Chair File: Leadership Dialogue — Harnessing the Power of AI in Health Care with Janice Nevin, M.D., of ChristianaCare

00;00;00;28 - 00;00;25;23
Tom Haederle
Since its launch in 2020. The American Hospital Association's monthly Leadership Dialogue series has brought together some of the best minds in health care for insightful discussions about the most pressing issues facing the industry. The tradition continues in 2024, with a kickoff discussion hosted by Dr. Joanne Conroy, CEO and president of Dartmouth Health and 2024 Chair of the AHA Board of Trustees.

00;00;26;00 - 00;00;48;08
Tom Haederle
The topic couldn't be timelier. There is a lot of discussion these days about the role of artificial intelligence in health care, and in this podcast we hear from a health care leader who "walks the walk" when it comes to AI.

00;00;48;11 - 00;01;17;20
Tom Haederle
Welcome to Advancing Health, a podcast from the American Hospital Association. I'm Tom Haederle with AHA Communications. The health care field has taken to artificial intelligence in a big way. The adoption rate is high, but debate continues about the best and most effective ways to employ it. In the first Leadership Dialogue of the year, the leader of a major East Coast health care system talks about what AI has meant to her organization and shares other thoughts and observations about where health care is headed right now.

00;01;17;23 - 00;01;47;15
Joanne Conroy, M.D.
I want to thank all of you for joining me today in my first Leadership Dialogue discussion. I'm Joanne Conroy, president and CEO of Dartmouth Health. Dartmouth Health is New Hampshire's only academic health system. We're the largest private employer in the state, and we serve a population of 1.9 million patients across northern New England. I'm incredibly proud to lead our system of over 14,000 employees that serve patients above New Hampshire and Vermont.

00;01;47;17 - 00;02;14;14
Joanne Conroy, M.D.
But the reason I'm here today with you is beyond my role as the president and CEO of Dartmouth Health. I'm serving as the board chair of the American Hospital Association this year. As a bit of a background, the AHA launched these leadership discussion series in 2020 when the pandemic began, and then my board chair predecessors have hosted insightful and valuable discussions since then.

00;02;14;17 - 00;02;45;17
Joanne Conroy, M.D.
I'm looking forward to continuing these. They seem to be of incredible interest to our membership because they allowed us to have conversations with health care experts from across the country, from within our membership, that focus on the most pressing issues facing our industry. The knowledge that we gain from this is really invaluable and our ability to listen and learn from one another is really the heartbeat of the American Hospital Association value that we deliver to all our members.

00;02;45;19 - 00;03;12;05
Joanne Conroy, M.D.
And that's exactly why we're hosting these Leadership Dialogue sessions to listen and learn from each other. It is my pleasure to introduce Dr. Janice Nevin. She is president and CEO of ChristianaCare, based in Wilmington, Delaware. She first joined ChristianaCare in 2002 as chair of the Department of Family and Community Medicine, and she later stepped into the role of president and CEO in 2014.

00;03;12;07 - 00;03;40;21
Joanne Conroy, M.D.
I actually met Janice when she was a CMO and we were in a number of stakeholder associations together, and I've always been really impressed by her thoughtful leadership. She is known as a pioneer and international thought leader on value based care and population health. And she's also been an early adopter of health care technology platforms and solutions that both advance health and improve processes for patients.

00;03;40;23 - 00;04;14;00
Joanne Conroy, M.D.
You know, I know just enough about AI to really be very dangerous, but Janice has made it really a focus of her organization to really understand the value that this can provide our industry. Now, I don't know how many of you know a lot about ChristianaCare, but it's a nonprofit teaching health system that includes three hospitals, an extensive network of primary care and outpatient services, home health care, urgent care, a freestanding emergency department, along with a number of centers of excellence in cancer care and women's health.

00;04;14;03 - 00;04;27;21
Joanne Conroy, M.D.
I've also had the privilege of working with Janice as we share time on the board together, the AHA Board of Trustees. So, Janice, welcome and thank you for joining me on my first Leadership Dialogue conversation.

00;04;27;24 - 00;04;33;07
Janice Nevin, M.D.
Well, thank you, Joanne, and thank you for having me. It's a privilege to be your first.

00;04;33;09 - 00;05;00;08
Joanne Conroy, M.D.
So, you know what? I know AI has been in health care for years and it's probably been involved in health care in ways that a lot of us haven't appreciated, like in a lot of our back office functions, especially in revenue cycle management. It's been actually commonly used, but there are recent adoptions of it, kind of a new type of AI called generative AI.

00;05;00;10 - 00;05;34;28
Joanne Conroy, M.D.
And I think it's all my physicians talk to me about is Chad GPT. But that's one aspect of where AI actually will be a valuable to our organization. So there's tremendous potential. The adoption rate is incredibly high, but there's also like a healthy debate about how do we actually harness it? You're an early and active adopter of these solutions, talk about how you started being interested in identifying the potential of AI.

00;05;35;01 - 00;05;39;19
Joanne Conroy, M.D.
And in what way did you first start to employ it within your health system?

00;05;39;21 - 00;06;33;11
Janice Nevin, M.D.
Yes, Joann, I think you you nailed it. We've been using AI across the industry for years. I think without actually understanding what it was and the impact of it. We've been using AI for a number of years, and particularly in the space of delivering care and also addressing health equity. In 2012 we received CMMI funding and that enabled us to establish a care platform which is now called Careveo. And it used sort of the AI that we've been working with for a long time, which is essentially machine learning, and it allowed us to take a patient population, use machine learning and artificial intelligence use to apply predictive analytics and essentially identify those

00;06;33;11 - 00;07;07;18
Janice Nevin, M.D.
patients who were at the most risk. And then created an opportunity for a care team to reach out to them before they landed in the emergency department or the hospital. So we talked about using AI to essentially change the trajectory of care, and we were able to demonstrate that improved outcomes. In fact, in 2016, I think it was 2016, we received the Eisenberg Award for safety and Quality for how we were using this technology.

00;07;07;20 - 00;07;35;05
Janice Nevin, M.D.
We've also used it to address health equity and to help us and disparities, which is really important to us as a health system. And again, like you, I am not knowledgeable particularly about the technology, but I understand how it can be applied. And in this case we are using AI to continually look at our data to identify disparities.

00;07;35;07 - 00;08;01;27
Janice Nevin, M.D.
And then when we see a disparity, we are able to apply that technology to understand what's the root cause of that disparity. And when we find a disparity that's due to race and ethnicity, it allows us to more deeply understand how we need to address it, create actions to improve the disparity, and then measure the impact of those outcomes.

00;08;02;00 - 00;08;22;00
Janice Nevin, M.D.
So that's just two ways where, you know, two ways in which we've used artificial intelligence to address issues that are so important to health systems across the country. Fundamentally, quality and safety and patient care. And then also the work that we must do to end health disparities.

00;08;22;03 - 00;08;55;11
Joanne Conroy, M.D.
So if I understand you, you've actually used AI as kind of a technology tool to allow your care management teams to actually be much more effective and focus their work on the population that will really benefit from it. And then also kind of doing that scan that, you know, it would take thousands of human hours to do to identify if there are any racial, ethnic variables that you need to consider when you kind of look at quality outcomes.

00;08;55;14 - 00;09;25;28
Janice Nevin, M.D.
Yes, that's exactly right. Now, you mentioned generative AI, which is a different type of artificial intelligence, and that's what's got everyone sort of paying attention. And so we too, have been paying attention. And the way that we have approached it, and I think this is really important, and as I've talked to other health system leaders and experts, I think, again, there's sort of these are tools that are going to be incredibly powerful.

00;09;26;01 - 00;09;53;16
Janice Nevin, M.D.
But as an organization, we need to be prepared to manage them and use them in ways that we know will have impact. And so like other health systems, we have a group, it's very much multidisciplinary interprofessional that is looking at how we should deploy generative AI. And so we have a stream of work that is really focused on governance.

00;09;53;16 - 00;10;30;24
Janice Nevin, M.D.
What are the policies that we need thinking about data, privacy and security, for example? How do we make sure that the people inside the organization are using these tools in ways that are going to be beneficial, whether it be to a caregiver or a patient? And then we have another stream of work that's really trying to understand what are some of the use cases, those early use cases that we can deploy so we can start collectively to appreciate the impact of these tools.

00;10;30;26 - 00;11;04;10
Janice Nevin, M.D.
So when we think about generative AI, we think about number one, how can we remove undifferentiated tasks that humans no longer need to do? So think meeting minutes, job descriptions. What are those sort of non-value added tasks that if generative AI performs them - it gives the humans in the system more opportunity to provide care. And then we think about how do we make the work that we're doing more effective and efficient.

00;11;04;13 - 00;11;28;09
Janice Nevin, M.D.
And I think you said something really important. Digital technology, generative AI, they are tools that we can deploy that can help people be more effective and more efficient at their job. So I think a number of systems, for example, have used these tools in the imaging space where we can bring to the top those images that are most concerning.

00;11;28;16 - 00;11;56;23
Janice Nevin, M.D.
So the radiologists can address them first and we get to a quicker outcome. Some of the folks at our cancer center have recently completed a study and are using AI to enhance detection of brain tumors and have demonstrated that this technology can increase the accuracy of detection by over 30%. Again, other ways in which we're using the technology to make humans better.

00;11;56;29 - 00;12;13;16
Janice Nevin, M.D.
And then I do think there's going to be an opportunity for us to do things that we've never done before and create sort of an all kinds of new using these really important tools. And that is an exciting part of all of this.

00;12;13;18 - 00;12;27;05
Joanne Conroy, M.D.
So for the layman or the uneducated health system leader, how would you describe the difference between AI and generative AI? Because I think they really are different.

00;12;27;07 - 00;13;13;18
Janice Nevin, M.D.
So the way that I understand it is they both rely very heavily on data and the access to that data. But sort of the way we've been using AI has been using that data, applying algorithms, getting smarter with those algorithms. And then it's been useful for predicting whether it's risk or moving tasks up. Generative AI is different because it creates new. And I've described the different ways that we think we can use it, but it's taking that data and it's packaging it in a way that is actually creating words, creating images, creating something new.

00;13;13;20 - 00;13;17;02
Janice Nevin, M.D.
So I hope that helps. That's one.

00;13;17;04 - 00;13;22;03
Joanne Conroy, M.D.
And that's what that's what most hospitals see as eyebrows go up, I guess.

00;13;22;05 - 00;13;57;21
Janice Nevin, M.D.
Yeah, yeah, yeah. It's one of those I think, you know, as always I say, you know, I, I work very hard to surround myself with people who are smarter than I am and I think this is one of those areas where having talent in the organization who can really help the organization understand, again, the power of this new technology and help us deploy it in ways that have have impact is really, really critical.  as well as understanding some of the inevitable risks that are always there with technology.

00;13;57;24 - 00;14;16;20
Joanne Conroy, M.D.
So looking at like the potential for this technology to really change how we deliver care and probably replace some of our vacancies with technology because we don't have enough people to do all the work we need to have done. What are you most excited about in it as clinical applications for the future?

00;14;16;22 - 00;14;42;11
Janice Nevin, M.D.
It's what you just talked about. I've long said we need more people, workforce is such a challenge, and we're all very focused on growing the pipeline and growing the people that we have in the workforce. But we know that will never have enough people. And so we've got to change how we work. And this is where I think the technology can have a fairly immediate impact.

00;14;42;13 - 00;15;16;26
Janice Nevin, M.D.
And also, you know, I worry and I know you do, too, Joanne, about the burnout that our physicians and caregivers in general are experiencing. And some of that burnout is due to the burden of administrative tasks. So I think there's a tremendous potential for for all kinds of AI, especially generative AI, to really sort of help support the humans that we will always need because health care is ultimately it's about people and it's about humanity.

00;15;16;29 - 00;15;43;29
Janice Nevin, M.D.
But if we can use this technology to support the work that they're doing, they will have a much better experience in whatever role they may play and at the same time, then they will be able to deliver a better experience for patients. I've talked before about Moxie. Moxie is a collaborative robot. When we first started to talk about Moxie, there were some nerves, you know, Is Moxie going to take my job away?

00;15;44;01 - 00;16;21;00
Janice Nevin, M.D.
And now we've had Moxie with us. We've got several Moxies with us. And everyone loves Moxie, particularly the nurses and the patient care technicians, because they can interact with Moxie. And Moxie then performs tasks that previously meant they were away from the bedside, away from the unit. Think the fetching and gathering that many caregivers have to do. And so Moxie has been a great way that we've been able to demonstrate, hey, this technology is going to help support you in your work.

00;16;21;03 - 00;16;58;22
Janice Nevin, M.D.
We've also deployed and we're sort of early stage, another type of robot into the home to help support our home health care providers. And that's a little robot called Astro. And again, the goal is to train Astro to work with the patient who has been discharged from the hospital, again, to ensure that that patient gets what they need is adhering to the care plans, taking their prescriptions, not going to replace that home health nurse or that home health care giver, but can support the work and make life better.

00;16;58;25 - 00;17;34;05
Janice Nevin, M.D.
And one of the ways that we have deployed generative AI is I mentioned we're we're testing out use cases. And so one of the ways again, that we're working to help our our nursing caregivers in particular, we created something called Blossom Chat. We partner with Microsoft Cloud. They are our cloud partner and we created Blossom Chat. And what we've started doing is loading all the nursing manuals, policies, all of those that currently live on our website and using large language modeling.

00;17;34;12 - 00;18;05;26
Janice Nevin, M.D.
So for example, a nurse can say, can ask the database how do I discharge a patient with diabetes? And Blossom Chat will provide the instructions drawing from all of that material: here is how you discharge a patient with diabetes. And so again, just another way that we're using this technology to really support our workforce and changing work and therefore changing how they work.

00;18;05;26 - 00;18;19;12
Janice Nevin, M.D.
And again, it just creates that sort of qualitative experience for a caregiver that we hope will connect them to their purpose, bring more joy and allow them to spend more time with people.

00;18;19;15 - 00;18;47;27
Joanne Conroy, M.D.
It's interesting when I think of the visual, it's actually Star Trek medicine. It's actually coming to life with many different forms where it can actually bring things together, seemingly almost out of the air. Yes. In order to deliver better care to the patient, the bedside. Now, you talked about your process that you're kind of making sure that AI actually helps the care process rather than inhibiting it.

00;18;47;27 - 00;19;11;18
Joanne Conroy, M.D.
And there's been a lot of discussion at the Biden administration about, you know, how do you control it? But the adoption is so fast. The question is, how do you get ahead of it to actually create a process where it's rolled out in a methodical way? Do you think your process is fast enough? Do worry about that at all because the adoption is almost vertical, I would say, in our organization.

00;19;11;20 - 00;19;35;07
Janice Nevin, M.D.
Yeah, it's a great question. And, you know, like most CEOs, I want all the problems that we need to be solved as quickly as possible, preferably yesterday. On the other hand, you know, this is new technology. This is about caring for people. It's about safety and quality, it's about privacy and making sure that we use these tools correctly.

00;19;35;10 - 00;20;15;16
Janice Nevin, M.D.
And so, yes, some of the challenge is how do we ensure that, but not to the extent. And so we have to go at the right pace. So we don't want to make it all about the challenge and the risk. We have to also really understand the opportunity. And I will be so bold as to say, I think that if we do this work as a field and we do it carefully, intentionally and really work together to find those opportunities, I think ultimately this technology can help us solve some of the really big problems that we have in health care.

00;20;15;22 - 00;20;51;00
Janice Nevin, M.D.
We've talked already about workforce, but I can see there's an opportunity here because we can become more efficient, because we can execute perhaps more quickly doing the right thing, the best thing more quickly. We can also potentially start to radically improve access, take cost out for people who have to pay for care. So it's a balance. There are definitely risks there, but we can't let the risks cloud the opportunity and our voice in health care.

00;20;51;00 - 00;21;13;10
Janice Nevin, M.D.
I know the AHA will do this. When it comes to regulation, we want smart regulation so that we can move forward, use this amazing new technology, these tools to solve big problems and not do it in a way that is so restrictive that we can't address some of these issues.

00;21;13;13 - 00;21;40;02
Joanne Conroy, M.D.
That was great. Thank you, Janice. And I know we could keep chatting about this timely topic for another 25 minutes, but I want to be mindful of everybody's time. So on behalf of the AHA, Dartmouth Health and certainly ChristianaCare, all of us in health care are really appreciative of all of our listeners, and Janice for you, for bringing your expertise to the table on this really important topic.

00;21;40;05 - 00;21;46;02
Joanne Conroy, M.D.
Thank you, everybody for finding the time to listen. And I'll be back next month for another Leadership Dialogue.