Dr. Nirav Shah, Chief Medical Officer, Sharecare
Historically, technology in healthcare has created greater inefficiencies. How will that evolve? Using real-world examples of innovations in technology, Dr. Nirav Shah, Chief Medical Officer of Sharecare, sees a future of personalized “team-based care” in which clinicians will collaborate with artificial intelligence.
Nirav Shah, MD, MPH, is chief medical officer of Sharecare, the digital health company that helps people manage all their health in one place. An experienced leader in healthcare and digital health, Dr. Shah also is a member of the U.S. Department of Health & Human Services Secretary’s Advisory Committee, an elected member of the National Academy of Medicine, a senior fellow of the Institute for Healthcare Improvement (IHI), senior scholar at Stanford University’s Clinical Excellence Research Center, and an independent director of public and private companies and foundations. Dr. Shah also cares for patients at Stanford Health Care.
Prior to joining Sharecare, Dr. Shah served as chief operating officer of Kaiser Permanente in Southern California and as commissioner of the New York State Department of Health.
Show Notes
Dr. Nirav Shah shares examples of how medical technologies have advanced over the past hundred years. [03:11]
Four areas in a patient’s healthcare service experience that will evolve by 2049. [05:47]
Team-based care plus A.I. will work together to eliminate redundancies in the healthcare system so that physicians can spend more time with patients. [09:37]
What are the human superpowers that computers can’t replace? [11:12]
Technology in health care has caused greater inefficiencies in the system. How is that changing? [12:47]
With new technology, how do we ensure that the world of 2049 is more equitable than now? [15:16]
What types of healthcare providers will go out of business if they don’t embrace new technologies? [17:15]
Will the hierarchy of health care, in which the dominant players of today be flipped upside down to be less important in 2049? [19:24]
Can this technology vision be achieved by 2049? [20:51]
Hearing aids can now quantify if someone is lonely. [21:43]
An example of a dry biomarker tracking for a person with a history of allergies. [22:18]
How can a federated learning model be beneficial and preserve privacy? [24:00]
Examples of the benefits of data that can be unlocked for public use, while preserving personal privacy. [25:43]
Seven key areas that can create a healthcare system that makes the world a better place. [28:42]
Transcript
Jason Helgerson: I'm Jason Helgerson and you're listening to health 2049.
Dr. Nirav Shah: Artificial intelligence is a different member of that team. So how team-based care plus AI will work in the future is that AI will augment the ability of that team that will take away some of the boring and redundant parts of our practice of medicine today, thereby opening up much more time for the human aspects of care. You want your doctor to spend time counseling you, talking with you, understanding what your priorities are. Not necessarily behind the computer typing up a note because that's what's required for billing purposes.
Jason Helgerson: [01:48] Today's guest has seen health and healthcare from many different and interesting perspectives. He has been a Stanford University researcher, Chief Operating Officer for Clinical Operations at Kaiser Permanente, and as the Commissioner of Health for New York State. Today, he's working on the cutting edge of health and science as the Chief Medical Officer of Sharecare, a health technology company enhancing digital health offerings with innovative AI solutions. To him, the future of health involves utilizing science to the fullest extent possible in order to make the world a better place. I'm Jason Helgerson and you're listening to health 2049. And it's my pleasure to welcome Dr. Nirav Shah to our program. Nirav, welcome.
Dr. Nirav Shah: [02:30] Thank you, Jason. It's a pleasure to be here.
Jason Helgerson: [02:32] Nirav, please tell our audience a bit more about your interesting background
Dr. Nirav Shah:[02:36] Well, I think you've pretty much covered it today. I am the Chief Medical Officer of Sharecare based out of Atlanta, Georgia. I live in Palo Alto and practice at Stanford University in internal medicine and find ways to play with my kids and go on hikes with my wife otherwise.
Jason Helgerson: [02:56] Oh, great. So we start where we always start with our guests, which is really to ask you to describe your vision for what healthcare looks like, feels like, what the experience is or will be for those living in the United States and beyond in the year 2049?
Dr. Nirav Shah: [03:11] Healthcare in 2049 is going to be very different than how most people experience it today. Healthcare will be much more digital, mobile, virtual, remote, real-world, ambient and continuous. And let me double click on some of those. When I talk about digital, mobile and virtual, what I'm talking about is today we rely on technologies that were developed decades or even hundreds of years ago, and the opportunity to leapfrog some of the limitations of those devices to what is capable today with our devices is exponential.
I'll give you a real example, when you use a stethoscope on a person, you can certainly hear the heart rate and a very experienced clinician can catch many different things about how the heart is beating and how it's working based on their interpretation of the sound. Well, today's stethoscopes, now, incorporate ultrasounds and other technologies so that it automatically gives you several more dimensions of data and knowledge in much less time. And that opportunity is not universally available today. It will be tomorrow.
When I talk about ambient today, we're still thinking about healthcare in very specific ways where it's related to a doctor, placing their hands on you in a clinical visit and then understanding how that disease has affected your body. Ambient means that there will be continuous sensing, without you having to even notice that healthcare is being delivered to you by artificial intelligence, by sensors in the environment that monitor and understand who you are relative to your environment. You know, the more we learn, the more we understand that you are defined by your environment. It's that whole nature versus nurture debate. For example, when it comes to how long you'll live. Your zip code is much more predictive than your genetic code in terms of how things turn out for you. We know that your weight tends to be the average of your three closest friends.
So the challenge is to change everyone's micro environment, to optimize for total health. Both physical and emotional wellness to best be able to achieve what every individual wants to do with their health.
Jason Helgerson: [05:32] So maybe we can think about this and your vision for the future in 2049. Can you put us in the shoes of the patient, of the individual, of the family and what would their experience be like in terms of the provision or the receipt of health and healthcare services?
Dr. Nirav Shah: [05:47] Today, we're stuck in mostly the play space provision of care. You often have to either go to a hospital or a clinic to get healthcare. Now we're moving towards video visits. But today the reality is we know that place agnostic care can be the reality for the vast majority of care. Of course, if you're getting a surgery, you have to go somewhere for that surgery, but tele-health has already replaced up to 75, 80% of primary care visits because of COVID. And we know that the opportunity in many other specialties is the same. So from today, going to a place and having bricks and mortars define where your healthcare is delivered to being place agnostic tomorrow.
Today, we often get our health insurance from our employer and that employer based insurance model has worked to varying degrees for the last 50, 60, 70 years. But more and more, we're seeing that government-based insurance, whether it's Medicare, whether it's Medicaid, whether it's the innovation that the bundles of care are creating, that government-based insurance will then probably create a great equalizing factor. And so I'm not a proponent necessarily of Medicare-for-all, but the fact is, and the reality is that more and more of our government is paying for more and more of our healthcare.
Today, we're still in a model of physician convenient care. Physician convenient care means the care was designed for and by clinicians, doctors, with their convenience in mind. The surgeon wants to have all of her patients in one ward of one hospital so she can run through at five in the morning and examine them on post-op day one according to her needs. And that made a lot of sense when physicians were a scarce commodity. But when you have many other augmentation of physician skills, tele-health, and other things, we can actually move toward a model away from physician convenient care toward person or patient-centered care. We've talked a lot about it, but we don't really even know what that means, yet. And I think that the whole paradigm of care around a person is going to be very different.
Today we're also focusing on the practice of medicine by focusing on individual diseases and individual organs, as opposed to focusing on the whole person. So you have specialists, you go to for your heart, your liver, your kidneys, and each of them works in relative isolation to try to optimize their organ, relative to the others and relative to your overall health. So you may get medications that compete with one another in terms of what their effects are on your heart versus your kidneys and unfortunately, that leads to a lot of problems. In the future, when we actually focus on a whole person together, it will fundamentally change the prioritization of care and what you get to optimize your total health based on what you want to do with your total health.
Jason Helgerson: [08:52] Interesting. So obviously today we talk a lot about team-based care, and I often like to say trying to make healthcare a team sport, getting providers from different areas working together as a team and sharing information amongst them so that you can get to better overall patient outcomes. But with the new technologies Nirav, do you see that that team will be less than just a group of individual physicians and providers and more technology, with maybe a single clinician interpreting the results coming out of the super computer, out of that technology that will be the interface with that patient? So instead of a team of many doctors, it will be a team of technology together, maybe even with the single doctor. Is that the future you see?
Dr. Nirav Shah: [09:37] Yeah, that's a great analogy and team-based care, it does tend to provide much higher quality care than relying on individuals or with all the handoffs we see from one person to another, the discontinuity of care when you don't have a team-based care model.
And how I like to think about it is that artificial intelligence is a different member of that team. So think of everything that's boring, routine, repetitive that you don't want to do as a doctor. I don't want to have to type up notes just for billing purposes. I don't want to just check the medications every single time to make sure they're reconciled. And yet these are important functions that can very easily be replaced by AI, technology and digital approaches that never get bored of checking medications. That never forget to do the safety checks of allergies versus what a patient has been prescribed. That don't make mistakes on the misspellings or the spellings of one medication versus another.
So how team-based care plus AI will work in the future is that AI will augment the ability of that team. That will take away some of the boring and redundant parts of our practice of medicine today. Thereby opening up much more time for the human aspects of care. You want your doctor to spend time counseling you, talking with you, understanding what your priorities are, not necessarily behind the computer, typing up a note because that's what's required for billing purposes. If I was in medical school today, I would actually think about those human superpowers that today computers can’t take over, things like critical thinking, communication, collaboration, and creativity. All of these are super powers in the future because it will be the hardest for computers to take over.
Let me give you an example, you've heard a lot about self-driving cars and how they're going to be driving us everywhere in the very near future. And they do a fantastic job. But today, if a self-driving car is faced with a little pile of snow versus a little pile of glass, It doesn't know what to do. Why? Because what computers don't have today is common sense. They don't have what every baby has, which is exposure to a whole bunch of field simultaneously so they can take things from one area and then apply them to another. That's a very, very hard problem for computers today.
So when we think about expert systems, they fail when it comes to common sense, humans don't fail with common sense. That's what makes us human. We can pull these threads together in very interesting and thoughtful and ultimately creative ways in ways that computers can't do today.
Jason Helgerson: [12:30] So Nirav, do you see this new technology and the evolving role of AI, do you think it's going to lower healthcare costs? Increase healthcare costs? Or have nominal effect, meaning the healthcare costs will continue that on the trends that are on today into 2049?
Dr. Nirav Shah:[12:47] Great question, Jason, and the history of technology in healthcare has been exactly the opposite of technology in just about every other industry. In every other industry when you have new technologies, they usually introduce greater efficiencies, right? When you introduce a computer, people become 20% more productive. Well, you introduce Epic into my workflow and what used to take me 45 minutes to admit a patient, now takes an hour and a half just clicking and pointing all across my electronic health record. So technology has failed us in one fundamental way in healthcare, and that's by increasing the barrier between the patient and the physician.
The future, however, could be very different. The kinds of technologies that we're now introducing actually augment the skills of clinicians. If you think about it this way, AI is replacing the senses one by one. So for example, in radiology or ophthalmology or dermatology or pathology, when you have an artificial intelligence read the scan before the human, they can catch all the things that humans won't. But then humans will also add value by recognizing patterns in different ways or catching things that today our computers can't.
If you think about it today, let's say a clinician is 93% effective, a computer alone is 85% effective. Together, they are 97% effective in reading these kinds of scans. That augmentation of the senses is the reality and when you do that and you think carefully of the value-based world we're entering where cost, price transparency, and all are driving adoption, those technologies that actually improve efficiency are, I believe, going to be scaled the fastest.
Jason Helgerson: [14:40] One of the concerns that's often raised about technology and healthcare is the potential for increasing the inequalities that exist today and that in this future state, even more empowered care or technology empowered care, will really create even bigger gaps between the haves and the have-nots of society. Obviously you're an optimist, you see the benefits of this technology. How do you respond to those critics? What do you think needs to be done to ensure that the world of 2049 is more equitable than the world of 2021?
Dr. Nirav Shah:[15:16] Great question, Jason, and the reality is we're starting to, only just beginning to understand the black box that is artificial intelligence in many cases. We’ve seen how algorithms, trained on a very specific population, can actually serve to disempower certain other populations. If an algorithm is trained on white males, it'll know what to do for white males and it may give the wrong answer when it comes to Hispanic females, for example. And so our first step is understanding the potential for bias and the potential for coming up with the wrong answers. That is widespread now, and many groups are spending a lot of time to make sure that we don't have such biases built into our tools of the future. And, you did say I am an optimist, and I do believe that what will most happen with technology is that all boats will rise.
As you know, for example, access is a problem to healthcare, and either it's like, you'll get admitted to the hospital and then you get Cadillac care with a team of clinicians working on you, or you don't have access to care at all in a rural area, for example. Or with the wrong kind of insurance or under insurance or no insurance and you fall off a cliff. Those are the kinds of realities we face today. There's a big step change between those who don't have healthcare and those who do have access to high quality healthcare.
If we go toward the future I envision, it'll allow us to actually decrease the inequities because it's not that giant step change between those who have versus those who have not.
Jason Helgerson: [16:54] In your future state, do you see that there'll be winners and losers? Will certain organizations, types of providers be clear winners, whether that's financial or otherwise, and do see others also being losers, meaning that they'll be financially worse or potentially even go out of business, in the future state that you envision?
Dr. Nirav Shah: [17:15] Well, clearly we know that the writing is on the wall that today, for example, the hospital should not be viewed as a revenue center, that hospital beds are like telephone poles, very expensive infrastructure that drives short-term thinking. Folks who don't embrace the fact that healthcare will be in the home, that healthcare will be delivered ambiently, that healthcare will occur wherever the person happens to be, not in the hospital, is going to be a loser.
So if you're building large hospitals, I would say in general, you're probably spending money on things that don't make the most sense for 2049. I'm not saying that there won't be any hospital beds, but you can almost imagine the hospital of the future as an ICU, on top of an OR, above an emergency room and everything else is delivered in the home.
Same thing with nursing homes. I think we've seen with COVID-19, that people don't want to be in nursing homes, but that if people are well supported, family caregivers can actually do a better job keeping people where they want to be, which is in their homes. Obviously there's a lot that needs to be done to get away from that nursing home model that we've embraced that is not what most people want, but I think that's an example of a loser.
I think the winners are those who are most agile and able to change and able to use these incredible sources of data to actually deliver high value, high quality care, wherever that person may be, whether it's at the job, whether it's at home, whether it's at church, wherever they may be.
Jason Helgerson: [18:50] You're really sort of painting a picture of a world in which the hierarchy of healthcare is almost flipped upside down, right? Where the hospitals and large insurance companies, which sort of dominant players in the industry today become much less important, and physicians or physician groups or community-based services empowered with technology, and even just technology companies themselves that are at the lower rungs of the industry today, rising. Is that an accurate description of this vision have for the future?
Dr. Nirav Shah:[19:24] Yeah, I think so. The reality is always slower than we hope it to be in the sense that we predicted that value-based care would be the norm about a decade ago and we're still waiting. Many parts of the country still don't even know what that means, even though everyone has heard the words.
Same thing, we’ve talked about how AI will become sentient and we'll be able to replace humans in terms of intelligence in five years, 10 years. I don't think it's going to be five years or 10 years, it’ll be 20 or 30 or 40, or maybe never at all in many ways.
So to the extent that there's a lot of vectors pointing in the same direction now in terms of what the future of the hospital is versus what the future of care in the home is. There's vectors around how integrated delivery systems are doing better, how value-based operators are doing better, how payment mechanisms are changing fast enough now that the whole insurance industry will have to rethink its model in some important ways.
I believe there'll be many changes. I hope that we see them on the ground by 2049, but the common thread is that individuals will be in much more control. That's the hope.
Jason Helgerson: [20:40] Great. So just along those lines, it sort of begs the question, how confident are you that your vision for health and healthcare in 2049 will actually be achieved?
Dr. Nirav Shah:[ 20:51] You can see there are a lot of optimists making big bets in healthcare today. The last five or six unicorns have been healthcare technology or digital health companies. And you've seen many big news announcements of people making big bets in healthcare, whether it's Amazon Health, or others. So I believe pretty strongly that healthcare will be very different in even 10 years from what most people experience it today. And by 2049, I hope that we're much, much closer to the vision I described at the beginning.
Jason Helgerson: [21:25] So let's double click into artificial intelligence itself. It's much discussed in a variety of different industries, but a lot of people, I think yourself, included very excited about its potential application in healthcare. Tell our audience a little bit more about how you think artificial intelligence is going to transform healthcare.
Dr. Nirav Shah: [21:43] You know, artificial intelligence allows us to think in very different ways than we used to be able to in just the near past. I'll give you an example. Today's hearing aids that Starkey makes, can now quantify an individual's level of social isolation based on how many words that hearing aid hears, based on how many different voices, based on the tenor and the tone. It can quantify if someone's lonely or not. Who would have imagined that possible? And that's possible today.
We now have all of these, what we call dry biomarkers. Biomarkers are little things like you draw someone's blood and you can see what is the risk of going on to have cancer, for example. Dry biomarkers are those that are taken from other data sources and allow you to help predict an individual's outcome.
Let me give you an example, let's say you walk to work every day and you carry your phone in your pocket and you have a history of asthma and allergy. One day you walk through the park and you have bad allergy symptoms. The next day you walk through city streets and your allergy symptoms aren't that bad.
Overtime, the phone's GPS data, by downloading local pollen counts daily, hourly, minute to minute weather and humidity data and where you're going can inform you in real time saying, you know, your allergies are pretty bad today. You may not want to walk through the park on your way to work. That's possible today with some of the work that we're doing.
Imagine a tomorrow in the not too distant future, where you could add other things. The model I'll give you something we call edge computing. All of that data, about who you are, about your personal symptoms, where you're walking, all of that lives on your phone. It doesn't have to leave your phone, that's called edge computing.
The model that predicts your asthma or allergy flare can live in the cloud and then be localized on your phone. The model on your phone gets smarter based on all of your local data. Your data never leaves your phone, but the learnings that AI model.that's smarter, goes back to the cloud and makes the mothership allergy prediction model smarter.
That's a system called federated learning, where we can take data from many different people, and preserve their privacy. Their data never leaves their phones. So they're never vulnerable to hacking or attacks or anything like that because you'd have to hack into every individual phone. On the other hand, with this federated learning model you can get much smarter, much faster and not worry about all those data issues that we worry about today. So federated learning, zero trust is another example. All of these new things that we finally started to apply to health care, and that have been around for a while in other industries, will allow us to preserve privacy and learn much faster and realize a vision of a learning health system in ways that we've talked about in the past, but we've never really done. You will have a digital twin who will go through your life and see what happens if you take one medicine versus another. And warn you before you take the wrong one, and which one's better for you. Those are the kinds of things that are possible today, but that will be scaled by the time 2049 runs around.
Jason Helgerson: [25:08] So are you at all worried? I mean, when we talk about when we're recording this, we're still in the heart of the pandemic and there's tremendous efforts going on around the world to vaccinate people. And survey after survey shows a lot of reluctance amongst people to get the vaccine. And, it sort of, also in just the rise of the anti-vax movement and just a lot of concerns about new technologies in healthcare, it’s potential impacts. Are you at all worried about this, or did you see it as a challenge convincing the broader public around the value of these new technologies so that they become widely adopted?
Dr. Nirav Shah:[25:43] Yeah. You know, that's a great question, Jason, and I think part of it is that we've had some bad actors in some real examples of where privacy has been violated or an individual's data has been exposed and that's led to mistrust on many levels and it goes back many decades in health care. And I think part of the opportunity here is that rather than focusing on the right to privacy, we may want to think about the right to information.
So for example, if we aggregate and anonymize that data from individuals. Google researchers looked at phone-based searches on things like stomach cramps, and then can use that person's GPS information to see which restaurant they'd been in and then have inspectors inspect those restaurants, which were found to be three times more likely to be unsafe compared to the conventional methods of looking for a food borne illness and which restaurants aren't keeping the food warm enough. That kind of opportunity from the right information allows public health and everyone to benefit while still preserving privacy. There's ways that we can, for example, Ways, this is an app that allows people to crowdsource traffic and accident information allows first responders to know about an accident 30 minutes earlier than their traditional systems. And actually get to an accident because of weaving through traffic better four and a half minutes faster than if they didn't use Ways. Kinsa Smart Thermometers allow you to predict where COVID is up to three weeks before any other system to understand where a hotspot of COVID is in America today. And that's because they have 2 million smart thermometers across America, which is a leading indicator of where fever and symptoms are versus a lagging indicator, which is when a person shows up in a hospital with COVID symptoms when test results come back or when someone dies from COVID.
So these are examples where there are tremendous benefits in data that can be unlocked without aggregating the data by keeping people in the driver's seat and keeping them as the owners of their data. We can do all of this today and we'll be able to do much better tomorrow as we have the privacy principles that are universal and don't lock down data universally, but allow us to unlock that public benefit.
Jason Helgerson: [28:19] So now we come to our last question, which is really sort of taking a step back and thinking about your vision and its broader implications. Why do you see your vision as really important to humanity, important to the world? And also, how does your vision, what you hope will be the case in 2049 in terms of health and healthcare if that is achieved, how will it make the world a better place?
Dr. Nirav Shah: [28:42] It's a big question, Jason. You know, in 2049, I hope that healthcare will first and foremost be equitable. We know the problems of today are because of the incentives and the way that system has been designed to perfectly deliver more and more healthcare and not deliver more and more health. So equitable will be number one on my list.
I think number two is person centric, which I spoke about earlier. It's not about the doctor, it's not about the hospital. It's not about the insurance company, it's about you and what you need and how you want to live your life.
The next idea is continuous. Today, we're very episodic in terms of how healthcare is delivered and if you think about it as a continuous linear function, you can continue to optimize health by little nudges that you don't even notice. With ambient intelligence around you, that can totally change the trajectory of your life in letting you do what you want to do, how you want to do it and with whom you want to do it.
I talked about privacy preserving and how people should be the owners of that, of their data and all of the data about them. And it's a very different approach than Facebook and others have taken to date. But I think at least in healthcare, it should become the norm.
Finally, I believe healthcare is a right, not a privilege and ultimately I think it will become free. We're moving away from commercial insurance for employer sponsored insurance is giving ground to more and more government sponsored insurance. And ultimately I hope that all of the quality evidence-based healthcare is actually free. Of course, there'll be a higher version, which, looks a little different, may have a little better branding, but the fundamentals should be free. And, you know, today cost doesn't reflect value. You, as a head of Medicaid in New York saw how two hospitals across the street had three fold differences in the price of a C-section, for example. With no differences in quality, that doesn't make sense and in the future, with a transparent pricing and costing system approach, we will be able to capture value much better. When cost does actually reflect the value.
I think that we talked about systems approaches and continuous learning today. We have that Swiss cheese model where a lot of medical errors happen because four humans failed in a row to do what they were supposed to do and in the future, when you take a more thoughtful meta or systems approach, we will have all those safeguards in there in ways that we don't today, because it will be cheap, easy, automated, and probably done by robots and others. And that'll facilitate continuous learning in different ways.
And finally care will become more and more human. It is still all about care. And empathy and compassion. The pendulum has swung a bit away from that because of all the burdens faced on carers, including physicians and nurses and others in the system. And we need the pendulum to swing back and stop moving on the side of actually delivering what people want more and more help.
Jason Helgerson: [31:55] Excellent. Well, thank you so much, Nirav for coming on the show.
Dr. Nirav Shah:[32:00] It was my pleasure.
Jason Helgerson: [32:02] And that was Dr. Nirav Shah's vision for health and healthcare in the year 2049. As always thank you for listening to health 2049. If you enjoyed what you just heard, please subscribe to us and share this podcast with a friend. Thank you and see you next time.
