Hi, everyone. This is Erica Spicer Mason with the Becker's Healthcare podcast series. Thank you so much for tuning in today. I'm thrilled to be joined by Steve Sutherland, the senior vice president of information systems at Saris, who will talk to us about AI and machine learning in the payment integrity space. Steve, welcome to the podcast. Thank you so much for joining us today. Hello, Erica. Thank you very much for having me. I appreciate the opportunity.
We're thrilled to have you here to talk about this topic today. And before we really get into it, I wanted to see if you'd like to share just a little bit more about yourself, your role, your organization, whatever feels top of mind for you. Sure. Sure. Well, I'm a native Texan. I was born and raised here in in the Dallas Fort Worth area, and I have been with, the Saris Organization now for going on 30 years, actually. I think I'm I think I'll celebrate my 29th anniversary in the fall or early
early next year. And I've had I've served in a various different roles within the IT spectrum here with the company, but I really just have seen the company, grow exponentially and and really grown along with it from a very small organization into, you know, what it is today, which, you know, we've got almost a 1000 folks within the within the Cirrus part of the company, and and we're, you know, a part of a bigger company, CorVel Corporation. So
Wonderful. Thanks so much, Steve. Appreciate you sharing a little bit more about your background. It's incredible. You've been with Cerus for about 30 years. So I'm sure you've seen so much change in in those decades. So I'm really looking forward to the perspective that you'll have here with technology and payment integrity. So I think we can kind of get into it from there.
So I'm wondering if you can share with our audience what you see as the role of artificial intelligence and machine learning in addressing payment integrity pain points. And in your experience working with payers, are there specific benefits that partners have realized in applying those tools to claims processes? Would love to know any success stories or case studies that might come to mind.
Sure. No. That's a it's a great question and really a relative and and hot topic today in today's industry. And and those technologies are really helping improve accuracy and consistency and efficiencies and in payment processing in general. I mean, they're applicable all across all sorts of other industries, obviously, as well. But in payment processing, you know, payers have a very specific time frame within which they have to make payments to providers based on contractual and regulatory
requirements. So they not only have a short window to do claim adjudication, but also, validate and apply various types of reviews and business logic and and all sorts of things that happen within the process while, you know, adhering to those very strict time time frame. So efficiency really is key. They've there's a lot to be done in a short amount of time frame, and it needs
to be done accurately. So artificial intelligence and machine learning specifically can really help improve these processes by, you know, automation. So automating certain very manual and sometimes inefficient steps. Claim identification. So helping identify and tag or flag problematic claims or trends that they might see that it can be taken off of the conveyor belt, if you will, or out of the process and looked at by a by an expert or some some other type of
of audit reviewer. So, those are just a few of those of, you know, examples of how it can be used. Very general, but we we have several specific use cases within our workflow where we've taken very manual processes like like data capture, you know, keying data from an image or something and improving that dramatically by applying this type of technology.
Yeah. I appreciate the example, Steve, and I think everything you outlined really speaks to what an impact efficiency in this process can play when it comes to compliance, timely payments, adhering to contractual agreements. There's a lot kind of on the line in that in that window of time as you mentioned.
And I know you also mentioned how payment integrity plays kind of this key role in health care where it ensures that claims adjudication is accurate and that health care organizations are appropriately reimbursed for the care that they provide. So as health care billing processes become more complex, what key challenges are there when it comes to the widespread adoption of AI in the PI space? And aside from complexity, are there other factors that are kind of driving those issues?
Sure. Yeah. There's always challenges, you know, with everything, especially anything new and and new technologies. So, you know, applying this complex technology to complex billing processes in general and then also health care data in itself can be a challenge. It's not always consistent. It's not always standardized based on your billing practices and those sorts of things. So the data working with the data in itself can be very complicated and challenging.
So there are certain pitfall or many pitfalls you want to try to avoid. A very important one is to exercise caution and and be diligent when it comes to data privacy and security, as well as transparency and defensibility. So those all kind of, you know, go along together, regulatory challenges in some cases. So really relying on your compliance and security teams to provide a sound governance and policy, when you go into these
deploying these types of solutions. Really, you need to need to look at that upfront and really have that established. There can also be a steep learning curve with staff and finding, expertise and resources to be able to to build these types of solutions and have the knowledge and and expertise. So really just have patience and be persistent as you try to recruit talent and expertise and and also train current staff on learning these technologies.
There can also be some gaps with, the output of some of these solutions and tools. So as you deploy these things, you know, what the analysis showed or what the research showed might not necessarily come to fruition when you put it in in an operational setting and the actual results might be off.
So really just having, a continuous improvement and validation process so that you're always making sure that, you know, the output is what it should be and that it's valid and continually improving whether it's a a machine learning model or, know, an AI, solution in itself. Mhmm. And probably the last thing I'd mention is just is just to avoid becoming too reliant on technology, on this specific technology or any really in general
over, you know, human knowledge and expertise. So finding that good balance between, you know, how the technology can be used and what parts of processes we can make more efficient and take and take the human element out of it. But then, you know, finding that balance of there's always going to be some of that that that really does need a human touch.
Yeah. That balanced approach sounds essential, and I appreciate what you said at the beginning of your response, Steve, that applying complex technology to already complex billing processes bound to have some challenges and barriers. So I appreciate you outlining some of those kind of pitfalls that organizations can avoid and acknowledging there will be a steep learning curve. The continuous improvement and validation is key. So, yeah, thank you again for for outlining
all of that. And I think that leads me into the next question that I had for you, which is, you know, just kind of acknowledging this idea of change management and demonstrating value anytime leaders are kind of encouraging their teams embrace new technology. We hear from leaders all the time that change management is one of the the most challenging aspects of introducing new tech.
So wondering if you have best practices or resources that you'd recommend to leaders who are planning to or have even newly adopted AI and machine learning for payment integrity. And maybe you can say a few words on how they can prepare for future advancements, and other changes ahead. Definitely. I can share some thoughts and and and some of my experiences in in that space. So I I always say and and have found based on our experience, identifying the right use cases upfront.
These technologies can be used to do a plethora of things. There's a there's a 1,000,000 use cases. Right? But really identifying what is the right use case to start with. So if you're just getting into this, how do I how do I get my foot in the door? What can I do? What you don't want to do is bite off the most complex project that that's out there that's on your radar. Keep it simple. Find a very simple, easy use case to start with. Don't try to tackle that most complex solution first, but
keep it simple. And then that way, you can get, you know, get your foot in the door, get your feet wet. You can you can get some experience and have some some quick wins without, you know, having to put in all of the time and effort before you're able to see sort of the fruits of your labor. Right? So the and then building proof of concepts, it's kind of the same it's kind of
the same thought process. Right? So let's build a proof of concept on this use case, prove it out, get your feet wet, you know, make sure that it that it proves out what your goal was to begin with and that and then you can lower your risks and and increase your success by having those proof of concepts because you're gonna have some failures. Right? So if you build a proof of concept and it doesn't quite work the way you thought it would, then, you know, you sort of back up,
make some adjustments, and and start again. So we we always like to do proof of concepts on especially complex projects like this or or with the potential to be really complex. One of the other things is to get
early buy in from key stakeholders. I mean, this may sound like a a no brainer, but you really gotta make sure that you've got your business strategy and your and your technology initiatives aligned so that everybody is on the same page as far as, you know, what the goals are and what the expectations are. Like a lot of current technologies out there, data really is the key. So having good sound data government management around that data and then the architecture.
The old saying, garbage in and garbage out definitely applies in this scenario. You've gotta have good data, and you've gotta have a good strategy and management around that data. And I guess the last thing I would say is just to educate, your staff and and your business units about these technologies, about data science, explaining these use cases and proof of concepts and, you know, getting folks to buy in and and get excited about these things.
Training always helps with the learning curve. But, I mean, having some positive outcomes and realizing value really will increase your excitement and building trust in these these new tools and technologies and the processes.
Thanks so much, Steve. I think this is really great advice, and I especially appreciate that point that you said about keeping it simple, especially in the beginning of deploying this technology, perhaps an organization is doing it for the first time and, keeping it simple can help help them get those quick wins, which I imagine would go a long way in demonstrating value upfront and kind of justifying the investment. So,
again, great advice for our listeners. And before we end our time together today, is there anything else you wanted to share that maybe we didn't cover already or any final takeaways that you'd like our audience to know? No. Maybe maybe just reiterate, you know, kinda what you just said, you know, making sure that your strategy and technology initiatives are aligned, getting that early buy in, and building bonds with your key stakeholders. Keep it simple. Keep it flexible.
You know, the right use cases, proof of concepts, those quick wins, build some momentum gradually. The quicker you can demonstrate value, the quicker you're gonna get not only management, buy in, and confidence, but also your end users, the folks who are actually going to be using these solutions, and show that value within the organization. And just just to kinda wrap it up, I mean, you know, Saris I've been, you know, in the business for, like I said,
for 30 years with Saris. I mean, we we are a technology leader in the PI space, and we've been doing this for a long time. So I would just, you know, extend the offer to reach out if if there are any questions or anything that, you know, I could assist with, and, my door is always open. Fantastic. Thank you so much, Steve. Really appreciate the time and all of the insights today. It's been a great discussion. Thank you very much, Eric. I've I've enjoyed it, and, it's a it's been a pleasure.
Well, thank you again, Steve. And we'd also like to thank Saris for sponsoring today's episode. You can tune into our podcast from Becker's Healthcare by visiting our podcast page at beckershospitalreview.com.