Hospital Revenue Loss, Cerner Blind Spots & How AI Innovation Can Fix These Issues

Hospitals lost $48 billion in revenue in 2025 from claim denials alone — a 25% increase year over year. In this episode we break down why Cerner EMR hospitals are structurally exposed to this problem, and how AI innovation is being built to fix it.

On April 22, 2026, Community Health Systems reported a $58 million net loss for Q1 — on nearly $3 billion in revenue. Leadership pointed to two compounding pressures: a challenging payer mix with fewer commercial patients, and a slow, steady ramp-up in claim denials that nobody's system was surfacing until it was too late.

Three weeks earlier, Kodiak Solutions published the most comprehensive analysis of hospital revenue cycle performance ever conducted — analyzing 2,300 hospitals. The number they found: hospitals lost more than $48 billion in revenue in 2025 from claim denials and uncollected bills. A 25% increase from the prior year. And the increases were specifically for lack of prior authorization and for medical necessity.

In this episode we break down exactly why hospitals running Cerner EMR are structurally exposed to these gaps — from the split between clinical and revenue cycle data models, to the inability to track true denial and appeal overturn rates, to the lack of machine learning needed to surface payer-specific denial patterns before they become systemic revenue leaks.

We then explain Bloom Value's patented enterprise visibility system — US Patent 20230260638 — and what cooperative machine learning engines, simultaneous real-time and historical data processing, and role-specific financial dashboards actually mean for a CFO trying to see where the money is going.

Transcript

SPEAKER_00 0:11

What if I told you that you worked all year and you will not get paid for it? Instead, you le lose 48 billion in that year. How would you feel? I'm talking about hospitals here. Hospitals earn this revenue, they deliver the care, and they just did not collect 48 billion of it in one year. And if your hospital runs Cerner, I'm going to tell you something very uncomfortable. You probably can't even see where your money is going. That's not a billing problem, that's a visibility problem. And today we're going to get into exactly where it's hiding and what AI innovation is finally doing to fix it. Let's start with the news, most recent news. On April 22nd, CHS reported a net loss of 58 million for the first quarter. Community Health System, one of the largest hospital operators in the US, running 69 hospitals across 1,000 care sites and 15 states. So what did CHS leadership say? They said there was a problem with the payer mix and the shift away from commercial payers towards Medicaid and Medicare. Secondly, denials were rising at the same time. But this is not a surprise to anyone who has heard what has happened with CHS before. 18 months earlier, on their Q3 2024 earnings call, the then CFO Kevin Hammond said, I can't say where there was one event during that quarter which could have pointed out this is where we are losing revenue. We just continue to see a slow ramp up in denials, as well as the time frame for adjudication process. And 18 months later, 58 million missing from a quarterly earnings report. And think about it, CHS is not alone in this. Just three weeks before this earnings announcement, Kodiak Solutions published the most comprehensive analysis of hospital revenue cycle performance ever conducted. They analyzed data from 2,300 hospitals across the United States. Hospitals lost more than 48 billion in revenue in 2025 from claimed denials and uncollected bills. That's a 25% increase net revenue leakage from prior year, which was 38.6 billion in 2024. So this is the state of healthcare revenue cycle report from Kodiak systems. So we're talking about this the way the hospitals are losing revenue here. Here is what makes the number even more striking. Kodiak found that hospitals actually got faster at collecting clean payments, even though on paper it looks good, but underneath that surface efficiency, the increase in clinical denials was specifically for lack of prior authorization for and medical necessity. The two most preventable, most upstream, and most expensive denial categories in the entire revenue cycle. Clinical denial rates rose over 12%. So final denial rates climbed, and Medicare Advantage plans were denying claims at a rate more than double traditional Medicare. That is the story of hospital finance in 2026. And if your hospital runs CERNA, which covers nearly one in four US acute care hospitals, it gets worse. Today we're gonna talk about why and what AI companies are building to fix it. I'm your host, Deep T from IdaHub, AI in Healthcare Podscast. Let's get into this. In 2026, class research, the gold standard for healthcare IT ratings, published a report on EHR market performance. The numbers for Oracle Health, which own Cerner, tells a clear story. Oracle lost a net 74 hospitals and their loyalty scores have dropped 10 points since then. So let's talk, take a moment to understand what this medical necessity actually means? Because it's one of those terms that get used constantly in healthcare finance. Medical necessity is the standard payers use to decide whether a treatment, procedure, or hospital stay was appropriate and required for that patient based on their own clinical guidelines, not the physicians are judgment alone. Every major peer has its own criteria, and those criteria change constantly. Medical necessity denials are the most expensive and the hardest category to deal with. And each one of this is made worse by how the way CERNA is built. So we'll get more into this detail. What Matt VP of Revenue Cycle Intelligence Kodiak Solutions said was payers are paying cleaner, simpler claims much faster, but the harder work is getting harder, with clinical initial denials continuing to rise and provided increasingly being asked to prove medical necessity, which is driving higher final denials. So almost nobody in hospital has a unified view of where the money is going because the system relies on a system that was never built for this. That starts with the reason one. CERN's data model makes it harder to see what caused medical necessity denials. The payer reviews the care that already happened and decides it was not necessary. A patient may have genuinely needed that inpatient admission, but if the physician notes did not document the severity in the specific language, the payer's algorithm was looking for, the claim gets denied. The relevant data itself is spread across both clinical and revenue cycle data models inside CERNER. So the data is there, but it is two separate architectures that were not designed to talk to each other. Custom reporting teams at hospitals frequently have significant challenges, acute accurately connecting clinical documentation data to revenue cycle outcomes. So the physician who documented the case and the billing team that received the denial are looking at two different systems here. And nobody can draw a straight line between them. There's another reason here, they require clinical expertise to appeal, and CERNAR gives you no reliable view of whether your appeals are actually working. Unlike the coding gear, which a building specialist can fix, a medical necessity denial requires a physician advisor or a nurse auditor to require a clinical justification letter arguing why the care was appropriate. Cerner customers are consistently report that tracking the true denial rate, true appeal rate, and true overturn rate over time is a significant challenge. Not because the data does not exist, but because it's scattered across the system that does not produce a single accurate picture. And without knowing your real overturn rate, you cannot tell whether your appeal process is which payers are worth fighting, or whether you even have the right staff allocated to the right accounts. You also cannot get a true snapshot of it of every account in its life cycle right now, which ones are approaching timely denials, which ones are pending, second, second level, second level review, which were written off anyone appealed. That invisibility has a direct dollar cost. So there are three financial reasons we're gonna dig into, especially with Cerner. The documentation cost problem here. An average physician spends two to three hours of documentation every single day on Cerner, and that's one of the reasons that's also driving higher claim denials. One of the blind spots here is claim denials with no root cause. Cerner tracks denials at individual billing image system. It does not connect a denial to the documentation that caused it and the registration error that triggered it, or the payer policy that drove it. So the same avoidable denial keeps happening again and again. The next is the data silo problem. Cerner has a clinical module, a billing module, an AR module, and population help mode. So they do have all these things, but they do not mean fully share data. A physician documents an encounter, coding generates a claim, a payer denies it, and the denial lands in billing, completely disconnected from the documentation that caused it. Nobody in that chain sees a full picture, and that costs hospitals millions per year. So here's a second blind spot: prior authorization fragmentation in Cerner. CERNAR has no unified dashboard showing real-time prior status across all payers. Staffs have to log into 15 to 20 different payer portals every single day just to check status. And Cerner cannot predict before you submit which authorizations are likely to be denied. That cross-payer intelligence simply does not exist. And month after month, without cross-system pattern intelligence, you're permanently in reactive mode. So you so this is the there's another visibility blind spot here. The payer understatements. Nobody is checking.5 to 2% of the net revenue lost annually to undetected underpayments. Payers sometimes pay less than the contracted rate. In CERNA, there is no native tool that automatically compares what was paid against what your contract says should have been paid at the claim level here. The last one we're talking about is the financial AI black box. So yes, Cerner has deployed denial prediction and prioroth AI, but once deployed, there's nobody who monitors these models are still working. A denial prediction model trained on 2022 payer behavior is still running for 2026, quietly applying old patterns to a market which has completely changed. So here's a blind spot. So denial patterns are hiding in plain sight, and CERNAR cannot surface them before they become systematic. If the original clinical notes did not support medical necessity, no amount of appeal letter writing fixes it. The documentation was a problem. Denial patterns are frequently varies, are very specific to individual health system, a geographic location, or a particular payer mix. A specific payer in a specific region may be applying a non-standard interpretation of medical necessity criteria for a specific procedure. And generating denials at a rate that is statistically significant across your claims population. But in order to see that pattern, you need to account for a large number of inputs simultaneously. You need machine learning algorithms that can process payer behavior, clinical documentation patterns, geographic variables, and claims history all at once and surface the specific combination that's generating denials before it becomes systematic revenue leak. CERNE cannot do it in its native, cannot do this natively. Without it, the same denials keep happening again and again. Now let's talk about the specific AI innovation, which is built to address exactly what we have described before. I want to look at it through the lens of their patent. This patent is by um Bloom Value Corporation. Their innovation is on system and method for optimizing outcomes for healthcare entities. Let me talk a little bit about this pattern. The core problem Bloom Value set out to fix is that healthcare executives lack integrated visibility, proactive exception management, and predictive actions, not department level visibility or module by module reporting here, but they're looking at integrated visibility across the entire revenue cycle in one place and in real time. That is a gap the pattern was built to close. The layer one, multiple financial engines need to be working together. Most analytical tools work on one problem in isolation. A denial prediction model looks at claims history, a prior art looks at payer rules, they don't talk to each other. What the pattern describes is a framework where multiple financial intelligence engines share data and inform each other. Payer behavior data feeds the denial prediction. Denial patterns feed documentation recommendation. Contract analytics feed authorization risk scoring. It's not one financial area, it's a coordinated system, and that coordination is what allows a platform to see across these data silos that CERNA creates. Layer two is real-time and historical at the same time. Most analytics we are talking about are do either one. They're either real-time or they're historical. This patent also describes how we can do both simultaneously. The system learns from historical patterns which payers deny, which procedures types, which documentation gaps cause, which denial codes, and it can apply those patterns to real-time incoming transactions. Let's talk about a practical example. A prior R request comes in. The system doesn't just check whether the payer typically approves this procedure, it looks at the specific documentation attached, the payer's recent behavior patterns, and generates a recommended action in real time before you submit. That's the difference between reactive revenue cycle management and proactive financial intelligence. That's the output which is enterprise visibility. A CFO sees the total revenue at risk, denial trends, contract performance by payer, a revenue cycle director sees today's specific claim queued by financial impact. A coding manager sees which documentation pattern generates the most denounce. So the right intelligence connected for the right person right now. Not just data here, it's a complete visibility. Three reasons for it is the driving right conversation for 2026. The Oracle transition is creating instability. And it has disrupted the billing and claims at even at 80 plus hospitals. Cash flow froze within 48 hours that affected organizations. Expose how deeply financial workloads are coupled into EHR infrastructure and how organizations with intelligence layers sitting above EHR have resilience that's only sharp stone. Medicare advantage pressure is intensifying. More than half of Medicare beneficiaries are now in Medicare Advantage plans. Payers are tightening authorization requirements and scrutinizing risk scores more aggressively than before. The HCC coding gap and the prior art fragmentation I described are becoming more expensive every single year. Organizations that close these gaps now gain a compounding financial advantage. The competitive gap is widening. Epic has been investing in financial AI, and some of them do see some measurable results in denial risk compared to Cerner. Cerner's market position is eroding. The gap between what Cerner provides natively on best-in-class financial performance looks like is not going to close on its own. It requires a decision to act. Three specific actions you can take as a VP of a revenue cycle at a Cerner hospital can take based on what we have discussed today. One, audit your denial root cause visibility, not just your denial rate, your root cause visibility. Ask your revenue cycle team right now, can you show me which specific payer procedure combinations are generating the most denials and whether the root cause is registration, documentation, coding, or payer policy? If the answer is we need a few meeks, we have the silo problem I discussed. Start there. Secondly, check whether your payer contracts are being paid correctly. Take one payer, one contract, one month of remittances. Have someone cross-referenced what was paid against what the contract says should have been paid at the claim level. Most finance teams who do this for the first time find discrepancies. They had no idea existed. Knowing your underpayment rate is a first step to recovering it. Three, demand performance accountability from every financial AI tool you run. For AI-driven financial tool linear revenue cycles, ask the vendor, can you show me how this model's predictions have performed against actual outcomes over the last six months? If they can't answer with data, you're running a financial AI that nobody is monitoring. In a regulatory environment focused on AI accountability, that is a risk you should not carry. Cernar Hospital has a clear visibility problem, not a data problem. The financial data exists. It is locked in separate models that were never designed to talk. That's what Bloom Values patent explains, and that's what they do. To cooperate you AI engines, real-time and historical data processing simultaneously, enroll-specific financial dashboards is exactly that. Not by replacing Cerner, but sitting about it and connecting things Cerner keeps separate. If any of this resonated, a denial pattern you can't explain, a Medicaid advantage gap you suspect you haven't quantified, reach out. At Idahub, we are curating innovative AI solutions for healthcare. And Bloom value is one of the innovative solutions for healthcare finance. The patent we talked about today is about one of Bloom's AI solutions called glass box visibility, a solution to visibility problems we discussed before. And glass box visibility is just one piece of a much larger picture. Bloom offers a full suite of solutions and a comprehensive platform built for enterprise-wide value optimization across the entire health system. So here's what I want you to do now. So if you're a CFO, a VP of a revenue cycle, or a healthcare technology leader at a Cerner hospital, and your organization reported a recent net loss like CHS, or you're seeing sitting on a denial patterns you can't explain, Medicare advantage gaps you suspect but haven't quantified, or payer contracts you're not sure are being honored. So this is your next step. Scan the QR core on the screen. You have two options here. Book a discovery call and get a live walkthrough of what this looks like for your specific health system, or download the CR Cerner Gaps and AI solutions PDF and start with that data. Either way, the visibility you've been missing is one just one conversation away. I'm your host, Deepti from IdaHat. Thank you for listening and see you in our next AI in healthcare podcast. Thank you, everyone.