Watch the recording

Hello and welcome to today's webinar for Trends Reshaping the Future of Payment Integrity.

Before we get started, I'd like to review a few housekeeping details.

Today's webinar is being recorded and an online archive of today's event will be available within one to two business days.

If you have trouble seeing the slides at any time during the presentation, please press F5 to refresh your screen on a PC or Command R if you're using a Mac.

You may ask a question at any time during the presentation by typing your question into the Q and A box that you see located on the right side of your screen and pressing enter there.

And finally, I'd like to remind you of a Hip's antitrust statement, which is located in the link just below the slide viewer.

We will, as always, comply with that statement.

Among other things, the antitrust statement prohibits us from discussing competitively sensitive information.

We're very fortunate to have with us today John Hoffman, Vice President of Payment Integrity Product and Stephanie Sojourn, excuse me, Shogren, Associate Vice President of Payment Integrity, go to market both with Optum.

So at this time, I'd like to turn the floor over to our presenters.

All right.

Good afternoon, everybody.

All right, so we're going to be going over today the four trends driving the next era of intelligent payment integrity.

And before we get started, just want to kind of tell a little bit about the presentation we're going to do today.

The important word in our title is not simply intelligent, it's era, because the changes that we have across the health systems are larger than just edits or analytics or even our review process.

Payment integrity has historically been associated with finding incorrect payments, preventing the overpayments and recovering dollars.

But those responsibilities, while they remain critically important, our mandate in payment integrity is widening.

So health plans are increasingly asked to contribute more in terms of affordability, provider experience, operational efficiency, fraud prevention and regulatory readiness.

So our discussion today is about how payment integrity evolves from a series of controls into a connected intelligence function.

So let's quickly walk through how we're going to approach that.

Thanks, Stephanie.

As mentioned, my name is John Hoffman.

I'm VP of Product and Payment Integrity here at Optum.

I'm excited to be here today, and thank you all for coming.

And thank you to AHIP for the opportunity to talk about what we see as the four print trends that will shape payment integrity over the coming years.

So we'll begin with the kind of macro forces behind the change.

It's important because we feel these trends are not occurring in isolation.

They're really responses to broader pressures across the healthcare ecosystem.

We'll then examine each trend individually.

And for each one, we'll try to discuss 3 questions, What's changing, why does it matter to health plans and, and what should payment integrity leaders be doing differently than they are today?

And then we'll finish by bringing the trends together into a broader operating model and give you a few questions to take back to your organization.

So I'm actually excited to see the mix of folks on this call.

We've got it looks like representatives from very different parts of health plans and that's a good thing, right?

Because some of you may leave payment integrity, some may sit in investigations, operations, clinical network finance.

The mix is valuable and is central to today's messages because the next generation of Pi cannot be built by a payment integrity organization alone.

So why don't we get into the macro forces driving change?

You know, and a lot of these will kind of seem obvious on the surface, but I want to dig in a little bit and give the context behind why we think these macro forces are really reshaping the role of Pi.

The first, of course, is cost and affordability pressure.

Health plans, of course, have always been responsible for managing costs, but today the affordability question is broader.

Leaders are not only asking whether claims were paid accurately, they're asking why.

Why a category of spend is increasing, whether an intervention could occur earlier, whether information from paid claims can shape future clinical network provider strategies.

The second force is really regulatory and policy complexity, so payment decisions are increasingly connected to transparency requirements, prior authorization processes, appeals, dispute resolution, data exchange, provider communication.

An accurate decision that cannot be clearly explained, supported or reproduced can still create significant operational and compliance risk.

And 3rd is really the changing nature of fraud, waste and abuse.

So potentially abusive activity can move rapidly across billing entities, provider entities, geographies, service categories.

At the same time, automated documentation tools can make records more complete in appearance even when the underlying billing pattern deserves scrutiny.

So our feel is these forces kind of widen the Pi mandate.

As Stephanie mentioned in the intro, cost pressure demands earlier intervention, regulatory complexity demands transparency and evolving risks demands connected data and faster learning.

A few numbers to add to the kind of context of the problem here.

These aren't new to any of you, but we all know this trend is unsustainable.

You know, National Health expenditures as we sit here today are approaching $6 trillion in 2026, projected to grow towards $9 trillion by 2034.

That brings it to over 20% of GDP.

So for Pi leaders, a take away is not simply they'll be more claims to review.

The bigger implication is that health plans cannot solve affordability by scaling today's manual processes at the same rate of medical spend.

So if claim volume, clinical complexity, policy complexity all increase, adding reviewers, creating more work, use eventually hits a limit.

So that changes the strategic question.

Instead of asking how do I review more claims, health plans need to ask how to make the overall payment process more accurate, more targeted, more preventative.

So a mature program will still recover inappropriate payments, but it will increasingly use the information generated from those recoveries to prevent the next incorrect payment and identify broader cost drivers.

Another interesting stat we're seeing here is the there are newer challenges facing the ecosystem.

So a great example of this is the independent dispute resolution process, which is the federal process that was created under the No Surprises Act.

The intention and spirit of the law is a great one, right, protecting patients from surprise medical bills.

The challenge with the IDR process is there's growing evidence that it's driving up healthcare costs.

That evidence includes providers winning some almost 90% of cases that go to arbitration and payments that go far beyond typical contracted rates, along with soaring dispute volumes and millions of disputes going through this process.

So you're driving up medical costs as well as administrative costs to deal with all the disputes.

Then of course, the new and emerging fraud schemes that are continuing to pop up throughout the industry with a renewed focus from CMS on fighting fraud.

So we see an example here of a, a $14.6 billion crackdown against 324 defendants across 50s federal districts.

We're seeing this more and more.

We expect this to continue in the age of AI.

So these are things that health plans are challenged with.

So as we think about heading into kind of the four trends, we have an audience question.

You know, I think there's no wrong answer.

You're likely experiencing all three of these, but we're asking the audience to try to choose just one so you can understand kind of what is the most urgent conversation within your health plan today.

So I think we can use the poll right, Chris, to get feedback on this.

And we're just curious what everyone is kind of struggling with in their organizations.

So while we do that, I'll hand it back over to Stephanie and we will dive into the four trends as we see them evolving.

All right.

Do we have any holes?

We want to just finish that up, maybe we could see what everybody's thinking.

It looks like we got about 62%.

Feel like cost and affordability pressure is the most macro force that's impacting the health plan.

23% regulatory, 15% fraud and abuse.

All right, that's and what we're going to go to here is getting into our four trends a little bit more in depth.

So the four trends that we have, the first, sorry, hold on.

The first trend is turning fraud detection into affordability strategy.

This means identifying signals earlier and improving the connection between payment integrity and investigative functions.

The second is transformation of the payment integrity operating model.

AI and connected data can reduce the administrative burden around expert decisions, but only if work flows and accountability are redesigned with them.

The third is shifting left with real time or more timely intelligence.

That means resolving preventable issues as close to the claim creation as possible or submission, instead of relying primarily downstream on denials and recoveries.

The 4th is addressing spend, where traditional payment integrity ends.

Human integrity frequently identifies signals that point to utilization, network, clinical policy, provider education, or ancillary benefit issues.

The question is whether those signals reach the function capable of acting on them or not.

These are all connected trends.

Earlier fraud intelligence depends on a transformed operating model and shifting left depends on connected data.

So when we extend our payment integrity's reach that depends on us all becoming a cross functional organization with governance.

So we're going to go a little bit into the next one here, which is our first trend.

Traditional fraud detection often starts with a pattern that we find after the claims have been paid.

That is retrospective work, and it does remain essential because many patterns are only visible after a period of time.

What is changing is the ability to combine those earlier signals and then identify activity that warrants closer review before the loss of the accumulate.

The phrase warrants closer review is also really important, and that is an analytic.

An analytic alert is not necessarily a fraud determination.

It's going to be a signal that helps investigators decide where to look, what evidence to gather, and how to urgently act.

So I want to consider this example that we have.

It's imagine a provider entity with a recent ownership change.

Shortly after the ownership changed, we noticed billing volume increases dramatically, the service mix changed and the provider begins receiving referrals from a much narrower cluster of previously unrelated providers.

While none of these indicators independently proved the inappropriate activity, but together they painted a much different picture which justified the review.

And this is where payment integrity can contribute to the affordability process.

If we sit close to the claims, the policy logic, the coding behavior, our documentation and financial outcomes, we then have that information to be able to connect everything back through enrollment or ownership and the referrals, and then create investigative data that our organization can act on sooner and then use those usually small resources more precisely.

So the objective is not to label more activity as fraud, it's to reduce the time that we have between the appearance of a meaningful signal and our review.

So what are the early signals that make prevention possible?

The first category that we have is identity and ownership inconsistency.

Provider enrollment information, the tax IDs, the practice locations, our ownership records and billing relationships may not align across all of our data sources.

And because that inconsistency, it's in and of itself is not evidence of fraud, but it may indicate that there should be a resolution or some additional verification needed.

The second is some sort of change or sudden change in the bill, the billing volume, or the service mix.

The key is to compare with an appropriate peer group.

Billing increase might be completely legitimate if a practice expanded maybe took on a different specialty.

But the analytic value comes from identifying changes that cannot be readily explained by the by the known business or clinical factors.

Our 3rd signal is the impossible utilization patterns.

Some people call these impossible days, but it could also not just involve time, but it's also geography, member overlap, the frequency, or even the combination of services that deserve review.

Our 4th is an atypical referral and documentation patterns.

Some of the examples in this might include unusually concentrated referral relationships, repetitive documentation records with limited patient specific variation, or even documentation that appears inconsistent with the build service.

Our mature analytic process that we need to create will evaluate all of these signals together.

It also incorporates context and gives knowledge to investigators who have access to the underlying evidence.

This is the difference between producing a large alert inventory and producing an investigation ready lead when you think about your current environment and which of these signals can you use today.

So our first recommendation based off of this is to bring Pi and the SIU closer together.

And it doesn't mean we're going to change reporting structures and it doesn't mean we're going to, you know, move our desks right next to each other.

But it does mean developing shared signals, common definitions, referral standards, feedback loops, as well as a clear criteria for when we escalate to SIU from Pi.

The I should understand what makes an analytic result useful to the SIU.

An investigator should feed those case outcomes back into the analytics so that the future detection becomes more precise.

The second recommendation that we have is to extend the connection to adjacent functions.

Provider enrollment may see the ownership changes, our network teams may see the contraction or the contracts relationships and the clinical teams may recognize unusual service combinations and then our appeals team may see the recurring documentation patterns.

Each of those functions holds a part of the picture.

So staying together and staying in constant contact is our third recommendation is for Pi to use that vantage point to convene the conversation together.

Because we have a distinct view, we wanted to.

We want to use it to connect the policy, the coding, the documentation, the provider behavior, and the financial outcomes.

The eye doesn't need to own all the interventions.

We're not trying to create more work, but it can help make sure that the right signal reaches the right owner.

So if you're a leader out there, the kind of a question that we have for you to think for yourself is do you currently have a process that converts those signals into actions and do those actions into learning?

I think it's skipped.

OK.

So our second trend is not simply the introduction of our favorite word AI into our existing process.

It's going to be the redesign of our entire operating model around where people will create the most value in a lot of our work flows.

Our experts spend a significant amount of time locating records, reading the documentation, extracting the evidence, moving cases between queues, manually assembling all that information for decisions.

And while those are necessary, they're not always the highest and best use of our clinical coding or investigating, you know, staff or expertise.

Near term AI use cases are practical.

They include summarizing our records, locating that relevant evidence, organizing case information and routing work.

We can use it to prioritize our cases as well as supporting quality review.

Connected clinical data adds another dimension.

When structured and unstructured clinical information can be associated with a claim, earlier reviewers may be able to reach more accurate decisions with fewer retrospective record requests.

Reporting has to evolve as well.

A monthly savings report tells tells us what happened.

But active intelligence should help our leaders understand what is happening, why it's happening, and what is likely to require attention, and whether an intervention delivered the intended outcome.

And finally, defensive AI and governance.

Those matter.

Pi organizations have to consider not only how they use AI, but how AI may affect the documentation and the claims they receive.

This slide establishes our practical division of Labor.

AI can carry more of the information burden, so it can summarize those extensive hospital records for our DRG reviews, it can identify potentially relevant passages in them, can organize our evidence, detect patterns, and put cases in a more useful order.

But our experts have to retain decision authority for those determinations.

For example, a model could identify documentation that appears relevant to a build service, but a clinical or coding specialist still has to review and decide whether the evidence actually supports the service under your applicable policy or the clinical circumstances.

This distinction matters for quality, trust as well as governance.

AI output can be incomplete, incorrect, or even overly confident.

That means users must be able to see the supporting evidence and understand when additional review is required.

The final responsibility, again, is governance.

Named owners should have defined acceptable uses, validation expectations, escalation paths, and human review requirements.

The most sustainable model is not AI versus people.

It's AI reducing low value cognitive and administrative effort.

So qualified people can focus on decisions requiring judgment, accountability, and context.

But want you to think about where are your most experienced people spending their time today and does that work require their level of expertise for a recommendation is first inventory where AI is already touching the workflow that you have.

So that includes any internal AI capabilities, but it's also going to include provider submitted documentation, the coding tools that we use, the automated appeal content or letters or other generally external material that you may receive.

Then redesign the workflow rather than automating just one task at a time.

If AI summarizes a record, but the reviewer still has to open 5 systems manually, re enter the same information, send the case through 5 queues and the overall workflow, it hasn't changed.

So AI has done nothing in this.

So you want to start with the desired decision, work backwards, determine what information is needed, what tasks can be automated or assisted, and who has the authority and where the quality checks belong.

Third, we want to move from static reporting to operational intelligence.

Leaders should be able to examine savings, realization or accuracy and appeal outcomes, provider patterns, turn around times, and emerging risks altogether.

A practical starting point is 1 workflow with a high manual effort.

Get the available evidence, the clear ownership, and then measure the outcomes.

Redesign it end to end rather than deploying AI broadly without defining the operating objective.

And once the operating model can process information more intelligently, the next question is where can we move it?

Can we move it earlier in the life cycle and where can that intelligence been be reused?

Thanks Stephanie.

That leads us nicely into our third trend which is shifting left now with real time intelligence.

So we all know the term shift left and payment integrity, It can be defined a little bit differently by different people.

To us it just means resolving an issue earlier than it would be resolved today.

And there are different flavors of this.

As of just recently here at Optum, we've been working really hard on how do we shift left to before submission.

So the strongest opportunity here is to provide plan specific guidance on coding, coverage, documentation, billing requirements in the provider workflow at the point where they can still correct the claim before submission.

So what I've done, we've got large payer footprint, a large provider footprint.

We've been working to drive connectivity between the two and we've started with claims editing and COB intelligence, delivering that intelligence to the provider through the EMR in their workflow to say, if you continue coding this claim this way with these two CPT codes, your claim is likely to be denied.

And we're doing that by delivering plan specific intelligence in that provider workflow to correct the claim all the way up front.

So that's, that's the ultimate sort of version of shifting left.

But I know every health plan is at a different place in their journey and not necessarily every health plan, every provider is ready to go that far.

So of course, you know, before adjudication, Pi can apply, apply the most current available policy, provider history, clinical information, and risk signals to improve claim routing and decision precision.

And then obviously before denial and appeal, the organization may have another opportunity to correct missing details, distinguish A preventable error from a a disagreement or route the claim for an appropriate expert review.

And all of all of this, of course is the contract to the the kind of old school at this point pay and chase model.

You know, once money is moved, recovery becomes more expensive, less complete, whereas if an avoidable issue becomes a denial, but the plan and provider may occur incur administrative costs, provider abrasion, etcetera, etcetera.

So you know, not every organization is at the same spot in their journey in the the shift left, but technology today is, is not preventing us from shifting further left ultimately starting to collapse some of the payment integrity and revenue cycle functions is where we see a lot of this going.

So again, we're starting with the kind of low hanging fruit of the more simple denials, if you will.

We're trying to empower our provider partners by giving them the right information at the right time before submission.

If you think about kind of the challenges we've been, we've been focused on, you know the payer challenges, but providers are big time feeling the pressure too.

You know, all 68% of providers saying submitting cleaner claims is more challenging than it was a year earlier, 41% reporting denial rates of 10% or higher, 54% saying claim errors are increasing.

So you've got a little bit of this escalating arms race on both sides, both providers using AI to perhaps code more aggressively, code more accurately, payers using AI to potentially down code more, deny more.

That's why we're excited about the shift left concept of you know traditional payment integrity is not going away anytime soon, but is there low hanging fruit we can get into the rider's hands at the right time to prevent a lot of this back and forth and unnecessary denials going forward.

And that to us is the concept of shift left.

So ultimately our recommendations are to start with the evidence is strongest.

So identify high volume area areas with recurring coding, documentation or billing issues.

Look for categories where the same type of error repeatedly leads to an edit denial, appeal or recovery.

These patterns create a focused opportunity for earlier guidance.

Second, make provider education specific and timely.

So, you know, we see things like broad newsletters explaining multiple policies.

It tends to get sort of lost in the wash of, you know, not going to the right person or not getting the right focus.

Really targeted communication tied to a recurring issue, specialty, service category or submission pattern is more likely to change behavior.

Third, this one's really important is you have to start to think about success and measure success differently.

So traditional Pi is pretty easy to measure.

It's focused heavily on identified or recovered savings.

So those can those metrics continue to matter, but shifting left requires really measuring differently more related to prevention and operational impact.

So examples of that include clean claim performance, avoided denials, reduced appeals, reduced rework, intervention accuracy, sustained provider behavior change and improved provider relations.

So a a successful intervention might cause a traditional recovery to decline, but because the error no longer occurs.

So leaders need a measurement framework that recognizes that as a success rather than a lost opportunity.

So I've had a lot of these conversations with payers over the past year and, and you do hear some pushback of well, I'm measured based on recoveries are measured based on prepaid savings and folks are sort of scared to let that go.

But if you can work with your finance teams to really start to measure behavior change on the provider side and how that might reduce some of your unnecessary denials.

There is an operational admin savings component here that is really important and it's, it's important to measure that accordingly as you continue on your shift left journey.

So we'll round us out here with our trend 4 and then we'll kind of go into some conclusion.

So trend 4, payment integrity frequently identifies problems that cannot be solved through claim edits or audits alone.

So we are seeing in certain specialty spend categories payment integrity is just not enough.

Some examples are a a recurring radiology pattern may point to a utilization issue or a laboratory billing pattern may indicate a network contracting ordering behavior issue.

Durable medical equipment spend may involve medical necessity or supply relationships or utilization or benefit design.

So this is what we mean by addressing spend or payment integrity ends.

We really see payment integrity as being part of a larger health plan affordability strategy.

And the leading health plans are starting to build really affordability teams where you see, you know, network and policy and payment integrity and utilization management working in conjunction and sort of a holistic affordability strategy where you're identifying spikes in spend and you're determining sort of the right lever to use across those four or five areas to make the greatest impact.

So Pi in this world may identify the signal, but the correct intervention may belong to utilization management, network strategy, clinical policy, contracting, provider engagement, even enrollment.

So these ancillary categories are a useful example because they often combined, combined this sort of rapid innovation, clinical complexity, fragmented delivery models and meaningful spend.

And so the answer is not to position Pi as the owner of every category.

The answer is to apply P is discipline of evidence, accuracy, pattern recognition, outcome measurement to a broader affordability strategy.

The second part of the screen here really is the agentic capabilities representing kind of an emerging orchestration opportunities.

So a bounded agent might monitor approved data sources, identify development pattern, gather relevant evidence, recommend an owner track the response.

If you think about all the data that health plans have, the best health plans now are really using AI to see spikes in their data very quickly, hone in on what's going on and determine next steps and really figure out where in their ecosystem of problems coming from and how it can be addressed through some of the areas I mentioned just a moment ago.

As we wrap up trend 4 here, our recommendations are first, incorporate ancillary categories into the broader cost containment strategy.

So this means connecting Pi findings with utilization, network contracting, clinical policy, and provider engagement rather than creating another isolated review function.

Second, be Enigentech adoption with bounded and audible use cases.

So good starting points may include opportunity identification, evidence gathering, case preparation, routing recommendations, follow up reminders, and intervention tracking.

These activities can really reduce the coordination burden while keeping accountable humans in control of the consequential decisions.

And then third, you know, we're all dealing with this.

You have to establish a cross functional AI risk framework.

The framework should answer practical questions.

What evidence is required?

What data made the agent access?

What actions may it recommend or take?

When is human approval required?

How will the results be monitored, what triggers escalation, who's accountable, etcetera.

These decisions cannot be left solely to technology teams.

So payment integrity, clinical investigations, compliance, legal data provider all play a role in this new world that we're transitioning into.

So the guiding principle is very simple.

The more consequential the action, the stronger the evidence, oversight and human authority should be.

That wraps up the overview of our trends.

So Stephanie, I'll pass it back over to you to kind of sum it all up.

Yeah.

So basically the short version of this whole presentation is that Pi, we got more to do.

Our scope is widening and Pi will continue to, you know, obviously connect and detect our errors.

We're going to try to prevent as many incorrect payments and recover those overpayments.

But our health plan leaders, we increasingly are getting that pressure from from them to get more timely affordability and increase our intelligence.

And that intelligence comes from us connecting several different views that are often separated.

We want to make sure we connect our claims, our documentation, our payment policies, our reimbursement policies, the provider behavior, our investigative information, our appeal outcomes as well as the final financial result.

And we need to move that intelligence upstream.

Our goal is going to not be to eliminate that retrospective work.

We always can find things after the fact or, you know, claims come in, you know, in orders that don't make sense.

It just is what it is.

So our goal is going to be use retrospective findings to improve our next prospective decision and analytic.

Our third message here is going to be let connection is going to be the differentiator.

Our powerful analytics that we need to develop will only have limited value if it just sits on a dashboard and nobody uses it.

We need to create that reach so that the functions have that have the authority to act on it, whether it be your SIU, whether it is your payment integrity or your clinical folks have that capability to act on those.

And finally, trust.

Trust is going to be the biggest part of our performance.

Payment decisions have to be accurate.

They have to be supportable and transparent and governed.

EI is going to be your affordability engine when it helps the enterprise make better decisions before, during and after payment and then turns around and learns from those outcomes.

We have one more poll before we kind of go to our next slide.

And if we could put that up, we're curious on which trend would you like to learn more about turning fraud detection into an affordability strategy, transformation of the payment integrity operating model, shifting left with real time intelligence or addressing spend where payment integrity ends?

I mean, regardless of which trend receives the most amount of votes, again, these friends reinforce each other.

Our organizations are not going to progress through them sequentially and we're going to make targeted advances across each of these trends probably not at the same time.

OK, let's see.

So, so far our answers are the highest is 38%.

So transformation of the payment integrity operating model, 38 percent, 23 percent is, I'm sorry, 31 percent is our next highest, which is shifting left with real time intelligence, 23% turning fraud detection into an affordability strategy and 8% addressing spend where payment integrity ends.

All right, so we're going to go to our next slide before we go into questions.

So the future model, it's going to be summarized into three different words, intelligent, connected and trusted.

We want to be intelligent enough to adapt, and that means our decisions are going to be informed by our current data, our current clinical evidence, our policies, the behaviors of our providers, as well as the outcomes.

Our system has to learn from what happened rather than keep repeating and identifying the same error downstream.

We have to be connected enough to act, which means signals have to move between the different functions of our departments.

The eye can't be and does not need to own every action, but it should be there to help ensure that that valuable intelligence does not stop because of an organizational boundary.

We want to be trusted enough to endure, which means our decisions are explainable, our governance is visible, and qualified people remain accountable.

These principles matter because speed without trust can create that provider abrasion, as well as regulatory risk.

And intelligence without connection can produce insights that no one acts upon.

And the connection without clear accountability can create activities without results.

So the strongest Pi organizations are going to find a way to balance all three.

You don't have to be the best, but they have to balance the destination is not simply more automation, it's going to be more accurate payments, earlier interventions with a better use of our expertise and with a stronger feedback loop between the decisions and the outcomes of those decisions.

I think we skipped over.

So before we kind of wrap up, I would love to see if anybody has any questions.

All right, thank you so much.

So at this time, I'm going to go ahead and address any questions that may have come in during the presentation.

First question is, looks like it's referring to when you were speaking about sitting close to the coding logic asking does payment integrity ultimately have to move into that AI mediated workflow as well rather than primarily operating on the resulting claim.

I'm just trying to read that question a little bit #8.

So I think payment integrity, when we think about AI that sits in, let's say, ambient listening within the provider, I think payers have to pay close attention to that technology.

And it comes down to one, is that just a, is it really AI that's doing it?

Is an ambient listening with like more of like a dragon dictation?

How is that really kind of filling in the blanks?

Is that AI assuming behind the scenes that something was mentioned?

Where is it filling in those gaps?

So from a pair's perspective, when I see ambient listening AI, I'm thinking I kind of want to see the transcript behind the scenes as well.

So I might say, yes, I I trust this, but when it says, you know, dictated but not read or you don't use with AI and, and there's just a signature at the bottom, I would want to know, is that provider actually using ambient listening in a way that is responsible?

And what is that software?

And then I may, as an auditor from a coding perspective, look at it from behind the scenes and say, yeah, I want to actually see that the real transcript if I feel like the AI is doing something maybe that's inappropriate or inconsistent with what was built.

So I feel like, yeah, as us as payers and and obviously with the members, we need to be a little bit closer to AI that is in the provider offices as well, Right.

Another question, it's a couple of parts.

If payment integrity is moving upstream toward the coding and clinical workflow, how far left does this ultimately go?

And the other part is, could AI eventually use the medical record, chief complaint and payer intelligence to model, simulate and guide potential care pathways before the counter even occurs?

I think from a payment integrity perspective, we have to be careful because remember, as health insurance payers or health plans, we cannot provide or dictate care.

We can talk about what are the appropriate, you know, clinical things and provide guidance, but ultimately we cannot dictate care from an auditing and payment integrity perspective, sure, we can look at it, but could AI eventually use that record to create potential pathways?

It's, it's possible it's, it's coming out there where, you know, something very simple of I have a chief complaint and it's created some, or I'm sorry, I have an assessment in my assessment and plan and it's already created, looked at those things and said, OK, I'm going to pull these after our or these care instructions based off of what you inputted as your, your chief complaint or your care instructions that come from, you know, the diagnosis that I've given you.

And it's just, you know, machine learning and things like that.

It could do that, but I don't know if we're at that point where a payer could really get involved in that care aspect to guide any kind of potential care pathways.

Another question is asking with respect to real time intelligence, how do you monetize avoidance savings that have displaced the need for recovery?

I'm happy to jump in here, Stephanie.

I think it's, you know, a couple depends on where you kind of sit in the ecosystem.

As I was mentioning earlier, you know, for your your health plan leader or payment integrity leader in a health plan, you really have to think differently.

You really have to change the incentive model.

You have to change the way savings are measured.

And that requires a lift that requires getting the organization on board to understand that, you know, more recoveries or more prepaid edit savings won't always be a good thing.

As the technology evolves and as there's more opportunities to, you know, deliver this shift left intelligence into the provider workflow.

We're seeing it already work already where if you get providers that are really utilizing the intelligence they're being delivered, you're starting to see some prepaid savings go down in certain categories.

And that that's good news, right?

You're avoiding unnecessary denials, you're avoiding the friction, you're avoiding the back and forth rework and admin costs.

But the flip side is, well, how do I, how do I prove that I'm adding value as a payment integrity leader?

So the measurement and the focus needs to change on the vendor side.

That's a change of business model, right?

So vendors, you know, largely Optum included, often get paid based on a percentage of savings and that model will need to change over time.

And we're starting to experiment with new models largely based on, you know flat fees or license fees where we're kind of orchestrating the shift left ecosystem using our tools, using our payer provider footprint.

But you can't really get paid on something that has never happened.

There's no contingency fee in that world.

So it's more about charging kind of an appropriate license fee to facilitate that avoidance as opposed to our traditional business models.

And I'll just kind of chime in from a payer perspective.

When I think about how we quantify cost avoidance, it comes down to we know like let's say we have an edit that required modifier, required a modifier right or left or an anatomical modifier.

I want to push that into a real time where the provider will get a flag to say, wait a minute, you need to append a right and a left.

And with that said, now I know that I'm not going to get corrected claims coming in.

I'm not going to get provider phone calls into the customer service.

So what I try to do is I try to quantify how many of this edit that I maybe have previously.

And if I had let's say 100 of them and 50% of the time I would get a corrected claim.

And 25% of the time I would get at least a phone call.

How much did those cost me?

If it cost me $100 to do an appeal, did it cost me $25 to do a customer service call, you know, end to end?

Is it maybe $10 per corrected claim?

Does it it cost me end to end figure out what it costs to be able to have somebody review the claim, review the denials, process the corrections and figure out, OK, so now if I'm pushing that to the point where I don't have to have a phone call coming in, I don't have to have an appeal, you know, submitted and I don't have to send out letters.

What am I now saving in admin work that becomes cost avoidance for something where maybe I knew I was going to get a corrected claim in or things like that, trying to quantify it that way, not just on the savings that no longer appears.

So that's that's one way You can also kind of say, yes, we're saving, you know, staff time, we're saving all this, you know, administrative work.

It still has a monetary value that you can report on.

All right.

Well, I think that wraps up our participant questions.

So Stephanie, unless you have anything else to add, we'll turn it back over to Chris to wrap us up.

Great.

Thank you so much to our speakers.

We thank you for that great presentation and for sharing your thoughts.

Thank you to the audience for different for participating in today's webinar.

And this does conclude today's presentation.

Thanks again and enjoy the rest of your day