Nancy Brown, Chief Executive Officer, American Heart Association 
David Hodgson, Co-founder and CEO, Project Ronin  
Dr. Jessica Mega, Co-founder, Chief Medical and Scientific Officer, Verily 
Moderator: Michal Lev-Ram, FORTUNE


Brainstorm Health 2022 - Big Data In Health Care: Promises, Perils, And Precision Medicine

Conferences

Nancy Brown, Chief Executive Officer, American Heart Association 
David Hodgson, Co-founder and CEO, Project Ronin  



Promises, Perils, And Precision Medicine


Nancy Brown, Chief Executive Officer, American Heart Association 
David Hodgson, Co-founder and CEO, Project Ronin  
Dr. Jessica Mega, Co-founder, Chief Medical and Scientific Officer, Verily 


Brainstorm Health 2022 - Big Data In Health Care: Promises, Perils, And Precision Medicine


I want to start out by actually just having each of you explain to us what your organizations do we know american heart association. I think it's pretty well, it's in the it's in the name. Um, but not only what your organization's mission is, but where you sit, where your organs sits within the big data conversation and I'll start with you. I'd be happy to start. So yes, I am the ceo of the american Heart Association. We are this country's largest health charity. Um but we are also at our core of scientific organization and we are the second largest funder of research outside of the NIH and heart disease and stroke. And we have been um collecting and creating datasets to help improve the quality of care delivered in America for 21 years. So um it is data is core and center to our work and through partnerships which maybe we'll talk about in a little bit later. We also are creating um clinical trial registries around specific populations like women to focus on women and heart disease. Dave thanks. Um my name is Dave Hodgson. I'm there Ceo of project Ronan Project Ronan is a clinical decision support system. So what we do is we take tons of clinical data, reference data, patient reported data and we give it to clinicians and the way we do that is by inserting a lot of contextual information, predictive models and reference and other statistical insights and we give it to the busy clinicians that are treating the most complicated diseases. So we're a young company and we're just getting started with our early partnerships mostly in academic medical centers and cancer centers where if you've ever witnessed the cancer clinic or watch the job of a cancer nurse, it's absolutely chaotic. It's an extremely complicated disease. The number of choices you need to make our bunkers and the jeopardy and the tension and the fear amongst patients is really high. So what we do is equip these clinicians who would have to deal with the rather clumsy to give me medical record systems that don't help with decision support at all and do not bring any of the current information. We take all of that data and we re render it and we put things on a longitudinal timeline. You can see the history of the patients treatment and then we help them with their choices. So we bring the reference data and then we show them statistically what may be relevant insights to then make good choices about clinical intervention, treatment, decisions, dozing and so forth. And then we also do something for the patient. We help the patient report their symptoms. Are you experiencing side effects on chemo, immunotherapy. And then we also help push education as well. So all of these data really put to work for the best decision making and the best of individualized treatment right now in oncology. Okay. Great Jessica. So my name is Jessica. Make. It's wonderful to be here. Thank you And verily what we're doing is bringing the promise of precision health with the goal of bringing it to everyone every day and we're a company that sits inside the alphabet ecosystem. So we sit at the intersection of technology and life science and healthcare And there have been three pillars that have really driven our purpose. One is that if you think about comprehensive health, there is a piece of it as you already heard that I as a physician put into a health record, but health happens every single day And it's especially to be about a 1500-fold increase in what we even call health information that's going to be available in the next decade. And we're working on harnessing that information. The second truth is that connecting people inside and outside of the four walls of a hospital or clinic is now becoming a reality That helps us all as we make decisions, not just when you see a physician but 365 days out of the year. That's a very data intensive program. And then the final pillars, what are the right incentives that you need to make precision health a reality? I already heard Nancy allude to some partnerships, partnerships are core to what we do. We're also interested in creating the right business model so we can get ahead of disease prevent disease. So we're not just always chasing a poor outcome. So that's what we've been focused on in data as you can imagine is central, but what's even more important is making it relevant to the individual who we treat. Okay. So I know we've had a lot of big data conversations over the last couple of days here. I want to like really hear from from the three of you. Um I know you have a very patient centric approach to this and but just starting with the data itself and some of the challenges there. Um Dave could you talk a little bit about um for the health care profession, for the applications that you're working on specifically? Um talk about clean data and how much of a challenge it is to get clean data and to clean it up the data I have to deal with all the time is mostly buried in the medical record system in the systems like Allscripts, Cerner and Epic and what If you take those systems, they're about 80% tax or a big bag of unstructured notes and some structured data. So one of the big channels we spend a lot of time on is cleaning that up and preparing it. So we're taking what's now fairly popular is an approach called data centric ai So instead of doing what we would call model, itis, I take a little piece of data to train a model and then for particular question of purpose and then sort of stick it out there and hopefully it works is we spend a lot of time upfront cleaning processing and mapping the data. So we focus very much on what we call the force of vis of data and we look at the volume of data, we can get ah pumped in the veracity of the data, how high quality it is, the velocity that it changes in how stale or otherwise it is and then sort of how how high quality is. So we take all of that and we apply a lot of machine learning and statistics and processing too, then map it to a central data model. So you can't rely on what you've got in these notes are all written by very different people and so we apply natural language models for that and then for the structured data, a whole ton of cleanup and processing and assessment algorithms to the incoming structure data. And then we have a corpus of high quality structured patient data and that we can offer re rendering of the data to the clinicians so they can see it very clearly across time, make really thoughtful interpretations of it heather up with the natural reference data and then use all of that to fuel predictive models. So there's a lot of work done mostly in the clean up and then the monitoring of that's where those pipes coming in and by the way, I know you you mentioned at some point when we were talking that um for the end user uh for for the clinician, um the app interface that they're interacting with is at like 1/4 grade reading level, I think. What tell us about that. So we um we have two different interfaces and interface for the clinical team and then interface for the patient and for the clinical in both cases, you really have to design thoughtfully for their workflow state for the questions they have and the places that they're asked and for the clinical team and the patients, they're sort of in very different places. So for patients you really want to have access as possible. So when you give a mobile app to a patient and if you look at our mobile app, it looks very simple, very big buttons, very simple colors and that app has 1/4 grade reading level. And so it's extremely easy for it to ask you how you're feeling today, how you tolerate your chemotherapy and pain. You have nausea and so forth. And then when we collect all of those symptoms, so we then return it back to the console team for assessment. We can also dispatch education. So if you're feeling nauseous or you have pain, we can tell you whether that's expected or not. And also, if there's any appropriate remedies, you could take to offset those symptoms. Now, the clinical interface is a whole other world. It's very sophisticated because we're trying to not only show them show clinicians what they need to know in the moment and that's itself a challenge because if you've ever spent any time in any clinic, it's very busy, it's very chaotic. And clinicians are just trying to know which interventions do I need to take right now? Who is at the highest risk right now? And so we present a lot of that data in that, in that mindset of where do I need to spend my clinical intervention time and risk? Okay, okay. I had my end users strong. Maybe someday you'll be able to use moxie as your interface. Pretty amazing. Um, nancy. I want to go to you another piece of of making sure that the right data and comprehensive data is, is being ingested in order to enable uh the other end of of of the cycle here, um, is making sure that different demographics are represented. You're one of the largest funders of cardiovascular research. Obviously heart disease impacts all demographics. How are you making sure that the research that you're funding is taking that into account. And I'm giving you a great segway to just absolutely. You sure. You know, it's amazing to see what has happened. What innovations happened to allow patient care to be better at the point of care. A couple of things to think about though a person's interaction unless they have sadly uh, you know, a dreadful disease with their healthcare provider is very brief. Um, in the course of the year. And when we think about clinical trials, overall, fewer than 10% of people in America are seen in an academic medical center where most clinical trials happened. And I know uh dr Megan will have a lot to say about that. So what our purpose is is to make sure that data that is being used, whether it's for research, whether it's for clinical care really speaks to. What are the factors that allow a person to live a long and healthy life. So we love what you are doing with electronic medical records. I will say however, that the electronic medical records don't contain information about a person's access to healthy food. They don't contain information about whether or not a person is one, you know, flat tire away from being able to buy their necks um groceries or not being able to pay their medical bill. There's a lot of things that impact people in their everyday life that must be taken into account. And as it relates to all demographics. We really pride ourselves at the American Heart Association. I'm thinking about all people and equitable access to trials and registries, which is why for our Um get with the guidelines program, a 20 year old data registry. Um if you have a stroke anywhere in America, 85% of every stroke strokes that happened in America are captured in the American Heart Association's get with the guidelines program. And from that we have very rich data on ethnicity race, you know, gender, all kinds of things rural versus urban. And that allows us as we then turn around and write the clinical practice guidelines, which is what we do with the H. A. To incorporate all of that information so that the full circle can happen so that patients receive the best care at all times. The other thing I will say is that we are focused right now on studying the best way to assure diversity in clinical trials. We recently um well actually tomorrow we will be announcing $20 million dollars in funding to four centers. Um four very prominent centers. I know that's okay. I'm not gonna tell you who they are. Just tomorrow, we'll be announcing who the four are. Um and uh these four centers will actually be studying how to best assure diversity and clinical trials. And we mean age, ethnicity. Thank you very much. Age, ethnicity, uh sexual orientation. All of these things really matter to make sure that what comes out on the other end really represents what people face every single day in their life Jessica. Can you jump in and talk a little bit more specifically about the partnership that that verily has with american heart association. So if you think about precision health and how important that is. We only are able to treat patients based on the data that we have. And it's really important for that to be generalized, able What we know in cardiovascular disease and many times women are underrepresented. It's estimated that only 20-30% of clinical trials, including interventional trials in this space include women. And so what we wanted to do is make sure to be precise. We need to be precise with the the outreach. And so we were very fortunate to partner with the american Heart Association and create a group called research goes red. And there's several things that are different about this program. The first one that starts with the individuals themselves, people. And we found women across the nation raising their hands and joining this research registry. It really was a movement. The second piece that was interesting is instead of as a researcher coming and saying, here are the top three issues I want to study. We asked the women themselves and it should be no surprise that one of the biggest areas that came up was menopause and obesity. Another area with inclusion. And so that is where our research started. And then from there, we've been able to continue to expand the registry and create a scientific cohort of some of the world's best leaders to support this initiative. And I always remind myself, the only way we get a new medicine or device or digital tool into the hands of patients and clinicians is through the best data. So you're going to see us I think continue to work in this area, particularly with the American Heart Association. And what was fascinating even during the pandemic. The same principles of putting patients in the center sharing those learnings has been really important. And we're excited this week to publish our results from a COVID-19 registry with that same patient centric city. Those results came out in the new England Journal of Medicine catalysts and I think I encourage all of us to get our learnings out there. It shares things that went well where the challenges occurred, the data and the underlying infrastructure. But what was helpful if we were able to create a network across the nation that went beyond one hospital system to get ahead of the pandemic and I hope we can do the same thing with cardiovascular disease and more chronic forms of illness. Right. And so that I think it brings up another question here. Another point which is on ethics and transparency in keeping the patient in mind as you're building out these, collecting both, collecting the data sets and and building out some of these systems. So how do you think about that from the get go? Like it's so interesting that you guys from the beginning went to the actual patients to the actual people who are going to be a part of the research and asked, what do you care about that transparency, you know, right there. But but tell me your your perspectives on how you approach that in other ways from the get go. Well. I'll just start by reinforcing the point that Jeff made, you know, these registries need to exist to serve the health needs of health consumers, You know? And so wherever we're collecting data, whether it's with apps through sensors in registries and clinical trials, you know, never forgetting, you know, the purpose is to help a person and people live longer and better lives. So starting by asking them what's important to them is transparency. And it's also, I would say the highest ethical question any of us can ask, you know, making sure that we have integrity and privacy of data that we are following. You know, all the rules and regulations letting people opt in. You know, I smiled. The research goes red as an example is an extension of our american Heart Association. Go Red for Women movement where million I see dr Susanne stein balm and the audience millions of women across the country, you know, advocate for their own heart health because heart disease is a woman's greatest health threat. And through that it's a marketing, you know, initiative where when it comes to science, you know, making sure that people are consenting and that we're following, you know, the protocols for keeping a person's data is very core in center but that none of that you shouldn't even get to any of that. If you don't start with the ethical question, why does this exist? And how does this benefit the people who we seek to serve and and real quick, very fast. Dave and Jessica can you have transparency and also keep it digestible for you know, healthcare consumer Yeah, you have to know I fully agree you the basics of rules of the game, our privacy and some basic ethical practices about how you safeguard the data and always having the key questions be the anchor of everything you do. The key about transparency particularly comes to in my world predictive model building is you have to really show very clearly whatever predictive model created on what basis did you created? What data was used to answer what very specific questions. So if you ever looked at the screenshot of our user interface, you would see immediately. Here's the data that was used to make that that conclusion and then we do a lot of work. It's a very big value for us to look at fairness and bias. So if we know a model was created in a largely white caucasian population that has prostate cancer, it's not going to generalize to other types of patients. And we were very, very careful about the input datasets and then how they're used and then checking for fairness and inclusion with and Jessica real quick. Anything to add on that. Is it is it a tradeoff or No. I think that to learnings that have been eye opening to me one is this concept of user centric design in health care is a must the second one is I've been blown away by how empowered people have felt when the data is relevant to them. And you heard us talk today about research. How does that translate them into clinical care? It's another area that we've been working on. We look at things like continuous glucose. And the minute you show someone here's how your body responds to oatmeal or popcorn, it becomes incredibly relevant. And so I wouldn't underestimate the condition and people's ability to absorb information if it's valuable and relevant. So I think good work going on and I think a lot more we can continue to do in this. Thank you so much. We didn't have as much fireworks as the wearables panel. I'm sorry. That was very contentious. But no, thank you so much, Nancy, David. And thank you you guys out there.