Join us on this exciting episode of Health Impact Live, where our host Shahid Shah engages in a compelling conversation with BJ Moore, the digital transformation visionary from Providence. With a rich background from his 27 years at Microsoft, culminating as the Vice President of Engineering for Azure, BJ brings a wealth of tech experience to the healthcare industry. He discusses his transition to Providence and how his journey has propelled forward the use of cutting-edge technologies in healthcare environments that traditionally lag behind other industries.
In this session, we dive deep into how Providence has embraced digital modernization under BJ's leadership, creating three strategic pillars: simplify, modernize, and innovate. We explore the significant role of ambient voice technologies in enhancing caregiver interactions and the strides Providence has made in integrating AI to transform patient care and administrative processes. BJ also shares insights on the ongoing developments in AI applications, from Providence Chat, a private version of ChatGPT, to Grace, a chatbot that reduces patient-provider communication overhead, and more. Providence is also one of the largest users of Nuance communication. They use this to automatically create clinical documentation from patient visits in seconds with DAX Copilot.
Don't miss this episode to learn about the future of healthcare technology and how embracing innovation can lead to improved outcomes in patient care and operational efficiency. We wrap up with BJ’s take on the current trends and his advice for other healthcare executives navigating the dynamic landscape of digital health. Like, share, and subscribe on your favorite podcast platform.
BJ Moore, CIO and EVP of Real Estate Strategy Operations, Providence Health
Shahid Shah, CEO, Netspective Media, Chair, HealthIMPACT
Shahid Shah 0:00 Welcome to Health Impact live digital health talks. We're here with BJ Moore from Providence. So BJ, I love your backstory. Tell us a little bit about yourself and where you were working on before you got to Providence. Sure, sure.
BJ Moore 0:14
So a long journey to get to Providence and healthcare. I was at Microsoft for 27 years, a lot of different roles. My last role there was Vice President of Engineering for Azure. Fantastic company, but after 27 years, I was ready to do something different, still too young to retire. So my wife convinced me to come to provider healthcare and, and help do the digital transformation. So I joined Providence five years ago, and it's been a fantastic journey so far.
Shahid Shah 0:41
Yeah, no, I love Providence, and especially the innovation arms that you guys have doing all kinds of new things. So we'll talk we'll get into the nuance and DAX conversation. But what what have you noticed that some of the biggest learnings from what you took from Microsoft that you brought into Providence? And then once you were at Providence, there's probably a lot that you've learned that man, I would have never known this? If I was at Microsoft? Yeah. What does that sound like?
BJ Moore 1:05
Yeah, so when I joined five years ago, the very first thing I did is I visited most of our hospitals, clinics met with administrators, patients, caregivers. And one thing that was crystal clear as providence and healthcare in general was about 15 to 20 years behind every other industry, what technologies we're using how we are using the technologies or engineering processes, the lack of use of cloud or modern tools. So it's crystal clear that we needed that catch up. At the time, my premise was that health care was just change adverse, maybe risk adverse, and that's why we were so far, far behind that my big epiphany at Providence was just we were missing a Northstar a vision. And with my tech background, I was able to bring that into Providence. And over the last five years, we've not only done a major digital transformation, but we've caught up that 15 and 20 years and past up, frankly, a lot of fortune 100 companies that be envious of the modern stack, the modern technologies that we've been able to adopt.
Shahid Shah 2:04
No, that's perfect. And so if you think about where you started at Providence five years ago, you could have done 100 Different things that when you arrived. Yep. So talk about what were some of the key priorities. And that'll bring us into where we are today, which are big announcement here.
BJ Moore 2:17
Yeah. So five years ago, I came up with three strategic pillars simplify, modernize, and innovate, simplify, we had 4000 applications, we weren't going to create great patient caregiver experiences of 4000 applications. So simplify is how do we retire the non essential apps? How do we go to standard platforms like epic, Oracle, cloud, Microsoft 365, stack nuance? modernize was how do we move to cloud native applications? How do we start leveraging the cloud for our hosting and data centers? We chose Azure. And then also how do we do modern engineering, you know, DevOps, agile, everything we were doing before where it was waterfall, 612 month releases, so we had to modernize not only our infrastructure, but the way we delivered services. And that really allows for the innovation, right? It's almost like Maslow's hierarchy of needs, you know, simplifies like eating, breathing, sleeping modernizes like sending the kids to school. And innovate is like self actualization. And that's really what we're able to do now. Because we've been invested over the last five years and simplify, modernize, we really can focus on innovate. And it's a perfect time right with with generative AI coming in into focus 15 months ago, we're now poised to take advantage of it instead of doing a lot of housekeeping that I think a lot of other has health systems are having to do.
Shahid Shah 3:38
So Providence has prioritized ambient voice going from voice to text, voice to EHR voice to many things. Yeah. Why is that so important to problem?
BJ Moore 3:48
Well, one thing I heard five years ago, I heard two things. One, the onboarding experience was horrific, it took our caregivers have to be productive. But the other common thing I heard was just interactions with electronic health record that was burning out our caregivers. And we needed to do something about it. And it was actually at HIMS. Five years ago that I met nuance, for the first time, saw their early DAX prototypes, it was just absolutely blown away. So I had heard from our caregivers, that, you know, the, the, you know, EHR and documentation was a core issue. And I was lucky enough to stumble upon nuance who hadn't been acquired yet by Microsoft, and really forged a partnership. So we, we started that partnership five years ago. And it's really been a five year journey to culminate where we are today.
Shahid Shah 4:35
What is the scale look like for your implementation?
BJ Moore 4:39
You know, I can't keep track of the scale. I know nuance, and Microsoft tells me that Providence has the largest tax copilot user in the world. I can't give me the exact count, just because that never changes. But we went live three weeks ago, and you know, really proud because, you know, we had to be on Epic hyperdrive to really get that rich capability. And then again, that nuance partners But for so many years paying off a tax copilot, so all the stars appear to be aligning, but it's really been a five year plan a five year journey to get there.
Shahid Shah 5:07
Yeah. And so as you are embarking on this mission, what are you surprised by as far as a maybe utilization by physicians, either acceptance or non acceptance? What are you seeing just on the ground
BJ Moore 5:21
as a relates to DAX itself? The adoption is good. I mean, there's, there's always people that are change adverse, but generally, we've got a set of physicians that are really bright, they want to make their lives better, right. And if we're bringing the technology to them, that makes our lives better, the acceptance has been good, because it was a five year journey, you know, the early iterations were tough. And so we had to pick the right doctors to partner with to help us on that journey as we evolved the product. But you know, five years later, the product is much more mature the adoption, you know, that change management, the acceptance of the tools much better. And when we survey our physicians, they would absolutely not allow us to take this away, it's now become an integral part of the work that they do every day. And if
Shahid Shah 6:03
you assume that ambient voice is in quotes solved, there's never been a journey. Correct. But if you think of that as being a now a met priority, right, your next two or three priorities that you're focused on? Yeah,
BJ Moore 6:16
it's definitely not a met priority. But but you know, what are the next set of priorities? It's really how do we take this this new, powerful AI tool, we've been using AI for years, you know, basic machine learning, predictive modeling, you know, natural language processing, we've already been doing that. It with generative AI and these large language models, it's just really changing the way we're doing things. So I'm happy to say we already have four things in production, I can share with your listeners, what those are. So the for our first one, we developed a tool called Providence chat. I can talk about generative AI till I'm blue in the face, but people don't get it. So what we did is built Providence chat, which is basically a private version of Chet GPT, we took the Azure open AI, wrapped it with a set of API's, and we allow then our 120,000 caregivers to interact with it safely in a secure way, the data doesn't go back to open AI or Microsoft, it stays within Providence. And now people can see for themselves the power of it, they can use it to do a recipe to write their job description, I use it to write my personal job performance review for the year, it's fantastic. And so people can see firsthand the power of it in a safe manner. But then my IT team sees all the prompts. And we know exactly what people are asking. So we can see the most common scenarios that we see people asking over and over again, to write job descriptions, where we can take that and partner with Oracle and bake it into our ERP. And so we can use that to create job descriptions within the workflow of of of the application. Or if there's, you know, sub clinical questions that people have, we can you know, figure partner with epic and figure out how to integrate it, you know, there. So allow a democratizes open AI, but it doesn't have a safe way that we can then capture. So that's one, two, we add a chatbot called grace that our patients interacted with. We've enhanced that with generative AI capabilities, we've seen a 30% reduction of the number of tickets or our messages that they send our physicians because they're able to self serve more effectively. So it's good for the patient has good for overworked doctors. And then we built a tool called probe ARIA for our doctors to actually manage their inbox. When we look at our physician inbox, it's taken about three and a half days to get through their inbox. This is a long time slot effort for the physician. But if you have a critical issue is a patient three and half days is a long time to wait. So we're using generative AI to triage those messages and putting the most important messages up front. A real example is a patient was having suicidal ideation. And the whole process that could have taken the physician three and half days to get to that with generative AI a new is a top priority put in front of the physician able to reach out to the patient intervene and potentially saving a life. And then the final solution we have is called Med pearl. It's for physicians and more effective referrals. You know, we're finding that the effectiveness or referrals or the quantity of referrals was kind of inappropriate. And with med Pearl kind of within the physician workflow. They either can self serve and solve their own problems or when they do refer to the physician they refer to the right specialty. So I'm proud that we're not only just talking about generative AI, but we've got for real solutions in production that are being used by hundreds of 1000s of people.
Shahid Shah 9:32
No so it sounds like you guys definitely didn't take the wait and see approach now. So how would you so now it now you've got two hats that you've worn one at a big tech company, then you've got the five years at Providence now. How would you suggest to executives at hospitals about this wait and see attitude? Lots of them have this wait and see right now? Is that dangerous? Is that a good idea? What What are your thoughts?
BJ Moore 9:55
Yeah, I think it's dangerous. I mean, that's it's such a hot technology. You can't Wait and see, you know, take chat GPT a lot of health systems are just blocking, blocking, oh, we don't want our employees using chat GPT. So they block it well, people are just gonna go home on the private network and use it. Right. And so that's why we built Providence chat, we know you're going to use it. So allow us to let use it in a safe way. So yeah, I would say embrace it. You'll notice the examples I gave none of them were making clinical decisions. They're all assistants are helping assistant to a patient or assistant to a doctor. And we've done that consciously. There's so many administrative functions that we can use this new technology for really perfected there have real impact to the productivity or caregivers outcomes for patients, unless perfect the technology and as it perfects, we can slowly move to more clinical scenarios. Because we're only going to get a first impression once with their patients, we're only going to get a first impression once with our doctors. We don't want to stumble with this new technology. So we're caught. So my advice to your listeners are embrace the technology that you can embrace in areas where it's low risk, if you do stumble, it, you know, doesn't have a high impact.
Shahid Shah 11:08
Yeah, and I get the sense that there are a lot of startups probably knocking on your door saying, Hey, we've got this AI, we've got that AI. One of the things I love about the AI platform shift, if you see it as a shift is that incumbents have more power because incumbents have all the data, right? Providence has all the data. And so introducing AI into your data streams, with your own developers with your own innovators. That's what I liked about provident, you're not waiting necessarily, for the next big thing to come from, whether it's Microsoft or anyone else, you're just doing it on your own. Is that something to only Providence can do or there, it's probably possible at other institutions, it's
BJ Moore 11:44
possible to others and it goes back to my first two strategic pillars around simplify and modernize, right? By having big partners like Microsoft or nuance Oracle ServiceNow epic, we can write on there. I call it the easy button AI solutions, because we're you know, we're on a single instance of epic, right, we fully embraced epic for all of our patients, as epic innovates on the AI front, easy buttons, we can just then choose to turn on those AI capabilities. We don't need to develop them ourselves, right? We're already on the version, we're already in hyperdrive, we can already take advantage of these. So we can just add our pay start to consume these new AI features, which is great, right? So. So right now I've got a long line between all those partners of all the new AI capabilities, right? So I talked talked about the four solutions were to have in production. But in six months, we're going to have a ton of Oracle things in production, we're gonna have a ton of epic things in production. We're using Microsoft copilot. So with very little effort, we're going to light up our enterprise with a lot of rich AI capabilities. Because we picked modern standard platforms and we picked good partners, good strategic technology partners that are embracing AI so we can ride on the coattails of their innovation. Yeah, that
Shahid Shah 12:59
makes a ton of sense. Earlier today, I interviewed John halamka from Mayo, okay. And he was talking about how they tried to reduce their risk in using AI in clinical by anonymizing and D identifying data. Sure. And using that in the AI engine, because it's anonymized there'd be identified it wouldn't necessarily, even if it went crazy, it wouldn't be giving anybody else's data out. What do you think about that idea? And that's something you would be trying at Providence as well. Yeah.
BJ Moore 13:26
So we we did created a data consortium called True Veta. Providence was a founding member of that about three years ago, there's now 35, other health systems. So that's kind of our arm where we're going to be innovating through Truvada and data. And they're taking that exact approach, right? It's it's the data from 35 health systems. But it's all been anonymized. It's all been standardized and cleanse and D duped. And then that kind of innovation can be built there, the large language models, the, the, you know, the correlations, the things that we as human beings can effectively do, and do it at scale, but it doesn't scale on the data collected data of you know, about 100 million patient records. And so I absolutely agree with the mayo approach, but we're doing it at a greater scale through through our Truvada investment.
Shahid Shah 14:12
Oh, that's a great idea of pulling, pulling together a whole bunch of others just to get exactly.
BJ Moore 14:16
Yeah, the large data set. And then I really love the diversity, right, we've got patients from Hawaii, Alaska, Texas, you know, all over the east and west coast. So not only is it a big data set, which hopefully your listeners all know, the bigger the data set, the more effective these models are. But it's a very diverse data set, right, the individuals and their diverse themselves now
Shahid Shah 14:36
fantastic. So just in the last couple of questions, then if you think about the way that Providence is attacking the next few priorities, what in your infrastructure the blocking and tackling kinds of stuff is needs to be upgraded? Yeah, before you can do all the modern stuff, you got to do some basic blocking and tackling.
BJ Moore 14:55
Yeah, so if I go back to my kind of framework around AI the easy button kind of things, if you pick the right partners, I think you can do a lot of AI innovation, you know, like through epic without investing in your infrastructure. But if you're doing the the bespoke kind of work that we're doing, you really need to be in the cloud, right? The cloud is the only place he really effectively do AI, it's the only place you have the compute and the storage needed to do that. So those are shortcuts you can't take as a health system, we've been on that five year journey, all of our data is in the cloud, all of our major systems are in the cloud. So you do have to make that investment to really take advantage of this kind of wave of technology.
Shahid Shah 15:35
And that's really the big advice is that at HIMS, we're seeing that everybody's so focused on AI, cyber, etc, that they're forgetting that there's a lot of regular day to day blocking and tackling day to day telecom day to day kinds of tech support that have to be done
BJ Moore 15:50
back to my Maslow's hierarchy of needs, right? If you're not eating, breathing, sleeping, you're not self actualizing. And same thing with innovation. If you haven't taken care of those foundational items in your technology stack. You really can't do the rich innovation that your system needs to do.
Shahid Shah 16:03
On the way out. What's something that you've seen here at the conference that maybe has wowed you or you're like giving a second thought to because you weren't hadn't thought about it before? And
BJ Moore 16:14
God I wish I said I had the time to have gone around and seeing that's my favorite thing to do and years passes the block all the booths I normally put on 1012 Miles unfortunately, so far, I've just been in interviews and meetings. So I've had a chance to do that. So I haven't seen the wild technology. But I hope I hope to be able to walk the floor and see that and be wowed. So nothing to share yet but hopefully I'll have something by the end of the week. Great.
Shahid Shah 16:36
All right. Thanks, PJ. Really appreciate your time and we'll catch up with you after the event so that you could tell us what you did learn.
BJ Moore 16:42
Awesome. Thank you for the opportunity to speak. I appreciate it.