Eric Topol On Deep Medicine, AI’s Healing Hands, And The Future Of Healthcare

Eric Topol On Deep Medicine, AI’s Healing Hands, And The Future Of Healthcare

Publication date: Feb 27, 2019

Artificial intelligence is the single most important opportunity to address all the major things wrong with healthcare today, according to Dr. Eric Topol, a world-renowned cardiologist, geneticist, digital medicine researcher, and author.


At NVIDIA’s GPU Technology Conference (GTC) on March 17-21, 2019 in Silicon Valley, Dr. Topol will present a talk on how AI and deep learning are beginning to affect medicine at three levels: clinicians, health systems and patients.

Dr. Topol, founder and director of the Scripps Translational Science Institute, answered a few of our questions on his upcoming book, Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again and discussed his perspective on the future of healthcare.

Can you share a formative experience you’ve had with AI in healthcare?

I had a very bad experience with a total knee replacement, in part because my data wasn’t incorporated into my plan of care. It’s making a case for how we can do so much better because we have the tools today, but they’re not yet implemented.

In your upcoming book, Deep Medicine, you discuss that AI can help with one of the major issues facing medicine, of it becoming -inhuman. The doctor-patient relationship-the heart of medicine-is broken.” Can you address the paradox of the perception of AI being inhuman, yet its capability to deliver personalized medicine?

Naturally, we wouldn’t expect technology, which is not a human, it’s a machine, to be able to rescue humanity, but that’s exactly what we could do. Medicine is so broken, inefficient, depersonalized, and shallow, that we have all the tools now to address every one of its deficiencies. We can make it much more efficient. We can deeply understand each individual to give precise care and prevention. Most importantly, we have the gift of time for deep empathy, which we haven’t had. We just don’t have enough time. We have keyboards that can be eliminated by natural language processing. Every single component of what’s wrong with medicine: whether it’s the cost, efficiency, lack of productivity, and mostly the human-human contact aspect, can be ameliorated.

In Deep Medicine, you discuss how AI has the potential to transform everything doctors do, from notetaking and medical scans to diagnosis and treatment. What are some promises and challenges of integrating AI in hospital data centers for analysis, especially with the recent discussions on data privacy?

In the book, I get into how data should be owned by the individual. When it sits on servers, it’s a target for cyber thievery and hackers. So, if we get to a different model where people not only own the data but also use sensors and other data, like genomics, to flow to their digital property, it will promote optimal cybersecurity. If you talk to cybersecurity experts, the smaller units of person data storage rather than millions, make it easier to maintain security.

As a physician, what would you say to other physicians who are concerned about AI in healthcare?

Yes, well I share the concerns because a lot of it is long on promise and short on implementation and validation. So what I’m espousing is that we have a remarkable opportunity here. If we really take this seriously and we are activists for our patients we will accelerate AI implementation, get the validation needed, and bring back the past to bring the future.

What do you hope attendees will take away from your talk at GTC?

The main thing is to get an honest appraisal of where we are today with AI in healthcare, the excitement, along with the caveats where this is headed and the potential for it to be a great answer to what are the woes of healthcare today.

Last April, you were invited by UK’s National Health Service to lead a review into the training staff would need to use AI technologies and robotics. How would you rank US medical facilities in terms of their AI implementation? What are some steps US medical professionals and facilities can take now to stay ahead of the curve?

The UK is ahead of the US because not only have they engaged with the planning, but also because they are already starting implementation of these ideas into the real world. The US has no AI healthcare strategy, there hasn’t been any national planning and it shows.

So what we have in the US is a lot of tech giants, startups, and companies that are going after AI in healthcare. We have academic centers around the country that are working on various research projects. We have FDA approved algorithms for companies, but we haven’t gotten any serious, cohesive, plan for AI in healthcare. The UK has that, China has that, but we don’t have that in the US yet. The UK has a whole education and training wing of the NHS-NHS Health Education England. The US doesn’t do anything like that.

I think this is going to be a big challenge for the future that we have in our sight.

We’re still early in the new year. What do you anticipate are the top three trends for AI in healthcare in 2019?

That’s a tough one.

Eye disease and radiology are two medical areas that are getting priority with lots of research and deep learning algorithmic development.

We have a lot of radiology algorithms that are already FDA approved, along with diabetic retinopathy systems. That’s important because diabetic retinopathy is terribly underdiagnosed which can readily be improved-and reduce the toll of blindness. We need lower cost approaches that get FDA validation for this indication. We’ll likely soon see sensors for potassium blood levels in smart watches that are bloodless. People with kidney disease have liability for getting high potassium in their blood, which can be lethal, so this will be a way to detect it.

The other thing is we’ll see much more use of machine learning in hospitals for things like preventing falls, better hand washing practices, and monitoring patients in intensive care units to prevent adverse outcomes.

We already have deep learning algorithms on the wrist with Apple Watch for heart rhythm (particularly atrial fibrillation) detection. That’s the first consumer AI medical device.

Did you know from the beginning of your career that this is what you wanted to do? What do you wish you knew when you were starting out your career?

In whatever I did, I’ve always gravitated to -where is this going to go?” You see something on the horizon, you know it’s going to be something right, and you feel it in your blood. Hopefully, you’ll find that too.

Topol’s talk will be followed by a signing for his forthcoming book, Deep Medicine: How Artificial Intelligence Can Make Healthcare Human Again, which is available for pre-order on Amazon. You can follow him for the latest in AI and medicine @erictopol.

The week is packed with more than 40 healthcare sessions that span medical imaging, genomics, and computational chemistry. Attendees can join two-hour, instructor-led training sessions on AI and accelerated computing, including workshops specific to medical use cases.

Check out the full healthcare track at GTC, and register today.

Concepts Keywords
AI Total knee replacement
Amazon Important diabetic retinopathy
Apple Watch Particularly atrial fibrillation
Artificial Intelligence Toll blindness
Atrial Fibrillation Diabetic retinopathy
Blindness Physicians concerned healthcare
Blood Imaging
Cardiologist Healthcare academic country
China Radiology algorithms
Computational Chemistry Consumer device
Cybersecurity Trends healthcare
Deep Learning Diabetic retinopathy systems
Depersonalized Healthcare sessions
Diabetic Retinopathy Medicine
Digital Eric Topol
Empathy Personalized medicine
England Diabetic retinopathy
FDA Machine learning
Future Sight Radiology
Geneticist Genomics
Genomics Computational chemistry
GPU Lot radiology algorithms
Hackers Artificial intelligence
Hand Washing
Intensive Care Units
National Lead
Personalized Medicine
Silicon Valley
Total Knee Replacement


Type Source Name
drug DRUGBANK Etodolac
drug DRUGBANK Nonoxynol-9
drug DRUGBANK Silicon
drug DRUGBANK Tropicamide
disease MESH diagnosis
drug DRUGBANK Etoperidone
disease MESH Eye disease
disease DOID Eye disease
disease MESH development
disease MESH diabetic retinopathy
disease DOID diabetic retinopathy
disease MESH blindness
disease DOID blindness
drug DRUGBANK Potassium cation
disease MESH kidney disease
disease DOID kidney disease
disease MESH atrial fibrillation
disease DOID atrial fibrillation


Original Article

Leave a Comment

Your email address will not be published. Required fields are marked *