percent is unused, and therefore does nothing to improve patient health. This volume of data is both a challenge and a burden, but also a huge opportunity, especially with advances in artificial intelligence (AI),” Beger says.
Three million deaths a year are due to healthcare errors or accidents, he adds. And as if all the elements detailed by Beger were not enough to demonstrate the value of implementing AI in healthcare, the ageing population, the rise in chronic diseases and disparities in access to care are further arguments.
“The gap between supply and demand is widening. We need to integrate useful technologies, rethink the way healthcare works with the help of AI,” he notes, pointing to AI use cases where a positive impact has already been seen: transcription, text summarisation via language models, medical imaging analysis... “The opportunities are immense.”
Jan Beger provided several figures revealing the interest of AI in healthcare. Photo: Paperjam
For Beger, artificial intelligence will not replace carers, but will offer them logistical and operational support. “Even if it’s not real-life medicine, AI is making rapid progress. It’s even capable of logical reasoning and is now tackling real cases, such as obstetrics and gynaecology. Another example: in intensive care, AI has reached a very useful level. In some cases, it is more effective on its own than in conjunction with a human expert. For example, we have a mobile X-ray machine that is used in hospitals all over the world, including emergency and intensive care units. We have integrated an AI capable of detecting a pneumothorax. Radiologists already know how to detect pneumothorax, but they are not always available. The AI can then immediately alert the staff,” he explains.
It would also make it possible to delegate repetitive tasks, like instantaneously analyse medical images. In other words, this would save precious time for what really counts in care: the human link.
The need for a care ethic
“Forty-six percent of patients are dissatisfied with communication with their doctor. AI can help to structure information and assist in decision-making. But humans are essential for empathy. And yet, some recent studies show that chatbots are sometimes perceived as being more empathetic than doctors.”
But be careful all the same, warns Beger. Technology must not be imposed from above, but must rather be part of an ethic of care. If AI can significantly improve patient outcomes, could we reasonably do without it? In any case, we need to “move forward more cautiously than elsewhere, because this concerns human lives.”
We need to redefine our jobs in healthcare. We need to do it now, before technology dictates our roles for us.
But there are still some obstacles, and because of a lack of training or awareness, the medical world is sometimes reticent. A kind of basic AI literacy is needed to involve everyone in the healthcare system, says Beger, including patients. “One of the major obstacles remains the lack of confidence in AI among healthcare professionals--especially in Europe and the United States. There is a general lack of education about AI. Many feel ill-prepared. Every healthcare professional, every nurse, every administrator should acquire a basic level of AI culture. Because, properly supported by AI, carers will be able to focus on what’s essential: patients.”
In this vision, humans would remain central, embodying those values that are dear to medicine: listening, moral judgement and empathy. The machine, on the other hand, would be the inexhaustible memory, the tireless analyst. Care would then become the point of balance between the two. “AI should take over repetitive and administrative tasks. But that presupposes that we redefine our jobs in healthcare. We need to do it now, before technology dictates our roles for us. It’s now that we need to rethink our functions in healthcare systems.”
This article was originally published in French.
