The 4th annual World Medical Innovation Forum convened experts from the fields of medicine, industry, venture capital, and technology to discuss the potential of artificial intelligence (AI) to revolutionize health care.
The sold-out event had 1,700 registrants who attended a series of panel discussions, fireside chats, and AI focused workshops.
The Forum also featured a series of presentations on new concepts for AI technology from early career faculty.
The Forum concluded with the presentation of the Disruptive Dozen—12 new technologies that are poised to have a significant impact on health care in the next few years.
The Disruptive Dozen included tools designed to reduce the amount of time physicians spend on the computer during patient visits, a machine-learning algorithm to reduce the rates of in-hospital infections, and tools for mining the electronic health records to predict disease risk.
Discussions over the three-day event touched on a variety of topics and perspectives, but four common themes emerged.
1. Data Is Critical
AI solutions in health care will rely on collecting good data, processing it efficiently, actively curating it and finding smart ways to use it to identify gaps in care and opportunities for improvement.
“Obviously, we are all aware that there is an explosion of data in every sub-segment of our industry—from discovery research, development, health care management, outcomes, real-world evidence, and so on,” said Jean-François Formela, MD, a partner at Atlas Venture. “There is also an explosion of computing power and the sophistication of the tools are growing rapidly, so there is great opportunity.”
“The thing that seems so inescapable to me is just that we are going to be collecting so much more data,” said Noga Leviner, CEO of PicnicHealth. “It’s going to become obvious that machine intelligence is just better at sifting through that and pulling out more salient points than a human could.”
2. Progress Will Come in Stages
While the potential for AI to transform health care is significant, forum participants cautioned that progress is likely to come in a series of small steps rather than sweeping changes.
“A lot of people talk about medical imaging because it is such a beautiful opportunity for AI to improve patient care in hospitals and health systems in general,” said Fabien Beckers, CEO of Arterys. “The problem we have today is that the structure that was built for medical imaging…is not well-suited for AI and big data. So, we think this is kind of applying iPhone apps on an old flip phone.”
“To get to that point of complete ubiquity is going to take a long time in health care,” said Keith Dreyer, DO, PhD, Chief Data Science Officer for Partners HealthCare and Vice Chairman of Radiology at Massachusetts General Hospital. “If you think that end goal is that ubiquitous state, we’re nowhere near that. But if you think about solutions that work to solve a particular problem, we’re clearly there every day.”
3. Solutions Have to Work for Patients and Providers Alike
Forum participants also agreed that AI technologies should be designed to fit into existing clinical workflows. The best technologies won’t work if they can’t be integrated into the everyday delivery of health care services.
“Part of the challenge—and this is really tough—is that you have to implement the technology in a way that is meaningful and acceptable,” said Anne Klibanski, MD, Chief Academic Officer and Forum Co-Chair.
“In the innovation world, you have two customers: You have the patient, whose life you have to make better, and you have the clinician,” said Atul Gawande, MD, best-selling author and Executive Director of the Ariadne Labs at Brigham and Women’s Hospital. “These tools have to be able to fit into the work of the clinician at scale and fit into the patient’s feeling and recognition that this is a tool [that will benefit them].”
4. There’s a Lot to Be Optimistic About
While there is much work to be done before AI technologies realize their true potential, there are huge opportunities to improve the entire spectrum of health care delivery, participants said. This includes drug research and development, medical devices, patient communications, diagnostic tools, and much more.
“It is up to us how we use it,” said Anthony Rosenzweig, MD, MGH, Chief of Cardiology. “It’s a tool, it can be enormously powerful, and it also potentially has pitfalls that we have to be aware of.”
“Ultimately, it is going to depend on health care providers, engineers, working together as teams to figure out how to use it best to improve care on a population and individual basis.”