- [Announcer] Production and distribution of City Club Forums on Ideastream Public Media are made possible by PNC and the United Black Fund of Greater Cleveland Incorporated.

(upbeat music) (bell dings) - Good afternoon and welcome to the City Club of Cleveland, where we are devoted to having conversations of consequence that help democracy thrive.

It's Friday, July 28th, and I am Robyn Minter Smyers, the immediate past president of the City Club Board of Directors, and a partner at the law firm of Thompson Hine.

I am so pleased to introduce today's speaker, Dr. Jag Singh, author and professor of medicine at the Harvard Medical School.

In the last year, conversations have swelled dramatically around artificial intelligence and what it means for how we live and work.

In his latest book, "Future Care: Sensors, Artificial Intelligence, and the Reinvention of Medicine," Dr. Singh looks into AI's effects on healthcare and what it means for doctors and patients.

Utilizing his years of experience as a medical professional, Dr. Singh's book chronicles the evolution of AI and healthcare from sensors to virtual care.

As artificial intelligence becomes a larger part of everyday life, Dr. Singh helps readers understand this emerging field and the long-term effects of medicine's digital revolution.

Dr. Singh is a professor at Harvard Medical School.

For almost two decades, he's worked at Massachusetts General Hospital, where he currently serves as a cardiac electrophysiologist.

Dr. Singh completed his internal medicine residency, cardiology and cardiac electrophysiology fellowships at Mass General.

He received his doctorate from Oxford University and a master of science in clinical investigation from MIT and Harvard.

Dr. Singh will discuss his book, and then join City Club CEO, Dan Moulthrop, for a brief conversation.

After which, we'll have our traditional City Club audience Q&A.

Next month, our forums are all free.

And on Fridays in August, we will be outdoors at the US Bank Plaza at Playhouse Square.

In September, we will welcome the community to our new home at Playhouse Square.

But now it is my distinct pleasure to welcome Dr. Jag Singh to the City Club of Cleveland for our last forum at our old home.

Dr. Singh.

(audience applauding) - Oh, my gosh, this is quite the forum, and I feel very unworthy.

I truly do.

(audience laughing) Well, it's a pleasure and a privilege.

And, Robyn, it's a deep honor to be here and speaking at this forum.

And thank you for that kind introduction.

I'm gonna start by a quote from a famous American author and futurist, Alvin Toffler, who said that change is not merely necessary to life, it is life.

And I think all of us here in this room today, it resonates with each of us because over the course of our lives, we have seen a lot of change socially, culturally, geopolitically.

The same thing as with medicine.

Medicine is, in itself, in a state of transition.

Medicine is evolving, it's changing.

And this whole digital metamorphosis is gonna transform the way we actually receive care and the way we deliver care.

And what I'm gonna talk about today is really very tangible technology.

I'm not going to be dystopian at all, and the book is not dystopian at all.

But before I do that, I'd like to just introduce myself in terms of where I'm coming from so you kind of understand what my vantage point is.

Grew up in India, did medical school in India, practiced in India for a couple of years before I got a scholarship to go to Oxford, practiced in the UK as a registrar in cardiology, and then made my way to the US and have been at Mass General Hospital for the last 25 years, and they haven't fired me as yet.

(audience laughing) And here, I trained to be an electrophysiologist.

And for those who don't know what electrophysiology is, just to kind of give you an overview, it's somebody who deals with heart rhythm disturbances, deals with sudden cardiac death, patients with heart failure.

What I do is implant pacemakers and defibrillators in patients.

I put catheters in their hearts and burn circuits.

So that's kind of where I'm coming from.

But over the course of this years and the training I've had, I've experienced some change too.

I've had the opportunity of working in a under-resourced environment in India, in a resource constrained environment in the UK, and a well-resourced environment here in the US.

Now, having said that, I think it pains me a little bit to say that our healthcare system in the US is big, but it's fat and sick.

And I say that because for a variety of reasons I would share.

Because we have millions of uninsured individuals, our costs are inexplicable and our life expectancy is nowhere where it should be compared to other developed nations.

And we presumably have about one fifth our GDP, which is about 17 to 20%, $4 trillion goes towards healthcare, and we are still indefensible when it comes to the quality of care and in our life expectancies across the board.

Now, why is this?

And I think it's because of the fact that we're probably to some extent inept, but more so I would say inefficient, and for sure, inequitable.

And bigger than that is the fact that we are very opaque.

We are very opaque in the way we reimburse medicine care.

We're very opaque in the...

There's no transparency when we talk about practice variance.

There's so much of variance within an institution, leave alone regionally, locally, nationally and internationally.

Now, beyond this, I would say that we are a living contradiction because the wellness of our institutions depends on the sickness of our patients, right?

(audience member laughing) Three of every $4 goes towards looking after patients who have chronic disease.

And that's not fixable unless we have digital medicine here to kind of help us fix that.

And I'll tell you how.

Let me start with the story of one of my patients.

This book is about 25, 30 different patient stories intertwined.

I'll just pick one.

This story's actually buried, but I think it'll kind of take you through the journey of what we're gonna talk about today about sensors, virtual care and artificial intelligence, and translating into outcomes.

So I met Paul for the first time in 2003 when I just finished training.

I was a newly minted electrophysiologist.

And Paul was 71 at that point in time in 2003, and he had four chronic diseases.

He had hypertension, diabetes, had chronic renal failure.

Actually was on hemodialysis for seven years before I met him.

Every alternate days, he had to actually go for hemodialysis.

He was referred to me because he had heart failure.

His heart function, which was normally expected to be somewhere around in the number of 60%, was down to 15%.

So one fourth of what the heart function really needs to be.

He was short of breath walking from just his bed to the bathroom.

Couldn't really survive.

He had actually been given a lifespan of one month.

He said the chances of him dying within one month was 90%, one year was a hundred percent.

So he came to me because he had heart failure.

And at that point in time, we had a new therapy, which the physicians in the room might remember or know about, something called cardiac resynchronization therapy.

This involves putting three wires in the heart.

And when you put three wires in the heart, you can coordinate the activity and actually potentially improve the heart function in these patients.

And so I did that much to, I would say, against the advice of many others, and Paul actually started getting better.

His heart function started actually improving, and from 15%, he became 45%, and nearly 60%.

Now, at this point in time, I just wanna take a break and tell you about heart failure because that fits into this whole paradigm of chronic disease.

It's one of the commonest causes of readmissions.

About 6.5 million in the US have heart failure.

And the drain on the national exchequer by 2030 is postulated to be about $70 billion.

So, huge, huge problem.

Alongside that, Paul also had a condition called atrial fibrillation, which I think many of you may have heard of.

Anyone above the age of 45 out here, you have a one in three chance of having atrial fibrillation during your lifetime.

And when atrial fibrillation occurs in conjunction with heart failure, it's something we call a double whammy, 'cause it drives the heart.

(audience applauding) Right?

It drives the heart rate really fast, increases the risk for stroke, and at the same time, makes patients actually land up into heart failure.

So what happened to Paul?

So, interestingly, he got better.

His heart function actually improved to 60%, as I said, so much so that he became eligible at the age of 75 for a renal transplant.

His anonymous neighbor down the street actually donated a kidney to him.

He's now at the age of 76.

Paul had a kidney, normal heart, he was a free bird.

He became a dancing instructor at the senior center.

(audience applauding) This is true story.

He became a dancing instructor at the senior center, met his new wife there.

(audience applauding) Got married, moved to Florida, and started an antique business out there.

(audience applauding) So what went wrong out here?

What went wrong out here was, everything that Paul had, hypertension, diabetes, end-stage renal disease, was all preventable, was all trackable, but we were not able to do that.

We can do some of that now.

Plus, the care he got was very siloed, was very transactional.

And he spent his entire youth dealing with renal failure and heart failure, right?

And now you can only imagine what the expense incurred by the healthcare system was for somebody getting hemodialysis every alternate day, and then getting repeatedly admitted with heart failure, right?

So putting that into the equation, what I really wanna say today is that medicine is not just about diagnosing a disease, is about coming forth with a diagnosis, but it's really finding a tenable life.

A tenable life around disease.

And we now have unlimited data.

We have super connectivity, we have massing processing power, which now allows us to transmit care in a way that we never could in the past.

And if there's one thing I can leave you with you today, is that the future of care will be partially virtual, not fully, partially virtual, will be sensorated, will be powered by predictive analytics and artificial intelligence, but this will become a part of sustainable workflows that we have in clinical medicine that will translate into better clinical outcomes while preserving the human touch.

So we'll come back to how this is all gonna be possible.

But let me first talk about, break this down into the four parts, and starting with sensors.

I'm gonna try to keep a watch on the clock out there.

So sensors I think is really important.

We all have these sensors implanted in us through electrophysiologists like me.

You can have devices.

These devices are pacemakers, ICDs.

These have the ability to actually measure fluid status, measure heart rate, measure respiratory rate, measure cardiac contractility, and can actually help us risk, predict and actually risk stratify patients.

And you'll be surprised that some of these implanted devices can actually predict which patients are gonna develop heart failure 30 days before they actually develop heart failure.

So they have the predictive ability already.

So this is called narrow artificial intelligence to some extent.

Now, not everybody has implantable devices, and now we have a slew of variable devices, and all these variable devices now, which could be earbuds, could be glasses, could be caps, could be fitted onto our watches that we all know about, they measure the same variables.

They measure right from heart rate, heart rate variability, autonomic tone.

They can measure your hydration status, they can measure the electrolytes off your sweat, the autonomic tone.

All of these can be now used to track individuals.

They can be used to track patients who have diabetes, heart failure, hypertension, and renal failure.

Everything that we just talked about is potentially trackable through wearable devices.

Now, alongside this, we have the mothership.

That is the smartphone.

The smartphone receives all this data, then transmits it potentially to the electronic health record.

But these smartphones also have the ability to provide us with high touch technology.

They can actually categorize what your mental state is by just looking at your usage, your activity, your sleep times, and actually can help better phenotype patients than we can in the past.

As one of my close friends said that phones don't lie, patients can.

(audience laughing) But with that comes the whole conundrum of data.

How do we kind of manage all this data?

And if I were to tell you, when we look at data, I'd use the iceberg analogy where you have stuff that is visible to us and you have stuff that is invisible to us.

The visible stuff we know within our electronic medical records, and I will refrain from getting into that, but there's some variable stuff that is now also visible to us.

But more importantly, there's this whole invisible data, which includes our genomics, our proteomics, our environmental data, our social data, our cultural data, all of which actually is more influential in terms of how we actually develop disease.

And this is where I think we need AI to help collate this information and help us to compute the data to help us understand it better.

There are several forms of AI.

There are narrow AI that I told you that within the devices that can do these predictive analytics, and then there's conventional AI that is machine learning and deep neural networks that can help us actually see things that the human eye cannot.

You can look at a watch ECG right now and diagnose if the patient is going to develop atrial fibrillation in the future, or patient's gonna develop a stroke in the future.

That AI can see.

We can't see it as humans in a watch ECG.

So that's something where now using conventional AI through machine learning and deep neural networks, we can actually predict which patients will develop heart failure or which patients can actually develop atrial fibrillation.

All of the things that actually Paul had and we could not track.

So I'm gonna get very quickly into the future models of care.

And I think that there has to be an evolution of how we understand this because sometimes folks think that we're digitized.

Our records are digitized, so that means we must be digital.

No, being digital is very different from being digitized.

Digital is a change in the culture of care.

Digital is a change in how we practice care, not being digital digitized.

Digitize is what our electronic medical records are.

So this allows us to transcend from what care is in its conventional form, what is we call as transactional.

So we see our patients at 6 months, 12 months, two years, but patients don't fall ill at those intervals, right?

Patient can fall ill at any point in time.

That's where continuous surveillance through wearables and implantables can allow us to continuously track patients and give them care when they need it.

There's a whole concept in electrophysiology that we call as exception-based care.

That means you see the patient only when they actually need to be seen, not see them at these periodic intervals unnecessarily.

And then there's this whole concept of MYOD which I elaborate in the book called manage your own disease.

The only way healthcare will become sustainable is if patients have access to some of this information and can be a part of their disease management process.

So the disease management models out here really need to change.

And for this to be possible, we really need remote monitoring on steroids.

We need to be able to have remote monitoring centers that can allow us to actually follow patients and have strategies using conventional AI, using generative AI, to prevent chronic admissions, to prevent readmissions, to manage our chronic care patients, and at the same time, alter the lifecycle of disease in many of these patients and actually track them through their lifecycle.

So along with this, obviously, there's this whole concept of digital twins, and I won't get into the details of this called twinning, where you can actually create a human replica, right?

Anatomy, physiology, molecular biology, and you can then use strategies to simulate certain environments.

You can fast forward disease or you can back up into disease and make some simulations to see what you got wrong and how you can change the disease structure in the future.

Now, the next era of care, so we all, I think many of you may be have heard about the whole concept of systemness, right?

All our hospitals, all our healthcare organizations are embroiled in this terminology of systemness where they're trying to bring all their medical facilities, their community hospitals with their specialty centers, with their ambulatory surgical care centers into one system to create some sort of efficiency flow within that system.

But the next level of care truly is going to be networkness.

It's not gonna be systemness.

Systemness may be a platform for networkness, but networkness is where there'll be an era of openness, there will be a chain in the social order of practice of medicine, where you can get care wherever you want, from whomever you want, for whatever it may be.

And this is going to happen because the expectations of our patients are gonna make this happen.

Now, having said that, I'm gonna try to close off very quickly before Dan gets to me.

So, I'm gonna get back to Paul.

So Paul is 91 years old today, 20 years after he was told that he had just one month to live, right?

And I think everything Paul had was trackable and preventable.

And I must say that even though I talk about sensor-based approaches and virtual care and powered by predictive analytics and AI, it is so important that this digital strategies actually remain complimentary to the human touch and the human bond.

And I'd like to close by a wonderful saying from Francis Peabody that the secret of the care of the patient lies in caring for the patient.

Thank you very much.

(audience applauding) - Audience, our guest today is Jag Singh.

He's just been speaking about his book, "Future Care: Sensors, Artificial Intelligence, and the Reinvention of Medicine" here at the City Club of Cleveland.

I'm Dan Moulthrop, and I'm gonna ask him a few questions before we get to the Q&A with the audience.

- [Jag] Sure.

- That's a remarkable story, Jag.

Congratulations on the book, and thank you for sort of doing it all through Paul's story.

You paint a vision of the future that sounds awesome, don't get me wrong, but I would also just, I just feel that you need to acknowledge that patients have been asking for change to the system for a long time, and they don't really, it doesn't seem to be trending in that direction, perhaps.

You said it yourself that the system is fat and lazy and other things.

What are the obstacles to the future that you want to create?

- No, absolutely.

So I think the status quo is the biggest obstacle and the obstacle for the status quo comes in multiple areas.

So one of the reasons why I wrote the book was to actually have this public discourse, is to kind of create change outside in, because inside out is a problem.

And inside out is a problem because I think as clinicians, we're very vested in the way we practice medicine, the way we've been taught medicine, and we're very well settled in how we like to practice.

And I think some of these changes can be brought in from outside.

And the administration also, I would say, the leadership within institutions and organizations are so vested in trying to just make sure that they're fiscally viable for the next quarter that oftentimes the long-term vision of how we can change the trajectory of healthcare gets lost in that.

And I'm hoping that conversations like this and educational forums like this can actually bring some of that change outside in, but also create the change inside out.

- You noted early in your presentation that this isn't about a dystopia, and a lot of the conversation around what we're calling AI, whether it's machine learning or generative AI or artificial intelligence or augmented intelligence, a lot of it is very dystopia-focused, right?

ChatGPT is going to ruin journalism, for instance, and similar models are going to ruin the creative workforce.

You really don't see it that way at all.

- No, I don't.

I know...

I will say this, I mean, you can kind of use the analogy of the internet when it came up and we thought this was going to be the end of the world and there were these cyber attacks on everything.

And I think we put our good people together and created the appropriate regulatory barriers and borders around that so things could remain safe and we could use it to advance society.

I think the same thing is gonna happen with AI.

I know there's a lot of hype.

There have been about the potential for generative AI can destroy civilization, and there's folks who are saying that it's the inorganic form of life that can actually overtake us, which is, you know, I think important to be cognizant of the fact that there could be nefarious activities linked to AI.

But I think all in all, the potential for actually changing the trajectory of clinical care is huge.

And I think we're already, we already have forms of narrow AI and conventional AI that work well.

I think it's this whole generative AI issue that we need to really get our hands around.

- How are you using AI in your practice right now?

- Sure.

So a lot of it is investigational.

The way we use it in, for example, let me give you an example of narrow AI.

So I implant devices in patients who are predisposed to sudden cardiac death, right?

What this device can do is, it's been coded to do a single task, can do it better than a human can.

It actually picks up arrhythmias or life-threatening arrhythmias from the heart, detects it, diagnoses it, and shocks the heart back into normal rhythm.

This is a form of narrow AI.

This has been around since 1985, 1990, so this is not new.

It's obviously got enhanced considerably since then.

So that's AI that already exists.

I think there are many forms of predictive analytics now that are finding their way into our EMR, and some of those are diagnosing patients off their 12-lead EKG.

So we use 12-lead ECGs from the heart to actually diagnose clinical conditions, but now there are ways to actually see into the future and say whether they have a high risk of developing atrial fibrillation in the next 30 days or next five years.

These are finding their way in as suggestions at this point in time.

But this will all eventually, I think the generative AI component is going to have a much greater influence in how we practice medicine at the back end, as well as at the front end.

- So one of the things that occurred to me, I mean, you're trying to do this from the outside in and trying to create these sort probably a lot of different pockets of disruption to ultimately change the industry and change the systems and the networks.

It struck me that one of the most complex things that people tend to think a hospital is doing when actually it's their responsibility is care coordination between this specialist and that specialist and so forth.

And literally, like, less than 1% of patients are probably capable of managing that for themselves or for a family member because it requires so much specialized knowledge.

Is that the kind of area you're thinking AI can provide a huge leap forward?

- For sure.

So let me give you the immediate priority areas.

One of them I think is, you know, one of the reasons when you go in to see the clinician and the clinician doesn't maintain eye contact, it's not that they're nervous, it's just that they're spending their time completing their note because otherwise they'll have to go home and start doing it at 9:00 PM at night.

So I think keyboard liberation is gonna be a big thing.

That- - Liberation.

- That AI.

(audience laughing) That AI and generative AI for sure because you can come and walk into a clinic room while you're actually talking to the patient, you have generative AI that is transforming this through natural language processing and converting this into a synthetic note that you can quickly edit at the end of it, but that really is something that will happen automatically.

And there's already examples of this happening right now as we speak.

So that's gonna happen for sure.

I think other than that, the mundane tasks that as clinicians we have to do such as billing, referrals, pre-authorizations, all that AI will actually already be able to do.

And some of it, pre-authorizations, for example, AI is already doing it in small pockets of the world.

But I think the bigger question you alluded to is, how is it gonna actually help patients manage their disease?

And I think there are several ways it can help patients manage their disease right from creating algorithms to actually help patients.

For example, diabetes, right?

Patients wear now a sensor that can help them detect whether they have diabetes.

Now you have algorithms that can help guide patients how they need to manage their diabetes so you don't have to keep finding a clinician to actually connect with, because we know access is a problem, those phone calls are a problem, getting connected to your physician.

And by the time you connect with your physician, your clinical state is completely changed, right?

So there are gonna be self-teaching algorithms that will come through these AI-based strategies to actually enhance care at the individual level.

- There's a danger that if care is enhanced, people get healthier, then the system gets sicker, right?

As you said before, the wellness of the system depends on the sickness of the population.

- [Jag] Touche.

So... - So you have to reorient the system.

- For sure, for sure.

I think much of that reorientation of the system depends on how do we incentivize the system.

How do we incentivize people to be better engaged?

And I think the incentivizations will be very different for patients as to physicians and for administrators.

And let me tell you what I mean by that.

I think the way you can change the system is to move from the conventional fee for service strategy where we're so focused on volume and not so much on value, is to create value-based payment strategy, which I know are a challenge.

- [Dan] I thought we were trying to do that in the last 10 years.

- But I think there will be shared saving approaches now that will necessarily come.

And if you can have some sort of capitated strategy where you have shared savings, that can then be saved between the clinicians, the hospital, and the third party vendor.

That's something we haven't talked about and I didn't have a chance to talk about it.

You know, for example, if you look at patient with heart failure, our hospital across sectionally, at any point in time at Mass General Hospital, which is a thousand bed hospital, about 120 patients carry a diagnosis of heart failure.

120 patients.

They may not be there for heart failure, but they carry a diagnosis of heart failure.

And the challenge with that is the readmission part and creating disease management models out there, which can be sublet.

This may be not go very well with administrators.

So, I'm gonna take a step back.

So we have seven heart failure doctors managing 4,000 patients.

How can you look after 4,000 patients on a day-to-day basis and provide them the best possible care?

And this is where I think third party vendors and disease management models will be involved.

So they can actually provide point of care more instantly, and then refer complicated care to the heart failure specialist.

So there's gonna be a change in the structure.

And disease management models will be AI powered, but will be run by third party vendors.

- The US healthcare system is sort of famously inefficient when compared to other systems around the world, and you alluded to that before, and it doesn't seem like a market-based system, it doesn't seem like there are market efficiencies in the US healthcare system.

Are there other systems in Japan, in the UK, elsewhere, that may be better situated to implement some of these changes?

- I think each one of them has their problems, I think.

And many of them are so dictated by their prevailing economies within their countries that it's been a challenge.

I think if any place is best poised to actually make these transitions and make this change and actually lead the way, I think it's the United States.

I think we have the ability, we just need a change in the mindset of how care needs to be provided.

And some of that is the expectations of our patients, but at the same time, is the willingness of the physicians.

And at the same time, in addition to that, visionary leadership in the organizations.

- How worried are you though that we're gonna get it wrong?

- I'm not so worried about we're gonna get it wrong, I'm worried about how long it's taking.

And I think that's the culture change that we need to really kind of put into place and hopefully forums like this will make a difference.

But that's what I'm worried about.

I think there are enough checks and balances that recognize that this is the way to go, that continuous surveillance strategies with as needed care is the future.

But I'm just worried about how long it's taking because the political will between the different organizations is very variable and all of them are interconnected, and you have to be able to make sure that they all can work well together.

- We're gonna go to questions from all of you in a second.

I would love to hear a little bit about how this is being received in the industry.

I think you have a room full of mostly healthcare consumers right now here who are like, "Well, that sounds pretty good," but I'm sure that it feels a little different inside the organizations.

- You know, I can tell you the industry that I work with is, for example, the device industry.

And I work with AI industry, as well as sensor based industry, and I do consult for many of them.

I know that the device industry is very excited about the potential for actually helping manage heart failure patients because they already have these predictive analytics.

So we've had some of these predictive analytics in the devices, but not everybody's using them the way we use them at Mass General or other places.

Not everybody's using them because there's this whole culture of, you know, it's different.

I mean, you now need to look into the data, respond to the data, who owns the data, how quickly do you do it?

This is where the industry's very excited because they feel that this is a huge opportunity.

So if you look at the number of remote monitoring companies that have come up in the last year, about 120.

Okay?

And many of them are all trying their best to leverage this information to see how they can become the, I would say, intermediary between the clinician and the patient to help monitor disease states out there.

Now, many of them are looking at information from devices and many of them are trying to become disease specific intermediaries, like heart failure and the like.

- We are gonna get audience questions in one second.

You had mentioned before about the human touch and the sort of tension between pushing too far into digital and losing the humanity, potentially losing the humanity of the physician-patient relationship.

And I want you to just kind of talk about sort of where you see that landing in the future.

- Yeah, no, absolutely.

You know, I think the system is a little conflicted at this point in time because it looks at much of the digital interactions you have with patients as fluff because it can't measure it, it can't quantify it, it can't reimburse it.

And I think that's been one of the barriers too for this, you know, how do you kind of balance the digital touch with the human touch?

But I think you can have the best artificial intelligence strategy, but you need a clinician to deploy it.

I mean, let me give you an example of cancer, right?

For cancer, when somebody has cancer, god forbid it, but when you have it, you have to meet with a multidisciplinary team where you meet with your oncologist, you meet with your radiologist, you meet with the radiation oncologist, you meet with the surgeon, you meet with the palliative care, you meet with the social worker, and this are usually combined meetings all at the same time.

You can't do that as of yet through an AI-based strategy.

I mean, AI tried to do it with cancer.

We have the whole Watson stuff, which failed miserably and happy to answer questions related to that.

But I think that's where the human touch is gonna be so important and so intrinsically tied.

I think we have to use AI to help us make better decisions, but we have to be able to have guardrails around them and then implement them because a lot of what we do is not just a decision, it's a shared decision-making strategy.

Everybody has a different philosophical approach to disease and they may want something different.

And this is where I think the balance between the digital touch and the human touch is gonna be really important.

- AI can't hold a hand.

- Can't hold a hand.

That's why the cover of the book actually has a hand out there because I think that is the most important part.

I think of really ensuring.

So the word care is the most important part of the title.

I think future care, sensors, AI, sounds really crazy, but I think it's the hand which is really important and preserving the human touch is really important.

- Jag Singh is the author, the book is called "Future Care: Sensors, Artificial Intelligence, and the Reinvention of Medicine."

Please join me and give him a round of applause.

Thank you so much.

(audience applauding) - A recent report by the Health Policy Institute of Ohio showed that infant mortality has gradually gone down overall, but for black Ohioans, the rate is still 164% higher than for white Ohioans.

And it said that the evidence is clear that racism is a primary driver of racial disparities in infant mortality.

My question to you is, what can AI do about that?

- That's a great question.

So, I think COVID, the pandemic really gave us a microphone in terms of understanding digital divide and understanding the social determinants of health and the variability and the inequity that exists.

It also kind of made us aware of the structural, as well as the systematic racism that exists, and we are well aware of that now.

And I think the beauty of AI is, if I can say this, it gives us an opportunity of prospectively trying to address this issue because we're still implementing it and not actually trying to backfill gaps.

We should try to backfill gaps also, but at the same time, it allows us to look at this prospectively.

I think the other thing about AI is that you cannot change bias or racism sometimes in the human mind, but you can ensure that the AI approach doesn't have any bias and racism.

You can kind of change it quickly in AI.

And therefore you can make sure that the implementation of AI-based strategies are not immersed in racism.

And some of the structural issues that have occurred over the ensuing years, centuries, can be set right.

And I think AI potentially has a role out there to set that right.

- So it sounds like data from wearable tech is very helpful to doctors in preventative care.

What about the for-profit health insurance and life insurance industry and other bad players who can manipulate the data to put more money from the hands of those who can't afford it?

Or can it be used to make the health and life insurance industry better?

- That's a really good question.

For folks who may not have heard it very clearly is, how do we use, if I may reframe that question to some extent, how do we ensure that the wearable data is being used appropriately and how do we kind of interface that with the life insurance company?

Some of that is already happening, and that's one of the questions you asked, you know, how do we make healthcare more sustainable?

And I think one of the ways of making healthcare more sustainable is, is for patients to have more skin in the game.

And when patients have more skin in the game, it takes off the pressure over the system and changes the cost equation within the systems.

And I think some of this is already happening because health insurances already now are offering lower premiums to individuals who have variables and are able to send their data to them to show that they're actually working out and actually performing the essential tasks that are expected of them because they know that these wellness strategies are gonna translate into less disease and potentially less expense for them and more profits for them.

So insurance companies actually now are actively using these strategies to get patients sucked into more wellness aspects of things.

You know, having said that, and I didn't mention this, I think when you try to sub stratify medicine, at least from, as an academic, I substratify it as an upstream layer, which is specialty care and care of complicated patients.

You have a midstream layer, which is care of chronic disease, which we can talk about.

And you have a mainstream, which is wellness.

And it's the mainstream layer, which I think really determines the top two layers of chronic disease and specialty care needs.

And I think if you can prevent that through wellness strategies, which may be incentivized by the insurance companies, I think that's a way that we can also potentially move the needle on expense.

- [Dan] Go ahead.

- Hi, Dr. Singh.

Thank you so much for being here today.

This is a wonderful forum.

My question is, I'll just read a little bit, the Vanderbilt Medical Center was recently being sued by its transgender clinic patients who accuse the hospital of violating their privacy and turning their records over to the Tennessee's attorney general.

What is the plan... Are people talking about this?

What is the plan to protect the rights of patients regardless of their sexual identity, however they identify?

Can you speak on that?

- I completely agree that everybody's rights really need to be protected.

I don't think it, as of yet, falls in the realm of AI and sensor-based approaches, but it's something that is close to my heart for sure.

And I think looking at it from the artificial intelligence aspect of things, I think the potential for AI to be misused on privacy end of things, on the security end of things, and the bias end of things, and the racism end of things, and same so for the transgender end of things does exist, and I think the appropriate guardrails necessarily need to be put into place.

- [Dan] Next question.

- Hi, thank you so much.

I've really enjoyed hearing you speak today on so many different avenues and really hearing from someone in critical care speaking to the system itself being kind of a problem when it comes to crisis care and chronic disease.

And I really would love to hear you speak more about the prevention aspect, especially when it comes to those who don't always have access to wearables or to really take the opportunity to invest into that themselves more so the financial disparities that some people have in getting access to taking responsibility or the education in that.

- Absolutely, absolutely.

So, some of the work I did just at the onset of COVID, within the first few months of COVID, was to look at video-based consultations within the cardiology division.

And we wanted to kind of look and see were there disparities kind of showing up at that point in time.

And you'll be surprised to know that 50% of patients from lower socioeconomic status, blacks, Hispanics, brown people, elderly people, were not getting access to video-based care.

And surprisingly enough, one of the other things we found that majority of these patients were actually heart failure patients because of the cross-section of all these covariates with those patients with heart failure.

So, yes, the digital divide is a big problem and I think it awakened in us that we need to look at this proactively and actually solve this problem proactively, not just kind of run away amuck with all these digital strategies, and then try to backfill the gaps later on.

That doesn't work well because then we're just repeating history again.

And I think that's something we're obligated to do.

I do agree about wearables and the expense about wearables, and I think many organizations like ours and others are looking at low cost options.

Low cost options to provide watches and low cost options to provide iPads to kind of help patients get more engaged in their own care.

Because you can only imagine that if you give a patient a $200 iPad and they're engaged in their care and they're preventing one hospitalization, which could be $10,000, right?

So the cost effectiveness of that equation really makes a lot of sense.

I think much of that is going to be dependent on the political will within different organizations and understanding that we're looking more across time and the trajectory of the disease state and not just at one point in time, you know, for the next fiscal quarter.

- So you said AI could help with like chronic illnesses and diseases, but many of our chronic diseases and illnesses come from the way we treat our food, like spraying pesticides and putting steroids in our animals and things like that.

So it's like, even if someone were to religiously work out and try to eat right, the coating that's on the fruit, the coating that's on the vegetables, the way we put bleach on our chicken, it's like chronic illnesses and diseases are still gonna come from that.

So what can AI do about that?

- Yeah, no.

That's a little bit out of my expertise, but I think, you know, generative AI, and this is a whole world of these large language models, are helping the industry to make decisions about a lot of the downstream things.

And I think just knowledge about this and discussions about this, coupled with LLM-based changes being instigated to kind of do the right thing I think will institute change.

But what you say really resonates with me because I think if I can expand on that point a little bit, and this is just going back into the prevention aspect of things, you know, we talk in medicine about secondary prevention.

Secondary prevention is you had a heart attack, now we're gonna look after you and make sure you don't die suddenly and we're gonna look after you so you don't develop heart failure.

That's secondary prevention.

There's another level of prevention that we know is primary prevention where you actually prevent the heart disease from actually occurring.

And a lot of focus goes into primary prevention where you say, you know what, I'm gonna modify, control the high blood pressure, control the diabetes, make sure the patient doesn't develop a heart attack or doesn't develop cancer or doesn't develop Alzheimer's or any other issues.

But what's really important, and that's where AI I think can really help, is primordial prevention, is where you actually start tracking people early enough that you even prevent the risk factors from occurring, right?

Diabetes is a curable condition.

Hypertension, obviously, if you have a genetically engendered disease, that may not be the case.

But otherwise, conventionally, type two diabetes, hypertension, atrial fibrillation, they're potentially preventable, they're potentially trackable.

And if you come back to Paul's story, right?

He had all these diseases that led to him having cumulative effect of disease.

But for the 20 years that he was between 73 and 91, he didn't have a heart failure episode, didn't have atrial fibrillation.

Everything was well controlled because you can prevent disease, right?

Same thing out here.

If you can prevent risk factors, you can downstream prevent disease.

And I think... And you are taking it one level before, is taking it right to the food we eat and prevent use strategies out there to prevent the possibility of developing risk factors.

So you're really talking about primordial prevention, and I think that's a really good thought and we need a lot of work to do that.

- You alluded to Watson at the end of your remarks.

And at one point, that was considered the future of medicine and didn't seem like it lived up to its potential.

So what do you think happened there and why do you think sensors and AI will fare better?

- Yeah, no, I don't know if sensors and AI will fare better, but I can tell you what happened, I think they will actually.

But for folks who are not that familiar with the Watson thing, IBM came together with Sloan Kettering Hospital in, I think, 2015, 2016, where they said, cancer has over a thousand chemotherapeutic agents, clinicians giving these chemotherapeutic agents, give them in different doses, in different cycles, in different combinations, let's kind of create a strategy of eliminating this practice variance so that we can have better outcomes out there.

And interestingly enough, they fed it with all this information, with all the trials and papers and research data and even genetic data, but there was a lacking, there were empty spaces there in the data, especially on the genetic side that found that it didn't really work that well.

It actually was a bust.

Two reasons for it specifically, and I'll give you one example, actually, not two reasons.

They looked at the efficacy of Watson to treat cancer in patients who had end-stage metastatic disease, and they looked at it in patients who had really early disease.

They found that an end-stage metastatic disease, Watson worked pretty well comparable to clinicians.

You know, patients was ill, they all had the same kind of line of strategy.

But when it was early disease, there are so many nuances, and this is where the human touch is so important.

There are so many nuances to care in the early part of the disease that they found that Watson failed miserably because this is where you need all that multidisciplinary approach and that shared decision making, and Watson really was not effective enough in providing the right care out there and went bust.

I think sensors will help us track the risk factors for disease and help us in mitigate disease.

I think AI, for example, you know, patients who have...

There's a story in the book about a young lady with pancreatic cancer.

I mean, you don't know how to predict pancreatic cancer, but now there's the hope that if you can look at 20,000 pancreatic images and find which ones actually progress to cancer, then using AI-based... - How does AI track basically heart issues from early age?

So I come like a long history of heart issues, and from an early age, I had problems.

So does AI kind of track that in the data and kind of make preventions or strategies to fix it before it turns bad, I guess, you can say?

- Oh, absolutely.

That's a really good question.

You know, for AI to have the ability to predict, it has to be able to work with large data sets that have that information, that clinical phenotype information, and have outcome data so that the machines can learn from that data and solve that equation and predict which patients are gonna develop that problem, right?

And some of that is already happening right now where you can look at a surface EKG, surface ECG, and you can predict which patients are gonna develop heart failure or which patients' heart function is gonna decline over time or which patients can actually develop atrial fibrillation or which patients are at a greater risk for sudden cardiac death.

I mean, we've obviously hearing about athletes and their predisposition for life-threatening arrhythmias.

Some of this eventually in the near future will be predetermined by looking at a 12-lead ECG.

You can then say that the potential for having an arrhythmic event is going to be pretty high.

Some of the work I've done recently, and I presented this as a late breaker clinical trial just a couple of months ago, where we looked at a patch monitor, which is a 14-day ECG monitor.

From the first day of the patch, by looking at that data, we could actually predict what was gonna happen during the next 13 days.

We could actually predict which patients were gonna have sudden death or have a life-threatening arrhythmia.

I just wanna take two minutes to answer this question a little more because it's really interesting because when I landed it in Oxford, this is, I'm dating myself, 1996, and I was starting my doctorate, they gave me a Holter monitor that is a 24-hour monitor of a patient who actually died while having the monitor on.

And they said, "Go figure it and figure out what you can do from this."

So I spent 16 hours a day for 16 weeks on that one patient looking at eight hours of their data, but really kind of measuring every possible thing to come up with a hypothesis as to why that transpired, why that happen.

That was just one patient.

Now you can look at the surface ECG with all the AI strategies and make that in seconds to decide which patients are gonna be more predisposed.

So I think there will be strategies of AI to help prognosticate, but it's also important to recognize that there's something called baseline risk and there's something called evolving risk, because your risk at one point in time doesn't necessarily determine that you're gonna have that disease, it depends on how you then intervene and what are the changes you make to your life that can change that risk across time.

Strategies.

You might be able to catch those diseases much earlier, whether it's colonic cancer, breast cancer, pancreatic cancer, that's where AI I think can help.

It's this systematic integration of information and trying to deliver personalized care through Watson that I think it didn't work that well.

(audience applauding) - Dr. Singh, thank you so much for joining us today at the City Club.

Today's forum is presented in partnership with Verizon.

It is part of the Health Innovation series, which is in collaboration with Medical Mutual.

Today's forum is also part of the City Club's Authors in Conversation Series in partnership with Cuyahoga Arts and Culture and the Cuyahoga County Public Library.

The City Club is very grateful for the continued support of all of our community partners.

We would like to welcome guests at tables hosted by the Center for Community Solutions, Medical Mutual of Ohio, and the MetroHealth Foundation.

Thank you all for being here today.

And we hope all of you will stay following the close of today's forum for a toast to the last 40 years here at 850 Euclid Avenue and all of you in the community who have made those four decades so successful.

As we prepare for our move to our new home in Playhouse Square, we are pleased, absolutely thrilled to announce the public launch of our fundraising effort in support of that move, the City Club's Guardians of Free Speech Campaign.

Information can be found at GuardiansofFreeSpeech.org, and I encourage everyone listening, all of our loyal members and friends, to participate in this campaign.

We have a very unique civic treasure in the City Club of Cleveland, and it deserves the support of each and every one of us.

Thank you again to Dr. Jag Singh, and thank you members and friends of the City Club.

I'm Robyn Minter Meyers, and this forum is now adjourned.

(audience applauding) - [Announcer] The production and distribution of City Club Forums on Ideastream Public Media are made possible by PNC and the United Black Fund of Greater Cleveland Incorporated.