PlayPlay Podcast downloadDownload Podcast Take Post TestTake Post Test

Transcription:

Melanie Cole (Host): UAB Med Cast is an ongoing medical education podcast. The UAB division of continuing education designates that each episode of this enduring material is worth a maximum of two five AMA PRA category one credit. To collect credit, please visit uab medicine.org/med cast and complete the episodes post test.

Welcome to UAB Med Cast. I'm Melanie Cole, and joining me in this panel today is Dr. Oscar Julian Booker. He's an associate professor in cardiovascular disease and Dr. Efstathia Adrikopoulou. She is an assistant professor of cardiovascular disease and radiology, and they're both with UAB medicine. They're here to highlight using empiric cardiac intelligence to address health inequities and help identify patients at risk for heart valve disease who might otherwise be overlooked.

Doctors, thank you so much for joining us today and Dr. Adrikopoulou. I'd like to start with you. Data has shown that people from racial or ethnic minority groups are less likely to receive preventive healthcare and across the board various ethnic groups have faced a disproportionate health burden. Can you talk to us a little bit about the unique challenges these communities have faced in terms of healthcare disparities? Why did you choose to study this topic of rural healthcare disparities? Describe the scope of the problem for us.

Dr Efstathia Andrikopoulou: Absolutely. First of all, thank you so much for having us today. It's such a pleasure and an honor to be able to share our experience with trying to transform and innovate the way we deliver care for our patients here at UAB. The problem is vast, sadly, and, I don't think it would need med cast to describe the problem in its entirety, in depth that it deserves. But our passion that has driven all of our efforts so far is how do we make sure all of our patients get high quality, equitable care at UAB.

And we quickly realized that there's a lot of disparities and inequities in the way that we detect disease and the way that we deliver care. Both within our urban center in Birmingham as well as outside of it in our rural areas. There's tons of data out there that, not just for the state of Alabama, but for the US in general, showing that people living in rural areas, face much more struggles in terms of getting access to high quality. And this reflects poorly on their outcomes.

These people live shorter lives and they have worse quality of life, and especially when it comes to patients with cardiovascular disease, patients with heart attacks, heart failure, diabetes, high blood pressure. There's a lot of data showing that patients living outside of urban centers, are doing worse. And this is what we set out to A, understand. using artificial intelligence and B, implement solutions to try to bridge the gaps in healthcare inequity.

Dr Oscar Julian Booker: That's a great point. It's well known that health equity is not just a function of the care that the patients receive within the health system, but is a complicated convergence of a number of issues some are patient-centric and some are health system centric. But if as an institution we can't start the process of ensuring that patients are not slipping through the cracks and those patients never have the opportunity, to receive care similar to those persons who perhaps are not burdened by, certain social determinants of help.

Dr Efstathia Andrikopoulou: Exactly and can I just, quickly add that, the goal is to look at both social determinants of health and as we gain more understanding on those, also, explore our patients and our communities, determinants of culture.

Melanie Cole (Host): And you both make such excellent points about the social determinants of health, and as there's gaps in cardiac care that sometimes can occur because of a patient's clinical or social status as we've just. Discussing Dr. Booker, I'd like you to tell us how you're leveraging the power of natural language processing and artificial intelligence to help address those inequities. How did you partner with third party vendor? UR cardiac intelligence to develop software algorithm that provides clinical decision support to help identify patients at risk for heart valve disease who might otherwise be overlooked?

Dr Oscar Julian Booker: Well, this processed started a few years ago, as we use our informatics resources to try to search through the medical record to find patients who made be slipping through the cracks. But one thing that we realize, and I think it's well known, more broadly. And that is that the majority of healthcare data is in a non-structured format, meaning probably only about 20% of all data is simple and codified in a way that is easy to search using standard search tools. Well, that's a lot of data that we're missing and realize that if we wanted to capture these people, we needed to think differently about how we mine and review the data so that we can better identify patients.

We were looking to partner with a company with an offering who could help us identify patients who have certain ular heart disease. And we were very intrigued by empirics strong utilization, use of natural language processing. Because we felt like. If we work together that we may be able to more deeply understand our patient population, both at a granular level, also in aggregate, to then engage with their providers, engage with the patient. Because we have a better understanding of our patient population as a whole, but also the individual patient.

And empiric has been a fantastic partner, and has shown willingness to work with us to refine and define algorithms to, again, allow us to look at the patients both in aggregate and at a more granular level.

Dr Efstathia Andrikopoulou: This is a great description of, the, very fruitful relationship we've had with empiric so far. And the only thing that I would like to add to that is that again, the goal is to ensure that we deliver high quality, equitable care to all of our patients and to our community. The only thing that we are very, very careful of, and very aware of is, the inherent bias that comes with any type of artificial intelligence system, including natural language processing. And this is something that we always we're very mindful of whenever we build our algorithms, we want them to be as bias free as possible.

Dr Oscar Julian Booker: I think that's a really great point and. I think that it can't be overstated that our algorithms understanding that bias is a very important concern whenever natural language processing and artificial intelligence are at play. one of our goals is to base our recommendations and algorithms as firmly to accepted guidelines as possible, but then also take to account our local clinical experts. As well as institutionally, those services to help us understand if we are imparting or if there are any unintended biases, involved in our data, or at least involved in our search for data. So I think that is an incredibly important point.

Melanie Cole (Host): I think you're both making incredible points and what a fascinating topic and study this is. Dr. Andrikopoulou. Please tell us how your team is in the final stages of developing an algorithm that can help identify cancer patients and cancer survivors who need cardiology services? We're learning more about the link and the cardiovascular implications of cancer and cancer treatments, how can this make it easier for oncologists to easily identify those patients and refer them to a cardiologist? This is an exciting collaboration and has it generated a lot of interest.

Dr Efstathia Andrikopoulou: I'm so glad you're asking me this question, Melanie, because this is something that I'm very passionate about. One of my clinical niche is within the field of cardio-oncology, which is at the intersection of cardiology and oncology, and focuses on caring for people who are either actively, battling cancer or who are cancer survivors. We are now understanding more and more that, cancer in and of itself and cancer therapies, can have an adverse impact on the cardiovascular system.

It's great that people with cancer now live longer and better. However, what we're now, seeing is. This population is developing cardiovascular disease. And they start developing heart failure or heart attacks or inflammation, affecting that heart muscle. And that made us, use our experience with empiric to reach out to our oncology clinicians and colleagues here at UAB and work with them to develop an algorithm that will enable us to identify patients with cancer who are either survivors or undergoing treatment.

And we can identify certain high risk features and flag these patients and notify their oncology provider to let them know that their patients would benefit from seeing one of our cardiology specialists in. And we're at the initial stages of this algorithm. We're hoping we can start implementing it early 2023. because based on the results that we've had so far with empiric treating our patients with Valvular heart disease, we are expecting that we're going to be improving access to care for these patients as well.

Dr Oscar Julian Booker: You know, I think it's important for us to note that there is so much medical knowledge that's out there and it's growing so fast. It's almost impossible or is impossible for an individual to keep up with this exploding, fund of knowledge that we have to have a grasp of to be able to provide the most current high quality up to date care. And then when you couple that with the sheer number of patients that are in our system that we have to keep track of.

There's no way that an individual can provide the highest quality care for all patients at all times. And I think that this is the space where artificial intelligence and natural language processing can help. This is a support tool that the fantastic discussion that you've had about the use of empiric and cardio-oncology is really only the first of many steps. Once we start to realize just how supportive technology can be to our ability to provide high quality healthcare.

Dr Efstathia Andrikopoulou: I totally agree. There's no limit to the lives that we can touch and the people that we can help. It doesn't have to be only people with cardiovascular issues or only people with, cancer who may have, issues with their heart, it could really be anyone.

Melanie Cole (Host): Well, that's exactly as my next question. Dr. Booker is anticipating future partnerships with other UAB medicine areas. You're just discussing this just now, surgery, obstetrics. So you have the opportunity to optimize clinical care both in the aggregate and in individuals. Where do you see this going?

Dr Oscar Julian Booker: That's such a tough question to answer because I can see artificial intelligence and natural language process. Supporting all aspects of clinical care in the future. I think really for us, the question is what is the next project and then what's the project after that? But at some point, I believe that AI and NLP or natural language processing will be fully integrated as a support tool for all aspects of clinical.

Melanie Cole (Host): I have a couple more questions for you doctors and Dr. Andrikopoulou, working with people from different backgrounds or cultures really presents unique opportunities for collaboration as we've been discussing, and creativity. In your personal experience, how have you seen this materialize at UAB Medicine? Speak a little bit about health equity, diversity in inclusion, and how you've seen that work at UAB?

Dr Efstathia Andrikopoulou: I have been very lucky and blessed to always have been part of very diverse themes, both culturally, ethnically, but also from just the background and a cognitive standpoint. And this just goes to show, how fruitful and productive our collaboration has been with empiric. Our team is extremely diverse. It's Dr. Booker, it's myself, and we are the only two physicians that are part of the core group of us working on those algorithms. The rest of the group is non-clinicians, data scientists, data engineers, biostatisticians researchers, who are very, racially, ethnically, diverse, gender diverse, and culturally diverse.

And I think this is to a big extent, one of the reasons why our partnership with Empiric has been so successful is because we are also very different and we all bring a different, viewpoint to the table, a different angle to the table. And this is important because like we've been saying this entire time, our focus is to be able to understand, connect, and help all of our patients. Every single one of our patients is going to be different. Their background is going to be different, their identity is going to be different. And the only way for us to be able to understand our communities and help our communities is if we make sure that our own group reflects our community. Otherwise, we are not going to be able to understand and help them.

Dr Oscar Julian Booker: You know, I think it's, very interesting you point out how culturally, diverse the primary group is that's working on this. and you pointed out all the different ways that our group, is diverse. And I think that it's very interesting, that when you have so many stakeholders who are working on such projects, many of which who are from historically disenfranchised groups, I think there is a higher level of commitment amongst those people to ensure that those gaps are closed. In addition to the different perspective that that diversity provides, the commitment of being from one of those historically disenfranchised groups, I think lights a little bit more of a fire to ensure that all patients receive equitable care.

Dr Efstathia Andrikopoulou: I couldn't agree more. there's me and I'm from Greece. Then there's you, there's Miguel, there's Chris, there's Lauren. So many of us and every single one of us is very different.

Melanie Cole (Host): What a great discussion This is Dr. Booker. I'd like to just give you the last word, as physicians, as you're just saying and speaking about your team, physicians play this critical role in addressing these public health concerns. I'd like you to speak just for a second about your recommendations for other physicians around the country to consider when they're treating underserved in minority patients, reducing those barriers to quality healthcare through patient navigation, through the multidisciplinary approach, and then wrap up the algorithm with empiric.

Dr Oscar Julian Booker: I think if I had one message to share to other providers that would be all of us got into healthcare to provide the best care that we can for our patients, but we are not infallible and there is no shame and no harm to getting help. And as we discussed earlier, there's so much to know and there's so many patients to care for that support will allow us to fulfill that goal and that obligation that we set forth in the beginning. And that is help patients and provide the best care that we can. And this is where technology is able to support us because the computer doesn't get tired, once we teach the computer how to read and how to understand, it's able to do those things in a very timely, very thorough fashion.

And it's really just there to allow us to be better providers. It's not there to replace us as providers. And I don't believe that computers can ever truly replace us as providers, but they can support us. And I think in the future there's really going to be the difference between those providers who don't use AI to support them and the level of care that they can provide. And then those providers who do recognize that technology can make us better at what we do. And we hope to, and UAB has been very supportive. And has shown itself to be an institution that recognizes the future is coming, and our ultimate goal is providing the best care that we can.

But to your question about the algorithms, the algorithms are the heart of what we do. Being able to search and identify those patients who meet certain criteria is at the heart of the provision of care that we're discussing. But we have to be mindful and we have to be careful not to introduce biases in our search and in our recommendations. And so that multidisciplinary group, either a, culture, gender, diversity or a diversity in clinical expertise to ensure that my limited scope does not inadvertently lead me down a path where some group is receiving less optimal care than another group.

This is why we have to work together and the nature of the work is expanding beyond just the traditional healthcare field, and we require all these other support teams, as Efy has pointed out, as far as data scientists and data architects and et cetera, ecetera, ecetera. the nature of this multidisciplinary clinical care team is larger and different than perhaps what it was in the past, but they all have those same goals that we do and that is providing the best care that we.

Dr Efstathia Andrikopoulou: I a hundred percent agree with everything that, Dr. Booker just mentioned. And the only last thing I'd like to point out is that obviously clinical care and providing high quality, care to everyone is in the heart of all that we do. And this is our passion and this is what drives us, and this is why we're developing and implementing those algorithms. At the same time, we're realizing that the next frontier is going to be ensuring that the data that we use, is of high quality and ensuring that there's high levels of data privacy, data security.

And to make sure and reassure all of our patients and our community and our healthcare allies and stakeholders within the UAB Health system and outside of it that, we take all of the precautions and some more to ensure high levels of data privacy and data security. And our next frontier would be to make sure that we built a structure of data governance and data oversight that will always guarantee a hundred percent rates of secure data.

Melanie Cole (Host): What an excellent thought leader conversation This was great guests, both of you. Thank you so much for joining us and speaking about this very important topic and telling us all about using empiric cardiac intelligence to address those health inequities. Thank you again, and for more information or to refer a patient, you can call the MIST line at one 800 UAB-MIST, or by visiting our website at uabmedicine.org/physician. That concludes this episode of UAB Med Cast. I'm Melanie Cole. Thanks so much for joining us today.