>>from around the globe. It's the Cube with digital coverage of AWS reinvent 2020 >>sponsored by >>Intel and AWS. Yeah, Welcome back to the cubes. Ongoing coverage of aws reinvent virtual the Cuba has gone virtual to. We're gonna talk about machine intelligence, cloud and transformation in healthcare. An industry that is rapidly evolving and reinventing itself to provide better quality care faster and more accurate diagnoses. And this has to be done at lower cost. And with me to talk about This is Dr Taha. Awesome. Who? Who is the director of machine learning at Amazon Web services? Doctor, good to see you again. Thanks for coming on. >>Thank you so much. Good to see Dave. >>Yeah, last time we talked, I think it was a couple of years ago. We remember we were talking about Amazon. Comprehend medical. And, of course, you've been so called so called raising the bar, so to speak, Over the past 24 months, you made some announcements today, including Amazon Health Lake, which we're gonna talk about. Tell us about it. >>Well, we're really excited about eso our customers. Amazon Half Lake, a new hip eligible service for health care providers health insurance companies and pharmaceutical companies to securely store, transform Aquarian, analyze health data in the cloud at petabytes scale, a Amazon health lake uses machine learning models trained to automatically understand context and extract meaningful data from medical data from raw, disparate information such as medications, procedures, Um, and diagnosis. Um Therefore, revolutionizing a process that was traditionally manual Arab prone and highly costly requires a lot of expertise on teams within these organizations. What healthcare Catholic does is it tags and indexes every piece of information on then structure in an open standard. The fire standard, or that's the fast healthcare interoperability resource, is in order to provide a complete view 360 degree view of each patient in a consistent way so you'll be able to curry and share that data securely. It also integrates with other machine learning services and a lot of services that AWS offers, such as Amazon Quicksight or Amazon sage maker. In order to visualize and understand the relationships in the data identify trends, Andi also make predictions. The other great benefit is since the Amazon health lake automatically structures all the health care organizations data into open standard. The fire industry format. The information now can be easily and securely shared between systems. Health systems onda with third party applications. So eso providers, health care providers will will enjoy the ability to collaborate more effectively with each other but also allowing patients and federal access to their medical information. >>I think now, so one of things that people are gonna ask is Okay, wait a minute. Hip eligible Is that like cable ready or HD ready? And but people need to understand that it's a shared responsibility. But you can't come out of the box and be HIPPA compliant there a number of things and processes, uh, that that your customer has to do. Maybe you could explain that a little >>bit. Absolutely. I mean, in practice a little bit. This is a very, very important thing, and and it's something that we really fully baked into the service and how we built Also the service, especially dealing with health care information. First off, AWS, as you know, is vigilant about customers, privacy and security. It is job zero for us. Your data and Health Lake is secure, compliant, and auditable data version is enabled to protect um, the data against any accident collision, for example, and per fire sophistication. If you are to delete one piece of data, it will be version it will be on Lee. Hidden from analysis is a result not believed from the service. So your dad is always encrypted on by using your own customer. Manage key in a keys in a single tenant. Architectures is another added benefit to provide the additional level of protection when the data is access and search for example, every time inquiry a value, for example, someone's glucose level if the data is encrypted and decrypted and and and and so on and so forth. So, additionally, this system in a single tenant architectures so that that way the data, uh, the key. The same key is not shared across multiple customers. So you're saying full ownership and control of your data along with the ability to encrypt, protect move, deleted in alignment with organization, security and policies. Now a little bit about the hip eligibility. It's a term that AWS uses eso for customers storing protected health information or P h. I A. DBS by its business associate agreement on also Business Associate amendment require customers to encrypt data addressed in transit when they're using area services. There are over 100 services today. They're hip eligible, including the Amazon. Health like this is very important, especially for, uh enabling discovered entities and their business associates subject to HIPAA regulations, and is be able to kind of and this shared model between what a the best protection and services and how it can process and store and managed ph I. But there's additional level of compliance is required on the on the customer side, um, about you know, anywhere from physical security thio how each application can be built, which is no different than how you manage it. For example, today in your own that data center, what not? But this is why many cats, growing number of health care providers, um, players as well as I, because professionals are using AWS utility based cloud services today to process, store and transmit pH. I. >>So tell us more about who was gonna benefit from this new capability, what types of organizations and would be some of the outcomes for for for patients, >>absolutely every healthcare provider today, or a payer like a health insurance company or a life. Science companies such as Pharma Company is just trying to solve the problem of organizing instruction their data. Because if you do, you make better sense of this information from better patient support decisions. Design better clinical trials, operate more efficiently, understand population health trends on be able them to share that that security. It's really all starts with making sense of that of that data. And those are the ultimate customers that we're trying to empower with the Amazon Amazon Health Lake. Um, >>well, And of course, there's downstream benefits for the patient. Absolutely. That's ultimately what we're trying to get to. I mean, absolutely. I mean, I set up front. I mean, it's it's everybody you know, feels the pain of high health care costs. A lot of times you're trying to get to see a doctor, and it it takes a long time now, especially with with covitz so and much of this, oftentimes it's even hard to get access to your own data s. So I think you're really trying to attack that problem. Aren't >>you absolutely give you a couple of examples like I mean, today, the most widely used clinical models, uh, in practice to predict. Let's say someone's disease risk lack personalization. Um, it's you and I can be lumped in the same in the same bucket, for example, based on a few attributes that common, UM, data elements or data points, which is problematic also because the resulting models produce are imprecise. However, if you look at an individual's medical records, for example, you know a diabetic type two diabetic patients there, if you look at the entire history and from all this information coming to them, whether it's doctor knows medication dosages, which line of treatment the second line treatment, uh, continuous monitoring of glucose and that sort of thing is over hundreds. You know, there are hundreds of thousands of data points in their entire medical history, but none of this is used today. At the point of care on. You want all this information to be organized, aggregated, structured in a way that you will be able to build even better models easily queried this information, aan den observed the health of the individual in comparison with the rest of the population because at that point you'll be able to make those personalized decisions and then also for patient engagement with the health lake ability to now emit data back on dshea air securely the a p i s that conform to the fire standard. So third party applications can be built also, um, Thio provide the access whether that's for analytics or digital health application, for example, a patient accident, that information all that is very, very, very important. Because ultimately you wanna, um, get at better care of these these populations better. In Roma, clinical trials reduce duplicative tests and waste and health care systems. All that comes when you have your entire information available in a way that structured and normalize on be able to Korean and analyze andan the seamless integration between the health lake and the arrest of the services like Amazon sage maker. You can really start to understand relationships and meaning of the information, build better, better decision support models and predictive models at the individual on a population level. >>Yeah, right. You talked about all this data that's not not really used on. It's because it's not accessible. I presume it's not in in one place that somebody can analyze its not standardized. It's not normalized. Uh, is that >>right, that is the biggest. That is the biggest challenge for every healthcare provider, pair or life science organization today. If you look at this data, it's difficult to work with. Medical health. Data is really different that I siloed spread out across multiple systems, and it's sort of not incompatible formats. If you look at the last decade, I mean, one of the greatest things is we witnessed a great transformation healthcare towards digitization of the record. But your data is scattered across many of these systems anywhere from found your family history, the clinical observation, diagnosis and treatment. When you see the vast majority of that data is contained in unstructured medical records like Dr Notes P. D efs of insurance, um, of laboratory reports or insurance claims and forms with the With With Covad, we've seen in quite a bit of uptake of digital sort of, um uh, delivery of care such as telemedicine and recorded audios and videos, X rays and images, uh, the large propagation of digital health, APS and and digital assistances and on and wearables and as well as these sort of monitors like glucose, monitor or not, data come in all shapes and form and form and start across all these things. It's a huge heavy lift for any health care organization to be able to aggregate normalized stored securely on. Then also be able to kind of analyze this information and structure in a way that zizi to scale. Um uh, with regards, Thio, the kind of problems that you're going after. >>Well, Dr Cox, who We have to leave it there. Thank you so much. I have been saying for years in the Cube. When is it? That machine's gonna be able to make it make better diagnoses than doctors. Maybe that's the wrong question. Maybe it's machines helping doctors make faster and more accurate diagnoses and lowering our costs. Thanks so much for coming. >>Thank you very much. Appreciate it. Thank you. >>Thank you for watching everybody keep it right there. This is Dave Volonte. We'll be back with more coverage of aws reinvent 2020. You virtual right after this short break

live from Las Vegas it's the cube covering AWS reinvent 2018 brought to you by Amazon Web Services Intel and their ecosystem partners hey welcome back everyone we're live here in Las Vegas with AWS Amazon webster's reinvent our 6th year I'm Jeff our table what they did six years two sets people rolling out of the keynote so much action we got another day coming tomorrow they're two great guests here we got dr. feci philomon is the general manager the machine learning and AI at Amazon Web Services and dr. Taha costs senior leader at healthcare and AI at Amazon guys welcome to the cube Thank You thanks itíd that you're here because I've been waiting to have this conversation Dave and I have been we just had an analysis of the distractions and glued up the stack around machine learning so much value now coming online that's been in the works around AI are really mainly machine learning that's creating a I like benefits and II just had to spend a lot of time with key nuts they almost a third of it around a I like capabilities and how Amazon integrates in from you know chipsets with elastic inference beautiful it's just good stuff so congratulations so what does it mean what does it mean for customers right now who want to kind of grok what's going on with Amazon and AI is that new sense the services coming online is that how long has been the works explaining yeah our mission at AWS has always been to take technologies that have been traditionally available for a few special technology companies and take that and make it available to all developers and we've done that I should say that we've done that fairly well when it comes to compute when it comes to storage when it comes to databases the analytics and we're doing the same thing for machine learning and AI and what we're doing because it's a new field is we've got to innovate at three layers of our stack to the bottom most layer as you saw in the keynote earlier has to do with frameworks and infrastructure so this is more for the people that fully understand how to deal with machine learning models and like to go in and tweak these models the middle layer then is for everyday developers and the data scientists and that's sort of where sage maker fits in and finally at the top layer of the stack is where we have our application services and this is meant for developers that don't want to get into the weeds of machine learning but they still want to use make use of all of these technologies to make their applications more smarter so they get the insight benefits get the insights have the day that without getting in town on the weeds exactly who want to get down in the weeds you can get down and dirty with all this other stuff yeah look at that right yeah and typically what we do with the top layer of the stack as we try and solve really hard problems and so customers can now take advantage of it because we've solved it for them and they can just take that and integrate it into their Apple quick what what's the hardest problem that you guys solve I mean traditionally speech recognition is a very hard problem that's one of the hard problems the other one is NLP natural language processing but I would say speech recognition is probably a hard problem and we just launched streaming transcription so you can now transcribe live as somebody speaks and of course you can connect it to translate and translate it as well live so great for our cute beers looking forward to having that on as a health care practitioner how does this all apply to that industry what kind of projects are you guys working on in that regard of course yeah so I mean to to posses point is want to continue to innovate on behalf of the customers across all layers of the stack machine learning in particular this week we launched Amazon comprehend medical particularly in a hardier heart problem where the majority of healthcare data is captured conversation and observations and unstructured formality so petabytes of data is stored across entire healthcare system that's a nun structure for form so to drive actionable insights and to be able to find the right elements to treat patients or to manage a population or even to do accurate billing it's been really an important that we can empower our customers with building blocks for them to build the right solutions to take advantage of that so Amazon comprehend Medical is able to understand the medical language and the context similar how clinicians understand the medical language and context for example if you're looking at a patient medical note Amazon campaign medicals able to with high accuracy extract medical conditions medications tests procedures being done on the patients as well as the relationship between those and understanding that context at this condition and this treatment go together as well as the nuances for example you know a patient has no family history of X or there's no smoking history all those are things in relation in the past or in the future or other members and this is really what we're really proud about launched an Amazon comprehend medical talk about how it works because you know I Healthcare has been a great field around where a is old-fashioned a is a queer when I wasn't doing it in the 80s early 90s ontologies were really popular and it's linguistics is kind of known but now that but you need that linguistics guru to do that he mentioned streaming the transcribed got metadata how do you guys get this kind of benefit when the balls moving so fast around these rapidly changing and verticals like healthcare because healthcare is got a big problem like other verticals where it's too many notifications what I pay attention to so much data how do you put the puzzle together let me first give you some context here as you probably we're at last reinvent we launched Amazon comprehend right comprehend is a text analytics service it helps you look into text and understand what's in there right we started out with general things that we could detect like people places things sentiment the language the text is written in and so on but when we started customers are picked on it and they're using it a lot but as they keep using it they came back to us and said hey it's great that you guys have this this you're giving us the capability to understand general language but some of our domains have some special language like jargon like yeah like take the legal domain for example right it's got charges and defendants and very particular things that are very relevant to the legal domain so they were asking us for a capability to sort of extend the comprehend to include their custom domain terms and phrases as well right so last week we actually launched a custom custom entities feature that allows them to bring in their custom domain into comprehend so the comprehend be extended to include their domain the so legal language is difficult to understand but medical language on the other hand is even more harder to understand that quick right acronyms jargon absolutely what is an entity looks like extracting that and extracting it uses alone yeah miss spells right but relating those entities together is super important because you could in one clinical note you could have multiple drugs in there with different dosages different frequencies and so you need to be able to relate those entities together right and that's the sort of thing that comprehend Medical allows our customers to do to solve some really so you're doing one of that entity extraction is under the covers is that right has it were I mean how does comprehending the medical work I mean just out of the box you have to train it there's no training meet needed know machine learning expertise needed so the algorithm extract these entities as well as the relationship between those entities and then also extracts any attributes that might be related such as negation or past and future or what's anatomy of the body relates one now all that is done out of the box and that's super important you want to know whether the patient's stopped taking a medication right yeah so negation things like that you want to know because that gives you the context just getting the terms alone doesn't really tell you much it each has had a great video about the f1 point of ethics imagine that for personal that's right you're not doing good right now take a break yeah so I feel like we're kind of now scratching the service of stress in the surface of health care yeah information yeah think about the health care industry for years it's been compliance-driven yeah whether it's hip Affordable Care Act yeah EMR and meaningful use right but the industry hasn't been you know dramatically transformed and disrupted and it kind of needs to be yeah how do you guys see that evolving I feel like you're now beginning to see that see change and that's going to take a while it's a high-risk business obviously but what's your sort of prognosis for that transformation and what's the vision as to the outcome yes now that's a really great question I mean one thing I mean one great things happen over the last decade is the digitization of your medical record so and that's really wonderful because before was all paper-based primarily unless you were an acute setting so now the majority of the US for example and globally there's this huge adopt adoption and propagation of these electronic medical records the issue there remains now when the majority of that data is observations and conversations as well as unstructured that that creates a different kind of roadblock for our customers and this is what we're hoping for service like Amazon comprehend medical that's HIPPA eligible means a lot of the early the compliance or help our customer meet their compliance needs that we'll be able to remove the heavy lifting of this undepreciated task about you know having in a large amount of time being spent on analyzing this text and extracting very low we're now with Amazon company and medical be able to really fast track that and be able to elevate it hit the nail on the head of the undifferentiated heavy lifting right that's the ethos of DevOps is that yeah let me give you some stats actually there are one point two billion medical documents that are generated every year in the US and 80% of them it's unstructured text so to make sense of that it's going to enable our customers to do some really amazing things one of the things one of the use cases that we see is its clinical trial recruitment so Fred Hutchinson which is one of the yeah the nation's top cancer research centers they recruit patients for clinical trials if you go to clinical trials.gov you'll see like 290 thousand four and 50 clinical trials open and typically from history we know that most of these clinical trials don't end up recruiting they don't end up meeting their recruiting goals because it's very hard to figure out which patients fit the clinical trial that you're actually trying to perform so comprehend medical helps these customers to very quickly narrow it down expand on the involvement of people in the community mentioned Fred hutch Roach has also been involved what I heard yeah what who was involved in this project sound it was a collaboration take a minute to explain that right I mean it's very similar to a lot of other services that we put it into the market we collaborate a lot with customers 90% of what we do is really coming from customers so we've collaborated with people like Fred hutch and some of the nation's top institutions to help us validate the service that we've built to actually make sure that its meeting sort of the requirements for those use cases that they are thinking of so we collaborate closely with them to get the service to where this today and we announced it as generally available yesterday ok so what's the use case I'll go ahead yeah I can expand a little bit some of the customers as well their use cases we're talking anywhere from hospital systems that when I use or take advantage of their unstructured text for things such as identify people who are for their follow-up appointments or stopping treatments or find an alternative routes to billers we're trying to identify it is accurate procedures were done if we account for all the procedures or care for all the billing which often time is hidden in those unstructured text and require a lot of manual process and often time the rules that can't really scale to things such as clinical trials recruitment how can you if example in Fred Hutchinson Cancer Institute use case for identify a patient and match them to the right clinical trial these patients often time have Harry Potter's worth of clinical notes down on the minute their longitudinal journey and to go from one institution another another and be able to really find it's no longer needed a haystack it's like a needle in the bottom of Atlantic Ocean and then be able to really do that match from hours and months down to a few seconds and that's really the beauty about the service John likes to talk about the 20 mile stare and I wonder if we could just look ahead how far can we take AI and machine learning in in healthcare and how far should we take it and maybe a more specific question as as a practitioner you know when do you think machines might make better diagnosis than doctors if ever how do you feel about that where do you see this all going I think I mean the whole idea about machine learning the beauty about it I mean the seta scope was introduced or how the thermometer was introduced in medicine and these are tools that we use to our advantage to really provide better care and and better outcomes and that's really what we're that's the mission that our health IT and customers and wanna are really driving tower's machine learning can do a lot of great things for routine things that human being can't can go and focus their attention to other things such as the Fred Hutchinson instead of going and mining these diagnoses in mountain amounts of data a machine learning will be able to identify that with a clinical staff can focus on care and that's really where I think I mean over the next decade and so we can see a lot of this advancement in in these building blocks as well as what Amazon's offering from forecasting and prediction algorithms Rana will be able to find you know fine-tune our capabilities to help customers achieve even precision medicine real-world impact because you're changing the workflow I mean someone's within the wrong line or the wrong process based upon their history yeah HIPPA HIPPA requirements really cause a lot of this record sharing thing to be a problem from what we've been reporting over the years it's kind of a solution to that so if I move to a service medical service I get all that records with me it's just kind of how you see going and how does other regulations that are holding you back that are blockers is that clear now how does that solve the industry challenge it's of privacy and if you look at the healthcare system today there are lots of inefficiencies in there right in the end this is all about improving patient outcomes and making sure that we reduce costs and that's what this boils down to and these are tools that allow our customers to do exactly that well guys thanks for sharing this insight comprehend medicals really awesome opportunities I think it's early days day one is you guys think right I think there's so much more that could be there I'd love to see the industry just from the personal is decided change it's just get out of the way of all these pretty broad hurdles get the data out there expose the data check the privacy box would be good right this is gonna change the game yeah maybe we should say a little bit about the how we built the service in terms of that right as you know at AWS security and privacy is number one for us right so this service is HIPAA eligible it's a stateless service what that means is nothing gets stored this is not the data is not used to improve the models or anything like that the only person that can actually see the data is the customer he's got the keys he's the only one that's sending the data to the endpoint and whatever he gets back only he can decrypt it so we've taken care to make sure that we can remove some of those hurdles that people have always been worried about well doctors take you so much for sharing thank you so much for having us here we are bringing you all the action here from 80s reinvent again as the compute power is increased as software is written with new apps a eyes changing the game of course the cube a lot of video we don't need some of these services to make these transcribes on the fly they succumb and I really appreciate it you think back on the more after this short break [Music]