 All right, our next speaker is Dr. Vinod Agarwal from MD Clone, speaking on overview of MD data platform. Thank you, Dr. Dev. Welcome, everybody. Once again, good morning to this wonderful conference on using VA data. My name is Vinod Agarwal. I'm a physician and part of the clinical team at MD Clone. Firstly, I want to thank Dr. Amanda Lee now, who's our host and sponsors the MD Clone Adams platform within the VA innovation ecosystem environment. I also want to thank Dr. Sandesh Dev for hosting us here and providing us the opportunity to show you why you should get involved in VA projects if you haven't already been involved and why the data at the VA is such a goldmine. I'm going to reinforce many of the concepts that Dr. Lee now has made and Dr. Sandesh Dev has made and Dr. Schwenke has made. And in addition, talk about certain special features of the MD Clone Adams platform that really have benefited clinical researchers so far. I'll focus in on these challenges faced by clinical investigators and why they use the MD Clone Adams platform to overcome these challenges. As you see this first slide clearly states it's a self-service healthcare data platform. MD Clone Adams platform is a self-service healthcare data platform. So let's dial into the challenge for clinical researchers. We hear this every day from the researchers that we're working with. There's often a need for them to work with analysts with coding skills to pull the data. Think about that. Then they need to get approvals for getting access to the data. Think about that too because many of them have ideas in the moment but then need to go and pursue a process to get access to the data. Often time what happens is by the time they get access to the data or learn how to work with an analyst to pull the data, the idea becomes old. They've got new ideas. So the original idea never really got pursued. They've got many ideas. They've got 10, 15 ideas coming in per week. Then when they get access to the data it takes time to iterate on the study design and analysis. Now that, why does it take time? It takes time because they don't have direct access to the tools to do the study design, to do the analysis, get the insights themselves and iterate on that process. On top of that, most of our researchers don't have protected time to do this work. They're doing this because of their professional interest and because of their curiosity and because they want to contribute to the knowledge at the VA and in the community as a whole. So let me introduce you now to the MB clone Adams platform. Adams is an acronym. The first A stands for ask. It's the query engine inherent in the platform that allows direct query by the clinical researcher. The D stands for discover. So researchers discover their insights through exploratory data analysis tools within the platform and they get insights through inherent statistical tools within the platform. This allows them to go from getting those insights back to the query engine and refine their process and their study design. The second A in Adams, the act, stands for the ability to take action on insights, action on certain patient care insights that they get, pathways that they discover can be changed. And the M and the S stand for measure, measure the impact of the action that they took and share those insights with other collaborators within the VA or other research groups. The MD clone Adams platform is housed within the arches innovation ecosystem. And Amanda had stated that this is within the walls of the VA that keeps the data secure. So now let me talk about why clinical researchers use the MD clone platform and how these challenges that they face are overcome because of the unique features and capabilities of the MD clone platform. You'll see this pyramid that builds in the talk that I'm about to deliver in the next few minutes and it'll lay out for you very clearly why they use the MD clone platform to go from an idea to their final outputs of taking actions. First of all, the breadth of data. Speakers prior to me have alluded to the fact that there's 25 million patients in the Data Lake. We have access to the same breadth of patient data that comes from the CDW or the corporate data warehouse. And then what we do is we've curated and categorized this data in the various domains of interest that clinical investigators are very well familiar with. All the different types of encounters, conditions, problems, diagnoses, measurements, lab results, every normal domain that clinical investigators use for their researches, for their quality improvement, and for their innovations. We build in a lot of semantic knowledge into the data that we've curated, meaning this hierarchical representation and categorization where needed, for example, under problems and diagnoses. So the ICD conditions have hierarchical representations there so that you can choose from the lowest level of granular information to groups and categories of information under problems or conditions or diagnoses. Not just that. We organize the information longitudinally. Clinicians think of their histories, the clinical histories that they're dealing with, longitudinally. What happens in time? Everything in a patient's life or a population's timeline is along a linear time span. Something happened in a certain year at a certain time, whether it's a diagnosis, whether it's a lab result, whether it's a procedure. So that's how we're organizing our data. So it's simple to think about. You don't need to think about the tables from where this information came, where it lived. We provide those definitions within the platform itself. The queries that investigators build are fully customizable. They do it themselves. And there's a repeatable study design that they can follow within the platform. They do it once. They can share it with other people. They do it once. They know the process of the platform enables them to do without missing any steps. There's steps to studying these and defining these variables, including not just what the variable is but when it happened and what output you want in the final output to get your insights. That's the platform almost forces you to do that so that you don't miss a step. And this is all through a friendly user interface. Something that you don't have to learn or think about. All our investigators have gotten as soon as they get access, they're able to start interacting with the data directly. Not much learning except for familiarity is required. Not much training. And even if they do, we're available to provide that support. It's only for them to get familiar. Probably takes a couple of hours on the same day as they get access. No coding skills either. So that's a big plus that all our investigators have talked about. Everything is in the platform. It's all in one. You query from within the platform. You gain your insights and discoveries from within the platform. And then there's also natural language processing within the platform. So when investigators need unstructured data and they need to analyze that, well, they can target certain phrases from within notes, from within reports, such as pulmonary function tests, such as echocardiograms. It's only when you extract certain target data elements and make them structured that they're useful for analysis, that they're useful for direct discoveries and insights. And we have an NLP studio within the platform that allows researchers to do that directly based on their own rules, based on their own judgment, based on fit for purpose intent. And then right at the top, they can use original or synthetic data. Synthetic data. I'll talk a little bit about what synthetic data is to make it clear. Synthetic data is the counterpart of the original data. The word counterpart, very important, it's not different from the original data, but it's statistically completely the same as the original data without any identifiable individuals in that data set. There's no IRB needed for investigators to get access to synthetic data. And synthetic data with proper approvals and agreements in place can be shared outside the VA. So that's why people, investigators, clinical researchers, use the NLP platform and more and more are getting enthused to do so. It's a complete paradigm shift in the way they do their work. This diagram illustrates how they can go rapidly from project idea to the endpoint of their project. They iterate through study design using synthetic data, evaluate feasibility of their idea while they're considering an IRB. As soon as they're confident about an IRB proposal that's required, or maybe a QI proposal that's required for getting to the original data, they can immediately turn the design that they've already done in the platform to now extract original data and then quickly proceed to get more insights, take actions for QI projects, or publish their studies. This several projects that clinical investigators are doing using the MD clone Adams platform, several different specialties, neurology, cardiology, endocrinology, and many of them will talk to you now and stewing presentations, at least three of them. And we can have more sessions like this and follow up contact points to talk about the various projects. Dr. Lee now has alluded to open challenge projects that are complex projects with large synthetic data sets that have been created in collaboration with MD clone for mission daybreak, suicide prevention, or the VCHAMP's open challenge for improvement in heart failure outcomes of the veterans population. Lastly, I just want to talk about certain key points that most of our clinical investigators have benefited from. Rapid speed from idea to value, easy iteration of their research design and their questions, and an anesthesiologist team from the Richmond VA, they've told us that they've waited for years to do this. They wanted to study outcomes from intraoperative interventions that occur, but they haven't been able to do so because of lack of direct access and the ability to interact with clinical data. They've also taken special data from the Pisces anesthesia system and integrated that with data from the CDW. So we'd like to talk more, interact more with everybody here and those that you would like to pass this message on to. We'd want to demo the capabilities of the environment, start interactive and collaborative projects, and spread this not only across Arizona, but really nationwide. We can do 50X to 50 to 100X of what we've done so far. Thank you so much. Thank you for your time. This last slide is important because we'd like you to contact our customer success partner. Her email address is here. It's malini.chrishnamurthy at va.gov to learn more and for follow-up questions and collaborations. Thank you so much.