Luba Greenwood John Halamka Jennifer Joe & Umbereen Nehal – SoDT 005

Show video

Welcome to the interview-series  on the socioeconomic consequences   of disruptive technologies  by Rethinking Economics NL.   In this interview, we will be focusing on  how disruptive technologies as artificial   intelligence, the internet of things, and big data  are transforming a very important topic: health   care! For this, four experts have  been so kind as to join US today:  Firstly with us today, is John Halamka. He is the  president of Mayo Clinic Platform, and prior to   this served as the executive director of the  Health Technology Exploration Center for Beth   Israel Lahey Health in Massachusetts. Previously,  he also was the chief information officer,   or CIO, at Beth Israel Deaconess Medical  Center for more than 20 years. As a Harvard   Medical School professor, he served the George  W. Bush administration, the Obama administration  

and governments around the world, planning their  health care information technology strategies.   Dr. Halamka is a practicing emergency medicine  physician and has written a dozen books about   technology-related issues, hundreds of articles  and thousands of posts on the Geekdoctor blog. 

Secondly with us today is Dr Jennifer Joe. She  is an entrepreneur, having founded companies as   MedTech Boston, Medstro, and Vanguard.Health,  and is an emergency room physician   in the Boston VA Healthcare System. She is a  cross-functional executive and innovation leader,   focused on areas as digital health ecosystems,  innovation challenges, and the telemedicine   community. As a daughter of poor immigrants  and having grown-up in rural Mississippi,  

she is dedicated to community building, providing  opportunities to the traditionally marginalised,   and providing basic health care to everyone.  She has over 70.000 followers on LinkedIn   and regularly teaches at Harvard Medical  School, the Harvard-affiliated hospitals,   and MIT, for example being a  mentor at the MIT Media Lab.  Thirdly with us today, is Luba Greenwood. She is  a Life Sciences, Tech and Health care executive,  

investor, and company builder. She  had over $5B+ in deals and investments   across multiple therapeutic areas and  life science, device, diagnostics,   and tech sectors. Her company-building experience  includes co-founding biotech and digital health   companies in the immunotherapy, women's health,  artificial intelligence, and microbiome space.   She began her career as a litigator at a  leading national law firm and recently served   as an executive at Google Life Sciences,  Verily, and as Vice President of Global   Business Development and Mergers & Acquisitions  at Roche. She lectures at Harvard University in   the School of Engineering and Applied Sciences. Lastly with us today is Umbereen Nehal. She is a   paediatrician and MIT Sloan Fellow  focused on strategic systems approaches   and utilizing human-centred design for improved  health and wellbeing outcomes. Previously,  

she was a Medicaid Medical Director, where she was  clinical lead on a $1.8b Delivery System Reform   Incentive Payment design, and a Chief Medical  Officer and Vice President of Medical Affairs   at the Community Health care Network. She has  been four-time “Top Voice” with over 235.000   followers on LinkedIn, and co-authored the United  States national curriculum, and contributed   to two state public health curricula. She is  also an Assistant Professor at University of  

Massachusetts, having received her training  at Aga Khan University Medical College,   Baylor College of Medicine and Harvard. And with that I would like to move towards   the very first question for Dr Halamka.  Dr Halamka, could you tell us more about   how the disruptive technologies as artificial  intelligence, the internet of things, big data,   are currently transforming health care worldwide?  And I was wondering, what are the most important   developments for students to know about  that arenot specialised in health care,   such as students and economics. Applicably, I  think, we just had an ambulance that drove by. John Halamka: I was going to say, you must  have orchestrated that, perfect. So I joined  

Mayo Clinic one year ago. And why did I join Mayo  Clinic? Mayo believes that by 2030 the health care   system of the United States, and really the health  care systems of the world, will be digital-first.   And if you're going to be a digital-first  healthcare system, there are certain modular   components you better put in place. Such as, if  you have, as does Mayo, 150 years of historical   data, structured, unstructured, images, telemetry,  [inaudiable], how are you going to use that in a   privacy-protecting, ethical way, to generate  new cures and new treatments, new algorithms   for patients of the future? And that's as much a  policy-question as it is a technology-question. 

So my first charge was, create a uniform  mechanism of data access for collaboration   across government, academia, and industry,  with all of Mayo's historical data assets.   And we did that. In partnership with Google Cloud,  and a series of de-identification algorithms, and   security protections that ensure privacy. Then I  was asked to create an AI factory, on top of that,   so that all the Mayo Clinic staff can say, oh  I woke up this morning and I have this notion.   Well can you explore that notion using this  de-identified data set to generate an algorithm,   and validate that algorithm. Well then, how do you  evaluate algorithms for bias? So hey, I’ll make   this one up but what if, if you've ever been to  Minnesota you may know, there are a fair number of   Scandinavian-Lutherans in Minnesota. So we create  a algorithm based on the population of Minnesota,  

and then let's imagine we're going to take that  algorithm to Spain, you know. Well, is it going   to work so well? Is it, is it fit for purpose,  right? So we've got to, had to build that tool.   We've had to build the capacity to ingest  novel telemetry, from the devices you wear   or the devices in your home, and marry that with  the clinical data that is from more traditional   electronic health records. And finally we've had  to figure out, especially in a time of Covid,  

how to deliver all kinds of care at a distance.  And what's been unique about Mayo in 2020,   is we began to deliver high acuity,  serious and complex hospital care,   in the homes of patients. And we've discharged  now over 200 critically-ill patients from their   living rooms. And that has taken a whole set  of technologies, which combine remote-patient   monitoring, and algorithms, and dashboarding,  and supply chain, and on-prem people. So I tell   you that as context, because as you look to the  digital-first healthcare system, manage your data,   create your algorithms, get new telemetry,  and deliver care of all kinds at a distance. Koen Smeets: I was wondering, you already  now brought it up to 2030 a little bit,   you already told a bit about that, but what do  you expect, do you have any ideas, of what we can   that we can expect over the  next decades, what we can see? John Halamka: Well, so. Whenever somebody  asks me to forecast a decade away,  

that's like saying it's 1993 and I hear this  web-thing, you know, is it going to go anywhere,   right? But a decade from now, we're all going  to have Facebook-implants, I mean, who knows,   right? But what you could say is this. The  experience that all of us have of health care,   well maybe not in your country, but let's say  in the U.S., is generally bad, right? That is,   we don't have the tools we need to navigate  the healthcare system for our families. 

And so what you would hope, is that, your workflow  is going to be quite different. You wake up,   your child has a fever. Aha, you pull out the  sensor, that is going to take a look at, not only   body-temperature, but tympanic membrane, tonsils,  you know, other things that may be appropriate.  

Initially, that data goes to an AI-chatbot, and  does an initial triage. Based on probability   of disease, you may or may not encounter some  health care practitioner, but the end result is,   there will be a very well understood way,  to get your child back to wellness. Call it,   ways for health care, so to speak. That's  got to be the norm, a decade from now. Koen Smeets: I think it is,  although it is a decade away,   I do think it is a very interesting  idea of what we can hope for, or expect.   Dr Joe, I was wondering, considering your  entrepreneurial background, what is your   perspective on the current and future impact  of disruptive technologies, as we discussed? Jennifer Joe: Yeah definitely. So, as Dr  Halamka highlighted, I think data and artificial  

intelligence are going to bring us into a new era  of understanding. I think it's just beginning.   John Halamka is a leader in the field and Mayo  also as a leader in the field they have massive,   John you have massive infrastructure and  you're brilliant at driving innovation,   driving teams, and putting infrastructure  in place. So I think you're a leader.   So where is artificial intelligence and big  data, I think in the US currently? I think   we talk about it, I think Dr Halamka and his  team are doing some really interesting things,   but I think we're just at the beginning. The  key points that Dr Halamka brought-up was  

data. So getting access to the data, what  kind of data is there cleaning the data, and   then the cross-collaborations, I think the cross  collaboration is profoundly important. Meaning,   you can't just throw a data scientist at data.  It's very difficult for them to look at the data,   know the meaning of the data, but after that,  come-up with hypothesis and understand how it   would actually fit into the healthcare system.  So I think we have a lot of work to do there.  I work a fair amount with the federal government,  Vanguard.Health holds a contract with them to  

drive innovation, and one of their big initiatives  is innovation and artificial intelligence,   and making a lot of the data that the government,  the US government, funds, publicly available   in a safe way for data scientists access, and then  also innovation initiatives behind driving it.  Where are we in terms of health care as a  practicing physician and adopting it? I think   artificial intelligence hasn't really made it into  the practice of medicine yet. Meaning, we're using   it, or we're seeing it a lot in claims. Or, were  seeing it a lot to potentially optimize workflow.   One of the early groups or organizations to come  out of MIT was optimizing patient no-show, right.  

It's very easy to predict if a patient's  not going to show, it's a huge burden to the   health care system, how can we maximize just  that simple piece? So, is AI in diagnostics?   Not yet, not really. Is AI in treatment?  Not yet, not really. I think that,   is it in clinical decision support specifically  with radiology? I think that's where we're playing   right now. And then how do we take that  and really elevate it to the next level,   hold AI accountable to all of the traditional  benchmarks that we hold healthcare accountable to. Koen Smeets: It's very fascinating.  And I was also very curious about,   perhaps how, not, perhaps even broader than  AI, general disruptive technologies that we're   discussing, how are they changing  the general workings of hospitals,   do you also perhaps have an example from your  own experience as an emergency room physician? Jennifer Joe: I work at the VA healthcare system  and I’m just gonna leave it at that. [laughs] And  

I love and adore it, veterans are wonderful, we  give good care, it's meaningful. We've had some   early innovations, but we are, you know, it's a  big, bulky system. So some of the innovations that   we've had for a long time is in remote patient  care, and high risk patients. So high-risk cardiac   patients, and high-risk pulmonary patients. For  the last 15 or 20 years, we've actually had remote  

patient monitoring of these patients. Where  we give them a vital signs measurement, blood   pressure cuff, pulse [inaudible], scale and they  actually remotely load that up into the system,   and we have a whole team that monitors that, flags  it if something is abnormal, calls the patient and   brings them into the emergency room. The VA is  unusual for, so in the US, the VA is unusual   because the payer is also responsible for the  chart, you know, the entire population. That's   unusual. So they were highly motivated to roll  something out like that, which I think actually   underscores that remote patient monitoring has  benefits to the patient, but also cost of care.

Koen Smeets: Ms Greenwood, could you also  expand on it? What is your perspective on the   current and future transformative impact of  disruptive technology? And I was expecting   curious about this from your experience  as both an executive and an investor. Luba Greenwood: Yes, absolutely. So it's  interesting Dr Halamka and Dr Joe are   really experts and leaders in the in the field,  when it comes to, especially when it comes to   how care is delivered. And I remember when  I was at Google, I was actually, was in the,   ended up in emergency room and one of the nurses  that was doing the intake, she said, hey what   do you, she was basically doing what she was  supposed to be doing, and she said hey what do you   what do you do? So I thought, you know if I say  investor or company builder that's complicated,   who the hell knows what that is, and I, so I  thought what is it that I do? And I said to her,   I do, I do something that hopefully in 10 years  you will actually turn around and look at me   when you're talking to me. And she, yeah, and when  you're taking care of patients. And it goes to Dr   Joe's point that, you know we do, and the big tech  companies and the amazing healthcare institutions,   are working tirelessly together, in order to  provide and change the patient-experience,   and in order to change, and help, and support  the physicians and their decision making,   and support other healthcare  professionals in decision making.  

But we're not there yet, right. So, and  we are looking at the tools to do that.  I do want to talk a little bit you know, about  a different angle, though. Because I come from   biotech, and therapeutics, and diagnostics,  and tool space. And that's a bit different.  

So when we think about what's going to happen  in 10 years, is there going to be change? I can   tell you with 100 for sure, the way that we make  drugs, and the kind of type of drugs that we make,   is going to be very different. People are shocked  today about, when we think about vaccines, and wow   how did they make vaccines so quickly, there are  some conspiracy-theories that perhaps they are,   they're not done correctly, but one of, some  of the things that people are missing, is that,   there's, the reason why it takes some time  so long to discover and develop drugs and   do clinical trials, is because we have enormous  inefficiencies. That with the use of big data, and   not just artificial intelligence, I’m not a big  fan of using ML, that it will solve everything,   just sometimes pure, simple analytics, and these,  basically, there are so many silos, taking away   the silos in development, can actually help you  tremendously. It can help you recruit patients  

faster, so that you can find the patients, it  can help you with actually doing the development   faster of the drugs, it helps you with the  discovery of the drugs, personalising the drugs.  And what do I mean by personalisation,  right? We know for patients, for example,   as an example with cancer, so you know, the same  patient with cancer with the same type of patient,   different, two different patients with the same  type of cancer, respond very differently to drugs.   So we now have amazing tools, and we're using  the data that we're collecting, again, in a,   in a responsible, de-identified way, so that we're  not using patient-data and patient information,   in a way so that we can actually predict  whether or not a particular patient is going   to be responding to a particular therapy. We're also making drugs, so this is on the  

treatment side, but we're now actually using  that type of data to, from the very beginning,   making drugs for the particular population that  will truly respond, and not just truly respond,   but particular population that will have limited  side-effects, right. Sometimes when you hear   advertisements for certain drugs, there's that  list of many things, many horrible side-effects   that could potentially happen. Well, if you make  more targeted drugs, they, there, you would limit   the side effects that people are have, that people  are having. So, there are some really amazing   technologies, and novel ways that we're right  now, doing drug development. The same thing for   another example is autism, right. We say autism  is a spectrum. Well, right now we're using AI  

[audio-cut] for example to actually find  the right population, find the right,   not only find the right population, but again  make the right drugs for that population.  So that's on the treatment side of the diseases  that we know, but there are also many diseases   that we don't know what in the world causes them,  we like to pretend we do, we like to pretend we   know the brain, but we don't. We don't really  still know, for example what causes Alzheimer’s,   how we can treat Alzheimer’s properly.  So we are using big data and AI and I’m   confident that in 10 years, as an investor  and company builder, and this is the kind of   technologies that I’m investing in, is to  understand the brain better, so that we can make   the correct drugs that actually truly work, for  everything. From Alzheimer’s, you know Alzheimer’s   and dementia, to anti-psychotic drugs as well. And in terms of the discovery of actual medicines,   you know as I mentioned the development can  be done much faster. You don't need to take a  

year you know, year or over a year to recruit your  patients, if we can find pa-, if we can find right   now you know, cut different you know, if you're  a target, and you can find your target-customer.   Why is it that, when you're running  a trial, you cannot actually find   the patient that you need to? So, that is again  another example of how we can use, pretty simple   basic technology that we already have right  now, that we as consumers use, and apply it   to a healthcare setting so that we can actually  match the right patient to the right trial.   And I would say, so that is I  would say for the medicines, right. 

And if you have been as an investor or company  builder in the therapeutic space or biotech   space for many years, you view these disruptive  technologies as something that expands your pie,   right. It's not something that, hey it comes in  and now you no longer need biotech, right. That   you will, everything will be solved with digital  therapeutics. That is not the case, right. That is   the case with some things, but that's not the case  with everything. So you see these amazing tools,   basically expanding the pie of possibilities. There are also many what, what's interesting  

is there are also executives, and inventors, and  investors, in the diagnostics and tool space. So   the disruption, in the diagnostic tool space  has been even more exciting, but it's a little   more risky than in therapeutics. So, it has been  historically a pretty tough business to get into.   It has low margins, the reimbursement is pretty  low, but actually when you look at the regulatory   landscape for diagnostics and tools, it's the  same as therapeutics, right. So you're going,   it's the same risk, risk profile, it's not,  they're the same regulatory hurdles, the same   barriers to entry with IP, but then once you're on  the market you're reimbursed at a much lower rate,   and there's a lot more competition. So, you have,  what's amazing has been is that there is this new   tech that is disrupting the diagnostics and tool  space, but however at the same time you know,   a lot of the investors and the start-ups are  recognising exactly what Dr Joe said, which is,   yes it's great that we have this amazing,  these amazing tools, but are they being really   incorporated as part of the workflow? So that's  a big question, right. You can be reimbursed,  

you can be on the market, but are they actually  going to be used by patients and by physicians,   that's a whole other question. But still, they do  hold a lot of promise, we have really neat things,   we were working at Google and at Roche, you  know incorporating new sensors for example into   devices to predict risk of infections,  we've used AI to you know, see and help   patients manage their chronic diseases you know,  you see companies like Livongo on Duo doing that.   Many disease apps, and disease management  apps, are being reimbursed by insurers,   and insurers have embraced this new model. So that  will actually continue, continue to go and evolve.  And what's been a very interesting I would  say, lastly I would say, is that for investors,   these are very true, everything I said is the  outlook and what investors that have traditionally   been in therapeutics and diagnostics. Now, what's  been exciting about this disruption, is that it   brought in investors and company builders  that have never been in healthcare before.   And they're now coming, coming in into healthcare,  and again building companies. So however,  

their space that they're investing in, their  space where they're building is a bit different,   you know they're looking more clinical decision  support tools, they're looking at telemedicine,   they're looking in ways to perhaps triage  patients, and, or even delivery of medicines.   So that's in itself, has been become a very  large market. Now it has a lot of landmines,   because many of the kind of the items that are,  you know in these pockets that has just mentioned,   will not necessarily be incorporated as part of  the practice of medicine, or the medical workflow,   but it holds a lot of promise, and  most importantly promise for patients. Koen Smeets: I think it gives a very  fascinating overview of the landscape,   especially for investors and how the technologies  are influencing healthcare. Dr Nehal, could you   expand on how you see this, especially from  your experience as a Medicaid Medical Director? Umbereen Nehal: Yeah thank you so much. I mean,  fascinating discussion, I’m thrilled to be a part  

of this. I think I’m just going to pick up on some  threads from what's already been, you know woven   into this discussion. You know, different things  that I’ve heard. So for instance, from Dr Joe we   heard about aligned incentives, when the VA is  both the payer as well as a provider of care,   and has a set population that is not because,  you know in Medicaid we often hear about churn.  

For those who are not familiar with Medicaid,  it's a government agency. If you're in Medicaid   you say, because there are 50 states, and that  each one's administered at a state level, getting   both federal and state funding. So we often say  in Medicaid, if you've seen one Medicaid program,   you have seen one Medicaid program. So there's  an unfortunate, that's another added aspect of   fragmentation, as Dr Halamka had referenced early  on, that you know a lot of people in the US,   we don't, haven't had very positive experiences of  healthcare. Majority of Americans get their health   care from their employer. But that means that if  you change employment, you can have disruption in   your care. And then, who owns your data, right.  We've talked a lot about data. The data, well  

is it owned by the provider, the hospital,  is it owned by the health plan, you know,   we've heard from Professor Greenwood about, you  know the business aspect, the investor-landscape.   A lot of times you'll hear about a  market's, market segment, you know we heard,   you know couldn't we speed-up clinical  trials, and get our patients in faster,   but then the question is, are we identifying  a very specific segment of the market, and we   have a very specific customer that we're serving,  and we design innovation in that space, for that   specific customer. And you know in business side  we talk about minimum viable product, or MVP.   But then, when you, let's say you are  successful in that, then you know, at MIT in the   there's the Entrepreneurship Center, there's a,  Bill Aulet is really famous as an entrepreneurship   professor, and he talks about land and expand.  So all of that though is about segmenting,  

it's about identifying a specific population, but  if you come more from the academic research side,   you always think about generalisability. Have  you, have you designed the product, have you   designed this study, such that it can be gen-  reasonably generalised to different populations.   We heard Dr Halamka mention, that yeah there's  a lot of like Scandinavian folks in Minnesota,   there's also a sizable refugee-population from  Somalia, right, and we know that AI for instance   doesn't work the same for, you know if you're like  saying doing dermatologic or facial recognition   software, it doesn't work the same across  genders, doesn't work the same across races.  So I think the the opportu, and actually I’m  technically not, I’m technically supposed to be in   class right now, so hopefully my professors don't  mind that I’m here if they see this, I’m actually   taking a risk management class right now. And one  of the things we were talking about was you know   supply chain, one of the things I haven't heard  mentioned yet is what can AI be used for in terms   of supply chain, right, that was a huge issue in  COVID in terms of whether it was PPE, whether it   was reagents needed for testing, whether it  was just simple supplies that we needed for   our everyday living. That's another application.  But one of the things we heard, because we were  

talking about which of the companies, which of  the industries, had foreseen COVID as a risk,   even though it was known, there was you know an  entire office of pandemic preparation, nobody   had necessarily prepared for it, because again  businesses often will think about their budget,   the likelihood, the probability something  will happen, and they'll often take that risk.   Whereas I heard from a classmate who was in the  financial sector, because of regulation from   the government, they were prepared. They actually  had had to test being able to go remote, and have   appropriate cyber-security for a pandemic. So they  were able to pivot on a dime. And so that's, so   as much as people hate government, it's like, oh  the regulations, we have like, it slows us down,   there's cost, there's a lot of benefit sometimes  to having that external authority, certainly when   I was a Chief Medical Officer, sometimes within  the kind of the gladiator-games that happen in   the c-suite, I, you know sometimes if I said it it  didn't necessarily have the same impact as being   like, well, we might get audited. I used to be in  Medicaid I know how this goes, we don't want to  

get audited, do we? So, sometimes you kind of need  that person on the outside wielding the big stick.  And then we've already mentioned around trust,  and I think that's a real challenge, and we   haven't discussed it, I won't go, get political,  but I mean, the elephant in the room is right,   that we're living in very, I’m like I would love  to live in precedent times again, like I’m tired   of the unprecedented times, but I, right now you  know there's a lot of distrust of government from   all sides. There's the historically marginalised  communities, who have, like things like,   Tuskegee, but then we also have, on the other  end of the political spectrum, people who feel   that government needs to be afraid of them, and  these you know, there's a lot of distrust. And   so as much as I would love for government to be  the answer, I also know that, it's complicated,   right. But I do think the role government can  play, is being that person with a big stick.  

Thinking about serving all populations, and  finding a way to maybe thread things through   different organisations, that maybe the business  interests may not necessarily take care of. Koen Smeets: Yes, I think it's very  interesting, and it shows really well,   as well for students in economics that wish to go  on in policy, the importance of the regulation and   the other governmental policies. Dr Halamka, could  you expand on what you think of the comments of   your colleagues, and how you see these things? I  was also wondering, especially for students from   economics, what they would need to know about,  for instance the, related to government and such. John Halamka: Well, glad to talk about that. And  so, I always am very careful not to use present   future tense. What that means is, we have  a product that will. Well either you have a  

product that does, or you will have a product that  will. So I’m going to be very specific about what   Mayo has already deployed. For the last several  years, with our large de-identified data-sets,   we have developed a series of AI-algorithms that  are used in real-time, clinical practice now.   And let me give you a couple examples of those.  So it turns out you could take a 12-lead ECG,   and this is you know, typical heart tracing, and  humans interpret them, machines interpret them,   but there's much more subtle data, that a machine  learning algorithm can find in that pattern than   a human or a traditional EKG-machine. We're  able to, with a very high AUC, above 0.9,   identify your ejection fraction from your 12-lead  ECG without having to do an echocardiogram. Now,  

just for your audience, basically what that means  is we can do a really cheap test, that is not   invasive, and have it give us, in effect, a good  idea of how well your heart is functioning without   an expensive, or invasive test. We are able  to predict hypertrophic cardiomyopathy, future   atrial fibrillation, look at progression of  pulmonary hypertension. So you heard from Luba   about this notion of pharmaceutical companies  saying, we're going to provide a molecule,   but we're also going to provide a sensor,  and then we're going to provide a service.  

What if that sensor is your Apple Watch. Because  we can look at your lead-one ECG, and tell you   through an algorithm today, that it is your  pulmonary hypertension is getting better or worse.  So I tell you all these things because  everything they've said is totally right,   right, it's the future is already here it's  just unevenly distributed, and there are,   certainly AI algorithms being used inside workflow  today for certain kinds of domains, but it is not   normal just to say, I’m going to walk into a  doctor's office and AI is going to see you now.   They've also said something really  important about workflow integration.  

So many entrepreneurs say, oh I’m going to create  this app, or I’m going to create this website,   or I’m going to create this sidecar to  Epic or something. Doctors hate that,   right. They're not going to go to a different  app or website, no. What they need, if you're   going to deploy an AI-algorithm, it needs to be  passively working behind the scenes inside their   EHR workflow, and help them at the point of care.  And so that's, very much as we've deployed these  

things in production what we've tried to do. My efforts in this Mayo Clinic platform I now   oversee is to scale what we've already done at  Mayo, to a much larger audience. So when you talk   about economics, this whole idea of platforms  is that, oh this is not just a Mayo specific   adventure, this is just look, effect like  Airbnb helps you rent a room if you have a room,   the Mayo Clinic platform is to say, oh how do  we link telemetry of all kinds, from various   manufacturers and a lot of interesting challenges  with lack of standards and telemetry today,   normalise that data, bring it to algorithms,  produce a result in workflow, that goes directly   to the physician, or the patient, or the payer,  or the pharma, or whoever the customer may be.  

And then from the economic standpoint is,  how do you ethically, and transparently,   monetise that and create value for doing  that kind of orchestration. I’ll be honest,   Mayo has thought of this three ways. Some is very  familiar to my entrepreneurial colleagues here.   We co-invest and start companies, and then realize  equity-growth as these Mayo generated products and   IP are promulgated. Sometimes we will license  algorithms, that will then go into the devices   as embedded software. Or sometimes we  will do revenue share if a transaction   is completed through a commercial associate.  So, our effort here is really to democratise   access to these tools, while also  asking the question, as the United   States moves from fee-for-service to value-based  purchasing and different reimbursement models,   how does Mayo position itself in what we hope is  this digital-first economy, to be a player not   competing with tech, but being  a part of that tech-ecosystem.  

So that, so I think everything is said is  right, and we're just trying to be a part of it. Koen Smeets: I was wondering, if there are  students in economics that have become now really   interested and wish to go into healthcare, and  especially also this tech-side of healthcare, how   would you recommend, Dr Halamka, to go into that?  Do you have any tips, or ways they can contribute? John Halamka: Well, so here's the sort of  interesting future, I think the United States   is going to experience, right. Fee-for-service  medicine is still the default in many of the areas   of the United States. I get paid more for doing  more. But on the East Coast and the West Coast,   there are a fair amount of risk-based  contracts, alternative quality contracts.  

So the question for your economics folk, is  how do you ensure there's quality, safety,   right care, right place, right time,  that's also affordable? And these are   very tricky issues, and at the moment,  it just seems like risk-based contracts   are the best way to go, where you're paid for  health, as opposed to being paid for sickness.   But certainly, we need a lot of analysis, and  we need a lot of refinement of those models. Koen Smeets:   Yeah I think there's a lot of work at least for  students economics as well to contribute here. Dr  

Joe, could you expand this on how you see this and  especially from the socioeconomic consequences,   as, how will this influence the affordability  of healthcare for marginalised communities? Jennifer Joe: Yeah definitely. And  I think that we're getting at where,   what do we think is the big opportunity for  healthcare and healthcare innovation in the US,   and reflecting I think on COVID and what it's  changed. So COVID has rapidly accept accelerated   innovation in the US. It's disrupted things,  but it's always also been an opportunity for   adoption and a change. The number one space  that we've seen this happen in the United  

States is telemedicine. So we were forced  into telemedicine across the United States.   The payments were the big regulatory initiative  that really drove that through. The fear of COVID   and delivering care in a safer way, both for  patients and for clinicians, was important.   So, telemedicine is now widely adopted. It has  many forms, and it's not totally clear what form   telemedicine will take. I think there's going  to be lots of different shapes. So when we're  

thinking about the future, telemedicine's  not going away. How much will it decrease,   it's a little unclear. The payments need to  be addressed, and we have good movement on   addressing the payments. There's lobbying groups,  there's physician groups, there's patient groups,   who are like, this is good, we like it, we  want it, we're going to keep it, we're pushing   on it. So I don't think it's going away. So I think when we're building, and you're   thinking of innovation solutions, you need to  build with that being the model going forward,   that has a lot of needs. We've already seen  that, with what platforms need to be chosen,  

how does it integrate into the EHR, but I think  we're also going to continue to see it in terms of   communicating with the patients. It's really  accelerated, there are just text-capability for   communicating with the patients, which is crazy!  [laughs] We haven't had that before, we do now.  But also the add-ons, right. So, I did my  telemedicine annual physical with my doctor   yesterday. She didn't have my vitals. How do I  get my vitals, you know, how does she know that   I haven't gained my COVID-19. I have gained  my COVID-19, and that should be a factor. But  

she probably should have weighed me, taken my  blood pressure, make sure that my weight didn't   increase my blood pressure, etc. So I think,  and I have not ever been formally trained,   people call that blue ocean, or blue sky, Dr  Nehal will tell me, that's a real opportunity.   Remote patient monitoring is a piece of that. Dr  Halamka has already alluded to Mayo apparently   has very significant remote patient monitoring.  I’ve already spoken to the experience of VA, that   we've been doing remote patient monitoring for our  high-risk patients. It works, it decreases costs,   it improves outcomes, and increases patient  satisfaction. We've had that for a long time.  

The rest of the US, because of the payment models  that have been set up, hasn't really had that.   We've had movement from a regulatory standpoint,  or a payment standpoint, so CPT codes, and also   the expectation and the fact that COVID is now  here with US for a while. I think that's also   going to be an area of opportunity to figure out  what kind of technology needs that we have there. 

And then a topic that is near and  dear to my heart, is that as we   accelerate innovation, and technologies, and  new care delivery models, we need to make sure   that we're providing good care to all Americans.  So lower on the socio-economic ladder, and,   or who has less access. So what does that mean,  what does that look like? That means, we have   to as a country get behind funding and supporting  innovators and entrepreneurs who are behind that   and committed to that, and it's not necessarily  a place that big business has always gone.   They can go there, I think if the regulatory  and payment models are there. In terms of   what are examples that I’ve seen and  really been excited about in the US,   I think the hospital systems have just been so  overwhelmed and focused on responding to COVID,   that we haven't really had that much  opportunity to integrate new innovation in. 

The most impressive or interesting  example that I’ve seen and I,   I’m writing a textbook now on digital health  and telemedicine, we're finishing it, the   most interesting example is Emory Grady. They're  located in the south of the United States, so they   have more disparities. They have overall lower  socioeconomic status. Georgia is right next to   Mississippi, which is where I grew up. Mississippi  is the poorest state in the United States.  

They have rolled out a tele-EMS system, which is  fascinating. They've worked with the public health   infrastructure. They've worked with 9-1-1, which  is unusual. They've worked with ambulances. And   they're facilitating a system where they can give  clinical support to the EMS that goes out to the   patient. Which is hugely important, because  that's how a lot of our lower socioeconomic  

citizens access healthcare. So if you can bring a  physician out there, you can assess the situation,   you can route them to the right place, but they're  also providing telemedicine visits. So a lot of   those 9-1-1 calls that are EMS, if you don't  have constant access to a health care system,   you maybe are accessing 9-1-1 at a time when you  don't need to go to the ER, you don't need an ICU,   but maybe you need your hypertension medicines.  And can we handle that with telemedicine? 

That's fascinating in terms of closing the gap  of disparities, but how many partners had to   come together to make it work, providing better  care, decreasing cost. So I’m very interested   in seeing how the United States can  get behind initiatives like that. Koen Smeets: It's very interesting. Ms Greenwood,  I was wondering, from your legal background,  

could you tell us a bit more about how this  relates to the regulation of such technologies?   And how does this affect your  perspective on these technologies? Luba Greenwood: Absolutely. I mean, one thing that  is clear and this probably would be interesting to   your students, is we are becoming you know,  the United States and other countries are   really becoming interlinked. And we're already  global economies and it's only going to be   growing. United States you know has quite a  bit to learn about the regulatory environments  

in other countries, especially in, for example  in Western Germany, as it kind of relates to   regulating of digital therapeutics. So US has  actually, did embrace digital technology. Actually   I believe Mayo was also quite involved. Verily,  Google, was certainly involved in that effort.  And and many other companies, so the government  did listen to start-ups, and large companies,   and providers, and payers about, what is the best  approach to regulation of digital technologies.   But we're still a bit behind companies like  Germany, that are actually making it fairly,   fairly easy and accessible for  digital therapeutic companies to   have their products on the market be  reimbursed, and then also get the data,   right. Because at the end of the day, if you  don't have the right type of data, you know you   can't really have a true digital therapeutic  that has any validity, clinical validity.  I would say on the legal side, so that's on the  regulatory side, so yes, still a lot of exciting   things are happening, regulatory authorities are  very knowledgeable now on digital and are open,   but we need more. We need more learning from other  countries. On the legal side, so that has been  

interesting, Dr Halamka mentioned licensing of  algorithms. So that is something that you probably   would have never heard about five  years ago. So, the legal community   was, you know felt very strongly that AI  algorithms cannot be patented. However,   that has entirely changed. So now, you know you  can actually, well there are some ways to do that,   but there are you know AI algorithms tied to  actual products are patented, and many of them are   being patented. So that's another thing that is  to, for both economic students and legal students,   to understand that there has been a huge shift in,  in how we see and how lawyers view intellectual   property, what can be licensed, what can be  patented, what can be protected with patents,   versus what can be protected with trade secret.  And absolutely, when you are an organization  

like Mayo and you come up with something  truly revolutionary, like they're doing,   you can protect it with patents, and other type  IPs. So that's also important, a pretty big,   important change, in how technologies are viewed,  patenting of technologies and AI are viewed.  Some of the other things that it are  important from the legal, regulatory   pieces, and this is actually, everybody  has touched upon this a little bit,   is ethics concerns. Ethics, data privacy, security  protection, and ultimately, and very importantly,   bias. So, when we talk about data and when we talk  about the use of AI, there is still when you look  

into who actually puts these the the algorithms  together, who puts the products together,   there's some inherent bias, not intentional  bias, not just in data itself, but even with   the software engineers themselves, and the data  scientists that are putting this together. So this   is actually something that I do each in my class  for, you know in the School of Engineering here at   Harvard, is to be aware of bias. So to be aware,  not just of the biased data but your own bias   and think very carefully because, this is no  joke right. I mean, when you train a model and  

if you train the model wrong, you can have the  best intentions, but you will actually be doing,   you'll be hurting patients. So this is an  area of both bias and ethics. It's another   one of these areas that in the next 10 years  is going to be an essential component of any   product design, right. Just as how today, we've  incorporated, for example design, which before,   if you talk to anybody in the diagnostics and  med-device, and you know, they thought a bit   about design but that was not the most important  thing for the patient or the physician to actually   really want to use the device, you just, hey it's  on the market, approve, reimburse, you use it,   right. But now we are incorporating of course  design into engineering, so the same way bias and   ensuring that you'll have a whole group of people  in any product design in healthcare that will be   looking at bias, that will be looking at  data privacy, will be be looking at security,   and looking at ethics, and the ethical use  of products, is going to be very important. Koen Smeets: Yes, yes. I think we see it  also in the other interviews that besides,   of course in healthcare it's a very, very  primary thing, but ethics is, I think always,   and the bias is a very important thing. So  considering time, I think we have to move  

towards the last question before the closing  statements. This is for Dr Nehal. Could you   expand for us, as well as on bias and  ethics, but I was especially curious on   how these technologies are affecting developing  countries, in, from a healthcare perspective. Umbereen Nehal:   Thank you. Yes, it is, there's you know with  everything there's always opportunities,  

barriers, and risks. And they they're always in  tension, right. So the opportunities are that,   there, if there is an understanding of needing  to train algorithms, as was just said you know,   to account for bias, then there may be you know  for instance like, I went to medical school in   Aga Khan University, which is part of the Aga Khan  Development Network, which has a global footprint,   both in different parts of Asia, and Africa, and  I know other parts. So that's an opportunity where   you could partner with other large systems  that care for for diverse populations,   and very intentionally develop the  algorithms, develop training models,   or train the, you know the algorithm that way,  such that what you know, to Dr Halamka's point,   when you do patent, if you do patent it, you do  license it, you want to make sure that you're,   then I mean, and this is an assumption of  mine, I’m not an expert in this area is that,   I’m assuming that there's not going to be a lot  of change to it that's the nature of a patent,   right. So you want to make sure that what's gone  into it a priori, is correct, and you've trained   it on the right population. And maybe you want  to intentionally build in ways, in which that,   even if you're licensing it, or patenting it you  can allow some customisation. Because, in theory,   machine learning should allow, allow learning,  right. So you should be able to apply a new data  

set, and it should be able to customise itself,  but then the inherent rules built into it need to   allow that accuracy across populations. So I think that if there's you know,   it from the design, again, underscoring what was  just said from the design-perspective, we talk   about human-centered design, as long as we're  thinking about the full spectrum of humanity.   And what, the other thing that's really critical,  so when we think about, I think you have to get   like really kind of in the weeds about data. You  have to be like a data nerd, where you have to   not just think about who is the data coming, from  but how are you parsing it out, right, how are you   defining the fields? You know, we know that for  instance gender is a spectrum, it's not binary,   right. We know that race is a social construct,  and because race is a social construct,   so for instance, what the status of those who  are black in America or are African-American,   may be equivalent to somebody who's of South-Asian  origin in the UK, just because of the historical   context, who's you know been the historically more  deprived populations. So I think, and and we have  

these populations, ethnicities, different  cultures, in different developed countries.   So if there is a, if people are genuinely serious  about addressing disparities, closing disparities,   closing gaps, being ethical, being equitable,  access, not just you have access to the doctor   as Dr Joe pointed out, but then, when you're  at the doctor are you being treated equitably?   Are the data, are the evidence-based guidelines  being used, are they are they appropriate for you?  And then you know, even like you know Dr, sorry,  Professor Greenwood had mentioned about you   know personalised medicine. So is that going to  become another disparity, where some populations   that have access, they have resources, get the  personalisation, everybody else gets the kind of   what's for everybody. That's another source  of disparity. But if we think about our   minority populations as being majorities in other  countries, is there an opportunity for partnership   in developing countries? As well as, you know  we know from COVID, supply chains got disrupted,   right. What's happening in China matters to what's  happening in the US, right. India is a place where   there's a lot of outsourcing of many different  professional services, there are many you know   manufacturing plants in different countries. This  workforce in those countries, they matter, I mean  

they should matter, just as human beings, but  from a business perspective, they matter, right.  So I think these are all opportunities  to for developing countries, for low- and   middle-income countries, but yeah, no, great  question, and we could talk forever, I’m sure. Koen Smeets: I think it's a beautiful way to  end the the series, the interview as well.   So for closing statements, we always have  the same question in this series, so it is,   if there's one thing you could say to students in  economics related to the topics discussed today,   what will that be? And I think  it makes most sense to do it in   the same order as with the starting  questions, so first for Dr Halamka? John Halamka: So maybe a quick statement about  economics. In this new world of virtual care,   some people say, oh, data is the new oil. I  would actually tell you, data is the new water.  

And that is, it's ubiquitous. What's the oil,  is the algorithms. It's turning the data into   wisdom and action. But as you've heard from my  other panellists, we're going to need labels on   those algorithms. This is a a bag of dried  fruit. Note it has a nutrition label on it,   tells me a little bit about its fitness. We're  going to need nutrition labels on our algorithms,   if we're going to eradicate racism and reduce  disparities, disparities of care across the world. Koen Smeets: I think that's a beautiful statement.  Dr Joe, what will be your last statement on this?

Jennifer Joe: I will resonate, I  think with what everyone's saying,   is that we live in a time of telemedicine,  digital, and everything is very interconnected.   So I think, what is going to lead to success  is cross-disciplinary, cross-collaboration   of different perspectives, and working together  to move innovation forward. So very excited to   have economists working at healthcare, and to  give us their perspective, and drive us forward.

Koen Smeets: Yeah I think  this is seen really commonly   in the series, it's the importance  of interdisciplinary work. Professor   Greenwood what is for you, one thing you  want to give to students in economics? Luba Greenwood: I would say, the most important  thing, just as actually both Dr Joe and Halamka   said, Dr Halamka said, is access to information.  Access to information empowers patients,   and we will finally, hopefully in the next decade,  get to the point where we can empower patients,   and not just patients with means, but really  patients that haven't even had access to care. And   that empowerment piece is very important, I think,  in how we practice medicine, not just in the US,   globally, sometimes this is you know, in a way  that we undermine the patient a little bit. And   so, having, actually empowering the  patient to speak for themselves,   the knowledge about their diseases, it could  help them to be knowledgeable about their health,   not even just the disease, that they don't  get the disease, is very important. And   also, just very briefly, I love the  dried cherries that Dr Halamka is eating,   I thought I was the only one that gets  them shipped from Washington State.  

They're very good, they don't have  sugar-added, they're good for health. Jennifer: What are these dried cherries?  Luba: you have to show them, you have to show,   they're amazing, they are incredible. Okay,   get them shipped, oh, you're on mute. John: Sorry, it's Chukar Cherries,  

in Washington, and I highly recommend the medley,  as opposed to just the tarts. Umbereen: Okay now   they should just ship it to us all free, given  that we've given them some free airtime. Right,   they got some earned media time. Koen: I will try  to see whether we can have a sponsoring-contract  

for the next series. Dr Nehal, what will be for  you the next, the last statement on this topic? Umbereen Nehal: So I think that you know, by  nature I’m probably not somebody who measures,   despite being at MIT, and like the non-data person  at MIT. I mean, to me data is a necessary evil,   right. I’m more of a humanist, I like to kind  of go with my intuition. But, I’m at you know  

at MIT getting an MBA in my mid-career, for a  reason. And that's because you know, we have to,   we have to be able to measure we have to be able  to have a common language. And then language often   is numbers, and language is values, but I think  that when we're quantifying things, when we're   putting a value on things, we have to also put a  value on trust, right. I’ve heard collaboration,   I’ve heard stakeholders at the table,  patients knowing themselves, so what is,   what is the value of human knowledge, like,  lived experience, what is the value of trust,   what is the value of collaboration, I don't,  I, we, not that we have to have a metric   for everything, but that has to be part of our  equation. Because when we only look at certain   numbers that are, you know usual metrics, usual  things on a balance sheet, we discount things   that are really important, and we erode that, and  then we end up in a big mess. So I would say that,   let's not let's not just assume oh we're good  people, we'll just do it in empathy you know,   like let's let's put numbers, let's put hard  numbers behind things, supposedly soft things,   that are important, so that  we remember to value them.

Koen Smeets: I think it's a beautiful  way to end the interviews today.   Thank you again, each of you for taking  the time, and we hope to see the viewers   for the next interview. Umbereen:  Thank you so much. All: Thank you!

2021-05-09

Show video