00:10:11 - 05:10:11
Jerrod Bailey 00:05
Welcome, everyone to Reimagining Healthcare: A New Dialogue with Risk and Patient Safety Leaders presented by Medplace. We're excited to bring you conversations with top risk and patient safety thought leaders from organizations around the country. Please subscribe to get the latest news and content. And if you found this valuable, particular episode, please share it with your colleagues to create some meaningful dialogues in your own community. And also, if you're interested in participating as a guest, please send us an email at speakers at Medplace.com. My name is Jerrod Bailey. I'm the CEO and founder of Medplace and I'm going to play hosts today. And today I'm joined by Julius Bogdan the VP of Digital Health Advisory Team at HIMSS, which is the Healthcare Information and Management Systems Society. Welcome, Julius.
Julius Bogdan 00:55
Thank you, Jerrod. Nice to be here.
Jerrod Bailey 00:57
Likewise, Thanks for being here. So Julius leads the digital health advisory team at HIMS. Right? for North America. His mission is to build a culture of trust in shared interest in data and technology's role in improving clinical and financial outcomes and to increase the capacity for making data driven decisions and healthcare. After a successful career at Microsoft, Julius became personally interested in the human benefit of technology to improve healthcare management. The interesting see come from a technology background, somehow you got suckered into the healthcare world.
Julius Bogdan 01:35
Yeah, I did, I found it always fascinating a lot of green space in healthcare, because I came up through financial services, manufacturing, where they were early adopters, and heavy utilizers of technology and advanced technology, pioneers in AI, so on so forth, and healthcare presented a much more interesting space, because you're actually impacting people's lives, right. So when I was evaluating my career, and where we wanted to take it, improving people's lives, seemed like a pretty good way to go. And I wanted to see what I could do in healthcare.
Jerrod Bailey 02:12
Interesting. So tell me a little bit about your role, like, are you an evangelist? Are you? Are you consulting with companies, you're kind of doing a little bit of everything? What, what are you doing there?
Julius Bogdan 02:21
Yeah, so in my role, I work with global healthcare providers academia, in the vendor community to really understand well, where healthcare is today, and where it's going, right. As an organization, we do a lot of research, and we take that research, and turn it into insights and build products, such as our maturity models, to assess where organizations are in their digital health transformation journey, and then provide guidance on the strategic roadmap, capital investments and validation of their decisions and strategy. Among other things one of the key areas that we assess and work with his analytics, right. And so that's what we're here to talk about today. But we cover everything from the more adoption utilization, supply chain optimization, diagnostic imaging, across the enterprise all of the main healthcare functions globally, we kind of is part of our scope.
Jerrod Bailey 03:21
That's great. We talk a lot about workforce and human capital optimization here. There's a lot of interesting things we can learn from other industries around we're a marketplace model. But there's lots of models that show just how efficient you can make a system by you hear the word democratizing its use too much, but by relieving some of the friction between entities, and, the people working for them and optimizing these highly trained physicians to do what they do best. And then, and then taking all of the other overhead that isn't optimized off of them. And I think doctors want that., I think the administrators want that. But it's going to take a level of thinking that is much more advanced, I think, in other industries before I see some more of that tangible benefit. But you're at least for today's conversation, you're pretty data. First, you'd have sort of a data first approach or you're very data centric, as far as where you see the opportunities to improve healthcare, right. Yeah. My little color on that like,
Julius Bogdan 04:36
yeah, no, absolutely. And we'll talk about some of the automation that's also under this umbrella as well. But yeah so data first is a little bit of an outdated approach. And let me let me explain why. With the advent of the EMR., healthcare really has become a data industry. We're collecting reams and reams of data. Data is the lifeblood of any or healthcare organization today. The challenge isn't me keen isn't taking a data first approach because we're swimming in data. But it's to be able to derive actionable insight from that data.
So a while ago very coming up through the ranks of psychology in having worked with data and analytics and advanced AI capabilities years and years ago, I remember doing data science exercise, data mining back in the day, where we analyze their survival rate of passengers on the Titanic, right? So we used all sorts of statistical methods to come basically did come up with a conclusion that women and children had a significantly higher survival rate of all passengers on the Titanic, since the norms of the time were women and children first, obviously, this made sense. Was this really this outcome of this analysis insight? Was it actionable? The point of the exercise was not to mathematically identify this relationship, but to show that analysis should lead to actionable insight, healthcare, identifying patients at risk for particular conditions is a great first start. But is that risk score that you've assigned that patient actionable? Do you have clinical pathways identified to mitigate that risk for that patient or that patient population? That's where we need to get to as an industry, not focusing on the data, but focusing on what we can do with that data to improve patient outcomes, thereby, patient population and people's lives?
Jerrod Bailey 06:35
Well, that certainly sounds good. We're going to have to create a lot more collaboration in the healthcare industry to see that actually come to fruition. And it's going to take some centralized bodies that are sitting in the middle of that data and waiting are actually incentivized to wade through it and find the best results. And I think, I think it's like another industries those bodies start to emerge, right? And then the players will start to emerge, and you guys are sitting in a nice little nexus of whatever, what the industry is doing with technology, and what kind of insights can be available to whoever taps into it? Right.
Julius Bogdan 07:11
Yeah, I mean, we research what the space is globally. And we provide thought leadership on where we see things going, where trends are what's applicable? What's not what the right time horizon for some of this is? So, absolutely, and that's why I love this space. I mean, I love working pins.
Jerrod Bailey 07:30
I wonder if I can, let me see if I can get you to predict the future. Truly. Okay. So when we look at the opportunity for data analysis, within healthcare, there's probably some, some lower hanging fruit, the more obvious places that we can start, right, those are maybe the practical places to start. And then there's like, the ones that if we get it, right, it'll take time, it's going to make the biggest shift in healthcare, right, but it's going to take out for it's going to take intentionality, it's going to take some of the smaller wins along the way to sort of get the budgets in the in the attention that we need. What can you help me what falls into that continuum? What's like, the low hanging fruit? And then what's the really hard but really big stuff? Yeah, so
Julius Bogdan 08:15
you know, like most other industries, healthcare can benefit from technology, more specifically AI in a number of ways, right. And we're seeing a lot of these is being implemented now. And I'll talk a little bit about kind of what the future holds. So, starting off with a lot of the early promise around AI was increased diagnostic efficiency. It's one of the main benefits initially and even longer term that AI can bring, lack of medical history, large case loads can introduce human error in healthcare settings, ai, ai algorithms can predict and diagnose diseases potentially faster than clinicians with minimal air in comparison to the human trials, right.
So, when we talk about there was a study, I think, back in 2017, that showed that deep learning AI models can diagnose breast cancer at a higher rate than pathologist very significantly, statistically significantly higher rate. So that that really diagnostic efficiency is some of the low hanging fruit that we can go after reducing the overall cost of running a business or sustainability, right, using AI to process diagnosis more efficiently, which can be run at potentially a fraction of the initial cost in terms of labor., if AI can analyze millions of images for a sign of disease, it removes the costly manual work involved from, imaging specialists or others and, treatment modalities can be tailored to those findings in a faster, more effective way, reducing readmissions, wait times improving veteran turnover, all of those things. I think health IT news recently predicted the cost savings by utilizing AI automation can be significant. So the top five ones that I think they've listed are robot assisted surgery in the $40 billion range, virtual nursing assistants and the $20 billion range, administrative workload assistants in the $18 billion range fraud detection, which is also paramount in insurance and other industries as well. $72 billion, dosage error reduction, and litigation with Duxiana $16 billion, we're talking about some serious money in the simple, more simplistic automation days of automating and optimizing business practices. Right. So that's,
Jerrod Bailey 11:06
you know, it's really interesting, you say that, a lot, a lot of folks miss this, a lot of people that are in patient safety and quality are in it, because they actually want to save lives. And that's what we want, right? We want them saving lives and improving healthcare and, and in making fewer mistakes. And I just talked to the Joint Commission yesterday, and it was they have a goal of zero errors from healthcare delivery. And those are all great goals. But I don't see enough sort of business justification to tight attach those and it's not to dehumanize those, it's to, it's to figure out how do you get these initiatives pushed forward, right, implementing AI takes resources, and it takes hires and it potentially depending what you're doing, and equipment and all this stuff, and there has to be a motivation behind it. And a lot of times, what our patient safety and quality people don't connect it to is all of the savings that it can drive. And, we'll look at efficiency and things like that I come from we met, please come from this interesting world that sits in between healthcare, and then insurance like med mal insurance, and all that we see all of the litigation, we see all of the settlements, and we see all the payouts of all of these errors that happen in, medical error still depends on who you talk to the third or fourth leading cause of death in the US, it's a massive problem, right. And here, we have AI that can like mitigate all of these issues, but it comes with a cost, but look at the savings, right? You look at like the activities that you're doing now as a hospital, five years from now, we're going to be lawsuits and, and in settlements, and, and all sorts of costs that that draw down on on the business. And I think that's a missing piece a lot with the justification of a lot of these technologies is like, look, we want to be doing this stuff now. Because we're going to be paying for it one way or the other if we don't catch a train?
Julius Bogdan 13:02
Well, I mean, to your point, I mean, all of the money that we're dedicating to the administrative side of healthcare, which is significant, especially in this country, is money that we're seeing way from clinical care and patient right. So the more we can optimize and for your capacity and funding from dealing with debtors litigation, administrative overhead, the more we can funnel into these more effective moving the needle in healthcare on patient outcomes, kind of the low hanging fruit intermediate to longer term things are better patient experience, right. So healthcare facilities everybody has experienced that these are typically crowded and chaotic kind of making poor patient experience, discharge processes to meet lack of communication. There was a study that showed, I think 83% of patients describe poor communication as the worst part of the patient experience. In dealing with hospital systems. Leveraging AI can help rapidly scan through data, reports, and direct patients to where to go and how to see quickly, avoiding the confusion and lack of communication, right. So the other advantage of AI is it's not dependent on schedules, it's available 24/7. And I can track specific patient data more effectively and efficiently than in traditional care, allowing more time for doctors to focus on treatments. Right. Patient Engagement is, one of the big buzzwords in healthcare today. This is how you can help get to better patient engagement. And then better preventative care, great AI and machine learning can assist with disease management and prevention. Its ability to handle vast amounts of data such as medical information, behavioral patterns, social determinants of health, or mental conditions means that it's an index valuable, that tool at your disposal. Remember, it's a tool, it's not the deal
Jerrod Bailey 15:06
to fix everything,
Julius Bogdan 15:08
exactly, it's a tool to help mitigate things like the pandemic. Right. Also according to the CDC, last figure that I saw was a 10. and a half percent of the population has diabetes. Right, there's another desperate need to treat and manage conditions in AI can help providers understand the disease through data, and also get data from different sources, right. So today, there's a lot of glucose monitoring systems, that allows for diabetics efforts to track glucose levels in real time access reports and manage review with of their progress with doctors was the 14th. So there's a more real time interaction communication and insights versus waiting for diabetic patients to crash and then end up in the ER, manage the episodic treatment, their release them and repeat the cycle, right. So you have that ongoing feedback loop, and you can actually better manage your care, then the traditional ways of doing that's all infusing AI to them. So I love
Jerrod Bailey 16:20
Julius Bogdan 17:30
So I, it's a burgeoning space, right. So AI really is an invaluable tool in this space, in particular, and there are two key technologies that are proven, that can really help drive this NLP natural language processing in ML machine learning have the potential to automate, classify, analyze vast amounts of unstructured data, right, summarize, label vast amounts of unstructured datasets leading better to better analysis and risk mitigation. This technology is being used in other industries, in particular, in insurance, to help uncover hidden fraud activity, or fraudulent activity, personalized policies and identify risk in populations based on external factors. I mean, this is, again is one of the early sweet spots for artificial intelligence, that has been proven out, it's more of a matter of creating the ontologies, for your specific because again, every, when you're talking about case reviews, peer reviews, documentation, reviews, you're reviewing them for specific things, and that ontology has to exist, in order for you to understand what the relationships are, and so forth. AI can then take that map those identify those relationships give you, hey, this is, what the common themes are across all of these, right? Create discrete elements, so you can analyze it more discreetly, there's an abundance of potential in this space, using those two technologies that are.
Jerrod Bailey 19:10
That's interesting. So it's I completely agree with that. I think that there's just some really obvious and immediate places for all of those technologies right now. What a lot of people don't realize a lot of risk managers quality people in hospitals, is while they're in the middle of their AI project, or haven't started their AI project, there is a tech company, I won't mention the name. But that tech company is an AI company. And what they do is they wait for the mistakes to happen at the hospital in for claims to be filed, or lawsuits to be filed. And then they will use their AI engine to comb through all of those lawsuits or all of those of those records and figure out which ones are the multimillion-dollar verdicts to go after. How do you how do you price a mistake, right? And that's essentially the business that they're in. It's fascinating work. And then again, they're waiting. They're like, great, don't anybody implemented AI, we just want you to keep doing what you're doing. And then we will come up afterwards and figure out how to monetize all those mistakes. So there really isn't. There are active players in the space, whether you're talking to somebody and they're not actively working on AI someone in their space is and it's probably, in some cases working against that. Right. So I think we it behooves all of us to know where we're at in the process and ask ourselves, how are we going to take a step forward this year versus wait another 10, and figure out what we're going to do?
Julius Bogdan 20:35
Absolutely. And again, from an administrative and operational perspective, to review case review, documentation review is a very labor intensive process. So you're spending money for people to do that. And that's not necessarily value, add work or patient directed work. So again, it's a cost versus the benefits that you can derive from implementing AI and, and allowing those resources to impact patient care.
Jerrod Bailey 21:05
Yeah, that's right. Well, you wrote an article, actually, it was in the CIO applications, I think it's worth it. You mentioned that a quarter of healthcare respondents were reluctant to implement AI because just a lack of understanding. So I have to imagine you're doing some evangelization as part of your part of your work, like how do you get out there? How do you start to chip away and get them activated around concepts like AI?
Julius Bogdan 21:34
Yeah, so clinical adoption has long been a barrier to implementing AI more broadly in healthcare. For a number of reasons.
One, at its core is trust, right, AI is viewed, viewed as a black box technology. And, frankly, its opaqueness hinders its understanding and utilization, bringing transparency and educating users on the technology and the algorithms, its impact its limitations, will go a long ways towards mitigating this.
Secondly AI needs to be integrated into the clinical and operational workflows in a meaningful, it's great to have an algorithm that can do early cancer detection, but if it's isn't integrated into the clinicians workflow in the EMR, with actionable pathways are worthless, then it won't be utilized. Right. So there's there are structural things that need to be put in place in order for to drive AI adoption implementation.
Finally, data scientists need to work with frontline clinicians in order to understand how to develop and meaningfully integrate these AI solutions into those workflows.. AI is not like I said, the magic bullet, it isn't supplanting the physicians or clinicians. It's augmenting their capabilities. I like to think of AI sometimes as augmented intelligence versus artificial intelligence, because it is a tool to augment the clinicians capability to, to treat the patient.
The last thing that I want to say on this topic is, okay, so the advent of the EMR has changed how clinicians are educated and trained in, traditional classroom settings, in universities, that education needs to pivot in incorporate AI going forward to teach the next generation of clinicians coming out of the foundational elements of artificial intelligence, because, AI is only going to accelerate, we need a base that understands what it is how what's utilized its limitations in order to be able to help drive adoption, it'll be a generational advance. And that's one of the things that's kind of holding back the adoption of AI is you don't have enough clinicians and administrators out there that understand AI and its impact. So start in earlier in the education process and build it into the curriculum.
Jerrod Bailey 24:05
And we've seen that in every other industry, right? And even healthcare has done this right. You put the technology into the graduates clap your hands, and by the time they get out, they say Why would I use a protractor when I've got a scientific calculator here? Right. It's great. That's key. Well, this is great. Julius any anything else that you feel like I know you're beating the drum out there for technology and healthcare, anything else you want to kind of leave us with or any other thoughts we should be contemplating?
Julius Bogdan 24:35
Yeah, no. So there's AI has been talked about for the better part of I guess, two decades now. In healthcare, we are just even as long as it's been kind of talked about, we're just scratching the surface. The potential to change the healthcare landscape is tremendous. And again, it's not an end all be all solution. There's going to be an evolution In your Trend, there will be some significant tidbits in the future of where it's really going to accelerate healthcare outcomes sustainability models and so forth. The future, it's hard to predict, but it's coming in, we're seeing some of the early advances in it. Now, the idea is, you need to be prepared as an organization, for what AI can do, because it's a can be a differentiator in the market. And without it, you're going to be left behind. And so you will become kind of the dinosaur and obsolete without AI., I know that's kind of, what's the word I'm looking for. But it's, so
Jerrod Bailey 25:50
it's not to say they need some tough love, sometimes release., it's true. I mean, everything's heading that direction. I used to hear AI, like, it was a buzzword. And then I started to see the product of it and see, the pretty astounding capabilities of modern AI. It used to be aI kept making promises, and then it kept falling short. Remember those years? Sounds like a good like decade, where we're like, oh, yeah, it's going to solve it all. And then you kind of saw what came out of it. And you're like, well, it feels more like an algorithm. But now it's really feeling like there's a lot bigger AI and artificial intelligence than there used to be. So I think it is inevitable, we need to know how it's how it's going to play. We also need to know what like liability and stuff like that is there's, there's still it's kind of the Wild West in terms of like contracts with some AI providers, and, who's going to own liability for when somebody gets misdiagnosed, right. So there's lots of really interesting problems to solve. But those shouldn't be shouldn't be impediments to us starting to walk down the path of adoption, right?
Julius Bogdan 26:52
Right. With any new technology, there are disruptive changes, there are things that we hadn't thought of, there are things that we need to address, one of the big challenges that is bubbling up now in AI, certainly in healthcare is health equity. Right. And so we need to consider that as we design these solutions, as we analyze the data understand the bias, implicit and explicit, institutional what exists in our data? Because that's at the end of the day, AI, the algorithms are mathematical expressions of relief. Yeah, they're not intelligence in other senses, and therefore, if they're trained on data that has inherent bias in it, is going to produce biased results. And if that's the algorithms fault, it's the data that, in understanding the nuances and carefully.
Jerrod Bailey 27:43
That's fascinating. That's great. Well, Julius, this was great. Super interesting, really nice to meet you. Glad to know there's a there's an ally out here in the healthcare world, running around talking about some of this advanced technology and definitely look forward to following more of what how do we how do we find you? How do people get a hold of you if they want to come find you? Obviously, we'll put your LinkedIn and anything else in the in the comments for the show. But
Julius Bogdan 28:11
yeah, absolutely. You guys can do the ins.org. And I'm on corporate growing companies with say, LinkedIn, you can reach me in to any of those mentalities. Gary, thank you so much. I mean, it was a pleasure meeting you and having this conversation. And I love these talks, because it helps me inform others about kind of what's going on and also to kind of understand where others are in their journey. So thank you.
Jerrod Bailey 28:46
Fantastic. Well, that was Julius Bogdan, everybody and thank you, everyone for listening to reimagining healthcare and new dialogue, in risk and patient safety, subscribe and share if you found it valuable. And if you'd like to participate as the guest again, just email us at speakers at Medplace.com And make sure you follow Julius and connect with him on LinkedIn. We'll put all of his information in the show notes. Julius is pleasure. Enjoy your enjoy my favorite state of Colorado. All right, we'll talk soon