Guest: Ken Hogue, Chief Revenue Officer, United Health Services Host: Praveen Chandran
Introduction and Revenue Cycle Leadership Background
Praveen Chandran: Hi everyone and welcome to another episode of season 2 of the RC Executive Lounge podcast series. I'm your host Praveen and I'm really glad you're all here today. I'm very excited to welcome a leader whose passion for healthcare began at age 15 working in a doctor's office on a military base and who has since built one of the most recognized revenue cycle organizations in upstate New York. Our guest is Ken Hogue, Chief Revenue Officer at United Health Services, the largest comprehensive healthcare system in the greater Binghamton, New York region.
Ken has led UHS's revenue cycle transformation since 2020, including a landmark Epic implementation that has positioned UHS among the top-performing systems in its class. Before UHS, he held revenue cycle leadership roles at Rochester Regional Health, Change Healthcare, and Allegheny Health Network with earlier roots at UPMC in Pittsburgh. He holds the CRCR and CPPM certifications and a BS from Clarion University of Pennsylvania.
What sets Ken apart is how he thinks about the work. He restructured UHS around a high-reliability culture, brought clinicians directly into revenue cycle workflows to address medical necessity denials peer-to-peer, and has championed the idea that revenue cycle is not a back-office function but a core part of the patient experience. Ken was nominated for the prestigious 2025 RCM Innovation Award, recognized by Becker's Hospital Review as a revenue cycle leader to know, and serves on the RCM Access Advisory Board. He is also a committed mentor to the next generation of RCM leaders through HFMA. Today we will talk about claims denial management, high-reliability revenue cycle culture, and the leadership lessons Ken has gathered along the way. Please welcome Ken Hogue.
Ken, welcome to the show today and thank you so much for joining us.
Ken Hogue: Thank you. I'm happy and honored to be here today.
Claims Denial Prevention Framework and the HRO Model at UHS
Praveen Chandran: Let's jump right in. Most organizations have claims denial management or denial prevention as a multi-year ongoing initiative. How does it work in your organization and what kind of frameworks do you use to determine the right priorities for a program of this scale?
Ken Hogue: Good question. For organizations that may not have started with a denial prevention or strategy focus, it is a multi-year initiative. Luckily here at UHS, when we went on to Epic, that was part of the foundation and the vision from the start. So it happened naturally and has been successful to the point where many people today ask how we stood it up and what continues to evolve and change.
When you talk about denials or claims prevention, it is very complex today compared to where it used to be. It really touches technology, people, and processes, and those processes continue to change with new code sets, quarterly new modifiers, billing regulations, and requirements. Depending on how payers or regulations change, it can vary by date of service or by date of claim submission. Trying to balance that from a technology standpoint is not cut and dry. When you have more payers within your organization, the complexity increases because you sometimes have to apply the 80/20 rule and manually work some of the fallout. Resources with expertise in denial management are at a premium for organizations today.
I have always said from a denial management perspective that it really needs to start as far upfront as possible. Working and consulting through Epic prior to UHS, our EMR is very much an upfront system. I always say that denials are often caused by processes or things that can be prevented upfront for at least 70 to 80% of them. The 20% is going to be the minutia that is just the normal course of business or regulation changes you have to address to get paid. But how do you balance both of those with technology and people?
For organizations approaching this, it really starts with researching the denial and asking: did it hit edits originally, did it hit errors or certain checks, and if it still originated as a denial, why? That is where you say, I have touched this two or three times through people, processes, or technology, and it is not just cycling through the revenue cycle life cycle. These are actual touches to try to prevent the denial. Why did I still get a denial? Just hitting edits may be where you need to use technology to electronically attach medical records or other similar things. It really starts with that foundation of laying out a root cause, your highest-volume denial, your highest-dollar denial, and then working through your remittance codes and remittance advisories to understand why each category originated.
On the HRO side, our organization is now on the high-reliability organization journey, which I think revenue cycle was ready for because it has been something I practice and preach within my organization. That has allowed us to really have that open framework for denials prevention, and within it we have a concentrated team. Instead of traditional insurance follow-up where after the first denial you work it from start to finish, we have a SWAT team. After the first denial and root-cause analysis, if it is not COB or something simpler, it goes to our denial resolution specialists who are at a higher level within the organization. They trend those categories and also help us with automation, as well as controlling and presenting the open denial inventory.
In today's environment it is not just a back-end function, because you need to understand everything that happened before the denial and then also ask: how do I fix this, how complex is it, and what automation can I use to work these denials? That automation may also include outside technology, claim scrubbers, and claim edits that surface service-level rejections back into our EMR. When we see those service-line rejections, we still treat them as part of our denial prevention committee because the claim made it all the way out of our system into the payer and then got rejected. You need to be asking whether your technology is talking to your core system, and that is more technical than ever when you talk about IT changes and mappings. It is not purely a finance function, and having prior IT background myself and in several of my leaders, whether from payer experience or prior IT roles, has allowed us to embrace denial prevention as a true technology, people, and process discipline. You also need to proactively think and sometimes verify before just letting claims flow. I would rather put claims on hold for a week or two and validate everything than work through thousands of denials at significant resource cost.
Top Denial Drivers and Data Patterns Across Organizations
Praveen Chandran: For our audience, Ken, you have worked as a leader across multiple organizations. Looking across all of those, what have been the top reasons driving denials, and what patterns have you observed from the data?
Ken Hogue: The payer landscape and its changes is one major factor. When we talk about policies, Medicare and Medicaid in particular, even on the managed side, part of that is knowing when contract terms are changing. The other thing I am seeing nationally from peers is the changing landscape within self-pay, the underinsured, and the uninsured. Healthcare premiums are getting more expensive both for organizations and for individuals in the marketplace. People are navigating that based on how sick they are or what they can afford and making the best-informed decision they can.
I also think we are in a technical era around data. People are now more responsive to using MyChart to send an electronic message to update their insurance than they are to a phone call. The standard 9-to-5 model, to use the Dolly Parton song reference, does not fit how people want to engage today given family, work-life balance, and commutes. The goal really is how do we have revenue cycle available 24/7, 365, as much as possible between people and technology.
One thing I can truly say is that with Epic specifically, or any other EMR, sticking to the foundation and making sure the automation tools are actually being used, and verifying that usage routinely, is critical. If you have turnover or growth and you trained people but you are not going back quarterly or monthly to verify, you end up with people who get busy and pulled in other directions and stop using the tools that would ultimately help them. That foundational approach encompasses both system and people, and I advocate strongly for the principle that an ounce of prevention is worth a pound of cure.
Realistic Denial Rate Targets, Benchmarking, and Clinical Integration
Praveen Chandran: Looking at your experience across organizations, what is a realistic target for denial rate that any leader can aspire to, assuming they are adopting agentic AI and whatever operational changes are needed?
Ken Hogue: Organizations need to look at this cross-functionally. You need to look at your initial denial rate, your secondary denial rate, your overall denial rate, and your final denial rate. You cannot take a one-size-fits-all approach.
When you look at the initial first-pass denial rate, top-performing organizations today are seeing roughly 6 to 6.5%, with some up to about 8%. But within that, it depends on how organizations report it, because COB and other factors, along with state regulations, all vary. It is about understanding what you can control versus what you cannot, and how you compare to peers. Here at UHS, we often compare to New York facilities, California facilities, and Colorado facilities because they have similar state-level regulatory environments.
Our denial rate may be a half a percent or a percent higher than some others, so I look at national best-performing, regional best-performing, and state-by-state, and blend that together. I always aim to reduce another five or ten percent of the net rate, not the rate itself, to set aggressive targets. We have been able to maintain somewhere around 6 to 6.5%, sometimes up to 7%, but balanced across the full year we are under 7% and at 6.5% a majority of the time. When you start to see the denial percentage increase, immediate action is required, reviewing what is coming in weekly and daily to prevent it from compounding.
Final write-offs are the other story I cannot emphasize enough. Leaders need to know the final denial rate across all of their billing and the final denial rate on what was billed within the last two years, because you need to make sure you are not just accumulating denials while generating metrics. The net collection rate on older denials tells that story. And retractions are another dimension: in this business we can get paid and two or three years later have money taken back. That is a soft denial, or it can post as a formal denial depending on your setup. Why does that happen? Clinical documentation. That is part of why embedding clinical involvement in denials prevention is more important than ever.
At UHS, we fully audit things and then learn from that. We have a champion model with a dyad approach within each department, both a clinical champion and a business leader, and then we balance what is happening at the network level for all of our affiliates, including critical access hospitals and rural health distinctions. Some denials happen within certain types of care delivery and you need to get into that detail while still balancing the whole. At least monthly we tear everything apart and put it back together to make sure nothing is hiding, and that has allowed us to protect revenue instead of just being reactive to denials.
Competing Operational and Technology Priorities: Governance and Prioritization
Praveen Chandran: This is a challenge we hear about from leaders time and again. There are always operational requirements from the operations team and innovation requirements from the technology team. Operational requirements are short-term bets that give you quick wins. Innovation bets are long-term, risky, and take a long period of time. For a leader such as yourself, the tension can be healthy or unhealthy between the teams. How do you navigate this?
Ken Hogue: Given the culture change we have gone through, embracing revenue cycle as a catalyst, we are a collaborative organization. We did have some friction early, some misunderstanding of what was needed or how significant certain changes were, but we have actually started to see several people from our revenue cycle team take IT as a career path and ladder. That is showing we did our job in developing people who can serve as the bridge between technology and revenue cycle.
The other piece is governance. When you talk about IT or data governance, it is about understanding those work streams for when something big and impactful can be turned on. It is also about monitoring after it is turned on: is it actually working as expected? Not a month or two later when it has already been causing denials and you are asking what happened with the automation. You need to track it all the way through to closed, completed, validated, and tested, with both positive and negative testing.
For prioritization, it comes down to compromise. Organizations need to quantify within buckets what they are able to work and what they are not. If the team can touch it manually and still get money in the door, that is not best practice, but it is a compromise, because there may be other things the team simply cannot recover or resolve without help. That is the first pillar in our negotiation with IT partners, and the compromise I sometimes have to remind my leaders of: yes, we may have to touch 300 or 400 denials manually, but if IT can address this, in the future we will have a thousand fewer denials to touch per month.
That really depends on the work streams and on leaders getting directors and VPs in the organization to understand how everything correlates, because not everything can be a fire or a top priority. Going back to the principle of what you are able to work and not, and then looking at each technology piece and making sure governance is in place asking what it is supposed to do and whether the leader responsible is verifying it is producing what is expected, because if not it can damage a workflow that is already in place.
The broader trend is organizations seeing more Chief Revenue Cycle Officers emerge, and that is really where strategy comes in, not just as a back-office function but how are we looking at the entire organization from start to finish. I want to look at how a registration occurred, how a referral came in for a service. Balancing that big picture with some core principles allows alignment. For the audience, the framework is: what is workable and what is not, and what is your monthly volume? Then create work sprints with IT, periods where we focus exclusively on initiatives the team cannot touch and bill at all, followed by periods where we work on some of the open things for efficiency. Creating those sprints in a partnership model is what works.
Praveen Chandran: Makes sense, Ken. Great set of insights. Let's take a short break. When we start the next segment we are going to jump into an initiative that Ken's team led, the challenges they were solving for, and how they went about the implementation. Please stay tuned.
Praveen Chandran: Welcome back everyone. Ken, let's dive right in. Can you walk us through a recent claims denial management or claims denial prevention initiative that you and your organization led? Let's start with what challenges were you solving for.
Recent Initiative: Alternate Payer Verification and TPL Data Quality
Ken Hogue: When we talk about alternate payers, such as the Veterans Administration or third-party liability situations, we needed to be able to verify the data within the EMR and confirm that it was being selected correctly. Questionnaires, intakes, and the fact that situations change over time: people retire, cases close, or someone relocates and brings a prior injury with them that may still be open. How all of that correlates is complex.
From our team's perspective, we really started asking: what are the key attributes within our Epic system, what is reportable, and what do those fields mean? We worked closely with our training team to make sure that knowledge was clearly understood across the organization, and then reverified some of those fields where they had changed, particularly with newer team members. That goes back to the core principle of verifying the data elements. Just because something has worked for five years does not mean it is still working correctly. You also need to look at it critically and ask: what if somebody made a wrong decision here? Do I have to rework and touch it ten times to get it to work? That led us to think through where escalation and re-education touchpoints should exist to make sure our workflows and system stay optimized.
Praveen Chandran: For this particular challenge, what did implementation look like? Were there any pitfalls along the way? And did you do this all in-house or did you have to go to a partner?
Ken Hogue: It was a combination. This did involve our clearing house, because some of the edits we saw within the clearing house were competing with what we saw within Epic. It was an interpretation issue around a specific field, for example the accident date, where data would be pulled. That also goes into how your clearing house handles updates or additional confirmation notes, what those mean, and how they pull within the system, because you do not want to triage and massage those manually every time.
For us it was working through that, confirming the decision, and using outside databases to verify whether a case was still open. New York has the workers comp board and other verification tools that are needed. Part of that was making sure we were sending records electronically, which did come out of Epic, but it also involved our clearing house being able to attach those and pass them to the payers. Behind the scenes, once a claim comes out of Epic with the claim definition file and the remittances, we needed to have the data part resolved and work with our claim partner. Understanding our setup, how things were gathered, and the detail of the record was the work: originally the system was looking for one thing, but once you traced it you found it was already attached or needed to pull from a different field.
Until you get operations, vendor technology support, and client support aligned with your own internal IT team, you end up talking at each other instead of working through the problem together. Once that collaboration happens, you also have that trust-but-verify model you can apply going forward, and things move much more efficiently. Working through an exercise like that and embracing that mentality actually makes future iterations much faster.
Measuring Success: The Touches Metric and Implementation Lessons
Praveen Chandran: For a program like this, how did you go about measuring success, and what has been the measurable impact so far?
Ken Hogue: A basic metric that many organizations may or may not look at is the number of touches. That is where we are able to say: what is the number of touches this claim took to go out clean and get paid, counting people touching it, our main EMR processing it, and clearing house related edits. We were able to move a lot of our normal touches from about 14 down to around 10 in targeted workflows, removing five or six steps. In some instances it went from 17 down to 11. When you break that down, it was automation and people work combined. The people touches were reduced by one or two efforts, but by surfacing the error or message upfront, so that if someone does not know how to handle it, it gets reviewed immediately rather than being touched three times to correct it. That is the piece that was most meaningful.
From a system standpoint, once you identify a common field that is being mispulled, adding an automation data check addresses it at that level, but education still needs to go back to the front end. Staff cannot just depend on AI or automation to fix their errors. They still need to be educated. That dyad model, learning from the incident and embracing both the system fix and the human education side together, is how we absorbed that initiative.
Praveen Chandran: Thank you so much for sharing all those details about this initiative. We are going to take a very short break. When we come back, we are going to talk to Ken about the role he sees agentic AI playing in claims denial management and denial prevention, both in the near term and in the future. Please stay tuned.
Praveen Chandran: Welcome back everyone. Ken, let's focus on agentic AI. It is a buzzword we hear at HFMA, MGMA, wherever you go. As a Chief Revenue Officer looking ahead in terms of innovation and your own organizational requirements, what role do you see agentic AI playing in claims denial management or denial prevention over the next 12 months versus three years or five years?
Agentic AI in Claims Denial Management: Near-Term Reality and Long-Term Vision
Ken Hogue: Over the next 12 months we have gone from, as you said, a buzzword, to piloting it, to a point where we have to embrace it. There is not a choice anymore to combat AI with AI. Payers are using it, so we have to use it, or else we are just going to be swimming along, working the work, and not able to keep up.
Where I see it going is toward something more intuitive. Beyond the basic if-A-then-B-then-C approach, more of an informed decision-making model: let me review if it could be one of these three scenarios, evaluate each, and make a recommendation. It still requires people. The analytical thought process and the innovator's brain, how does that brain work to drive success? That is really what AI is doing, functioning as an internal analyst to gather and synthesize information, but we still need to trust and verify before we accept a decision or allow a submission to proceed. And then there is refinement: if data fields change, if there are new codes, if there are new regulations, you cannot just tweak one thing in AI. It is much more of an architectural approach because you also need to reverify through positive and negative tests that you are still getting the same output.
Today, if AI is working 500 edits, that is what you staff to. But if you make a change and it fragments from 500 down to 300, and then another payer policy change brings it down to 100, you have lost the efficiency of that automation. Your team starts filling the backlog and wondering where the volume is coming from without any actual new growth. Balancing that is about making sure your true experts, those who know the work, are able to supervise and supercharge the workflows and the simpler if-then situations, with human eyes confirming that it is actually working.
I look at it this way: supervisors and managers used to manage people. The modernization going forward is managing people and technology and the workflows that connect them. In some instances there may be managers who manage one or two analysts or experts but are primarily managing automation processes. The evolution is very much toward people plus system, and people leaders who are strong in that dimension also need to be able to synthesize data and technology. If not, those teams are not going to be as high-performing as they were five years ago.
Step-Function Improvement and Adjacent Technologies
Praveen Chandran: When you look at agentic AI, are there areas where instead of incremental 5% or 10% change you think there could be step-function improvement?
Ken Hogue: A 5 to 10% improvement is not out of the question, especially as we learn more and understand what AI can do. A lot of EMRs and vendors are adapting very quickly. We have gone from wanting it and needing it and pressuring development, to now having a lot of development and asking how do we trust and test all of it and say yes, this works for us, or hold, we need further validation. The methodology has evolved such that a lot of leaders have to synthesize whether it is actually going to work because the tools are available. And I do not think 5 to 10% is out of the question. It could even be more, but it requires that high-reliability approach of turning over every stone to make sure it is still performing. You need analyst or human eyes on the technology every day, week, and month. Organizations that do not maintain that oversight are going to struggle with AI and automation rather than be helped by it.
Praveen Chandran: One last question for this section. When you look at claims denial management, we have heard about newer technologies like ambient AI. Are there specific technologies that you and your team are watching with respect to RCM and claims denial management?
Ken Hogue: Ambient learning is always going to be one. How do we reduce the burden of physician documentation and others? Providers' time is valuable. They need to see patients, and in some instances there is a shortage of providers. We need them treating and seeing patients. The other component is AI in the two-way communication space: chat agents, augmented AI for conversational or customer service related approaches. We are not saying we no longer need customer service staff. Automation will handle the simpler questions, but the questions left over after that automation are harder. That is actually where you have to scale and adapt your team to be able to answer those, because if technology handles the simpler things, the questions that reach staff are going to be either more compounded or situations where the patient had five questions, got two answered, and still has three they could not resolve and needs to speak with someone.
You have to make sure the answer the AI gave was confirmed, and then handle the more complex items with a level of patient satisfaction in mind. We need to make sure we do not lose the human element or the human touch, because if we do, it becomes a frustrator rather than an accelerator for the organization.
Praveen Chandran: Totally makes sense. Ken, great set of insights on agentic AI. Thank you so much. We are going to take one last break before our final segment. When we come back we are going to talk about leadership challenges and how a Chief Revenue Officer of a large system navigates these challenges, especially around building momentum for change management. Please stay tuned.
Praveen Chandran: Welcome back everyone to our final segment of the show today.
Key Metrics, Daily Pulse, and Team Temperament
Praveen Chandran: Ken, every leader tends to have one or a few metrics they look at every morning, or once a week, or with some frequency. Is there a claims-related metric that you obsess over?
Ken Hogue: The clean claim pass rate. I kind of live or die with that from a blood pressure standpoint. The other thing we look at is our built and accepted claims, or the open AR awaiting payer acceptance, not just what has been denied, because some of those service-level rejections come back two or three days later. It is really about understanding that full throughput of revenue cycle.
Beyond the traditional metrics, there are also the less conventional ones: what is the team temperament, what are the work queues like, are we working the work we are receiving in the same week we receive it or are we starting to see backlogs build? Days work on hand is another metric some use and some do not formally track, but looking at every department's workload, having analysts look at it weekly and report up where there may be backlogs, that shows when someone is unexpectedly out and opens the conversation about capacity. It is all about balancing the team and making sure they do not burn out. So to me it is the clean claim rate, but also the pulse of what are we working and are we working it within the week it comes in, or are we building a backlog that may need some resource rebalancing or a collaborative look as a team.
Building Momentum for Change, Leadership Advice, and Closing Reflections
Praveen Chandran: One thing we have heard from leaders time and again is that in a complex RCM environment, especially within claims, building momentum for change across cross-functional teams like operations and technology has been really challenging. For this kind of initiative, for your organizational priorities for the year, how do you go about building momentum for change, especially if there is a step-function change in the way they operate or in the way they talk to patients about claims?
Ken Hogue: It is a three-step approach. The first step is making sure your team is informed enough to actually explain it in detail, not just a generalized quick answer, because patients will poke holes at a quick answer and ask why and how this could occur. The second is informing about regulation changes. It is not always the hospital or physician office's fault. There are regulation changes and timing issues, and it is okay to say there was a recent regulation change, we have made the adjustments, and we are working with your insurance company to get this resolved. You do not need to get involved, but we are actively handling this on your behalf. The third is when something is competing, being clear about what we are able to verify and where we may need the patient's involvement. If you are seeing a lot of rejections or COB-related issues, it may be faster to have the patient call and resolve it directly because of what organizations can and cannot see on their side.
You also have to make sure you are protecting payment long term, not just billing something and getting money in the door. Is it accurate? Because two or three years later having payments taken back because due diligence was not done is a much harder outcome. Making sure teams know that, educating them properly, and then running routine lightning rounds with those teams to gauge how many questions and how much chatter there is, will show whether there is hesitancy or whether there are policy changes that need scripting or a more intuitive conversation to further empower team members to speak confidently about it.
Praveen Chandran: If you were to give one piece of advice to another CFO, CRO, or RCM leader at a VP level and above who are tackling denial management challenges or denial prevention challenges, what would that advice be?
Ken Hogue: The advice I would give is that most of the time your experts at knowing how to make those changes are already within your organization. A lot of times you need to be able to convene those cross-functional teams. Sometimes the question is how do we make them focused enough to give us those answers. Embrace the team you have, ask those questions, and then work to have those working groups polish up the thinking, because a lot of those great efforts are already sitting on your team. From there you can partner with vendors or technology to make those decisions faster or execute those change-control items. Do not take it personally. Listen to the facts to be able to make an actionable change for your organization.
Praveen Chandran: That is a really powerful leadership insight. Thank you so much. One last question to close the show. Is there a recent book, podcast, or video that has significantly influenced your leadership thinking in the last one to two years?
Ken Hogue: One of the books that resonates most with me is Joy Notes, written by Neoqu, a personal friend and mentor. I am very into music and I think music is a universal language. It is something I have used with my leadership team to think through periods of trauma or when change is needed, which is what her book is about. For people who do not want to read something purely technical, each chapter has a song association with it. That made it fun and interactive as a leadership experience, and some of those songs actually spoke to the heart and soul of going through what the chapter described, creating that moment of recognizing I am not in this alone, I am going through something similar, and I am now able to relate to what she wrote. It really reinforced the idea that we are all in this together. Every organization has very similar challenges, and there is no challenge that someone else has that no one else does, which is why there is an ecosystem of RCM leaders, experts, and organizations to collaborate with.
Praveen Chandran: Makes sense. Ken, you have shared some powerful insights on two very interesting topics, claims denial management and RCM leadership. Thank you so much for joining the show today. As discussed earlier, I really hope you can join us again on a different topic like patient financial engagement. Until then, thank you so much for joining the show today.
Ken Hogue: Thank you for having me.
