S1E15: Capturing Marketshare with AI (ft. Nic Nguyen, Tufts Medicine)
Healthy Uptime Podcast Nic Nguyen (Tufts) _ Jordan Cooper (Rackspace)-20261103_145940-Meeting Recording
November 3, 2026, 7:59PM
23m 57s
Jordan Cooper started transcription
Jordan Cooper 0:03
in the Chief Clinical Officer and Chief Integration Officer at Tufts Medicine. Nick, thanks for joining us today.
Nguyen, Nic 0:09
Thanks for having me, Jordan.
Jordan Cooper 0:10
So for those who don't know, Tufts Medicine is a health system headquartered in Burlington, Massachusetts, in the greater Boston metro area. Now, the Tufts Medicine Health System has 1,100 beds spread across 5 hospital campuses with an additional 140 additional ambulatory outposts supported by 3,800 physicians and APCs. So
Nick, you're in a competitive urban environment at a time where gaining market share is crucial to improving margins, and you can't rely on a big pool of referrals as do larger health systems in the region. Instead, Tufts has been differentiating itself with value-based care, being hyper-focused on keeping patients healthy and leveraging technology.
Can you give me a few examples of how you've been leveraging technology to gain market share, particularly in the more profitable commercial payer population?
Nguyen, Nic 1:01
Yeah, it's a great question. And you're absolutely right that we are in a very competitive environment. And so we need to understand how to reach patients in the community in different ways than the traditional health care delivery models, which is providing care in an office or practice and then referring to a hospital or a specialty care. So a couple of examples where we're thinking outside the box and trying to deliver the care and get the care closer to patients.
more easily at their fingertips is our virtual care offering. So in that modality, we offer primary care virtually. So virtual, we call it virtual first, but not virtual only. And so patients can access preventative, chronic, acute care needs through a virtual care platform, which functions as their new primary care clinician.
And then we pair that with a very dedicated navigation system so those patients can get connected to ancillary diagnostics, as well as specialty care within a very close time window from the initial onset of their interface with the virtual primary care program. So that's one of the ways that instead of relying on brick and mortar and, you know, actual sites of care,
We are delivering care in a way that is closer to the patient, easy for them to access, but also helps them navigate a very complex healthcare system.
Jordan Cooper 2:17
So I love how the answer is almost always both. It's not just digital and it's not just doubling down on brick and mortar and in-person visits, but it's integrating the two. I'm wondering what sort of kind of cultural adoption issues you've had with both your clinical staff and almost more importantly with your patient population.
Nguyen, Nic 2:38
Yeah, I think you answered the question by asking the question at the beginning, which is a yes and, it's really everything. I mean, healthcare is extremely nuanced and there's so much demand out there. One of the nice things about Tufts is that it's not just blending AI and digital and other modalities with also our traditional ways of delivering care, including
Jordan Cooper 2:41
Hmm.
Nguyen, Nic 2:58
you know, our ambulatory brick and mortar practices. But we're also thinking about this very pluralistically, which is that we are partnering both with employed and private and independent practice across a vast integrated network to work with each other. So for instance, the example I brought up is the virtual primary care model will partner with also not just employed assets within
the Tufts Medicine system, but we'll also work with private independent primary and specialty care practices and understand how we can work together to deliver the same level of high quality care to a patient that may live far in the community. And then to answer your question, that's all change management, right? Partnership management and understanding, well, how do we reconcile
where the digital starts and where it ends with where our more traditional ways of delivering care start and end. And the same with our collaboration between our employed primary and specialty care assets, as well as our independent private primary and specialty care partners. So it's really collaborating across all of those areas.
to make sure that we're approaching a patient as holistically as possible. I'll say one more comment. I know we need to move on, but taking the patient lens, you might imagine that instead of them getting lost in the maze of healthcare, they might be able to say, this is a really great concept. I'm able to engage with my virtual primary care, but then I'm able to be referred to a private
cardiologist in this part of the Tufts network that's still part of the Tufts family, but in a different framework, even though they're still part of the family. That just enhances the ability for us to offer different care modalities, options, regionalities to our patients by leveraging all of our assets, both in person
digital, as well as private, independent, and employed.
Jordan Cooper 4:41
So I have two follow-up, I mean many, but two immediate follow-up questions on that. One, are you seeing overall utilization of the health system going up or down with all these additional referrals? I'm wondering if one, your partnered employed and independent practices, are they having more volume, more revenue? And also,
kind of, you know, you're trying to do value-based medicine, which ideally is reduced encounters, improved health. So how's utilization being affected here?
Nguyen, Nic 5:10
Yeah, it's a good question. I mean, utilization in health care generally is pretty high because there's so much demand. The truth is we have a high health care literacy population in Massachusetts as general. So there's a lot of demand for primary and specialty care services, even though there's a lot of competition. So there's a lot of utilization.
I think the key to value-based care, or one of the keys to value-based care, and this is a huge part of the Tufts Medicine culture, is making sure that that utilization is at the right level of acuity. So the things that we want to think about when we're leveraging AI or virtual first or other tools that help navigate the patient to the right place is to make sure that we're matching
their needs and their acuity with the most appropriate setting. So what I mean by that is so many things of what we can do for patients should be on the front end of things, in a preventative spectrum, thinking about health maintenance, thinking about caring for chronic conditions before they become acutely exacerbated. So the whole point of leveraging all these AI innovative digital tools, as well as leveraging
a more diversified and varied network is to say to the patient, let's take care of your congestive heart failure in the primary and cardiology office rather than waiting for it to end up, you ending up in the emergency department or the hospital. Let's make sure that we are able to have access so that you can take care of that sprain or strain or urinary tract infection.
as opposed to you having to go to the emergency department. So I would say the utilization is always high. Our goal now is to move that utilization from high acuity to, you know, in high sort of high risk and high medical complexity to a place where we can manage it more on a preventative and maintenance standpoint.
Jordan Cooper 6:52
Yeah.
I appreciate that perspective on where you're trying to get the right care to the right patient at the right place and front load the preventive care. I'd like to ask an additional follow-up question on how you're supporting both employed and independent practices. Now, I know that you're making investments in a great deal of IT resources, which they independently, they being these independent practices, probably can't financially support.
Nguyen, Nic 7:01
Yeah.
Sure.
No.
Jordan Cooper 7:17
How do you extend these tools and best practices to those practices without their own IT resources and governance structures?
Nguyen, Nic 7:24
Yeah, it's a really great question, and I don't have a clear answer for you. All I can say on that front is a lot of the different innovations and technological advances that we're either doing or exploring, and I can name a few, we try to pilot them in places that are
that have some readiness for it, both from a culture operation standpoint, but also from an EHR and to be frank, legal standpoint. There are things that we can do in our employee practices that we can't do in our independent practices and vice versa. So we have to be thoughtful about where we launch an innovation based on that quote un quote readiness.
But wherever we can, we try to do things both on the employed and the private independent side. Because again, if we're stronger as a network and more consistent as a network, that will enhance the care coordination for the patient. So some examples I might bring up on this is, for instance, we have an AI-generated scribe tool that we use.
And we are able to deploy that to any of our practices that have Epic. And so we have different versions of Epic for our employed and independence. I don't want to get too much in the details of it, but needless to say, those are the types of things that we try to share more broadly than just the employed complement. We are working on different triaging tools that leverage AI. And again, when
We're working on something that needs to be in a very controlled setting, a testing environment. We might try an employed part of the organization before we figure out how we can deploy more broadly. So we do try to support both our employed and our private independence, and sometimes trying something new that's innovative and that involves technology, it's almost a level of
where the readiness is from all the standpoints that I mentioned, both technology, operations, and even culture.
Jordan Cooper 9:18
So you've mentioned that Tufts has made substantial investments in consolidating its technology environment on AWS. Has that transformation made it easier to scale initiatives like the AI triage innovation that you mentioned or the Microsoft DAX ambient documentation? Are there still barriers preventing broader adoption after you've consolidated into a single cloud environment?
Nguyen, Nic 9:41
Yeah, I think the answer is yes to both, right? In the sense that, you know, the consolidation and integration piece has certainly helped us have the opportunity to accelerate where we program, pilot, or expand some of these initiatives. And I think there's still challenges. You know, we all know that as health systems come together,
their coming together is usually a multi-year, if not multi-decade journey of full integration. So whether it has to do with nuanced operations or still separate functional entities that are connected under one emblem or one umbrella or different EHR utilization, I think that
variation creates some challenges to consistently push initiatives across the entire health ecosystem. So the short answer is I think the integration consolidation does get us on that journey and helps us move along and gives us more opportunity as we move forward. But there are still challenges, quite frankly,
because as you know, a system integrating takes multiple years, like I mentioned.
Jordan Cooper 10:48
So Nick, listeners to Healthy Uptime Podcast love hearing concrete examples of exactly how your solutions are functioning. So if we take AI triage, for example, and perhaps the condition of a urinary tract infection, can you talk about how that works potentially in relation to messages in the emergency department and
Nguyen, Nic 10:58
Yeah.
Mm.
Jordan Cooper 11:08
maybe its impact on per member per month performance and what data you need to provide to payers, et cetera.
Nguyen, Nic 11:11
Yeah.
Oh, it's a great question. It's a good example. So the place where we're using, where we're planning on using AI triage for clinical triaging is twofold. It's in pilot pockets. One of them is tied to our virtual primary care first model to give patients access
more around the clock guidance, especially after hours. And we are looking at whether we can use that in our traditional primary care practices as well. So all of this is in motion. So to use that UTI example, you know, in our ideal world, what would happen is that whether a patient's trying to reach our virtual primary care first or our brick and mortar traditional primary care practices,
Maybe they are not able to reach a person because it's after hours on the weekend, but they are able to access this AI triaging tool. And in an ideal world where we are ensuring that all the quality and safety and clinical best practice and standards of care are incorporated and built into this AI, a patient who is at risk of pyelonephritis, which is more severe advancement of a UTI,
would be guided to an emergency department or hospital, and that's what they need to be. The other patient might be triaged or redirected to an urgent care where they could get, you know, absolutely adequate care. And another patient who thinks they have a TI but doesn't quite line in terms of symptomology might be able to be scheduled for the next available outpatient appointment.
And that patient, if they didn't, weren't able to get with some guidance, might end up in the ED anyways, even if they didn't need it. So you can see that a complex and safe and robust AI triaging tool could help the patient land in the right care setting, which at the end of the day, let's take the example of the patient that does not need to go to the ED or the hospital.
they would end up getting guidance to redirect to their primary care clinician, which from a total medical utilization, the PMPM performance should overall and aggregate, if you multiply that by thousands and thousands of patients, deliver on our performance for value-based care and on PMPM.
Jordan Cooper 13:20
And not only would that be a boon to the financial performance of the Tufts Medicine organization, but it would also improve patient experience because nobody wants to sit for six hours in the weekend in the emergency department just to be told they never had to arrive there in the 1st place.
Nguyen, Nic 13:34
That's exactly right. You're spot on. It hits on two points, which is it helps the patient experience, it gets the patient the appropriate level of care, and it also helps the system in terms of the access that it needs to provide to the patients who do need to go to the ED or to urgent care. You can imagine that instead of that patient being in the ED for hours and hours, the ED staff can now take care and triage the patient who has chest pain and MI that's evolved.
It needs to be triaged more quickly and end up, you know, at hospitalized. So it load balances the organization. Now, I will say, you know, it's an exciting proposition to think about that, a value proposition to think about that, but we still have a journey to go through in terms of the safety, the clinical robustness, and the continued expansion of a tool like that to.
feel really, really confident that it's doing what we would otherwise do or better.
Jordan Cooper 14:24
So Nick, we've been talking about getting the patient to the right place, the right time for the right acuity of care. I'm wondering about if we could pivot to getting data, reliable and timely data into the hands of clinicians and care managers at the right time in order to provide the right level of care. So I understand that Tufts is transitioning towards a data lakehouse architecture to support predictive analytics and machine learning.
I'm wondering what clinical use cases you're most interested in that will enable be enabled with that investment to get the right data to the right clinician at the right time.
Nguyen, Nic 14:57
You know, it's an excellent question. You know, there's so many examples. I'm someone who always is a data geek, so I love to leverage data and understand how we can implement it and share with our clinical teams in a way that can impact change on our patients. I think one key example that we're thinking about is in the area of something we call ambulatory care sensitive conditions.
That is an initiative wherein if a system is working really well across its different outposts, primary care, urgent care, emergency care, hospital care, primary and specialty care, again, independence and employed, that we should be able to reduce the amount of patients that are, that either go to the ED and or admitted to the hospital for conditions that could be cared for otherwise, either
upstream from a timeline perspective or upstream from a location perspective. So it harkens back to the phrase you said about right care, right place, right time, right people. But the only way to do that is to ensure that we're leveraging data, being able to get data out of where the patient's journey has gone.
So getting the right diagnosis data accurately documented, making sure we can capture that and that episode of care and the location where they're at, navigating that across the board so we can see by location, by certain diagnosis, where are areas of vulnerability and where are the areas that we should focus on based on that data to say,
we need a better COPD management program here, or we need a better diabetes care program in the ED over here. So there's different initiatives that we can do, but oftentimes we're relying on our guts and what we hear in stories we see in the ED. But if we can leverage diagnosis data and patient care journey data,
marry that together and see where we can target our initiatives that will certainly help with a large endeavor like ambulatory care sensitive condition reduction across the system.
Jordan Cooper 16:54
So given that you have this combination of a growing AWS footprint, the consolidated Epic environment, the emerging data lakehouse architecture, the increased SAS adoption, that kind of raises a question. I mean, you have a lot going on. Who's accountable for the ongoing operation performance of the entire ecosystem?
Nguyen, Nic 17:14
Oh, I don't know. You tell me. I'm going to ask you that question. I just got overwhelmed, Jordan, just hearing all the things you enumerated. You know, it's a great question. And I think what you're asking is a question about governance. And at the end of the day, the way that I think about health care systems, and this is at Tufts and beyond, and Tufts, I think, does this exceedingly well,
is that in healthcare, any governance model that oversees, whether it's AI initiatives, innovation, interoperability, clinical excellence, it has to be multidisciplinary, multifaceted. So you always have to think about, you know, who are your operational leaders, your clinical leaders, your financial leaders, your
technology and IT leaders, bring those all together to make sure that they have a lens in the different things that you're trying to govern. So I guess my short answer is it lives across all the disciplines that manage healthcare. And I do think that whether we're talking about UTI triaging, virtual primary care first, data analytics warehouse, all of those things need to have
a governance lens that includes all of those different facets of healthcare. Otherwise, you kind of lose a little part of what's important to patients and what's important to a healthcare system. Ultimately, I would say the thing that that governing body needs to think about are the following tenets and principles.
no matter how fancy something is or how complicated an integration is, is number one, are we doing right by the patient and the communities we're serving? And #2, are we doing it in a way that sustains the people that are caring for those patients? If we can keep those two tenants in mind and we can have a governing body that is multidisciplinary across the key facets of healthcare,
I think we're on the right track, but it's exceedingly complicated. And I answered your question with a general statement because I don't know the answer myself.
Jordan Cooper 19:07
So, I mean, that's fair. You know, so we're providing solutions from many different data sources, many different stakeholders, many, and a lot's going on in a complicated health system. I'm going to kind of go, and that's what's working. And we presented an example of what's in the best case scenario, what happens to a patient's journey when technology does what it's meant to do.
Nguyen, Nic 19:29
Yes.
Jordan Cooper 19:29
Now, I mentioned that Tufts has made major investments in creating a unified cloud environment, which presumably gives you a stronger foundation for digital care and AI. So now we've talked about what has worked. From your perspective, what's the biggest gap between what that technology platform is now capable of delivering
and what clinicians and patients are actually experiencing.
Nguyen, Nic 19:52
It's a great question. You know, again, there's a number of examples where I think that there is a gap where AI and technology could solve that. Yet at the same time, there is some...
understandable and healthy anxiety and caution about leveraging a certain technology to meet that gap. So the things that I'm thinking about that will come down the pipeline eventually are leveraging AI to make sure to help with radiology findings, and then leveraging our integration and consolidation to say,
that that AI generated radiology finding is something that is observable, shareable, and verifiable across all the different parts of a system that's still integrating. So you can see that that is a very complex continuum that includes both, you know, EHR integration,
consolidation of clinical protocols and safety, as well as marrying that with AI. So I use that as an example to say you can see how complex it might be to fill that potential gap. And along with the complexity that you can just imagine what I outlined, there's also a healthy reticence.
from patients and clinicians to say, to fully trust something that may feel extremely automated. And I don't say that in a negative term because I do think that as healthcare people, we have a huge responsibility, right? This is not so much leveraging AI and being okay with the mistake we made in the scheduling for.
you know, some fancy service that is not healthcare related. We're dealing with people's lives, we're dealing with people's trust, and we're dealing with clinicians' belief that we should leverage some of these technologies. So what I'm saying as a barrier that is a normal barrier or challenge is overcoming that and being able to
produce something that is not just automated, that leverages technology, AI and integration, and also engenders trust amongst the patients and clinicians that are pulled into that.
Jordan Cooper 22:01
So Nick, we're at the end of this podcast episode. I'd like to pose a final question to you. We have covered a lot of ground today. And I'd like to ask you to synthesize what we've covered and offer one or two lessons learned that you think would behoove listeners this episode to keep in mind as they try to confront their own challenges in their own health systems.
Nguyen, Nic 22:07
Sure.
Yeah.
Yeah, good question. We did cover a lot. I would say you have to start somewhere. That's the first thing. So no matter how complicated and how big or small your system is, if you have an AI or technology-driven solution, before you get scared away by, you know, the
what's the word I'm looking for, the skepticism around it, as well as the clinical robustness, I would say lean into that and explore it. And there's always a place to pilot it, and there's always a place to spread it beyond there. So that's one. You got to start somewhere. Second thing is think about what you're trying to solve for. You're trying to help patients. Is it the right thing for patients? And is it something that you're
your teams can sustain. And then the last thing is even something very specific, like a very, you know, isolated AI driven initiative that seems only specific to one specialty or one part of your healthcare system, the more parts of your healthcare system that you can include in the oversight advisory or governance of that,
I would say that's the way to go because there's always going to be either operational, financial, a population health, a value-based, a clinical, a risk management lens that impacts the way that you do this and will make you better for it in the end.
Jordan Cooper stopped transcription
