How Technology is Changing our Healthcare



you Oh hopefully my slice come up so I'm gonna talk to you today about something that I think can be transformative for healthcare so I just want you to suspend disbelief for a moment if a then B so I'm telling you I can do this deliver high-quality medical imaging better quality than MRI Pat CT all of these things we call by initials make it a thousand dollars she a thousand times cheaper and put it in a wearable or a portable and so if you could do that we could save a lot of lives in MRIs saved my life 24 years ago today finding my brain tumor has it saved anybody else's life diagnosing it so heart disease and so forth it's expensive though the cost hasn't changed in decades but if we could make a thousand times cheaper way of seeing inside of our body in high resolution in a surgery we could figure out do we have it all out rather than just there's a very few specialized surgical suites globally where you can like roll the patient into the MRI or and then pull it out pull the patient out and see if you catch it and back in and back out like why that's pretty awkward it's cool but it's expensive and there aren't very many of them we could put diagnostic technology like this inside of ambulances in fact the number one cause of disability in America right now is complications due to stroke because there's two basic types of stroke there's the clot type stroke and then there's the burst type stroke if you have the clot type stroke you can take a drug to burst it if you've got the burst typed stroke you can take a drug to clot it but if you get the drug wrong patient dies so today that means access to medical imaging within two hours of a stroke if you live too far away tough luck you're not going to walk or talk for the rest of your life so if we could diagnose the type of stroke closer to where the stroke happened we could make it so the patient could walk and talk again and there's a million things we can do in the ER in ICU with monitoring a patient if we can lower the cost of medical imaging in fact a lot of technologies are being developed with an assumption that an MRI or a CT is $1,000 in the case of a CT you also get radiation you can only have a certain number of them in your life or they actually give you cancer because of the radiation but why not have this at home like why not be able to see inside of our bodies like Angela – Angela Jolie cut off part of her body because she had the braket gene I could envision that we could make a bra where we could monitor anybody for high risk for breast cancer or any other type of cancer where you could check at home like you do for hypertension now if you've got a blood pressure cuff or a thermometer you could just see the answer to basically three questions is it getting bigger is it getting smaller is it saying the same size that's totally different than thinking about diagnosis now because the cost is so expensive and then there's the biological basis of brain disease right now the way we diagnose depression for example is you answer a series of questions are you sleeping all the time have you gained weight do you have thoughts of suicide if you answer yes to that and the litany of other questions you're clinically depressed and the way you're not clinically depressed anymore is if you start answering no to those questions it's very subjective but a study published last summer with 1,200 F MRIs that's a video form of MRI of patients with clinical depression showed patterns of different types of depression like anhedonia that's the absence of hedonism like no joy versus anxiety different kind of with these patterns we could see if the therapy was getting and making the patient better or worse much more quickly the average clinically depressed patient loses 12 years of employment from the depression and your chances of depression over a lifetime are about called 25% but if you've lived in a war zone in more than doubles it's a huge problem and many others this can have an impact to help us understand our minds by letting us see and high resolution in a precision way inside of our minds and then you know like I never had a conversation with my grandfather he had a massive stroke before I was born so his speech center was hit we've shown that you could actually communicate speech telepathically to a computer just by monitoring certain areas of your of your brain but there's more and maybe the most radical part of this is not just what it can do to your body but what it can it can enable in the next communication medium which i think is brain computer communication this is a about a decade old this is work by Jack gallant at UC Berkeley University of California Berkeley where he he started using rats then he moved up to macaques and then he started using graduate students as lab rats putting them in MRI scanners for hundreds of hours making them watch youtube videos recording how their brains lit up how they reacted to those scans what fMRI measures is oxygen use millimeter by millimeter in your head and making maps of that with the images the students looked at then a new video clip was shown and the computer inferred with just the scan library of the fMRI data of the graduate students what it thought the student was looking at that's decade old and it's pretty uncomfortable to lay in an MRI scanner for hundreds of hours and so with higher resolution and worries of you know comfort in wearing something we could do so much more in fact I showed live on stage at Ted last year us achieving a billion times higher resolution than FMR so this is this is coming and they're profound legal and moral and ethical implications but the goods are really the goods I feel outweigh the bats in a profound way cuz it's the opportunity to understand our brains and maybe control ourselves better and maybe you know if we take it even further like if we understood each other would it end war because right now we have a dialogue problem and ultimately you know here's the thing I bring this up you might say stay away from that telepathy stuff that medical imaging stuff is good but sorry like if you make a law that you can't use it on your head you're not dealing with the the 2 billion people in the world that suffer from brain disease so I think this will go pretty fast if we think of lowering the cost of medical imaging people will use it on their their heads to communicate and we the reason I talk about it is because we have to define what it means to be responsible there also we can focus the ultrasound down I'm going to show you how this works I haven't showed you yet we use ultrasonic pings and camera chips and red light to replace the functionality of a 2-ton magnet with liquid helium and superconducting magnets and we focus ultrasound down and we can we can actually oblate tissue so surgery without the knife opened the blood-brain barrier deliver micro dosing of drugs at the right intensity at the right spot so imagine a chemo dose at a hundred x lower but the right intensity wherever you need it so MRI into a portable a wand a wearable or something the size of a smartphone with pretty profound profound implications I was running I have to admit I'm and three years ago advanced consumer electronics at Facebook and I knew some advances were going to come down the pike in the trillion-dollar manufacturing infrastructure where I've lived and breathed and shipped billions of dollars worth of consumer electronics on the hairy edge of physics for last decades after my brain tumor because I needed health insurance I'm American I would have probably done a multimedia art but I'm going to show you how this works have my cell phone out for a reason your body is translucent to red light you see that light that's just my I don't want to blind anybody but um yeah so goes right through x-rays gamma rays the fields of 2-ton magnetic bores also go through your body but guess which one's cheaper like by a lot the problem is the red light scatters and everybody thought that scattering was random gonna skip this further so red light right and then we focus these ultrasonic pings down and where we focus the ultrasonic pings the light changes color ever so slightly because we use a laser instead of a flashlight we can't change the color and then we record the color shifted light with a hologram we record a hologram on a bear camera die very similar to the camera die in your iPhone that has pixel sizes that are the size of the wavelength of light which enables us to record Holograms so just to go back for a moment this is super important because three-quarters of humanity lacks access to medical imaging and that access hasn't changed significantly in the 25 years since I had my brain tumor nor has the price and so there's a lot that can be done if we completely rethink it so we have and we've made three new components a new kind of laser a new kind of ultrasonic chip that can be made in existing silicon factories in Asia and new kinds of camera chips very similar to the type that are shipping in in the smartphone's you all have for the past five years we've had camera chips with pixel sizes that there's that are the size of the wavelength of light but nobody's really use them on mass to record the waves in the wavelength of light I'm a physicist by training I'll explain to you a bit more about it but a common question I get asked is what about skull and bones so here's the sort of short course in optics black absorbs light white reflects light or and so the white just scatters the light through the bone so that's real human skull you can buy it at skulls unlimited calm we really did and the light goes through it and so here's just a little cartoon of how this all works we focus in ultrasonic ping down we do the ultrasound first for a reason that's gonna sound obvious when I say it but um sound travel slower than light so you see the Lightning and then you hear the thunder right and so then we bring in the light the light that goes through that ping changes color ever so slightly and so what we can do is bring in from the side color changed light to beat against this and we make by recording these two beams of light together what's called a hologram or these interference patterns and we can decode them just like rosalind Franklin decoded this iconic image of x-ray diffraction to reveal the structure of DNA for the first time it took her a while she didn't really get full credit for it till she was dead but we can do that at a million times a second and so here's where we are we started I put my job not knowing if this would work just just thinking boy every brain cell in consumer electronics is focused on B RNA are and so I saw these things coming down the pike to enable augmented reality which the only moneymakers so far in augmented reality has literally been Pokemon go so these improvements that enable us to do this or put in place in the factories tens of billions of dollars are spent to enable next generation Pokemon go and I just thought great you guys go do that I'm gonna go check out if we can do something here and it's it's come out pretty well so this is where we are this year in an alpha cat and then assuring to enable you know like a tricorder I had a wand get these things in ambulances so here's some of the first images that we got well actually these are images from about a month ago where we compare these are crossed toothpicks in a phantom tissue a phantom tissue optically mimics real flesh-and-blood as well as ultrasonic Lee mimics and that's the MRI so you know there's it's a different image quality this one was done last weekend here's just some slices that we took we don't slice the kidney we image inside of the kidney which is very blood Laden and so we looked inside the kidney and then compared it to mr and ultrasound and the image quality that we're getting is is arguably better we're not at the end of that we're in small rats right now small animal imaging rats right now and we'll be in human trials next year but doing the rats first to you know basically test this so here's the thing what this means is false positives I got a question about false positives and it's like okay and like you do your blood pressure cuff and you get a weird reading you just do it again it takes almost no time or you you know you measure your temperature and you get a weird reading you do it again you really care about the answer to those three questions if we can change the cause structure and the complexity of doing this and you know a lot of people come up after me when I talk about this and say you know they really hate the noise of the MRI and they're glad I'm getting rid of it and I'm just gonna say this right now I don't care about that I think MRI is amazing technology tough it out wear ear plugs seriously or it's so cramped it's amazing the problem is it's super expensive and so here's um yeah so we'd better with better understanding of disease progression by also seeing deep in their body we can have a better understanding of early things so where we are right now is we're in prototype and we've made the new kinds of lasers the new kinds of camera chips and ultrasonic chips we're scanning small animals and and we'll have an alpha kit at the end of next year human trials next year and probably beta kits so this is coming fast because it's consumer electronics and we're here just trying to figure out really you know I'm in Silicon Valley it's tech week every week where we are perceived as myopic trying to come here talk about it and figure out you know how to collaborate and make this stuff happen this is all way bigger than our little company it's how we work in the infrastructure of the world to make it happen my most previously most successful startup was called One Laptop Per child as an MIT professor I created a architected $100 laptop and made with my much more famous co-founder Nicholas Negroponte a a multi-billion dollar not-for-profit startup that catalyzed 30 billion dollars of revenue for our for-profit partners and became the fastest growing consumer electronic category ever recorded and most importantly transformed the lives of hundreds of millions of children in the developing world and so part of that was as a start-up figuring out how to work with NGOs with the health industry like in this case health industry in that case education industry but but the big companies the little companies everybody the universities and so forth and so really it won't happen unless we all enable these things to happen so medical imaging for a thousand x lower-cost and that's all I have to say I think the panelists have I've got two minutes left on the clock but I cede it to the panel and thank you okay remarkable and hopefully gonna talk about something a bit more exciting than Pokemon go but that really is up to dr. jack reuler second dr. Jack come up on stage and welcome his fellow panelist kwon do sonal Wadhwani johanna shields thank you good afternoon everybody fascinating every time I hear Mary Lou speak I am always aghast at how how medicine is gonna change has everybody seen the recent stuff on TV with a very kind of ginger gingerly looking Trump standing next to the Queen at various sort of things well apart from that being quite funny this marks the 75th anniversary of the d-day landings this year which were remarkable because what the d-day landings started to do was to end a war a battle against something which we thought was potentially unbeatable but what we learned from that and the analogy of that and what we're going to be talking about in this panel is that it wasn't because of brute force that we ended the war that we overcame this grand challenge it was because of data it was because of intelligence and then more importantly not just having data and the permission to use it but actually doing something about it it's data its intelligence its planning and then it's convincing people like me to actually act and this is going to be the crux of of our discussion today and to have that discussion we have obviously Mary Lou who has already introduced herself but also Joanna shields from benevolent Sunil Wadhwani who's come a very long way as well from from the igate fat and and I WI AI foundation and and also Kondo from Samsung who's also traveled a very very long way and so thank you first of all very very much for taking the time and traveling such a long way for us today please welcome our our panelists my mum thought it was good that I studied medicine because she thought it must definitely be a job for life you know North London Jewish mum great so I'm a bit worried that I'm out of a job pretty soon tell us what you guys are doing and and what is going to possibly put me out of that wonderful career that my mum thought I'd be doing till our 65 years old Joanna why did you tell why didn't you start oh yeah we've got microphones I I'm CEO of an exciting company here in London called benevolent AI and what we do is we are we're looking at the challenge of you know for all of you guys who've suffered the horrible experience of someone you love you know being diagnosed with the debilitating disease and then realizing there's no treatment for it if you think about it there's 13,000 untreated diseases that are out there today and what benevolent AI are doing is we're trying to organise all of the world's biomedical data and to link that to proprietary datasets to clinical data patient level data and really open up the possibility of understanding better the underlying cause of disease and addressing the other big issue in in medical care and that is that the drugs that we do have have such low efficacy rates that the top ten selling drugs in in the u.s. right now the the kassir eights are between 30 and 50 percent so what we're trying to do is unlock the underlying cause of disease and develop a treatment that is targeted to that particular patient and a type that addresses the very specific mechanism behind their disease and then we design a molecule to hit that target okay so nothing that's particularly gonna put me out of a job just possibly treat a casket that's good my mum will like you come what are you what are you doing to put me out of a job so just to share a couple of minutes of background is I'm from India originally I'm an engineer went to the US went to Carnegie Mellon and since then I've started a series of companies in the technology space the largest of these grew to over 34,000 employees and sold it a couple of years back so you learn something about how to scale technology based organizations with that what's relevant for the discussion has in addition to the companies that I have I also have two nonprofit foundations both in India one is a health care nonprofit that I started five years ago and we use we basically use innovation data analytics to transform primary health systems in really low-income settings in rural parts of India urban slums and so on so today we're running several hundred clinics in the worst parts of India last year we treated over five million people with free healthcare and that number is growing dramatically every are relevant to this discussion a year ago my brother and I started and artificial intelligence Institute in India also a non-profit and the idea is how do we how do we apply AI for social good why do we help people earning less than two or three dollars a day in areas like health care education and so on so I would say jack not to worry medical background huge need for folks like you and even more so the developed world we have a giant need for people with a medical background in low and middle income countries and even in those countries especially when you get outside the metro areas there is a huge shortage it's going to be there for a long long time but AI can play a dramatic transformation role in helping us overcome that shortage of medical personnel just fantastic I've just been seeing the ticker for my pension fund which has just gone up as a result of big companies Lenovo Tyco Electronics I'm Berk known as MSD outside the US and I currently lead the global strategy group for Samsung for all of Samsung's businesses started five companies right and the context for today is actually company I started called care visor where we use artificial intelligence to help make health care real time 24/7 through a completely different interface right so patients should not have to wait to call up Jack in the middle of the night get his answering service and then hopefully somebody will call him back the patient's you they would just talk to an avatar that then should be able to help him or her middle the night weekends whatever it is so Jack are what we're hoping to do is actually not putting you out of a job but actually make me more effective doing what you do instead of actually having to spend a lot of time ineffectively on things that shouldn't be done in your practice so I can be playing more golf at the same time like I was like the reservation is an excellent um you know you made a very very important point and that is about technology being applied not just to countries where we have greater resources but also to those where we are lacking resources and where there are far fewer of me per capita and I think this is a very interesting point I I recently was given the privilege of interviewing the outgoing chief medical officer for the United Kingdom Dame Sally Davis and she said that despite all the great things she stood like like setting up genomics England and really putting antimicrobial resistance to the forefront of the world stage that the greatest gap that she saw having been in office for ten years was the growing gap between health span and life span so we have we have become we were living longer but we become sicker at the same time but in fact the lowest to deciles of wealth are actually getting sicker earlier so net-net all we have done in the last decade with despite everything that we have tried is increase the gap between health span and life span with your financial position being one of the key problems that is driving that that greater gap we will help people like me I am sure do more but what fundamentally do we need to do to direct our efforts towards the developing world obviously Mary Lou I mean you've got a lot of things to say about this because your technology is radically cheaper than what a lot of people I comes from a country where one in five dollars spent is on healthcare it can't keep getting more expensive of course the poor are suffering I mean of course and and the bureaucracy they have to navigate through to get care is so difficult that when you're sick you can't do that well if you don't have a family member to do it for you lots of luck so but the key part of it is just lower the cost right then we can change the rules if we just lower the cost right even like guidelines right now for breast imaging globally breast imaging is I'm not sure yeah maybe he's that too that instead yeah so um 10x more scans should happen just to follow existing guidelines just for breast cancer well just just for one thing it's just we aren't meeting even what we think the guideline should be let alone changing the guidelines so in in in reaching we have the internet everywhere that happened right so this is something that you're maybe you should speak about it but yeah I mean we know that the that in the case of radiologists there's some new stuff for them today to understand disease progression but if it's just diagnosing you know the cancers we know that most most AI machine learning programs are better at it so great that's cheaper fantastic let's stop the fight like we can understand it about early a disease progression and you know I think at that same meeting where you were interviewing the outgoing head of NHS somebody said the biggest thing you could do for cardiac health which gets half a half of us got very round numbers cancer is gonna kill us the other half heart disease I don't know you're the doctor not that kind of doctor um so like if you could see what that cheeseburger you ate did inside of your arteries if you could see so how do how do we actually work on our own health maybe if we could see what was happening in our bodies we could could work on it better maybe there's a lot of there's a lot of efforts towards that I think many of you are trying different things but how do we work on yeah of course to your point I totally believe it's all about reducing the cost of health care I think we do see cost of health comes in two different ways one is actually reducing the per incident cost but the other is actually how do you avoid the cost all together if you can treat something earlier to avoid something more complicated later on and for example we talk about medication not being very effective but the realities right now most patients are not adhering to whatever care plan that they've been given by their clinician whether it's to exercise taking medication and so forth so I'd like to show you you something real quickly hello this is Khan I'm getting slightly worried about my job right now well the whole point here is that this this demo this way what should I do if I cannot feel my baby kick terrible it's definitely it's definitely a demo well this is likely like we're right now at the whole point so you can use artificial intelligence to sense on a real-time basis what's happening with a patient right if a patient's trending low or high and their glucose you can actually intervene get them to take some insulin drink some juice whatever it is if a patient after a rehab a cardiac rehab program is not walking get them up to walk very very basic things reminder take their medication we can do this 24/7 that's actually how you get patients to adhere to whatever guidelines or care plans they're on that's how you help to avoid the cost that's actually how you help to reduce the overall healthcare system Gus let me add another dimension to this so every year close to 6 million babies and children under the age of 5 die of preventable causes ok that's a horrific statistic it's a statistic think of all the tragedies behind that 6 million families every year with teeny little babies and little kids who die of causes that are preventable this happens by and large in sub-saharan Africa parts of Asia especially South Asia and parts of Latin America and for each baby that passes away there's a hundred others that suffer who don't die but who grow up stunted with medical issues mental issues and so on and so forth the issue there is less cost it's the bigger issue is lack of access because when you get into areas like this you're talking object power poverty no electricity for much of the day healthcare clinics that don't work medical personnel that don't exist in these places and if they exist they don't show up for work so it's access to care its cost of care and it's quality of care those are the three things we have to hit AI can play a big role in helping to overcome these so some of the things for example that we're working on at my AI Institute and these are different kinds of issues than the ones we deal with in the US and the UK and so on so tuberculosis is the largest infectious disease killer in the world today kills close to two million people a year more than HIV and malaria combined so we're trying be using AI to help detect tuberculosis earlier by analyzing coughing sounds on a smartphone using AI algorithms we are using AI to help detect tuberculosis patients who are likely to fall off treatment similar to what you just mentioned because those patients are the ones that become multiple drug-resistant to TB then it's really tough dealing with them and even though overall globally they the incidence of TB is kind of flattening out the cases of multiple drug-resistant tilly are exploding in countries like India multiple mdr-tb as is called accounts for only 5% of the cases but it accounts for over half the cost incurred nationally on tuberculosis so those are some of the kinds of applications we're working on high-risk pregnancy you know in these very low-income settings you have a number of mothers who because of lack of care lack of nutrition etc have high-risk pregnancies and those are the ones where the children are likely to die or the mothers likely to die in childbirth we are using AI to try and help those identify those cases earlier and even within high-risk pregnancies which are the cases that are more extreme which are less extreme how do you intervene what do you do etc so I think as we talk about advancing healthcare around the world and meeting the United Nations sustainable development goals for 2030 and health is a huge part of it if we don't find a way to address this bottom 2 or 3 billion people in the world who really suffer then we're not you know we're not we're not serving humanity properly I think that's fascinating it's there is there an argument that in fact though that we should also be thinking not just about this is like almost on a charitable basis but actually if we do transform the cost to treat somebody or the cost to image somebody or the cost to get somebody to stick to their their meds that in fact we are opening up two thirds of the world to a new market that we can make it cheaper but also democratize it is that not the ultimate objective of what we're trying to do in health care you're saying the doctor the number one question I got from people as you're taking my job away when I meet medical doctors it's just really what's better ascent or the ai+ medical doctor in the loop is great especially on all the corner cases there aren't enough doctors in the world we do have the telecommunications infrastructure and having there's a great value there if if we if we can pull it pull it off and then there's plenty of work for you to do in all time zones 24/7 all the time if the value of the half of the world is the market yeah I mean maybe you have to just for the economists we could just justify it in ninety percent of adolescents are living in the developing world so it's the future too we're not having enough babies well you know I just wanted to bring up a statistic that I think is is important it costs two point five billion to bring a drug to market in 10 to 15 years now there's 300 million people that are suffering from rare diseases that given that economic model there's no chance that they'll ever have a treatment you know in their lifetimes and that has to change and I think what's really couraging right now is when you look at a problem from a multidisciplinary approach like the human mind can only process so many pieces of information and no matter what it how much of an expert you are you know as a biologist a chemist and you're trying to you know design and develop the next drug if you can imagine bringing together teams like cross-functional teams with domain expertise across AI machine learning and all those other disciplines that normally develop drugs you bring all that together with you know powerful algorithms and you you you know find correlations and connections within that data that weren't you know that no one was able to find without technology and you have the potential of coming up with personalized treatments that work for individual patients and really trance you know really changes the outcome for that the 300 million people for instance that suffer from rare diseases and under this economic model unless there's major disruption will never have a treatment in their lifetime absolutely I want to talk a bit about you mentioned the big thing personalized medicine which I think a lot of us are expecting is going to become the norm of course if we live in a pants panacea where we have managed to everybody irrespective of where they were born managed to get to the point where you don't drive tuberculosis you don't have a one in three chance of dying giving child giving childbirth in Shaba a lot of these all of these sort of fundamental things that we take for granted I've gone you know like for everybody we're left with cancer and the diseases of aging okay it sounds great but here's something that I've been worried about it's that we are then left with the diseases that are gonna take more than simple blockbusters to treat us all there going to be multiple adaptive combinations of therapies that require very frequent scanning that require you know every time we go to the loo we'll probably have something sequenced about us in order to then change what we do or whatever supplements were then meant to take um but the question that I've got is okay that's great but how do we how do we even distribute the that the value there how do we how do we actually get the money back to the people are developing those those innovations because it's very easy for me is let's say missed a drug company and I say okay I've given you a thousand doses of this pay me that but how on earth do we actually make personalized medicine a reality in there in a commercial sense right now we have the world's largest installed capacity to make biologics drugs we also have the world's largest portfolio of biosimilars and the thing I spend a lot of time on right now is gene and cell therapies we're looking forward to the next decade or two right and I personally have a point of view that we talk too much about personalized medicine thinking that we're developing drugs for an end of one for one person it's never gonna happen in fact the first two gene therapies that are available out there right now cell therapies are available out there right now is for cancer it's Carty we're basically you take blood from the patient's body you fix the drugs fix these cells essentially and can put it back right now that therapy cost anywhere between a quarter of a million dollars to two million dollars depending on which indication it is and the limitation right there is just a processing of the blood right and then getting it back to the patient it's actually not really the cost of getting the gene inserted into that patient cells so the real challenge I see going forward in the next two decades arounds gene and cell therapy and an enabling personalized medicine is how do we actually create the large enough segments of these different kinds of patients and then hopefully with some basic genomic sequencing as before if you can say Jack you're in segment a Mary Lou you're in segment Z and then we'll have 20 different segments and 20 different kinds of medications that it's more tailored than a one-size-fit-all treatment that we have right now right we're still very very far away from a single treatment for a single person I'm struggling to keep up with one size fits all so if you're now saying to me that I have to think about 20 cohorts or maybe proper end of one medicine which you're saying and I tend to agree probably in the short-term might be impossible to actually do completely personalized medicine but personalized enough how do we change the way I'm taught how do we change the way I'm practice what are you guys just developing technologies or you're developing training schemes what what does the future of me look like I think we've heard some really interesting things over here about what's what's happening at the genetic level protein level etc it's two other I think really interesting developments that are worth mentioning so today if I think of some of the most innovative healthcare companies you know who I think of not just Merck and GSK and so on I think of Google Samsung Apple Microsoft which are doing some really interesting things in terms of collecting your personal data to help figure out what your health care so there's a lot going on in that space and over in a very short timeframe in the next two or three or four years you're going to see these companies again Google Microsoft and so on coming out with some really interesting offerings in the healthcare space not worry about your job that's not safe okay I'm still worried the second thing is there are some very large foundations and also people like with the World Health Organization working on getting personalized medicine to the point where it's not just about your DNA and what works on your DNA certainly that's a big part of it but where do you live what's the water quality in that region what's the air quality in that region what's your diet like what are your neighbors like and all of this is available in different datasets and if we can put this together it takes personalized medicine to a whole new level it's your DNA plus your environment your habit your diet etc which I think gets this to a whole new level now you raise a great point track training for all of this it'll come as technologies develop training schemes get developed these are sharp companies that know what they're doing and they know that the only way to get this technology actually used in the real world is to get people trained to get people comfortable these technologies I just add something to that the human body is like the most complex machine that we know of 37 trillion cells so when something's wrong with the human body it becomes the perfect machine learning problem so what you were describing is all those aspects that you know represent the environment which I grew up in and combine that with my DNA and my medical history if you could have access to all that data you've a much highly likelihood of number one diagnosing what's wrong with me and then also coming up with a suitable treatment so I think that that's the we're here to talk about AI and machine learning and the power of that and the impact on on medicine something really exciting happened last week along these lines there was an announcement around a technology that was deployed that allowed you know competing pharmaceutical companies to use the same algorithms to train and use to machine learn on each other's data without giving away any proprietary secrets so it's really exciting use of blockchain and you know the capability to understand and train on other company's data so that you have access to more information about patients and then you know enabling that to be you know used for a global benefit versus just the benefit of one pharmaceutical company there's a lot of potential but the human body is perfect machine learning opportunity hmm I mean I'm still I'm still slightly worried I'm getting less worried but as a result this is a few minutes but I want to just go back to this point about like training because at the moment you know most of my colleagues if they're presented with an open-water scan which they may well do in a not so distant future you never know somebody might buy a headband and start doing MRIs on themselves is very very true or some yeah could well be twins that coming out merrily when can I get my MRI headband beta we have an example optical coherence tomography went through FDA in six months twenty five years ago in fact I was in the ER with a friend last weekend and they wanted to do an ultrasound on his eye and I didn't want to be picky I'm like can I ask a question I'm not challenging your authority I just for my own edification why are you not doing a CT while you're doing ultrasound and the answer was because we don't have OCT in the ER I'm just gonna say that again the answer was because we don't have it in the like last weekend at UCSF Hospital nine of a10 MRI scanners worth down so there was a long line to get into like it's so expensive the liquid he lit like it's the no brain like to read the image that not a hard problem and got a significant amount of AI talents if the doctor can't read the image fine we had if you want the radiologist great I think one of the things is if you see something you trigger in the US as something where you have its interpreted that you have to do a biopsy if you see something but that's because it's so expensive what if you just watch it for a while like sure I'm I'm still I'm still getting lots of I'm still getting lots of your scans like that I'm picturing three years from now right I'm getting lots of scans sent to me on whatsapp by a patient who's monitoring something inside their head this is a whole new sort of level of me having to interact because I'm still the gatekeeper I'm still the one whose neck is on the line it's my kind of insurance if you like if I get stuff wrong about whether I operate on you on they're the ones that die and as a patient you know that's my issue is I don't want to die even if anybody makes a mistake in the system and so like like you can do like the time lapse like boy that's getting bigger maybe I should talk to Jack boy that's getting smaller I mean we're already starting to do that now I'm people are using Apple watches to or various other things that preceded that where there was vital connect or cardio and saying hey look his data and this is actually clinical grade data but the point I'm trying to make is that we also I think need to help folk like me to be able to actually understand and act upon it because your job is on the line for not understanding the data exactly that right like because you know one of the things I watched one of the Elizabeth Holmes movies it turns out with that board of directors in the state of Arizona anyone can get any blood test they want without a medical doctors permission we have fair enough to thank for that it's one interesting thing that they did because as a patient I'm speaking now as a brain tumor survivor who takes a dozen medications every day and I'm not compliant because I'm too lazy to fill the subscriptions prescriptions I'm not compliant because it's too hard to get them filled and that's a long story but here's the thing is like if you present with four or five symptoms 80% of the time it's one of five things let it be and 20% of the time it's one of four million things and you remain undiagnosed for a long time with a lot of pain and people think you're a hypochondriac that's what I suffered with for a decade before my brain tumor was finally discovered what a lot of people suffer from and cuz Jack you can't figure it out it's okay like it's hard if it's almost impossible to figure this stuff out that's what all of this data can help us sort of at least lower the number of those and think of is there a way to test in between the ones that are most likely it's gonna take a while to get there but getting there is important with the data and getting you getting your neck off of the chopping block is also very important or it won't happen I have two answers for you jack one is an analogy I mean like in early 90s I helped a company to launch minim minimally invasive surgery laparoscopically so literally within a span of a year we took penetration from zero to almost 80% around the world right because we built training centers around the world and made it to their clinicians incentive to go and get trained and how to do these particular procedures right so if it's important enough these companies will figure out a way to train the clinicians on how to do it the second part is frankly with this thing the AI will help you do your job I'll give you an example if you're an orthopedic surgeon nowadays you tend to spend your morning doing procedures by the time you get back to your office three or four o'clock in the afternoon you're seeing a bunch of patients sitting in your practice waiting to see you because they're concerned about a procedure it was done in the past the first two things that you do when you go in is you take off the dressing to see if there's an infection of the wound right and then the second is you're checking to see if there's a range of motion that is being regained after a hip for a knee replacement for example so with this bone with this apt i just try to show you instead what we do is we have the patient just take a picture of the wound that image then either gets analyzed by a machine learning algorithms basic say it's fine or it's it's not or that image could go to your practice where an assistant can look at that picture and say come in or don't worry about it and range of motion this this phone here has all kinds of sensors so instead of having to put a whole tractor up to my knee to try to measure how much I can move just do it like this the phone sensors will help you measure the angle and therefore if you're getting your range of motion back some technology can help the patient's just the same way technology can help you with your job and train you and how to do it too so again just to highlight the differences and differences in both challenges and opportunities between the developed world flying AI in healthcare in the developed world versus the developing world so we talked about some of that a little bit earlier but in terms of challenges here we have huge amounts of data you know in all the developed world medical data that we can analyze etc when you get into these low and middle income countries a you find that a lot of the AI based solutions that are developed here don't work quite as well over there because diseases express themselves differently in different populations right different genetic characteristics different environmental issues etcetera etcetera so you have to develop solutions locally the big problem in countries like this is you've got very little data that's actually accurate and reliable so medical data even in a big country like India 1.3 billion people very tough to come by and when you do come by very high degree of inaccuracy a lot of fudged data etc so how do you build the algorithms initially not easy but there are ways to slowly overcome that even when you build the right models the next problem and you Jack alluded to this earlier is trying to find commercial models that are sustainable commercial meaning come up with a for-profit company that can sell this stuff I'm trying to sell this to families that are making literally two or three thousand pounds a year not a month 70 or 80 percent of that income goes into buying food one health issue and the family basically wipes you out and all these families have a have the shoes because they have nutrition problems and bad water and this and that so very so I've come across and I've been doing this for a while and I've spoken to people at the Gates Foundation and w-h-o and so on all of us have come across very few commercial models that actually work at this level of income and that level of income below two or three dollars a day is where the real healthcare challenges are so so again there's a huge opportunity and I I truly do believe AI gives us a way forward over here that we've never had before but we need to be very intentional we need to pick the right problems we need to pick the right partners and then over time hopefully the right commercial models will develop we we've got a few minutes left so I've got a few thoughts which I'll come to at the end as a result of what you guys have said but before we do I want to talk about a question that has become quite like talked about a lot in the last in the last year and that is in and around aging we've touched a little upon a little bit about complex chronic diseases and so forth but a lot of people are really kind of in our circle are almost fighting more for immortality than they are for equality and basic things I wanted to know as people who are in this very privileged setting that we are in who've worked for big companies who don't really have the same problems as those who get malaria and tuberculosis and so on what is your gut feeling on this kind of seemingly growing obsession we have on reaching that escape velocity and living forever should we be directing our technologies to those things even if it is a matter monkey you have more billionaires now than most any time in history I guess if you do dollar correct it and skilled at age so that's what a billionaire wants so that's that's why it's being funded but maybe we can all benefit from it but you know your point like half of humanity is doesn't address down but I don't know if you can stop it so how can we harness it mmm senescence is a factor in a lot of disease you know stem cell death and I think that you know like you said it I don't think it's moral that we're spending you know like people are focused on longevity as an outcome but if you can understand why a cell goes senescent and why it dies you can treat a lot of diseases like you can maybe reverse heart disease and you know there's a lot of good things that can come out of this but you know we've got to get to a world where we're not all you know just in our microcosm bubble here you know the we really live in a privileged environment all of us here and we need to think about that beyond you know our our world you know I am personally very optimistic I mean in spite of all of these challenges that we have I think we're living at a unique time when there's a huge amount of innovation happening and not just in the developed world there's a huge amount of innovation in countries like India and Brazil and China obviously etc there's a huge amount of entrepreneurship happening in these same countries and entrepreneurs like Mary Lou and you know many of us are the ones who drive commercial models and I think there's an increasing sense everywhere that I go of people that we can't forget the half of humanity that's really not that doesn't have access to all the things we do so there's an increasing number of people focused on this there are transformative technologies like AI coming up coming through and now more than ever before governments in these countries like India and Nigeria and so on both at the federal level the central level as well as at the state and local level and now and more than ever before the leading technology coming they would resist it for a long time yeah so I think we've got this great confluence of positive factors and I'm confident that over the next five seven ten years we will see a huge amount of progress come closing a closing thought I think we stand at a very interesting time in healthcare both the life sciences the things that are being done in the labs and so forth are helping us improve the quality of their years that we are here right I have never really focused any of my research or anything like that on extending life all of I've tried to do is to improve the life the quality of the life that one's here and I think the advent of a I really just turbocharges and enhances that by a hundred times well I think it was was its spiderman's said that with great power but can't comes great responsibility and I think with a lot of these technologies we will have the opportunity to do the responsible thing and not just have a few of us live longer or live better but actually make that a new market and something which brings the quality for a great many of us more than just those in this room and I'm really pleased to say that there are four people here who are working towards that and I thank you very very much Mary Lou jo Sunil and Kong for spending the time with us and also traveling an extremely long way whilst you're trying to do this really important work rather than just talking about it please join me in saying thank you to this wonderful panel and thank you for listening you you

Leave a Reply

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