Um, you to today's Brown Bag. Quick administrative announcement next week. We currently do not have the brown bag speakers scheduled. Check the website for any updates. But at this point I'm expecting that we will not have a speaker for next week, but we are delighted this week to have Dr. Mark Braunstein as our speaker. Many of you already know Mark. He has been a professor of practice in the School of Interactive Computing. He's served as a founder of a company and the health IT space. And for years He's taught Georgia Tech's introductory Health Informatics course. And he's also just recently completed the condition is Health Informatics on fire. His talk today will highlight some of the most interesting innovations he discovered and y to get there. So with no further ado, I will turn it over to you and mark your camera's off just so you know. And it looks like you're unmuted also. Xlab. I hear you but I don't yet see you. I love that. They try and should be working yet. We see it now. Okay. I'll let you take it away. Are you seeing my screen as well? Yep. There's no I'm not saying it's the size. So we're seeing your PowerPoint. Are you sharing the entire screen or just the app? I'm hearing? Yeah, Let's try that again. Sorry, folks. We tested this out. I got started, but sometimes we have these minor technical difficulties. We'll get it sorted out here shortly. All right. I see your slides. Yep. Yep. Okay. We're all set. Yup. All right. Thank you again. For whatever I wasn't around that happen. I have given you Brian bags in the past and through them we've traced the history of health informatics. As I recall, the first time I presented, I could talk about was a vision of where things might go. And now in present, where things are in a pretty exciting glimpse of where things really and go where they are ago. He said, I recently completed the second edition of my textbook, 8100 COVID isolation, doing that. Looking around the world are interesting examples, the highlights of it. First of all, a couple of definitions. Realize many of the health informatics is the application of informations out technology to the healthcare delivery system really about taking patients either one at a time. As a group, say all the diabetics practice or across regions, states, nations of the world. I divide it into four subdomains. Electronic records. Population health, which is the groups within a practice. Mhealth only exciting things going on. Biases and public health. For all pretty familiar with these days is resulted. Fast Healthcare Interoperability Resources, really the focus of our arrears. Now, focus of this talk is the first successful technology representing very outdated is inordinately complex. Human body is more complex than any device we created. And we have a relatively incomplete and ever-changing understanding. Um, so it's a real challenge. Come up with this. Those are representing health, health care and human body in a structured way. Buyer is not really about that representation, was about how you package better representations being able to use into a granular data model and make it shareable using a rest API. I know this may seem pretty obvious to you because everybody else but health care was doing it for years. But this was a big innovation in healthcare when we first introduced about eight or nine years. There will hopefully fires about our package data. How to share that. This is a surprisingly Oldfield. Most people probably wouldn't assume that. Help him familiarly says really dates back to the nineties, eighties. It's been around quite a bit longer than the first use case was high, you turn illegible and written notes into computerized notes. And here we see Dr. octo Barnett, who spent many years Massachusetts. He's not with us anymore. But he developed one of the first systems for doing that. If they're laboratory mouse. In the 1960s, ibm and the akron Children's Hospital, the first hospital electronic health record system. Any of you who were all from Earth, the feel that probably heard about Epic and Cerner, the two major hospital to try and health record systems. But the very first one was this hospital in 1906. That's me right there. Back with my 60s hairstyle. I developed one of the first ambulatory electronic health record system that's used by doctors outside of the hospital. Going into literally a visiting hunger's house. I've been seven is the other major early use-case political decision support. How do you provide patient-specific information? Elegantly filter at appropriate times and today at intelligent filtering and hated them a lower right ear bow many times now involves data which is very complicated. Information. While this field is even earlier than the electronic health record, the first to do clinical decisions or scarring 90. Hi the late Dr. Homer Warner at other nice things as Utah. And he developed knowledge rules. You see an example of what life does this patient have one consolidation of finding that is suggestive of the vote pneumonia, a halo with six physical exam findings that if you had one of the first two and to the next or the diagnosis of lung consolidation is appropriate. Notice they're all physical exam findings these days. We use imaging, MRI and graphy and those sorts of tools to do this sort of on the back of his office. And he built this knowledge base over many years by interviewing experts and all the various domains of health care and built it into an early electronic records system helps us, was actually used at later-day since the early. Another example was mice and some antibiotics, and their names end in MYCIN, like erythromycin. Hey, telescopes the purpose which help physicians use the right antibiotic and developed by Dr. Ted Shortliffe, who's still active this VM for university. He built a system of about 500 production rules, rules that were applied to Asia then AP tests and even BATNA. And he showed that mice him operated at roughly the same level as infectious disease specialist doctors who were trained to do just this. Other better. It says general practitioners. Now, you'd think that that kind of performance, this stuff was used many years ago. The matter is, or was. This is a graph of electronic health record adoption. And you can see that there was it 2008 wishes when the Obama administration best at $30 billion of stimulus money, you get hospitals and physicians to adopt electronic harking back to the very small percentage have fire came along about here. When it became clear that these systems didn't have any way of talking to each other, sharing data. And there was no good way of accessing the data. They can, they develop. Apps prove their functionality. Buyer is the first, not the first attempt to do something like that, but they all have only one base. So just to emphasize the point, government classify EHRs in three ways. Only the most sophisticated were theoretically capable of providing one. Or wishes argue arguably, most importantly, the HR can do. And although we had pretty broad adoption of EHRs by 56 years ago, only 40% of them even have the potential. Not many of those were actually, uh, why, why is this interesting industry up until then, at least proceeding like the proverbial waters. Other, there are several reasons. The first are perverse financial incentives out here in the United States has always operated under a PE or procedure. This more things hospitals and doctors do, the more money than me. That's not to say that everything is governed by that. Put, how are you going to go to hospital who invest tons of money? Electronic health record system, which, which might actually reduce the revenue. The government fix that by giving them the money. Essentially Navy mandatory do. The other reason they were reasons they weren't broadly adopted more. They didn't fit well into the integrated into the workflow and process thing here, patients, that workflow process, as you see here, is itself pretty complicated. And a study in 2015, the EHRs were widely deployed, concluded that they really didn't improve efficiency and even mind reducing. Third reason they were accepted was duplicate data entry. You might have an EHR, but if you wanted to do is use clinical tools like the ones shown here, to calculate scores that are useful in making decisions about caring for patients who are on the right to submit quality measurements. You have to do this as a separate process, reentered data. You don't know why now? I contend it's all happening now. Why why why isn't tours become the hair? Well, going back to the financial basis, the government since 2008, about the same time as the HITECH program was introduced, has been working on changing that model to something that I would call value-based care. There are a number of phrases for. Basic idea is that hospitals and physicians are rewarded financially if they produce better results, outcomes at lower cost. Combining the two is very powerful book is, you know, they're not using cost. I not doing needed things. When needed things. The outcomes of fire came along and it sells workflow, process issue. This is a real fire up those we are asked adults in Australia, hey guys, pharmacists, physicians and the management of complex occasions. Here on the left, here, that it's menuItem, those me highlight built into this. Amy, I knew that everything that users use to access everything else. So to them it looks like it is the EMR and you wouldn't know it, but that big white area where it's running is where everything else is, time where information is charted and tare the information. So put the two together is write it in the workflow and process. It actually appears to the r to the n1. More importantly or as importantly, because of the platform that are run on, we'll talk about that platform. It can access the data in the EMR so no more duplicate them. And lastly, the government or this, which in 2019 at the meeting, how big industry trade shareholders before COVID had 40000 people below this year who I just deal with Burma and the administrator of CMS program that pays for Medicare and Medicaid, the biggest By far in the United States that we're going to use fire to implement our policy. Basically said to the industry, you gotta get more. And the most important example of that was the way the government has implemented the 21st Century Cures Act. So this was a law by President Obama in 2016. Near the end of the second term. The idea was to give patients digital access to their medical record data in hopes that they would use it to help think better care of themselves. Would share it with other physicians that are seen by other physicians. And they might even contribute the data. Research. And the law said that hospitals and planets had to provide well editor via gateway. Digital gateway. Patients can get that data. The implementation of the 21st Century Cures Act, which occurred here once after seeing Rubbermaid Then state and said, and you've got to use fire to do that. Well, if hospitals and clinics Hampshire provide a fire gateway, so patients data, now these are provided with electronic medical records is after provide their gateway. They all do. And you can see recent, this is from 2019. Even, even then, couple of years ago, a 100 percent of hospitals have an operational fire gateway. You a lot of regions of the country now, red and this next year, when demand they actually takes effect, it's April 2021. And fires being adopted globally as well. This diagram of the graphic to the happy fire. Fire is the most widely used implementation of the fire standard or a web server done in Java. Things Agnew who developed it was kind enough to provide me. Show just how widespread adoption Israel, the dark blue in India, probably means there are a lot of contract program is it doesn't fire. Probably doesn't all represent US and India, although the government is also sorry, fire. Understand why it's been successful in as, in one small way. I'm going to go pair and out of the prior interoperability standard, CCDA. You see LFCC day it was document vase and you can see documents contain bodies, inspection entries. It mimics the structure of a paper document, not granular. If you want to see patients record, you get the whole thing. The CCDA document, maybe a 100 pages long. It is pars below with some effort, you can go in there and pars and find what you're really interested in. Oil painting is Georgia Tech students back when my course are using CCDA because that's all we hear, is even broader than CCDA. Hidden call. This is everything from the technology base you need to implement a web, a standard, all the way to the data model. You need to support clinical reasoning. Yeah, it's all done in a consistent way. That data model is built by something, all resources, but we'll look at those in a second. And it is granular in yet just the resource you want. So all you care about is the patient's medications you can get just that. You can ignore the rest of it. You don't even have to know the rest of it works. But let's understand what a fire patient resource is. We'll use the patient will have a fire resources will use the patient resource. This is the machine readable part of it. An XML JSON is more typical. I wanted to show you that you can do it next now. And if you stare at that for a few seconds, you can go to figure our catch. This, this is Henry V. Is the seventh male born on September 24th, 1932. He gets his gear, the good health clinic. For a particular use cases you can extend the resource and the extension for a renal clinical trial. There's also a text representation of the information in the resource that's provided. So that if a system actually capable of digesting Fire Resources, guess the resource. Programmers can just grab the text description, provide next venues. Finally, there's metadata, of course. A key piece of metadata being the resource ID, every resource as the unique ID on each server. It's not unique beyond that, certainly those. So if you start working with multiple fire servers, even though you may get the patient resource for the same patient from several different servers, they're going to have a different ID. We lack a universal patient ideals, country. Obvious way. The bottom there I mentioned the good health clay. More carefully and you'll see that's a reference to a different kind of fire resource to care provider resource for Fire Resources, reference other fire resource and look at that. Or I move on. I said fire isn't about how you represent health care and if I package it and how you access and share. There are standards though that deal with how you represent it. Well, let's import our ICD-10 CM. Hello, We'll stay there for Rural Health Organization about how you represent diagnoses. Diabetes has an ODE or failure, has a code, so on and so forth. Or laboratory tests like a blood count or clinical observations. This is the doctor. Is that your skin is yellow. The system of codes called loin, LOINC. That occasions here in the United States have a coding system called RxNorm. There is no global standard for medications, although I the next group stove it is working on one. We being used. Navy of the United States is the Uber ontology of healthcare. It represents medical concepts and their relationship. Well over 300 thousand concepts, over 1.3 million relationships. Very big, but it's very rich and use. Look at it in a second. Before. I said that fire resources reference other resources and create network. Here's a graph of one medication request decryption network. Obviously it contains the medication. There's the part of the resource since the medication. Okay. And you'll see the name of the medication, which is the Zehr from mice and it's an antibiotic. Ending is shortlist, the famous clinical decision support system. Right above the name is it's a little bit of code and URL that tells you if it's a snowy. You look more carefully a little further down the dive in now, which tells you this is actually a Dutch formatted resource. That's something called a profile, which we'll talk about that we're going to go to. So we're going to use the Dutch or the browser to look up the color. Do I found that? Yes, indeed it is. As if from mice. And there's other information. As to in moon. Obviously you need to know for which patient the prediction was prescribe. And there after patient preferences, their resource ID on that server. Id can be alphanumeric, has this one is in it can be numeraire. Really completely up to the local server were implemented. Do I really want to politer to the patient, visit or encounter where this prescription was written to the practitioner who wrote the prescription and the reason it was prescribed. And in this case it's traveler's diarrhea. And that itself has a snowman. But we'll just see there's a rich interplay between resources, in between resources and the Coding System. Or I can provide examples of references to all the coding system. Now we'll go to the happy server URLs up there. Those are your ish that should go learn and play with it here I asked for a list of all the patients. Server. Of course. And I liked the Read button for the very first one. Notice Asian resource ID on this server is number 1, 6, 8, 7, 2, 3 zone 1, which is more common actually, um, I click the read button and go here, and here's the URL. Api fire or regenerates in response to that. And this is a good way to understand the way a fire rest URL, this structure, the green part is the server. Happy fire.org, and this FHIR.org base R four so high there are several generations of fire out. There are four is the latest. How to revise or it's our far so the versions of the server that support all the earlier and that's part of the URL. What resource you mentioned it in there. Over a 100 fire of resources that this White. And then there's that number 1687237. The resource ID precedence says I'm actually in whole patient history still getting not one resource, but a lot of them. But the first one that shows up is the patient resource a is, in this case in JSON, which much more typically used. And it begins with these two lines which confirm that yep, this is in fact a patient resource and it has the id. Then I asked was Everything is another element. Is the value six O's that have been defined by. In this case a country HL7, the organization that helps to develop Fire, built a lot of value sets and permission profiles earlier. So this is a Belgian buyer patient resource. You're getting a sense of this. This is quite global. We've looked at one from Netherlands, Belgium. And resource tells you that it has been constrained by a profile. There's the URL to get to the profile. You can actually go there unless you have credentials. Here's the URL that I was able to get it somewhere else and that's what it looks like. Here's just a little taste of what a fire profile does. The first line there, it says that you must use National Patient ID as the fire resource identifier so you don't have the problem. Yeah, It's a different different servers have here in the United States where we don't. Back it says that you must constrain marital status to the values that said by HL7. And further now you could probably not legal in Belgium to do it, but you could specify an extension of the fire patient resource indicated for patients race here in the United States. Almost always done but in many countries. So that's five. But we're really interested in what you can do with it is largely about fire ax. In order to have fire apps that have to deal with the problems that if a vacation and authorization because these apps are getting protected health information, very sensitive information. Oh, group at Harvard, Boston Children's Hospital. Smart, which they converted the fire with. Fire came on. And it uses technologies you're very familiar with. Do establish identification and uses Open ID Connect pane technology would use boggle your website using removal of Facebook credentials. This case, it's going to use the hospitals credentials. And it uses Auth to indicate what data you have access to. The app, not user has access to. Identify the user. And then we're giving access to the app. And this little diagram just shows that the apps that have both read and write patient data. Does this happens when you install an app on your smartphone? Sometimes you're asked to prove the app, getting the specific access to data that you want. And the right here is how the OpenID Connect actually appears when a patient is asked to link their medical record, day, the sun to back to the four domains of health IT. And we're going to focus now on the first. Many of you probably have an iPhone. And now history is is correct because I do this every time it gave Can you say Oh, the students? Yes. Hello hands, how many of you know that you can get your entire digital medical record and your IPhone. Well, I see names. I'm not seeing videos or how many days? Oh, you can't see that. Well. Usually it's a small percentage of people in the audience, but in fact you can. And this is an example of that. I'll show you on Android. We're doing it later on. This is real, this is for my friend theater fear, fresh off. Happy to share. I need medical record data. Importantly was the inventor of Netscape and was widely used phishing website. Now, his hair comes from Weill Cornell Medicine in New York. It shows the left is what the patient sees. I'm going to write a second. And ears their patient view of this particular he was immunization for the zoster recombinant vaccine. When it was given is completed. The PBX owed for the next thing on the right is what Apple guy going to the fire gateway for Weill Cornell. And that is, that's how they do this. They didn't invent the infrastructure that makes possible the iPhone, they're just taking advantage. And this is a fire resource, looks a bit like the patient resource we have our earlier, but this one is for immunoassays. Now in how Kit Apples data store for health data, they have their own version of fire resources. They all begin with. Hk. Well, these HK, I don't know what, I should know what Apple calls the host, let's call them modules. These modules can be mapped to fire, and in fact, Microsoft has developed a Swift library to do just that. It's open source, it's on GitHub. So if you want to access the health data stored in the HealthKit, austere and an iPhone, and run a fire app on their iPhone, something that Microsoft Apple is now supporting. You have to go through the approval process. You can use Microsoft Cognitive. Ml Theta has It's as though validity. Parents have been a number of studies that show there's much more variation in step counters and even things like rice is they take blood pressure, then there should be a volume issue. Patient could obviously you could take step counts 50 times a day. Their physician who sees them every 90 days is finished. And now they want some sort of bottom line. And as a result of all this, there is a need providing data that has real value. Z position. Open mHealth developed at the University of California, disco in Georgia Tech's down involved with this, socket is a standard for packaging and data. And making were useful. Here is a standard for providing a mean, an average over a specified time period. On the right is an actual smart fire have developed at Georgia Tech, Dr. John boots. Through it, leverages all that and the ability to wrap in the house the fire is a useful app that a physician I have within a group that created it's smaller and it will essentially, this hospital is also interested in this problem and they develop a PG HD patient-generated health data and metadata repository that has mapping tools. Are there tools that allow you to, in theory take anything from structured surveys of patients to active tasks that patients that assign to their activity data. Data from devices such as blood pressure monitoring, data that their phone may have collected and map it all. They arise at all and represented all as phi resources. They currently have. A number of examples of this is actually their whole range. They are hoping that entire list. And that encompasses all of it. We're going to actually look at an example of a survey. Later on when we look at a wildfire app, though, at hammering. Purveys are increasingly important. Understand patients are doing. They're not in the hospital or the doctor's of all patient reported outcomes. Pros. Paying more attention. I'd say mHealth. That real quick. No electronic records. Here. Fire is also still really crop up. Duke, he is an AI driven or fire at the incubated at the American Academy of Family Physicians Incubation Center. Physician picks up the phone and box the SUV and says, non-being, this is where reveres as medications and there are fish in my notice that two of them are diabetes medications, but we might say Sukh, HOW graph of hemoglobin A1C over time, even though you love an A1C is a lab test, it is very good way to track that. Well, how well control the patient's diabetes last three months. Obviously, they can also dictate core is yeah. Connects to their EMR, whether the mother Rhea, why isn't far. Here's a mark fire out and developed at Emory. Emory is now a center for app development partnership with Rimini, the database company that began ATGC. Now a graduate out of their own. And they along with Emory, are actually developed some commercial firearms. Going to look at those here, they're being offered. This one allows position, this case, an orthopedic physician. Look at their own patients that they, that fall into some Parameters, they specify how their outcomes for this case, it's some aren't being surgical procedure. And you see down the list of outcomes of a right pain. So pain control is an important measure of success. As is this SF 12 V2 nervous this image, That's the patient survey. And I say, mention when we talk about, here's another, our app developed that every deep brain stimulation is a very sophisticated surgical neurosurgical procedure that has complicated pre and post surgical regimen, keeping the patient straight, making sure they do the right things, right time, keeping them in sake with a dirge and making sure the H and O what they're supposed to do when and actually do it allocated. And they've automated the whole thing using a patient by our app on the left, running on a phone and a provider firing app, right? Running within sir. Here is an example of a IRF ready within an EMR. In this case it's a levy or more because nobody MB cout and more than provide a screenshot really are more earns. The EMR vendor might get upset. But these, these are various scoring systems. This one is the risk. How to go about assessing the risks the patient might have atrial February a stroke. Well, given that they have atrial fibrillation, I'm in cardiac arrhythmia. Based on that, risk, going to be more aggressive or less aggressive. And the oxygen providing these scores for all time as follow layouts, but they're not connected via HR. Position as well. Grab and pick up their phone and they have to enter data that they've already enter. Much more attractive when they're built into the war. They can go grab that do. Here's another example of helping Dr. Crowe know her well regarded EMR. That's actually built up IRR for the patients to the point that provides all sorts of useful services to the patient making. With appointments. They can invest their doctor, they can look at their health information. Thank you. Even get alerts to the initial problems like an interaction, drugs their day. There are fire App Gallery's the Absalom, those galleries maybe free or you may have to pay for the gallery on the left is epic EMR that's used in a variety of awards teaching hospitals around the country. Emory is now switching from Epic. They can play with that all the major health systems land and will be on Thursday. Another major EMR company. All the military and veterans hospitals either have or will answer. And medicare Federal Insurance Program for people like me who are over 60 high as a fire up gallery. They don't have the medical records. Medicare server, but then you have all the claims. Medicare pay for all the care. They don't have the level of clinical detail, the humerus down what? They are in one place. They have access to everything that was happening. But there are lots of apps where that smoke maybe more important than having. Will now move ahead. And we'll take a look at population health. Population health is rounds of the new areas if people actually doing it, although we were doing back in the clinic where I work, it means that it's just looking at each position as fragments with looking at groups of patients need to be monitored image. In this case, Asians are high risk for COVID. And they all been vaccinated and they had a COVID test. Is there anything either a profile that needs to be brought to the attention case manager who is taking care of them. Or can you see my mouse? Now? This is this app that was developed by remedy, the same company miss working with them. Every commercial app they provide to the health system is the alert comes about because of a technology called CDS Hooks, which comes from the same smart grew in Boston. We talk to them and I've got several slides that say activity in gray. They all have a URL. These are public bikes that any of you students might know to play around. This one's a demo of how CDS though it's works. Lower left there you see that information, Then it in what's called a card. In this demonstration reaches their RX, you, they have a patient do the imaging view. The user's selected hypertension, they prescribe brand-name drug for hypertension. And CDS Hooks, which is logic that's built into the EMR, comes back and says the provider, if you prescribe generic, you can say $79. And if you push this blue button, will, I'll change it for you. Here's another interesting EMR technology from the same smart group and Watson bowl bought boat data. Or fire was developed really for the care of patients one at a time. What about if you're doing a research project that you ought to coal through thousands and thousands of patients and Emory looking for a particular subgroup of patients. Well, then I'll let you do that against their live EMR because it would kill response time. However, the while data is loaded with offline process that runs when the ABR is busy. All right, Can you run against the data warehouse? Were a copy of all the data in file format is stored and what we're using. And there's a lot of interest in this. Aws, Health Sciences South Carolina, which is the Google bot major health systems in South Carolina headed by EM USE in Charleston as a statewide data warehouse. And they're using more data than nine patients in that warehouse who need to have more aggressive, that's the not reach them because they're qualify for a vaccine. There's our public fire data or data access server. You can play with this URL. The results come back in a format called AND line delimited JSON, which is designed specifically for data represented in JSON. And then there's a little bit of the results for this query I did against the mart server looking for pace for imaging studies for patients who have a particular Blue Cross, Blue Shield play. And the JSON, He's easily converted to more useful forms like CVS, and you can then do an analysis of it. And there are a number of public service. We'll move on to the last of the four subdomains, public health. This is the CDC and the World Health Organization and all the people who are very busy these days trying to do with the COVID pandemic. But there are important we use cases to which fire, fire apps could be use. There are medical problems that need to be reported that probably certain infectious diseases, sexually transmitted diseases, COVID. And the process for doing that right now is pretty labor-intensive. Often putting in, you could obviously Bill fire apps. It's the EMR. Cds Hooks, the detected this page it as something that needs to be reported. Bring up the fire out, and ask the provider any additional data that the app can't get automatically from the EMR. And build the electronic report. Then back to public health and did some prototype projects to do that. So far, this is not yet being done in real-world. World Health Organization is deploying fire apps, not smart apps. These are not connected to transform their more guidelines. And here this is their term, paper guidelines. It is more guidelines. They don't mean the smart me were talking about earlier. This is one of their guidelines and antenatal care guidelines. These guidelines are used in third world countries by health workers who are not trained, they're not nurses or physicians. And they need the support of these guidelines. Knew I knew what to do. Fire app makes that more dynamic. It tells them what needs to be done based on what they know about occupation. And the guidelines becomes smaller, which is why WHO? They're in use in some initial use cases. Another boundary out of the smart group and Boston, we're back to that depth. Our health courts, which contain parts of the digital record you see on the right there. Or digital or paper card that contains QR code. Or they can be storeroom phone. The obvious use case is your COVID vaccination status. And I'm showing you here from left to right the actual sequence and screens that a agent using common health, which is the Android equivalent to Apple's HealthKit. Have an Android phone out there, overtime and health. And you can download an app and you can connect to your health system. You put in your portal credentials. If you don't have Portal credentials, can't do it. So you need to go to houses, get some jazz. And you can, you can get it on your phone. If you have an iPhone, apple as the iOS 15, which is now either now or comb. It, which is now supporting our health cards for vaccine amines project, the open source community to develop common health with you just love that. Already has an app in the Apple store. It verifies. Um, yeah, well, we have the four subdomains and health urbanized. So we'll just look at some stuff that I think is DICOM is the universal standard for representing medical images. The actual images are stored in the DICOM file along with a lot of metadata, contains important information such as who is the patient. What they don't contain as clinical data about the patient that's in their EMR. For Microsoft has developed something called DICOM on fire. Process is DICOM metadata and information for the patient's medical record. Pinker creates a fire resource, be attached through the image that provides a comprehensive view. Clinical spectrum. Spectrum around this EMR is particularly useful for patient care and for me, sir. This is yet another open source project. Anybody go to the GitHub repository? M code is a HL7 initiative that takes the codecs project, which is representing cancer care in that set of Fire Resources extensions. Alright. This is more broadly in order to increase patient involved in clinical trials. Currently it takes forever to develop and test the occasions more clinical trials, more participation of hill. Yet only 3% of all cancer patients dissipate in one. Part of the problem is that EHRs and patients with EHRs have them. But couldn't you go fire routes that use an extended version of fire? This is a roadmap of data. They're going to add hiring codecs. This is the overall theme of the ontology that they proposed was pretty complex. Answers, pretty complex. But the vision right there is that patient portals and send him codeine it into matching the services using fire. Be able to present clinical trial mentions back to the providers who speak. Why EHR vendors can help your clients or patients. That individual cancer patients were electronically the match to clinical trials, which they're a good fit and enormously complex process of figuring all that out right now. Hopefully, that would increase manage a patient. Management. Another cool part of the fire technology and one that wasn't developed by the group in Boston is clin fire. Those of you who are familiar with XPath. This is a very similar language. Britain are fine. Air, clean fire. Will visit flaming fire hopefully. Are certain. Feeling great, Ryan. A one of the things we'll do is create a fake passion for you. Here are gone all the way down. I've gotten back hole patients record. There's a fire bundle aggregation of Oliver Fire Resources. And I can write, and there's the bundle for here. I can write a fire path expression. It's pretty easy. I was able to do it. And it produces whatever. 19 reduce by exploring the bundle, in this case, looking at the LOINC code that ice, I said it should look for in grabbing the patient's pulse value. In this case, it won't go boats for and presenting just those though this is really useful stuff. For example, people who are writing apps are all you really want is the Pauls Valley, Michigan crazy graph of me by this simple fireproof brush and go through an entire medical record, get exactly what you want. Here's an example of this in the real world. This is, I'm sure a Health in New Zealand where they go through the patient's record using fire path. Great, what's called an adaptive questionnaire. This is a questionnaire that patients are asked. Yet their past medical history, lifestyle, and where all that information is already stored in an EMR. It can be presented to the patient as both toys. That's not correct. They can change it this out. Part of a very interesting project called the Somme gene network, which wants to give Everyone in the third world digital laxness to their medical records. Part of the project is called Turning paper into fire. Presentation made by this guy, the VA error sera, who's the head of the work project in Rwanda. Friend of mine in this way, he's received the norm is notoriety for his work and he's using Mark, I just want to do a time check. We're pretty far over time at this point. I'm only got to feel free morpheme. Okay? You've gotten that India has said we're going to use higher. We're going to give all the Indians a visual record. This is a prototype patient. Here create your European Union is working on something similar. This is a provider out of those areas, has been in genomics. There is a big complicated project to figure out how to represent genomic sequences in fire. Importantly, answer here. And maybe the most ambitious project out there is an attempt to use fire, turn medical evidence into computable fire resources are currently focused on COVID. They can turn any clinical trial and that governments clinicaltrials.gov site into computable fire resources. Even more interestingly, they can take the results of that trial using their goal and turn the results into a fire resources. And here's the example of that statistic. The number in this particular clinical trial, 541 females were placebos in this trial of a mess. I can never pronounce that a fire drawn. If all heard of men and 35 percent of them have the Vedic, we're fine. Finally, I think this is the last slide. If you're interested. This stuff, There's a website called Lantern. There's its URL maintained by the federal government that shows you where all the fire endpoints are out there and gives you a lot of useful developer information about them. This is the last one. Well, Apple, Microsoft, Amazon, and Google have not, this has not gone unnoticed by them. They've all got made reference. Already talked a bit about what Apple's doing. Microsoft and Amazon are trying to be the fire development platform of choice. And Google's do a lot of stuff with Firebase then that's it. Thanks for a great talk. We're starting to lose people in the audience. I think given that we're about to, that it's over. Maybe I'll just take one quick question if you could briefly address it. And there were a bunch of other really good questions, I'll ask you, maybe just to send those directly to Mark afterwards. They can try to answer those for you. So first question we had the chat was as mater legislation like HIPAA actually made interacting with patient data more challenging. But in the 70s, talked about some of the plus sides, a fire, they're sort of downsides. The other legislation, I guess, I guess it has of the sentence. The same time. It's granted the rules that allowed us to have digital access. The reason that EPA was created was because the federal government, why the low points processing here in the primary example of that, the digital technologies and they knew if they were going to do that, they had to create a set of rules that probably access to that thing. It was actually created to facilitate digital access, then I don't know how you would be able to have the kind of visual access we're having today. If there weren't worried am rules. I know it's frustrating. Well, thank you very much. That is. Great talk with a lot of, a lot of interests and good attendance from the, from the students. So let's all thank Mark or for the great informative presentation. And I will see you all in two weeks. So remember lowbrow back next week. So I'll see you in two weeks. I believe dopamine, it might be our presenter then to see you then.