Episode Transcript
[00:00:00] Speaker A: You.
[00:00:02] Speaker B: The idea of being able to eliminate the administrative waste that comes with getting a claim processed and paid is a dream of both providers and payers. Getting there is going to be, I think, the challenge of the next five to ten years. I think in lieu of that, what we're seeing now is a lot of what metrics is focused on where you can use predictive analytics, you can use AI and machine learning and also natural language processing, quite frankly, to start to do things in an automated way that we've never dreamed of being able to do.
[00:00:45] Speaker A: Hello and welcome to the Alvarez Marcel Healthcare Industry Group, Human Capital and Workforce Management podcast. In this podcast series, we discuss the most pressing workforce issues facing healthcare leaders today. We're joined by Bharav Gupta and Ross Steele from Medmetrics as well as David Dydance from a M's healthcare industry group to chat about automation trends in healthcare revenue cycle management. How about we chat a little bit about each of your roles? Sound good? I'm really curious to hear about what you do and how you help your clients. And then we'll jump right into the topic of the hour. The thing that I'm constantly asked about, which is are the robots taking all of our jobs. So I'm hoping that we can jump into that. David, how about we start with you? Let's talk a little bit about your role at a m and how you've partnered with organizations and clients to help their revenue cycle process.
[00:01:40] Speaker B: Sure. My name is David Nidus. I'm a managing director within the revenue cycle practice at Alvarez and Marsal. I've been with the company for a little over five years, but been doing revenue cycle consulting for the better part of about 16 years. We do a lot of work partnering with both hospitals and healthcare providers to improve revenue cycle through process improvement, technology enhancement and overall general strategy to enhance not only cash flow, but efficiency of operations, which obviously is the key here. We also partner with revenue cycle services companies and collaborate with them on ways that they could both improve their processes, but again, also look at workforce strategies and technology enhancements to do more with less in this current environment.
[00:02:32] Speaker A: So I'm really excited to introduce a m.
Can we start Garab with you? Do you mind sharing a little bit about your background and also tell us about Medmetrics?
[00:02:43] Speaker C: Yeah, absolutely.
Garav Gupta, I'm the senior vice president of product strategy and performance management here at.
You know, I came from a background in turnaround consulting, designing process management tools, reporting tools, KPI designs. We really took that to heart here at know it's in the name metrics. At Medmetrics we're an RCM management firm with a deep technical and measurement, or KPI driven bent. Right. What we really want to be able to do is report on operational performance while driving as much efficiency into the process as we can.
The preponderance of our business is in the services world. We have a couple of thousand folks who are functioning as the revenue cycle on behalf of our clients. And then we also license technology into the marketplace. Automation technology, workforce management technology, and decision support tools that help our clients understand how they're performing, what the drivers within.
[00:03:58] Speaker A: Their businesses are really excited. And Ross, you work with him, right? Can you tell us a little bit about your role?
[00:04:05] Speaker D: Sure. So Ross Steele, VP of denial's process and strategy. I came to medmetrics a little over a year ago via an acquisition. So before that I've been with the company that's worked within the denial space for a little over eight years. I came to it a couple of months after I finished law school in a health and policy fellowship with the Ginsen Institute of Health Law and Policy and got into this provider space into a retro denials population that was focused on a specific area of the revenue cycle. And we engaged in a couple of years long effort just on standardization of process, utilization of all data available, and really fitting well within the medmetrics infrastructure to serve as an IT solution to specific revenue cycle challenges.
[00:04:55] Speaker A: I'm so excited to hear about what you're implementing at medmetrics that our listeners would be interested in maybe leveraging enabling technology to help create more efficient processes to maybe fill the gap of their current skilled worker need, as well as to support staff in giving them the tools they need to do their job. So we have a really meaty juicy conversation ahead of us. I'd like to start by setting the foundation of what we all think is going on in healthcare and what we need to do to do work differently in the revenue cycle space. So today let's talk just about revenue cycle, since we have three experts here would love to hear your take, David, on some trends that we're seeing in healthcare revenue cycle.
[00:05:48] Speaker B: As everyone probably is aware, the whole healthcare industry has a cost problem. Everything is costing more, whether it be drugs, supplies and human capital and human capital, both from a provider perspective, but also from an administrative perspective. The american healthcare system obviously has a very heavy burden on administrative tasks to ensure maximum payment, and that obviously gets us to the revenue cycle. And within that there's been a drive to see how we can reduce that cost. I think initially, I would say three to five years ago, there was a very big push to just move everything offshore, outsource anything you can to India, the Philippines, or even Central America and the just as a pure cosplay, thinking that we can handle that. I think what we're seeing, however, is those that have just done straight outsourcing have sacrificed quality, and that quality has impacted revenue in ways that I don't think a lot of people initially planned. They thought it was a commodity. You could train anyone to do it. They could work ar work denials, and it should all be fine. And now we're seeing people start to think maybe offshoring isn't the best answer for all of your problems. And now we're working with clients to really understand what do we need to keep onshore and what can we do offshore. And really at the end of the day is how do we just reduce the number of people we need to do it. And that's where companies like menmetrics come in, where we could leverage technology, leverage machine learning, leverage a lot of different tools to only work the accounts that we need to work and trying to automate those that we don't. So if we are just checking up on claim status, that's something that we could probably automate. We don't need a person doing that. But if we need to call back to figure out why the claim status isn't updated, then maybe we do need a resource. And it doesn't necessarily need to be a skilled resource. We could do that offshore. However, if we need to write a complex appeal, because there is a level of care denial that an IPA in Southern California is fighting for a very nuanced reason, then you need someone who's really smart and knows how to handle those claims and is very skilled at handling those claims follow up and actively own that, and that's maybe someone you want onshore. So right now, we're really seeing a lot of our clients look for how do we use the least amount of people, but also use the right people? So when do we use offshore? When do we use onshore? When do we use a clinician and making sure that the right person is working the right account at the right time?
[00:08:26] Speaker A: I think that it's really important for us to understand how organizations are specifically using different types of resources. So, David, thank you for kind of delineating the different buckets. Know this is one you would use a clinical resource or a highly skilled or licensed resource. How is medmetrics leveraging different types of resource groups are you doing offshore work? Do you have certain pockets of individuals that are supporting a specific function, that are highly skilled that you're investing in, and then skimming off the top some more transactional type work?
[00:09:07] Speaker C: A lot of what David highlighted earlier in the trends for the last few years has shaped our approach to offshoring. We have consistently been on the higher cost, higher quality side of the market. Traditionally did not do a lot of offshoring. And then over time, we learned that we had to. What we did find is that what was key for us was actually owning the offshore resources, right. Turnover rates can be higher offshore. There's a deeper investment in getting those folks up to speed and trained up on our systems, our processes, our reasons for why we have to do things the way we do. And owning those offshore resources has made a huge difference for us. And typically we are segmenting the work, right. What we find is that a wholly offshored model versus a complex functions, onshore versus simpler functions, being offshore model yields better performance. Right. The most expensive part of the revenue cycle is when it doesn't work. And that's what providers saw as they went to the purely offshore model, is that the net return from a lower cost product was a negative ROI. Right. The loss of revenue resulting from inefficient processes and lack of understanding and untrained individuals was driving revenue loss. Right. Leakage, as we measure it in our world. And when that leakage overwhelms the cost savings, the market has to think differently. So we've really deployed a great deal of technology and had to go out and actually acquire a firm in the Philippines to properly resource the human capital and talent that we need to accomplish the appropriate functions. And we do find that you can often source more complex talent offshore or higher skilled talent offshore at more attractive rates for some of these administrative functions, authorization initiation, something where someone needs to be able to read a medical record in order to communicate to the payer why something is being done.
Nurses offshore can be very attractive from a price point perspective relative to onshore to accomplish some of that work. So there is a great deal of segmentation of job functions that we're doing and then bringing a lot of the right technologies into play to optimize. As David said, when we're having folks touch on accounts and what types of work they're performing on those accounts as they get routed and automating as much of the truly row work that we possibly can. And we started off with a couple of core technology tools, one of which was a workforce management tool that we call workflow, which is ingesting all of the data from various different subsystems and performing a whole series of boolean logic type analyses, rules, engine work to arrive at a very optimal distribution of work into folks, and automating as much of the rote tasks as possible as the capabilities of technology, of RPA, of APIs, into payer systems and other elements have come up to have grown in capability over the past, call it five, seven years, we're bringing more and more of those types of things into our workforce management tools. Right. So we have bots that are able to screen scrape payer websites for claim status. So saving us four minutes per touch for those reps out of the gates. Right. In terms of looking up the current status of the account, all of these tools help us mitigate cost and then balance the appropriate resources from the appropriate skill levels to match the type of function that we're trying to accomplish.
[00:13:41] Speaker A: Yeah, it's a good. Dave, you get to use the word boolean in the conversation. Right. I noticed.
Quick question for you. Why the Philippines? Why did you select the Philippines as your offshore.
[00:13:58] Speaker C: Know? I think we were looking at multiple locations. I think the Philippines is an excellent location for a talented pool of human capital, as well as english speaking, strong ties to the american economic system.
And there was a strategic acquisition for us in terms of capability that also had philippine operations. But there was a multitude of reasons. And to be fair, I don't think that the Philippines is the end of our offshore human capital.
You know, I think that we will end up with multiple offshore locations over time.
[00:14:45] Speaker A: Well, let's pivot. Let's talk a little bit about that enabling technology, whether it's used by your staff offshore or in the US.
David, you always use this great analogy of the evolution of the revenue cycle business and how to leverage technology. You make a connection with the auto industry. Do you mind sharing your thoughts about kind of where revenue cycle, healthcare revenue cycle is right now versus where it was a while back and kind of that evolution over time?
[00:15:20] Speaker B: Yeah, happy to. Being a Michigan native, it seems that we always try and draw things back to the auto industry. But I do see it in a very similar way where when revenue cycle is really at its heyday, call it 15 years ago, you have a lot of resources, each having their part along the revenue cycle, all doing the work manually, you're working off paperwork lists, and everyone's doing it. And over time, we've evolved where technology is able to automate things the same way on the assembly line, you have robots now that are able to do the mundane tasks, but you also end up with a need for a different type of labor. So instead of having someone who can just push paper, you need that skilled laborer who is able to solve that complex problem, but you need less of them. So same way in the auto industry now, where the assembly line has a lot of robots doing the automated tasks, but you have highly skilled workers who can solve the problems or do the really complicated pieces of building the car in the revenue cycle. It's very much the same where there's enough technology out there to do the basic blocking and tackling, but you still need a different kind of resource, one that is skilled at solving complex problems to manage your trickier cases or cases that might involve a little bit more complex resolution.
[00:16:51] Speaker A: So let's talk about what that technology is.
In the healthcare industry, there's two places that we typically point to for where enabling technology, especially automation, AI, et cetera, seems to be focusing on. It's the clinical side, so the delivery of care, so robotics, et cetera. And then RCM is always pointed to as a place where enabling technologies are being leveraged to either fill a gap with missing labor, scrape off the top more tactical type tasks so that you can focus on more sophisticated work if you have a limited resource pool, and also for cost savings, efficiency. I'm very eager to hear what are the specific technologies that are being leveraged. And Garv, do you mind sharing a little bit about medmetrics and what you've implemented from the technology standpoint? I'm just always blown away hearing about how this technology is being used in practice.
[00:17:55] Speaker C: Yeah, absolutely.
We have to start with a base foundation of bringing together what is often siloed data across our clients data stack. Right. They're getting EDI data from the payers. They have patient accounting data. They often have other subsystems that are calculating multiple components of expected reimbursement eT. So, you know, I mentioned earlier, and it's funny that you highlighted the word boolean because it really hits on what Dave was saying earlier with the evolution. Right. The core of our technology for many years was this workflow engine where we were bringing together this information and leveraging Boolean logic to drive a great deal of automation where we could. And then we started to integrate with RPA bots and API connections into payers to drive again what we would call highly exception driven workflow, eliminating as many of those rote tasks as we possibly could and upskilling our team. Right. To address the more significant issues that go beyond pushing paper. And we've driven towards a great deal of optimization in that realm with what I would again call Boolean logic. I think that it would be great as we came towards the denial space, the clinical and the administrative start to come together. Right. And truthfully denials represent a sizable portion of the revenue cycle work. Right. You got to get a bill out the door and if it pays it adjudicates and subsequent bills need to be dropped to secondaries and patients. But things go well, right? And the systems can handle most of that standard processing without human intervention. And so a sizable portion of what needs to be worked on is that denials, population, things that are more complex. And even there, our boolean logic initially helped us route those things and drive that work in an optimized format to our end users. But in that denial space, Ross and the team in our denials management business have taken it to the next level, right. With the machine learning and AI capabilities, bringing together some of that clinical picture and where it meets the administrative denials and clinical denials that our providers are experiencing and represent the preponderance of the revenue loss. So I'll turn it over to Ross to really talk a little bit about what we're doing around the denials management space and deploying automation there as well.
[00:20:50] Speaker D: Yeah, and it's an interesting area that starts to implicate that binary on the back end, on if an insurance company is going to deny something or not. So taking those distinct areas within the institution and standardizing not only what you're going to do with the account that comes through, but how you're going to utilize all the information that's available within the system, within the application that's being used and within the different departments that are providing inputs into the information that's kind of implicating what the subsequent processing of the accounts are going to be becomes very important, much more important for those assessments on what was done in a standardized way, how that impacts subsequent processing, then the base performance of the function. So that was really the labor intensive part of what we do comes in that initial classification, the initial binning of relevant variables, different features that go through, different features that get kicked out by the analysts on the back end that you thought were relevant, different people who you think were high performers that really aren't providing some of that back end return on what they're doing. Everything gets put out in the open. So there's not really like we're trying to eliminate some of the areas for people to hide and identify anything that can be done in a standardized way. Because the more we're able to do that and provide, utilize some of that boolean logic that comes through not necessarily as the end goal, but as a product of what the outcome will be, the more we're able to start utilizing that in a prospective manner on both the propensity to deny side and then also on the propensity for an overturn on the payer side as well.
[00:22:45] Speaker A: I think this is just so cool, the amount of detail and the sophistication that you all are at now leveraging this technology.
I wish that in other areas of our healthcare space that we would be able to develop this so quickly and broadly.
[00:23:03] Speaker C: I think Ross explained quite a bit of the process of how we got to the automation in the denial space. But fundamentally, what our technology is doing is predicting the probability of denial, but also predicting the probability of overturn and overturned. By which user, which representative would be the most effective at overturning this type of denial and providing the right clinical justification and really automating quite a bit of the denial work. So more complex denials are routing to humans, but there's an entire population of very low dollar, but high volume denials that our clients experience.
Lab denials, outpatient mris, things that didn't get authorized, where the cost of appealing a denial through a clinician, which can run into 100, 200 plus dollars when you include all the follow up, makes it very difficult to make that investment in appealing those denials. And so what the technology that Ross and his team have developed really enable us to do is truly automate. Right. The system is identifying those denials, generating an appropriate appeal letter and submitting it back to the payer, and achieving overturn rates that are comparable and in some cases higher than what our human overturn rates are. And that's where we're really delivering a great deal of value.
[00:24:48] Speaker A: Yeah, no kidding. So in aggregate, all these death 5000 cuts really are material. It's just the processing of that is so significant.
There's something that I wanted to ask you about.
You just shared. You can predict likelihood of denial. How does that work?
[00:25:11] Speaker C: So, based on a lot of the claims history, and Ross can speak to this in a lot more detail. Ross, why don't you take that one?
[00:25:19] Speaker D: Sure.
So when we're able to integrate the available transaction sets that implicate the patient's coverage, and we're able to look at the historic performance within the institutions we have locally for that specific account, and take the nationwide data set of institutions that we have. What we're able to do is identify relevant features that are going to impact whether or not when that claim, when that off submission goes out the door, whether it's going to have information that's going to allow a pass through, or if it's going to prompt a denial. So when we standardize and approach the data in a uniform way and have that back end understanding of what's going to happen, we're able to take that to prospectively understand and build our own propensities to deny and help avoid and find action items for institutions to take to make those steps and avoid those delays in revenue.
[00:26:22] Speaker A: So you're constantly learning from the information that you have.
Well, let's talk about some of the sexy things that are happening in the future. David, you've mentioned to me blockchain is going to take over the revenue cycle world.
Need you to explain that to me. And also anything else that you're seeing clients thinking about, clients investing in where you think the future of enabling technology is. Would love to hear your perspective.
[00:26:52] Speaker B: I think that the dream for everyone would be if blockchain can solve a lot of these problems.
There's this idea in a pilot that was tested back in about 2017, 18 ish with a few payers and providers where if we house both the provider list so you don't have to worry about credentialing, and you house the authorization and claim on the blockchain, you should be able to instantaneously verify that the provider is the right provider, instantaneously have your clinical records uploaded to support the authorization, and then once the service is rendered, you should be able to adjudicate via the blockchain in a very expedient manner. That pilot actually went decently well. But unfortunately, in talking with a few folks who were involved in it recently to prepare for this podcast, it sounds like the funding for that dried up. The government really hasn't been interested in funding it, and the payers and providers are kind of looking at each other being like, we know that this is the future, but it's going to cost a lot of money and neither of us really are interested in footing that bill at this current time. So some of that has been dead on the vine. Now, that being said in talking in some insider circles, that's still the dream. And I think that eventually we'll get there. The idea of being able to eliminate the administrative waste that comes with getting a claim processed and paid is a dream of both providers and payers. Getting there is going to be, I think, the challenge of the next five to ten years. I think in lieu of that, what we're seeing now is a lot of what metrics is focused on, where you can use predictive analytics, you can use AI and machine learning, and also natural language processing, quite frankly, to start to do things in an automated way that we'd never dreamed to be able to do. Appeals. Writing that Dave mentioned is a great example of that. We'll continue to see those in the future, but with some future legislation being proposed by the current administration around making authorizations easier, I think some of the hope is, as authorizations and claim adjudication becomes more straightforward, you can automate more of it and allow for advanced technology to help process claims in a way that doesn't necessarily need as much human interaction or follow up.
[00:29:25] Speaker D: We're preparing for that in a pretty targeted way. So we have a lot of document classification models that are inputting information into systems and help driving subsequent workflows. Those are all integrated into applications and can help reduce some of that administrative burden and reduce the need for some of the resources to come in and classify some of that information.
It also allows us to leverage external data sources and reluctant payers that may not be fully integrated into that auth validation process, but throughout the next four or five years, they should be integrating into communications and interoperability processes that will allow more efficient processes. Our job right now is to take the information that's available, utilize some of that ML to drive that process, and get ready for kind of the things that David was discussing towards the government side, finalizing what they want to do.
[00:30:31] Speaker A: So this is what the future looks like with technology.
I'd like for us to look in our crystal ball for a moment and think about what does the future of the workforce look like?
I joked, are the robots taking over? I genuinely want to know what happens to the worker. Are they taking a backseat when it comes specifically to revenue cycle? Garv, what's your thinking? What does the future look like for the revenue cycle mean?
[00:31:07] Speaker C: I think that we are going to see a continued advancement and evolution, per David's auto industry model, where more and more rote tasks, or what used to be semi complex tasks, become automatable, right? They become part of the EDI landscape and or there are bots or ais accomplishing significant components of it.
I do believe, though, that in the long run, that upskilling process, right, the fundamental need to have clinicians or someone who can handle escalated pathways, someone who can identify the trends and bring the appropriate resolution to bear. Right. I don't think that we'll ever get away from having some human attention to the revenue cycle. There's probably just a few too many nooks and crannies of the adjudication process that either exist or will come to exist to make that possible. But I do think that the automation, the engine, and it's, particularly, as we're seeing now with some of these large language models and so on, really opening up a world of capabilities I hadn't imagined would be possible quite so soon. So I should hold my tongue and I'll let Ross speak.
[00:32:44] Speaker A: Yeah, you said that Ross differs in opinion. Ross, how is your opinion different?
[00:32:50] Speaker D: So we've been involved in a lot of system and workflow design throughout the last few years. So as we start to look at that, it's kind of this offshore concept and setting the expectations of what to expect from the workforce, what we expect the workforce to do moving forward, all kind of come back down to how strong the system is and how strong the data that's going into the systems and uniform it can be. So moving forward, that's only going to improve. What we start to see is the need for people will be to resolve ambiguities in a targeted way that helps the broader purpose even. There's areas now where we have enough standardization through automated processes, where having users provide information really provides more confusion to what's happening and the analytic back end assessments of what we do than the value we gain from having them interact with the account.
So as the more skilled workers entered into those analytic back end worlds, and the engineering becomes stronger and stronger, and the processing power for these becomes quicker and quicker, resolving those ambiguities and data gaps becomes the most significant thing that we can do, all the way from our clinical decision support system, specifically focused on level of care and clinical policy, bulletins and input of data into emrs and live decision making on what gaps mean, the model, understanding that these will implicate downstream processes, and having a workforce and an application that is able to be nimble and react to those kinds of things becomes very important. So it's not a zero staff concept, it's just a more focused workforce on resolving the problems, identifying the issues, and ensuring that we are operating in a standard and efficient process to get where we need to go.
[00:34:58] Speaker A: So no matter what the revenue cycle worker is of the future is a more sophisticated, flexible, dynamic resource. It sounds like no matter what the future holds. Well, David, any parting thoughts on what you think the future of the revenue cycle workforce looks like.
[00:35:23] Speaker B: No. Yeah. I think the only thing I would add is it's probably never been as exciting of a time to be in this space with all the technology that is being leveraged and the different types of workforces that are being employed. It's an exciting time. It's fun to see how different companies are evolving, what's working, what's not working. And I think that we've seen more creative thought and more strategic approaches within the revenue cycle than I have ever seen in my 16 year career.
[00:35:57] Speaker A: Thank you. It is an exciting time in revenue cycle. You all make it sound very exciting as well. Well, thank you to all three of you for being wonderful guests. I really appreciate your perspective. So glad that you could join us today. Appreciate your expertise and.
