Transforming Healthcare: The Role of Automation and AI with Scott Lundstrom at OpenText World 2024

This is Textron tv. Hello everybody. We're back at OpenText World in Las Vegas and we're talking to my friend Scott here about what's going on with healthcare life sciences and IT and ai.

Scott, welcome to show. Thank you for having me link. There are a lot of priorities in those vertical industries by different organizations.

And what are you hearing from folks about what's top of mind in those sectors? Sure. So I think especially as we, you know, we look at the healthcare space, um, real challenges there.

Um, we have cost containment pressures. We have staffing, significant staffing shortfalls in that industry. Um, and we have a real desire to improve patient wellness and patient experience without dramatically increasing cost.

And the, and the use of facilities. So automation, um, in, in many different forms really contributes fairly significantly to that. Um, we work with customers that wanna become true digital enterprises.

They wanna eliminate paper. And paper is so prevalent in healthcare. So it begins with capturing those communications, applying AI and machine learning to digitize that printed content and turn it to data and make that data actionable.

So think about this virtuous cycle where once we have digital content, we can apply intelligence to it, and out of that intelligence we can drive the user experience. So in the US, costs are overwhelmingly driven by patients with chronic disease. Chronic disease patients require management, right?

They have to stay on their meds, they have to take their tests, they have to adhere to their treatment plans. Well, without automation, it's very difficult to do that. It becomes a really manual process.

We get big gaps in care and people suffer as a result of it. So we capture that data, we automate the response, let's say, to a bad test result or a missed prescription or appointment. We try to recover and reengage that patient.

Um, and especially with chronic disease, it's not one thing if we think about diabetics, right? Well, we have new diabetics, they need education, they need to understand how to manage their disease. We have, well-managed diabetics, they need to be encouraged, right?

More exercise, better recipes, better management, uh, of their blood sugar. But probably most importantly, we have the train wrecks, right? People that don't understand what not treating their disease is gonna do to them in the long term, the cost, it's gonna create the lifelong difficulties.

So if we can quickly address those gaps in care, if we can try to deliver care, even considering things like social determinants of health and special access, maybe we need to provide transportation, maybe daycare, maybe home health assist. We can drive out costs, we can reduce utilization and we can improve wellness. And that's the real goal of our technology in healthcare right now.

One of the things, at least that I've seen is that when we digitize a process, we tend to take the legacy process and just try to kind of move it lift and shift it into some digital format. And we don't really re-engineer the process. And so sometimes it comes across this feeling as cumbersome as ever.

'cause it's just like I'm using a tablet instead of filling out a piece of paper. But it's the same kinda disconnected workflow is, are we kind of putting a cart before the horse? Do we need to re-engineer the workflow first or do we kind get it in there and then try to make it massage it into something later?

Sure. And, and in, you know, in healthcare and life sciences, it's very difficult because we're so tradition bound, right? In, in life sciences, we have very rigorous requirements about how those processes need to op to operate.

So often, you're right, we are replacing a laborious manual paper-based process with a digital one. And that can provide some significant improvements right off the bat. But these are conservative industries, right?

We're treating individuals, we're responsible for their health. So the proven process, the current process has been proven to work. So the challenge is how do we create incremental evolutionary change as the organizations adjust to the technology?

So it is an ongoing process. Um, we do some things though that really jumpstart that, alright, by virtue of the fact that we're multi-cloud and we work with data in place, um, we have direct access to most electronic medical record systems. That's very rare for an IT vendor, right?

And we have certified integrations with all the largest health record plans. Um, the other is, um, you have to be able to pass regulatory audits, right? And that's another significant advantage of a complete digital system.

Um, in, in, in our demo booth upstairs, we show basically an aviator assisted FDA audit for a life science product. Um, saving thousands of hours of document preparation. Um, in effect you create an auditor's workbench and they could ask questions of aviator about how this process is running.

So there is, it is tradition bound, but those traditional processes can still be dramatically improved through automation, um, and analytics, um, through experience and, you know, uh, patient communications. So it will be an evolutionary process, but there are significant early opportunities that we can address. We've been talking about AI here all week.

Um, what will be the impact of AI in these sectors? Particularly, are you seeing some interesting use cases for some of this stuff already? And, you know, what can folks expect?

Sure. And, and health and life sciences are really different. Um, you don't see what we would consider consumer grade large language models.

Nobody's trying to do any of this with chat GBT or, or Claude or any of these new LLMs. Um, in healthcare, the real focus seems to be on small language models. It's not about replacing the doc or the nurse, it's about providing an assist.

It's about augmenting their ability to quickly make a better decision. And, you know, one of the challenges we have in healthcare is doctors learn through reading journal articles and journal articles are produced through clinical research and they can't read all of them. Um, lots of evidence that suggests it takes about 10 years for a significant change in practice to really make it into the market.

Um, by automating those processes, by providing that AI assist, we can dramatically accelerate the introduction and new knowledge into the care plan of, of current patients. So it is a real, again, an acceleration of the traditional approach. And, and that's probably the first big win we get in healthcare.

I don't think there's any secrets that some kind mistakes are made in healthcare. So, um, do you think we'll get more accurate prescriptions and more accurate, uh, recommendations for surgery as a result of ai? Or, um, is the machine, as we've seen, tends to hallucinate sometimes?

So, you know, how do we kinda strike a balance there? Again, I, you find hallucinations happen most in consumer grade, large language models. When we begin to look at, at some of these very narrowly defined and trained small language models, there's, there's no hallucination, right?

They produce a, a kind of a validated outcome. So that's probably one of the first things that we'll see is, you know, it's kind of a lead a thousand flowers, bloom kind of market. There won't be one Uber AI that runs the hospital.

Um, you know, the other, uh, significant thing we see is, um, we can check for errors, right? We understand the best practice, the, the physician world, uh, the world of, of drug discovery. Thousands of compounds, right?

Hundreds of, of deltas or differences in process one product to another. It's easy to make a mistake, right? And, and one of the things we'll see, AI and machine learning deliver is in effect these virtual checklists that makes sure we've never missed a step, that we've always considered the best practice or the best inputs.

Um, so it will, it will provide some guidance. Now, I think also we see there are some challenges, right? We see payers now doing, uh, denials with ai.

Uh, that's dangerous, right? And, and I do think we'll see regulations especially around use of AI and the financial and clinical sides of healthcare. But it is a, a technology that's really delivering significant advantages already.

Do you think we'll see a wave of innovation because they'll, we'll be able to maybe find cures for things that we couldn't previously find. 'cause the data was just too immense and we could actually start maybe, I don't know, identifying causes of cancer clusters in places that have always kind of been a mystery, right? Right.

And yes, I, I do think we'll make some advances there. Um, you know, in in life sciences, um, there's a, a term called fail fast, right? And when you think about it, we're looking at thousands of compounds to address a particular disease.

So how do we grind through those and find the most promising? So we now have the ability to evaluate compounds digitally to fail in seconds instead of in months to very quickly narrow down on, on the right compounds, on the right, uh, formulations. Um, the same with treatment plans, right?

Um, you know, we get into, uh, treatments that are done over and over again, hip replacements, right? Cardiac valve surgery, um, tremendous variability in in, in the outcomes of those procedures. And a lot of that is around the process and, and how the physician approaches that.

So I think, you know, one of the most promising things we're gonna be able to do is to very quickly surface best options, best practices, and deliver them directly to the care provider. Mm-Hmm. What level of priority do these organizations put on the IT system?

'cause especially in healthcare, when I'm talking to those folks, most of the budget for technology is for things that are gonna be more, you know, better surgeries, robotics, things that touched the patients directly. And not a lot goes into the IT side of that equation. So, and that's really changing, right?

And when you think about the old model fee for service, right? It was about attracting the patient to the facility for their care, right? And even on a one-time basis, and you know, we saw hospitals really investing in things like big MRI suites, surgical robots, things designed to kind of dazzle the consumer and attract, you know, brilliant young doctors.

Well, with the emergence of value-based care, the CFO has moved from the back of the house to center stage. It's all about making sure that we can survive. Now, about 30% of our patients are state patients through Medicare Advantage or Medicaid.

Um, many of them are covered under value-based care agreements, where we pretty much have a, a stable, uh, per patient per month reimbursement that we have to manage within. So all of a sudden, healthcare has a whole nother set of constraints on it, around cost, around adherence to value-based principles, and around actually improving wellness and health over a patient in the long term. So that really is changing the role of it.

We see lots of upgrades now in, in operational systems, um, big improvements in scheduling. People are really beginning to embrace remote and home health as a lower cost, uh, vehicle of care. Um, and often this is a result of, um, the CFO and the financial side of the house coming back and saying, we have to change the way we deliver care if we're gonna survive.

It almost sounds like they're actually trying to keep people outta the hospital. They are. They absolutely are.

Um, and, and one of the core premises of value-based care is we manage wellness, not not disease. I mean, well, we manage disease, but it, it's different, right? You think about the old model, you got sick, you went to a doctor, if you needed a procedure, you got admitted, and maybe next time you went to see a different doc, maybe next time you went to a different hospital, right?

And now the, the goal is to really get you into a consistent, engaged relationship with your principal physician, with your PCP to have them really own your wellness and proactively look forward in what could cause costs, what could interrupt your wellness, you know, how, what is the best long-term care strategy for you? Because now you're an annuity that they have to take care of. Mm-Hmm.

Um, and the only way that they get real upside is to keep you outta the hospital, is to keep you well. So yeah, the, the process has really changed and, um, and you'll see that become even more evident as more and more the population is covered under value-based agreements and under Medicare and Medicare Advantage. Hmm.

How do you manage and store all the data that's being collected? Because now there's, you know, watches and all kinds of things that they're hooking up to people, and it seems to me like there's a real cost of storing all that data. So what are the strategies for making that manageable?

Sure. And, and so one of the, you know, the very first strategies that we embrace is that we use data where it resides, right? So much storage is a result of doing extracting transformation and loading, right?

We pull data from the medical record system, from the, the surgical scheduling system, from the, the, you know, um, operating theater equipment and supply chain. We have to put all this together. And you think about a complex surgery, there's a lot of moving parts there and a lot of systems are involved in actually getting that scheduled and making sure that, that there's high quality outcome.

So first off, to be able to use the data where it resides in its current form, totally up to date, huge savings for the average institution. Um, you know, another area is, um, archiving healthcare organizations never throw away a bit of data ever. Mm-Hmm.

So if you're utilizing one of our content management products, um, we can easily create an automatic archiving capability that'll move that data off to a lower cost tier. Um, and then finally, again, using data where it resides really makes it much, much easier for us to do things like advanced analytics. Um, and honestly, digital content is real, a real precursor to any kind of machine learning.

So the ability to capture everything, to digitize it, to access it where it lives, um, you know, that creates great savings in storage, but you have to have appropriate technology where you're not giving up capability to save money on, on storage, right? So again, this using data where it lives, being able to create machine learning pipelines outta data from where it resides, being able to capture and recognize content internally into data, these are all really critical capabilities in healthcare. Um, and, and that's really what we're focused on.

How, again, how do we augment and assist that electronic medical record system that's already in place? And last question, security. We have seen healthcare institutions are primary targets for ransomware.

All kinds of bad stuff is happening. Um, can we secure this data? Sure we can.

And, and again, this is a kind of an investment challenge in the healthcare space. Um, but again, I think your CFOs really begun to understand the cost here, right? If you think about the cost of a breach, okay, first you have to deal with a ransomware gang, right?

You have to pay 'em off, you have to hope you get your data back. Um, and often now, uh, it's not even just about getting your data back, right? They, many of them don't even encrypt your data.

They just offer to sell it, right? So you have to pay them just to keep your patient data outta the market. Then you go through a long process, seventy five, ninety, a hundred twenty days to recover your systems and get everything up and running and any office of Civil Rights comes and tags you with a multimillion dollar HIPAA penalty.

So it's, it's a lose, lose lose or a hospital that isn't focused on security. So we work very hard. Um, we have a strong commitment to move to biometrics for all of our products.

Um, we really want to get away from passwords in two factor. Um, we embrace zero trust, okay? Which I wouldn't wanna get into the technology here, but pretty much zero trust assumes that you can have security without firewalls, right?

That, that every bit of access is authenticated, that every idea is validated. Um, and, and the combination of that, of having more secure applications, more secure cloud platforms, more control of the identity of your users, and then eliminating some of those challenges around password changes, that creates a much more robust security environment. And, and we see a lot of interest in that narc from our customers.

Alright. Hey folks, if you wanna make a difference and short of becoming a doctor or a nurse, go work in IT and healthcare because it's changing everything for the better, we hope. Scott, thank you, Mike.

Thank you. Pleasure. All right folks, we'll be back in a minute.