Episode Overview
In this episode of Value-Based Care Insights, host Daniel Marino explores the evolving role of artificial intelligence (AI) in population health. As AI continues to dominate industry conversations and drive vendor offerings, healthcare leaders are faced with questions: What is real, what is hype, and where does the value lie?
Dr. David Nash, Founding Dean Emeritus of Jefferson College of Population Health and a nationally recognized thought leader in value-based care and population health, joins the conversation. Additionally, Rick Howard, a seasoned Chief Data Officer and AI Strategist contributes to the conversation with his deep expertise in leading data-driven innovation across healthcare organizations.
Together, they break down common misconceptions, highlight the most promising AI applications in care delivery, and offer practical insights into how health systems, providers, and payers can responsibly integrate AI to drive meaningful outcomes and return on investment (ROI).
LISTEN TO THE EPISODE:
Daniel Marino:
Welcome to Value-Based Care Insights. I am your host, Daniel Marino. As we think about the evolution of value-based care and population health, boy, there's a lot of activity that's going on. I've been doing population health now for probably around 15/20 years, and with now the increase in new technologies and in particular artificial intelligence, boy, there's just a lot of activity that's occurring. And along with that activity, I think there's a lot of misnomers as to what artificial intelligence is, how we believe it can support some of our business services, manage our population, all of that.
I had the opportunity to go to the HIMSS Conference, oh, I believe in March, and again spoke to many of the vendors. And I think 90% of the vendors that were there had AI within their name. It's certainly the new, the new buzzword. So many of the questions I often get asked as I'm having conversations with leaders around the country is help in separating fact from fiction. Right? What can AI do for us? How do we make sure that the artificial intelligence that we're starting to incorporate within our processes is real, can support our outcomes, is worth it, is going to create some type of ROI on it if you will.
Well, I'm really excited today to have 2 wonderful guests, certainly experts in this area. 1st is Dr. Nash, Dr. David Nash. Dr. Nash is a board certified internal medicine physician. He's the founding Dean Emeritus of Jefferson College of Population Health. He's worked in the population health space for 25 years. Certainly, you know the expert when it comes to population health.
My other guest today is Rick Howard and Rick is a senior leader, Chief Data Officer and has worked with a number of different organizations, and if there is any expert out there on artificial intelligence, it certainly is, Rick. So Dr. Nash, Rick, welcome to the program.
David B. Nash MD MBA:
Thanks! Great o be here. Great to have Rick with me, for sure.
Rick Howard:
Thank you, Dan, and same, and likewise great to be here, and great to be sharing the stage with Dr. Nash.
Daniel Marino:
So, Dr. Nash, maybe we could start with you. When you think about where population health is going right now, you know one of the big challenges that we've seen is, it really falls on a couple in a couple of areas. One, how do we aggregate all of this data to make sense of it. There's always a challenge in having providers, physicians, care managers react to activities instead of proactively managing the care. And I think just performing to the right level of outcomes. You know, when you think about where we're going with technology, how do you feel like it's going to influence what we're doing on some of those areas that I mentioned?
David B. Nash MD MBA:
Well, great question, Dan, and wonderful to be on a program talking about the future of population health, since I had the privilege of helping to start the 1st college in America called Population Health, edit the textbook, and edit the journal. So, now it's part of our everyday lexicon. Let me reassure you that 15 years ago, even on our own campus here in the great city of Philadelphia, nobody knew what we were even talking about. So, what a hoot and a half. Right? Okay?
So now to try to answer your question. Well, look, you can't improve quality and safety. You can't improve anything unless you're measuring it. So even the measures are a mess right now. And then, you can't improve anything unless you have the right data in a timely severity adjusted non-punitive way. And the current electronic medical records, not to mention any names, but the outfit in Madison, Wisconsin, doesn't give you the data that you need.
Daniel Marino:
Right.
David B. Nash MD MBA:
It's basically a billing system, no matter what they say. So I'm involved with a host of fantastic companies around the country that are trying to do exactly what you described. For example, an outfit like Innovaccer in San Francisco, California, which sits on top of all of these medical records and literally sucks the data out of the EMR, out of the lab, out of the outpatient arena, out of billing, out of clinical details, and puts it all in one place for the clinician, and there are competitors for sure to Innovaccer.
But we're only going to be able to improve care by giving clinicians timely non-punitive data about their own performance relative to a peer group. And then we have to teach them how to use the data to improve by standardizing care by reducing variation.
Daniel Marino:
Yeah, and integrating it into their process, right?
David B. Nash MD MBA:
Right.
Daniel Marino:
It has become more of the norm than the exception.
David B. Nash MD MBA:
And then we'll get to it, I'm sure. But we've got to then build the newer clinical decision support tools into the workflow, because I can't open more than 2 screens at a time. It's impossible, and most especially for the folks who are busy clinically still, not me anymore after 25 years. But you can't have the time and energy it takes away from the doctor-patient relationship. So, I'm really excited about all the new technology coming our way that will give us patient level data and population health level data. My only concern, Dan is, I hope I'm around to see it because, you know, we've been talking about this at HIMSS for 10 years, and I'm on the other side of the mountain. So, I really hope we get to see this technology. But the short answer is, you're right. We've got to pull the data from multiple places.
I would add one other thing at the patient level, we have to really harness those predictive analytics we've been talking about for 10 years, too, to prove that you're sicker than me. And you're going to need extra resources, and you might need a discharge planning, and you might need a social worker, and you might need, God forbid, a nursing home, and we ought to be able to know all that just as soon as possible. Let's remember the World Champion, Philadelphia Eagles had great predictive analytics on the field against those lowly Kansas City Chiefs, probably better than most hospitals.
Daniel Marino:
Right, right? And it clearly worked for you so that.
David B. Nash MD MBA:
And it clearly worked. That, that and the tush push clearly worked.
Daniel Marino:
Clearly worked!
So, Rick, let me turn this over to you. A couple of weeks ago you were on the program, and you made a couple of really interesting points. Right? You talked about the data. You talked about the opportunities that we have using artificial intelligence to really aggregate and make sense of that data. And one of the questions that I asked was, how do you know it's real? Right? So how do you know how to take this information, and you know, sort of I don't know, filter out the bias and really ensure that the information that we're getting out of our intelligence is something that can be then integrated within that workflow. Speak to that a little bit.
Rick Howard:
Would be happy to, and again great to be on, and certainly great to be here sharing the stage with Dr. Nash. But Dr. Nash is absolutely right. We've got a significant amount of data, even at the per-patient level that's out there. Any system is going to have to have clean data going into that system to make any sense out of it. So, data normalization is pretty much the step one. And it's not necessarily the easiest step when you're talking about data coming off of an EHR.
But once you get past the normalization layer, then you start to really get into the exciting components. So AI can help with the ability to build a clinical model supporting a specific chronic condition and giving you those prescriptive and even predictive measures that need to be taken, or that you can start to anticipate, and how you're managing that patient, that specific patient. I'm not talking about all patients with a chronic condition. I'm talking about a specific patient.
AI has the ability to do that by bringing the patterns together, understanding those patterns and then applying some of those patterns with additional information to a specific patient and their condition to be able to manage appropriately.
That's the exciting area of AI, but it gets better. You start to think about that application of AI. Well, now, you've got digital assistants out there that can sit and communicate with a patient, have a dialogue with the patient. And there's been a number of studies that state, the digital assistants may actually get a more transparent data set from the patient than an individual would, because the patients are going to be concerned about being judged, or having bias against them for their behaviors, their habits, their specific norms.
The more transparent that data is, the more rich that data becomes to be combined with the clinical data. And now you're having a legitimate conversation about not only clinical, but behavioral activities for that patient that will better manage their specific chronic condition.
Daniel Marino:
So that so that key, as I hear you describe this, right, is to take this data and to use it in such a way that we're much better informed about the patient. Right?
Rick Howard:
Right.
Daniel Marino:
So, Dr. Nash, as I'm hearing Rick speak to this, then it really becomes,you know, the role then, of the primary care physician, and in particular the specialist. Because I think the specialist is the one that's really been challenged here to use this to really provide more proactive care, right to think about that within their care model, and then frankly wrap around that care management service to help them drive with the patient what the potential servicing, the clinical service should be, and particularly the post-care outcome, should be.
David B. Nash MD MBA:
Well, I think that's partially right, quite frankly. But I see the interface of technology and specialists, maybe in a different way. Let me answer your question first, and then I'll give you my sense.
Daniel Marino:
Sure.
David B. Nash MD MBA:
So sure, in ACO reach, which is largely, you know, ambulatory and then team, TEAM, which is mostly inpatient. Better data is going to mean more efficient care, less waste, better outcome, lower cost.
Daniel Marino:
Sure.
David B. Nash MD MBA:
But the data requirements are formidable. Now for Part 2, as it relates to the primary care doctor and the specialist, here's what makes me excited as a primary care internist. I want to have good data on the performance of the specialists in my network. And I wanna pick and choose who's gonna be in my network. Because, look the way we currently do this is the primary care doctors are in the network, and we might own one or 2 specialists, but most of them are out there doing what they've always done. Private practice fee for service. And we largely have very little idea about their performance.
So, what I'd like to see Dan is incoming data on the performance of specialists. And then we decide who's in the network and who isn't with the ultimate goal of creating a high value, clinically integrated network, not a narrow network. I hate that term, but a high value network that can really deliver population, health.
So I think, you know, that's where the data flow, the incoming information back to primary care on specialist performance which we could do today. We got to have the moral authority and the economic incentives to make it happen.
Daniel Marino:
So if you're just tuning in, I'm Daniel Marino. You're listening to Value-Based Care Insights. I'm having a great conversation with Dr. David Nash and Rick Howard, talking about population, health and AI, and kind of the next generation of value-based care.
So what I'm hearing you say, Dr. Nash, one of the things that you see as a real opportunity here is to aggregate this new, you know, this set of data. These new opportunities, using AI to help to inform us about what's happening with our network.
David B. Nash MD MBA:
Exactly.
Daniel Marino:
To drive that awareness with our specialists in the hopes of making some type of behavior change.
David B. Nash MD MBA:
Right. So let's look around the marketplace right now, right, in the middle of 2025. There are organizations that are stealthily committed to doing this right now. We had 3 of them on stage at our 24th annual Population Health Colloquium here on our campus at Jefferson, Philadelphia, just a month ago.
So for example, Risant right? So Geisinger, Cone Health. These are outfits that know what they're doing. They're going to benchmark. They're going to compare data. Then, when they take additional delivery systems under the Risant umbrella, somewhere around one per year for the next 4 or 5 years. Imagine the data capabilities, of 4 or 5 places under the Risant umbrella. That's 1 great example.
How about Craig Samitt leading longitude health with the 5 amazing, gigantic delivery systems? Imagine being poor Craig with 5 big bosses, but beside that they'll be able to benchmark, too. Depending on how level the playing field is.
And then how about Populance? You know. Henry Ford thinks that it's got the secret sauce, and is so convinced of that that they're going to let Chris Stanley go out and sell it and deliver data and tools and care coordination and guidelines and care processes to other places. So look, this is happening.
Daniel Marino:
Right.
David B. Nash MD MBA:
I think, you know the folks out there who don't track these kinds of things. And, after all, you know, this is what Rick and I do for a living in part. I think folks are going to wake up one day and say, Holy Mackerel, I'm behind the times, let's get with it, and we're not talking about just an EMR. Let's just once again. That's, that's old hat.
Daniel Marino:
Going beyond EMR, absolutely.
David B. Nash MD MBA:
And then the final piece for me, which is really exciting, especially because I have forgotten more medicine than I ever learned.
I'm excited about all the clinical decision support tools that are being developed that could be embedded into the workflow, and tools like OpenEvidence. Tools that folks like Walters Kluwer are working on combining stuff like Avid and UpToDate. I mean, it's just super exciting.
Daniel Marino:
It is. It’s super exciting and just tremendous amount of momentum from some of those organizations that you mentioned.
So, Rick, let me turn this over to you, though, you know. So Dr. Nash just went through a whole host of different organizations. And you know, embedded functionality.
When you look, and you're seeing the landscape of where artificial intelligence is going, now. I sort of put it into 2 categories right? There's the advancements of the business services, and how we can do more with less and create greater performance. But there's also a lot of clinical support services out there that AI has a big opportunity to really provide some advancement, solve some problems.
Where do you see both of those at? Is one further along than the other?
Rick Howard:
Well, I think maybe the administrative side might be a little further than the clinical side. But Dr. Nash just made an important alignment around how you support a clinical workflow.
We need to look at things in the form of, “How does a physician get an instruction inside of their sourcing system so that they're following a best practice”? The industry calls that a computerized clinical practice guideline. AI driven computerized practice guidelines that are already interfaced, or, more importantly, interoperable with an EMR can slide that instruction into the EMR, even to the point of a change. Think about making a change as simple as going from a IV based antibiotic to a tablet based antibiotic. Most people don't understand. There's a significant cost alignment between those 2 particular distribution points of an antibiotic.
If we can trigger that decision inside the EMR to say it's time to move, based on all of the data that the AI engine has now formulated and understands where the patient is at that particular point in time. You can now make that switch between the IV based antibiotic and the tablet based antibiotic.
That is a simple example of a very much of a much broader application of these capabilities to drive these best practices. And the beauty of AI is, it's continuously pulling in data. So those best practices are changing continuously based on the fact that it's pulling data through.
Daniel Marino:
So then, Dr. Nash and I know this is your area of specialty here. It just leads me to believe that you're hearing a lot in the media about the advancement of greater diagnosing, of, say, cancer and other illnesses that are really impacting our population. From a clinical perspective, what are you seeing in terms of that advancement of a lot of that clinical support, a lot of the clinical servicing.
David B. Nash MD MBA:
Well, I think we're seeing a quiet revolution. I think our listeners are super sophisticated, and they know in fields like radiology and pathology, where machine learning has been in play now for 7, 8 years, we're getting better diagnostic accuracy. We're improving our efficient approach. We're reducing waste. We're reducing error when it comes to making diagnoses. And I'm excited about what the future potential is when we can put all of this stuff together.
Yeah, on my phone I've got, I alluded to OpenEvidence. I mean, these are incredibly sophisticated. It's ChatGPT 4.0, almost, on the clinical side. If I had a tool like this, when I was a intern and resident, would have been the most fantastic thing.
Daniel Marino:
Oh, my! Gosh!
David B. Nash MD MBA:
Saved me a lot of time. I would have slept a lot longer, so I can just imagine what my, you know, 38 year old Doctor daughter is going to have in her fingertips in the next 5 years at the bedside. So I'm excited about all of this, but I also think, you know, it's still culture change and and getting physicians to implement this into the workflow. We're only gonna do it, if it doesn't disrupt what we do every day, and if we believe in our heart that it will enable us to do a better job. And so all the technology in the world, we know culture eats technology before breakfast. So while it's great, I think we have to also think about the human machine interface and the workflow.
Daniel Marino:
How about some of these companies out there, you know, and I'll bring up one. You know. We had talked about early on Artera.
David B. Nash MD MBA:
Yeah.
Daniel Marino:
Really allows for just a vast number of aggregation of data as they begin to look at different clinical outcomes and so forth, to create diagnosis and probability of diagnosis, and probably probability of treatment.
David B. Nash MD MBA:
Yeah.
Daniel Marino:
Speak a little bit about that.
David B. Nash MD MBA:
Sure.
Daniel Marino:
Because I think that’s really a game changer as we see it.
David B. Nash MD MBA:
Sure!
Daniel Marino:
And that revolution that you mentioned.
David B. Nash MD MBA:
Yes, and again, you know, I'm excited about Artera and many other clinical databases which use AI enhanced capabilities. Artera is on the West Coast, and they have a 250,000 person data set of, you know, believe it or not, prostate biopsies.
And so for guys like me, who have had a prostate biopsy, and I got it read by an academic great person at Jefferson. But then I said, hmm, 250,000 patients versus 10,000. I'd love to have my slides read by that outfit, and so I managed with my academic urologist as my great doctor, send my slides off to Artera. Of course there's Jefferson graduates at Artera to help put all this technology together. How exciting! And look! It's not the end, all or the be all, but it's certainly as from a patient perspective, I found it very reassuring to learn that, yeah, I'm in a super low risk category, based on tens of thousands of men my age with my diagnosis.
And let's extrapolate that to pancreatic biopsies, lung liver, kidney. I mean, we're going to be able to do all of this and do it efficiently and effectively at a pretty low unit cost.
So these technologies are amazing. And I believe that companies like Artera and many other GRAIL being another great example, you know right now, the gallery test is, I think, something north of 40 potential cancers. Well, you know that they're going to add 40, 50, it'll be 60 cancers, both, you know, liquid biopsy for solid tumors and others. This is going to be continued to be super sophisticated.
Daniel Marino:
And I think just building off of that. It just creates another information point where it allows providers, specialists, even patients, to be much more informed.
David B. Nash MD MBA:
Yes.
Daniel Marino:
Going back to that, that overall philosophy and population health. If you're really going to make the change, you need to be proactive.
David B. Nash MD MBA:
You need to be proactive.
Daniel Marino:
Totally supports it. Right? Totally supports that drive.
David B. Nash MD MBA:
I would say. The last thing that I'm really excited about is what AI is gonna do for clinical trials. Right? So I mean, that's a topic we could have another hour on, that. We ought to be able to reduce the number of folks in trials we could do n of one trials do them, you know, so quickly, so efficiently. Get people of color, get women get all kinds of diversity. I mean, this is gonna blow apart the sort of, current CRO structure.
Daniel Marino:
Oh, it's gonna allow us to move so much faster. So, Rick, you know a lot of listeners from the provider community, right? We've got you know IT leaders on here, CEOs, CMOs, many hospital leaders. You know, it's challenging right? In my earlier comments I said “It's hard to figure out fact from fiction”. Right? What advice would you give to some of these senior leaders, the CMOs of the world, who are who are excited about this, and to be careful not to get wowed by the technology, but to really focus on those things that are going to create a lot of good value. Any suggestions come to mind?
Rick Howard:
Yeah, and I actually understand where you're coming from, Dan. There's a lot of noise in the signal out there, and so we've got to separate the signal from the noise. And I think, though there are some areas where there has been some very tried and true AI based solutions. Ambient listening is one of those. What a game changer from the perspective of allowing the doctor to actually be face to face with the patient and have a conversation about the problem instead of being face to face with the computer.
Dan Marino:
Sure.
Rick Howard:
So I think ambient listening is one of those. I think digital physician tools, and I think Dr. Nash just spoke to one of those. Digital tools that allow the mass of information that's being created quite frankly, hourly, on the healthcare front, to be at the physician's fingertips so that they now have the information to make the right decisions. And then I'll go back to the digital assistants, giving a doctor a digital assistant so that they can not necessarily even type a question, Ask a question. And get that answer coming back to them.
So I think there are tools out there that are going to make the doctor's life easier. And I'm sure Dr. Nash will agree with me. I'm excited, because not only does it allow the doctor to focus on the patient, it's also optimizing the doctor's visit timeframe to where we can start to eliminate some of the burnout that's being caused by pajama time, as I call it, with all of the work that has to happen after they leave the clinics, or after they leave the hospitals.
Daniel Marino:
I think we're gonna see, we're gonna see this technology moving fast and furious. I think I think at the end of the day it's gonna come down to what makes sense for the organization, and how you could truly incorporate that change, those opportunities into your population health strategy.
Well, gentlemen, I mean I appreciate your time. Boy! Quick half hour! We could have talked about this for another hour, hour and a half at least. I'm so excited to see where this is going, and I'll tell you I'd love to have you back on. And particularly you, Dr. Nash, as you start to kind of work through some of these organizations and seeing some real results of what AI can do. I mean to tell those stories is just so exciting.
David B. Nash MD MBA:
Super. Well, I look forward to it. Hope you have us both back.
Daniel Marino:
Absolutely! Absolutely, Rick, thank you as well, too. Appreciate it!
Rick Howard:
My pleasure!
Daniel Marino:
And I want to thank all of our listeners for tuning in. Without you, Value-Based Insights would not be as successful as it is. Until the next Insight, I am Daniel Marino, bringing you 30 minutes of value to your day. Take care.