Beyond Accuracy: Rethinking Inpatient Coding Quality in 2026 Webinar

Ella Jeffries: Hello, everyone. This is Ella Jeffries with Becker’s Healthcare. Thank you for joining us for today’s webinar, “Why Inpatient Coding Quality Is Now a Governance Issue.”

Before we begin, a few quick housekeeping notes: we’ll start with a presentation and leave time at the end of the hour for a question-and-answer session. Submit questions any time using the Q&A box on your screen. Today’s session is being recorded and will be available afterward using the same link you used to log in. If you have any audio or visual issues, try refreshing your browser, or submit a technical question in the same Q&A box — we’re here to help.

I’m pleased to introduce today’s speaker, Laura Legg, Vice President of Coding at Coronis Health. Thank you, Laura, for being here today, and I’ll now turn the floor over to you.

Laura Legg: Well, thank you very much for that nice introduction — I really do appreciate it, and I want to thank all of our attendees for joining me today. It’s my pleasure to be here.

We’re going to talk about Beyond Accuracy: Rethinking Inpatient Coding Quality in 2026. What’s changed in 2026 is that accuracy no longer protects the organization from the things that cause major disruption — clinical validation denials, pattern-based targeting from payers, quiet case mix erosion, and, one of the worst outcomes, executive loss of confidence in HIM and coding departments.

So today we’re going to reframe inpatient coding quality as revenue integrity governance. We’ll define governance maturity in plain language, walk through a maturity model, and translate it into three practical pillars: completeness, defensibility, and severity integrity. Then we’ll make it real through three clinical stories, because the fastest way to understand risk is to see how it shows up in the actual patient narrative and payer outcome.

As you listen today, keep a question running in your head: if my CFO or CEO asked me, “Are we protected?” — could I answer with more than an accuracy percentage? By the end, you’ll have the language and the structure to do that.

Here are our objectives for today: understand how payer analytics, audit scrutiny, and documentation complexity are reshaping inpatient oversight; recognize why accuracy benchmarks no longer provide the protection they used to; explore a governance-based model; and identify practical steps to strengthen coding oversight without increasing unnecessary audit burden.

A little about me: I’ve been around the HIM coding world for a long time, I’m very active with AHIMA and HFMA, and I’ve served on the advisory board for Briefings on Coding Compliance. I’m really happy to be here presenting today, so let’s go ahead and get started.

Laura Legg: Let me ask the question the way your executives would ask it: are we OK? Most HIM leaders answer with the best metric available — coding accuracy. But executives aren’t really asking whether the coding team followed all the rules. They’re asking whether the organization is exposed or protected.

So the real question isn’t “are we accurate?” It’s “are we protected?” Protected from validation denials. Protected from payer pattern analysis. Protected from a slow decline in severity capture. Protected from reputational risk when mortality or quality reporting doesn’t align with clinical reality.

This is where coding quality becomes leadership work. The higher up you go, the less people care about the mechanics and the more they care about stability. If we can’t translate coding quality into protection, we leave a substantial gap between operations, reporting, and executive confidence. Today’s framework closes that gap — it gives us a way to talk about quality that sounds like risk management, because that’s exactly what coding quality has become.

Our oversight environment has changed in three big ways. First, payers have gotten better at finding patterns — they’re not randomly pulling charts, they’re running claims through algorithms whose analytics spot outliers: high sepsis rates, unusual respiratory failure capture, disproportionate MCC use, or severity shifts that don’t match peer benchmarks.

Second, clinical validation has moved from occasional to routine. It’s no longer enough that a provider documented something — payers look for clinical indicators, consistency of treatment, and documentation clarity. That’s why we can see stable accuracy and still see growing denials.

Third, public reporting and internal quality programs make severity integrity visible. If severity capture drifts down, mortality ratios and quality measures can move in ways that create leadership concern, even when the care itself is excellent.

This environment rewards consistency, defensibility, and disciplined governance. It punishes variability, silent workflows, and education-only fixes. So our quality model has to evolve.

Laura Legg: Why? Because accuracy alone creates false confidence. We’ve all had occasions when we looked at our quality rate — 95 to 97% accuracy — and felt that was great. But it’s a backward-looking compliance check. It tells you whether what was coded matches what was documented. It does not tell you what you failed to find, what you failed to query for, or what you will fail to be able to defend.

That’s the trap. You can maintain a 95 to 97% accuracy rate and still lose millions to validation denials because the documentation wasn’t clinically strong enough. You can maintain accuracy while MCC capture quietly declines because documentation review discipline has eroded under productivity pressure. You can maintain accuracy while the same DRG error repeats because monitoring controls were never added.

Executives don’t experience accuracy. They experience denial spikes, appeal workload, revenue variability, and uncomfortable payer conversations. Governance maturity aligns our internal measures with the real-world experience of risk.

[Polling question: attendees were asked to rate their organization’s current coding quality maturity. Results skewed toward organizations already measuring completeness, with roughly 28% reporting a fully integrated CDI and coding oversight dashboard.]

Accuracy is genuinely valuable — it gives us a baseline signal that coders are applying the guidelines correctly to what’s documented. It catches sequencing issues, PCS and code selection issues, principal diagnosis errors, and other technical mistakes that absolutely matter. It also provides a shared language for improvement: without it, we can’t calibrate auditors, compare teams, or verify whether education is working.

But let’s be honest — accuracy is a compliance precision metric. It’s typically retrospective, and it depends on what the audit tool and audit scope define as correct. Accuracy is the foundation, it’s our floor. What we’re adding today is the protection layer — metrics and controls that measure what accuracy cannot.

What doesn’t accuracy measure? First, it doesn’t measure what’s absent. If a coder missed an MCC that was supported in the record, accuracy may never flag that unless your audit methodology explicitly looks for missed opportunities. That’s why an organization can look good on paper but still be underperforming financially.

Second, accuracy doesn’t measure clinical defensibility. A diagnosis can be documented and coded accurately and still be denied because the clinical indicators don’t support it. That’s not a coding error — that’s a defensibility failure.

Third, accuracy doesn’t measure recurrence. If a team receives education but the same error keeps repeating, the issue isn’t knowledge — it’s control design and accountability. So the problem isn’t that accuracy is wrong. The gap is that accuracy is incomplete as a governance system.

Think of it this way: accuracy asks “did we follow the rules?” Revenue integrity asks “will this hold up? Will this stay stable?” Revenue integrity requires upstream controls — risk-based pre-bill review for high-risk DRGs, a defensibility checklist, and a monitoring system that detects drifting behavior before it becomes a financial event. It also requires integration: if denial analytics sit in one silo, CDI sits in another, and coding quality sits in a third, you end up with fragmented learning and repeated surprises. Mature governance pulls those threads together into one operating system — that’s the shift we’re outlining today, from reporting quality to engineering stability.

[Polling question: attendees were asked about the trend in clinical validation denials at their organization. A significant increase and a moderate increase each drew 34.7% of responses; 23.6% were unsure or not currently tracking the metric.]

Laura Legg: Silent erosion is one of the most costly failure modes because it doesn’t trigger any alarms. It looks like stability — you’re hitting your accuracy target, there’s no major denial spike, everyone believes the system is working. But underneath, small changes accumulate: fewer queries, lighter documentation review, more reliance on summary notes, less time spent in consults and nursing documentation. Over months, CC and MCC capture declines by a few points, CMI slips, severity reporting drifts. Then the question suddenly arrives: why is case mix down? At that point, you’re chasing history.

Mature governance prevents this by making completeness visible in real time — capture rate trending, missed opportunity audits, and query consistency monitoring. The key idea: if you don’t measure it, you can’t protect it.

Laura Legg: What is governance maturity? It’s simply how disciplined and predictive your coding oversight system is. Low maturity looks like this: we audit randomly, we report some accuracy scores, we educate, and we hope it goes away. High maturity looks like this: we identify where our risks are, we design targeted controls, we monitor recurrence, and we integrate denial data and CDI alignment so the system improves continuously. So maturity isn’t more work — it’s better architecture. It answers: do we have a repeatable way to prevent the same issues from returning? Do we have completeness in place? Do we have a defensibility standard that CDI and coding share?

Many organizations respond to pressure by increasing auditing, more education, more pressure on coders. These actions may be necessary temporarily, but they’re not a mature model — they really represent volume. Maturity is designed smarter. Oversight means you audit what matters most, not everything. Targeted risk controls mean you build checklists, pre-bill reviews, and escalation pathways for your high-risk DRGs. Structured accountability means repeat errors are tracked, addressed quickly, and prevented from becoming cultural norms. Executive alignment matters because leadership needs risk framed in financial and operational terms, not just accuracy. And predictive modeling is the future — using trends to anticipate where denials and erosion may occur next. The message is simple: we’re not increasing pressure, we’re increasing control.

Let’s walk through the five-level governance maturity model, like a ladder. Level 1 is where many teams start: accuracy reporting — useful but limited. Level 2 adds trending: which DRGs, which error types, which facilities, which coders, and calibration to ensure consistency. Level 3 adds completeness tracking: MCC/CC capture rates, missed opportunities, query rates, and chart review discipline — this is where silent erosion becomes visible. Level 4 is where organizations become truly defensible: CDI and coding share criteria, denial analytics are integrated, and high-risk DRGs have structured pre-bill oversight. Level 5 adds predictive intelligence — automated flagging, risk scoring, and early warning signals. The goal isn’t to jump to level 5 overnight; it’s to move up one level with intentional controls.

It’s possible to run a very busy audit program and still be immature. If audits don’t translate into controls, you’re just documenting problems repeatedly. Maturity shows up when audit findings lead to structural change: a checklist is built, a workflow is clarified, a monitoring dashboard is created, escalation thresholds are defined, and denial trends adjust audit focus. Education supports the system, but education is not the system — if the same error repeats after education, that tells us the issue is process adherence and accountability, not knowledge. High maturity reduces audit burden over time because you reduce recurrence. Low maturity keeps you trapped in a cycle of discover, educate, and repeat.

Executives care about exposure and predictability. If a payer targets sepsis, respiratory failure, malnutrition, AKI, or encephalopathy — the usual culprits — those reviews create volatility, and that volatility turns into write-offs, appeal costs, and leadership concern. Mature governance identifies high-risk areas and adds protection: a defensibility standard, a pre-bill review model, and denial integration into audit strategy. This is also where HIM leadership influence grows, by translating clinical coding risk into financial and operational risk. Instead of saying “accuracy is 96%,” you say “our validation exposure has decreased and our defensibility controls are stable.” That kind of language builds confidence.

Governance architecture is the operating system — what you monitor, how you define defensibility, how you prevent repeat errors, how you escalate issues, and how you report risk to leadership. Activity is what happens when the operating system is weaker: you run audits, send education, hold meetings, but the same problems keep recurring. In a mature system, controls do the heavy lifting — people know what good looks like because expectations are codified, leaders have visibility because dashboards show completeness and denial exposure, and escalation is predictable, so behavior stabilizes.

Laura Legg: These three pillars are very practical — they translate maturity into what teams do every day. Completeness is about finding the full story: secondary diagnoses, CC/MCCs, and appropriate queries. Defensibility is about ensuring what we code can withstand scrutiny. Severity integrity ensures the record reflects true complexity, supporting accurate severity of illness, risk of mortality, case mix index, and quality reporting.

If you strengthen only one pillar, you’ll still have exposure. You can improve completeness and capture more MCCs, but if defensibility is weak, denials may still rise. Or you can be conservative and avoid denials, but completeness declines and your CMI erodes. Your governance needs all three.

Completeness. Complete clinical story capture starts with discipline — what documents are reviewed, in what order, with what expectations. When teams skip consults, nursing notes, and diagnosis details, secondary conditions get missed. Completeness is also behavioral: it’s easy to capture the obvious. Maturity requires consistently evaluating opportunities — asking, is the respiratory failure really supported? Was the malnutrition clinically documented? A mature team measures capture trending and runs missed opportunity audits, creating visibility into what accuracy misses. Leadership has to align productivity expectations with review time; if you reward speed without review discipline, you train incompleteness into the system.

The financial impact is real: a small decline in MCC capture across thousands of discharges can equal millions in reimbursement. It also affects quality reporting and mortality benchmarking, because severity is part of risk adjustment. When completeness is weak, the organization doesn’t just lose reimbursement — it loses predictability. Finance sees drift, leaders start asking questions, and teams scramble to explain why. That’s why completeness must be measured like a financial KPI. Completeness isn’t about maximizing acuity — it’s about representing the full clinical reality reliably.

Defensibility. Defensibility is the modern battleground. A diagnosis can be documented and coded accurately and still be denied because the clinical indicators aren’t strong enough or the documentation is inconsistent. Defensibility requires shared criteria between CDI and coding — if CDI queries sepsis using one standard and coding applies another, you create inconsistency, and payers notice. The practical tools here are simple but powerful: an indicator checklist for high-risk diagnoses, a pre-bill review model, and a denial-to-audit feedback loop. When denials rise, your audit focus shifts; when documentation patterns improve, your controls relax.

Defensibility is how you protect revenue without swinging into under-coding.

A defensibility control framework is a repeatable system, so you’re not relying on individual judgment case by case. Start by defining your high-risk diagnoses and DRGs — sepsis, respiratory failure, malnutrition, AKI, and major complications. Then define what “defensible” looks like: objective indicators, treatment consistency, and clear provider documentation. Next, apply risk-based pre-bill sampling — you don’t need 100% review, you need a smart review where risk concentrates, and when a case is borderline, escalate early for CDI alignment, physician clarification, or secondary review. Finally, fully integrate your denial analytics; denials aren’t separate from quality, they’re the most honest feedback loop you have.

Severity integrity. This pillar is about accuracy at the system level. If you under-code, you compress reimbursement and distort risk adjustment. If you over-code, you increase validation risk and pattern targeting. Mature teams monitor severity trends — SOI/ROM distribution shifts, mortality benchmarking, CMI drift, and DRG mix changes. They ask: are we representing patients consistently? Are certain service lines drifting? Are queries declining in higher-acuity areas? This pillar also requires alignment between CDI, coding, and quality — severity isn’t just a coding output, it’s a reporting reality. When you manage severity integrity, you protect revenue, growth, and credibility.

Laura Legg: Here’s our first clinical story — one we see every day. A patient is admitted with pneumonia, deteriorates rapidly, and the record documents sepsis and acute respiratory failure. Both diagnoses are coded based on the provider documentation. But payer review determines that the clinical indicators aren’t consistently documented to support the systemic infection or the respiratory failure thresholds. Across a set of similar cases, this documentation gap triggered targeted clinical validation denials. The lesson: accurate coding alone isn’t enough — diagnoses have to be supported under payer review.

When the team dug into the root cause, it wasn’t that coders didn’t know sepsis. The organization had never shared a standard for what sepsis documentation had to look like to be defensible. Coding and CDI were operating at different thresholds. Auditors were reviewing documentation presence but not indicator strength. There was no checklist for high-risk diagnoses, no routine pre-bill sampling for the DRGs most likely to be targeted, and denial trends lived in a separate world, so the quality program didn’t adapt quickly. That’s governance maturity — it’s not incompetence, just immaturity. The system relied on individual judgment and retrospective accuracy checks instead of a defensible control.

Once you see it this way, the fix becomes obvious: you don’t educate harder, you build the standard. The governance shift here was structural. The team built a shared clinical indicator library for sepsis and respiratory failure — one standard everyone used. Borderline cases were escalated quickly. They implemented risk-based pre-bill review for the top targeted DRGs, and, most importantly, denial analytics became part of the quality operating system: when denials rose in a category, audit focus shifted; when documentation improved, controls relaxed. The result was fewer surprises and fewer denials — the system learned, adapted, and prevented recurrence.

Through all of this, accuracy stayed stable and audits looked fine — no crisis — but CMI slowly drifted down. Documentation review got lighter, coders relied more on discharge summaries and less on consults, and secondary condition capture became more inconsistent, especially for renal, nutrition, and respiratory comorbidities. Nothing screamed “problem” until finance noticed the CMI trend, and by then everyone was looking backward to explain a pattern that had already been happening for months. That’s a classic case of silent erosion, and it’s why completeness must be measured alongside accuracy — accuracy alone can’t warn you early enough.

Laura Legg: Our second clinical story involves a dialysis-dependent patient with chronic kidney disease and malnutrition documented in a consult. Over time, reduced documentation meant secondary conditions weren’t consistently captured during coding. Coding accuracy remained stable, but query rates slowly declined and secondary diagnosis capture became more inconsistent. The hidden impact, despite stable accuracy scores, was a decline in MCC capture leading to measurable CMI erosion. The problem only became visible when finance identified a drop in CMI — another classic case of silent erosion.

The root cause wasn’t that coders forgot to code MCCs — it was that the system incentivized speed without protecting review depth. There was no dashboard tracking MCC capture by service line, and coder audits only focused on what was coded, not what was missing, so the organization had no early warning system. Because the quality program wasn’t measuring omissions, feedback to coders focused on what they did wrong rather than what they may have missed. That created a false sense of stability, because the error rate stayed acceptable while omissions grew. That’s another governance issue: you get what you measure, and if you only measure accuracy, you only train for accuracy.

Once the team quantified the impact, the conversation changed. A 0.08 CMI decline across a high-volume system isn’t a rounding error — it’s a material financial event. They added MCC capture trending and a missed opportunity audit category, recalibrated productivity expectations to protect documentation review discipline, and standardized which documents must be reviewed rather than relying on spot checks. Because they could now see completeness in real time, leadership could intervene early instead of explaining drift after the fact. The outcome wasn’t just more MCCs — it was stability, predictability, and confidence. That’s what maturity delivers.

Laura Legg: Our last clinical story is about recurrence. A surgical DRG sequencing issue appeared during audit findings and was addressed through coder education, but the same error kept recurring. The underlying problem wasn’t knowledge — it was the absence of structured repeat-error monitoring and an escalation control. It became a continuing problem that education alone wasn’t fixing, which meant the organization needed to move it into a governance control.

The root cause was simple: there was no mechanism to prevent recurrence. No dashboard tracked repeat errors by DRG, coder, or facility. No thresholds defined when a repeat error becomes a performance issue. No time standard for remediation, so feedback came late. Education was treated as the solution, but education is not a control — when a problem repeats, you need monitoring plus an accountability architecture, or you’ll keep educating the same error over and over.

The governance shift: the team redesigned the control system. They started tracking DRG and coder groups, applied a zero-repeat expectation to high-risk DRGs, set a three-day remediation standard, and defined an escalation trigger. The tone changed too — from “we’re pressuring coders” to “we’re stabilizing the system.” Coders knew expectations, managers had clear levers, and auditors saw improved follow-through on recurrence. Errors dropped, and audit burden decreased because fewer issues kept returning. That’s the payoff of maturity: less rework, fewer escalations, and a calmer overall operation.

[Polling question: attendees were asked whether their organization tracks repeated coding errors. 37.7% track them formally through dashboards or reports, 31% discuss them informally in meetings or education sessions, 8.2% don’t formally track them, and 23% were unsure.]

If you track it, you control it. If you don’t, you’ll reteach the same lessons endlessly. Tracking repeat errors also has a cultural benefit — it removes ambiguity and creates fairness, because expectations are consistent and visible.

[Polling question: attendees were asked how their organization handles high-risk DRG review before billing. Risk-based sampling led at 37%, followed by 100% review of select high-risk DRGs at 33%, review only after a denial occurs at 15.7%, and no structured process at 13.7%.]

This poll separates reactive systems from proactive ones — if most answers land on “only after a denial” or “no structured process,” the organization is reacting after the financial impact occurs. You don’t need 100% review of everything; you need a structured review where risk concentrates. Laura noted that her own organization saw real gains after starting a pre-bill review of select high-risk DRGs.

Laura Legg: When governance is immature, the financial impact shows up in predictable places. CMI drifts down when completeness erodes, compressing reimbursement and creating budget variance. Validation denials rise when defensibility is weak, creating write-offs and constant appeal work. The operational cost of appeals — physician engagement, audit resources — is significant time spent generating no new value. And severity drift affects public reporting and internal quality dashboards, creating reputational risk and leadership anxiety. So the financial story isn’t “coding needs more resources” — the story is governance. Maturity reduces volatility, and that’s why executives support it when it’s framed correctly.

What do CFOs expect from HIM leaders now, in 2026? Most don’t want a perfect accuracy percentage — they want predictability. Are we going to get hit with a lot of denials? Is our case mix index stable? Are we exposed in high-risk categories, and if so, do we have a plan? Is it disciplined, and is it measurable? When you present to finance, lead with protection: validation trends, defensibility controls, completeness monitoring, and recurrence reduction, with accuracy included as a foundation metric — not the headline. That’s how HIM gains executive influence: by speaking in risk, stability, and financial impact.

A maturity dashboard is different from a reporting dashboard. A reporting dashboard shows metrics; a maturity dashboard shows signals plus actions — where risk is rising, who owns the control, and whether the control is working. Include accuracy, but also MCC capture trending, query rate stability, validation denial rate, repeat error index, and high-risk DRG monitoring — then connect each signal to an action plan and an owner. That’s how you prevent the common leadership failure of watching a problem for months and not acting until it becomes a crisis.

Building governance in 90 days has to be realistic and targeted. Days 1 through 30: establish visibility — add completeness trending, define high-risk DRGs, and connect denial data to quality review. Days 31 through 60: align CDI and coding criteria, implement a defensibility checklist, and start risk-based pre-bill sampling. Days 61 through 90: implement repeat error tracking and escalation, and create an executive reporting cadence focused on stability metrics. The goal isn’t to fix everything at once — it’s to stabilize the high-risk areas and build a system that can scale.

Year two is about sophistication. Once the basics are stable, you can add scoring — defensibility indexes for high-risk diagnoses, automated flags for pattern review, and predictive analytics to identify where exposure is likely to appear next. You can also formalize governance forums, integrate CDI, coding, and denial review with clear action and ownership, and begin reducing manual audit burden by focusing human review where it adds value. Maturity is progressive — it’s not a one-time project, it’s an operating model that moves from compliance to monitoring to revenue integrity.

That’s the headline shift: from compliance to revenue integrity. Compliance says we correct errors after they occur. Revenue integrity says we prevent predictable errors and protect stability. When governance is integrated — completeness, defensibility, severity integrity, denial analytics, and recurrence controls — you move from surprise to predictability. That’s what leadership feels, that’s what clients feel, and that’s what payers respond to. It doesn’t require 100% audit; it requires smart control and discipline. Accuracy is the floor, as we’ve discussed, and governance maturity is the differentiator.

Let me close with a question you can take back to your leadership team: are we measuring performance, or are we protecting revenue? If your program can only answer with accuracy, you have an opportunity. A mature model lets you answer with stability — completeness is monitored, defensibility is standardized, severity integrity is trended, denials are integrated, and recurrence is controlled. Accuracy is still the floor — we never abandon it — but maturity is the differentiator that determines whether your organization is calm and predictable or constantly reacting.

I want to thank all of you for your time today. I’d love to hear what maturity level you think your organization is in, and what your next move will be. Let’s turn it over now to our moderator for Q&A.

Ella Jeffries: All right, thank you, Laura, for a wonderful discussion today. We will now begin today’s Q&A. Audience, please feel free to submit your questions via the Q&A box on your screen. With that, our first question: can you comment on the recent settlement by the DOJ against Kaiser and a health plan for allegations that they violated the False Claims Act by submitting invalid diagnosis codes for their Medicare Advantage plan in order to receive higher payments from the government?

Q: Can you comment on the recent DOJ settlement involving invalid diagnosis codes submitted for a Medicare Advantage plan?

Laura Legg: I don’t have a lot to comment on regarding that specific case — it’s a situation we’ve all seen and heard a lot about, and one some of us may have been involved in adjacent to. It’s a difficult situation, and invalid diagnosis codes are a big issue in the risk adjustment arena especially. But on that particular case, I don’t have information I can comment on. Thank you for asking, though — I do appreciate it.

Q: Many HIM directors focus too much on DNFB and won’t hold high-risk claims for a second-level review prior to billing. How do you get HIM leaders on board with a secondary review?

Laura Legg: That’s true, and if you can architect that hold to be as short as possible, there will be great dividends from that pre-bill review on high-risk diagnoses. Hopefully you have data on how many of those high-risk diagnoses are generating denials — that’s a good way to build a predictive case for what you can improve over 90 days. I feel your pain: holding DNFB is not a popular thing to do.

Q: How do we deal with situations where a payer has more stringent criteria for a diagnosis than what’s found in general coding guidelines?

Laura Legg: That’s a good question — I’ve had that situation myself. I did a lot of appeal work during the RAC era, and what I would do is write an appeal letter that included specific places in the chart where I found the supporting documentation, along with examples from Coding Clinic and other guideline sources. There may also need to be some negotiation with payers about their guidelines: if they’re provided to you, you can take them to places like your medical staff for discussion about how to implement them, or push back on overly stringent payer guidelines through contract negotiations. I hope that helps.

Q: How would you structure a work group to address better governance controls for inpatient surgical coding? What positions and people should be involved?

Laura Legg: For inpatient work, I’d involve my inpatient coding leaders, my CDI leaders, my quality leaders, and someone to represent denials — and I would not forget to have compliance present as well. That group should focus on putting a plan together that includes all of those departments.

Q: What are the reasons for increased denials?

Laura Legg: One of the biggest drivers is the speed at which payers can now re-evaluate charts — their algorithms can scan thousands of charts in minutes, flagging claims that look high-risk or where a code requirement isn’t matched by another code in your coding summary. If you can run that same kind of check on your own claims pre-bill, that helps. The other driver is slower documentation review than we’d like, which causes us to miss diagnoses we should be capturing. The whole playing field has changed — I’ve been coding for many years, and coding and preventing denials hasn’t gotten easier, it’s gotten more complex and more difficult. Between payers’ ability to scrutinize claims so rapidly, ever-changing coding rules, and the growing volume of documentation inpatient coders have to review and apply critical thinking to, that’s what accounts for the increase.

Q: What are your thoughts on autonomous medical coding?

Laura Legg: I have a lot of thoughts on that — I recently attended an AHIMA virtual summit on AI coding as well. I think we’re going to keep seeing it piloted, and likely soon, on the outpatient side, if it isn’t already being piloted in your organization; I’ve seen it piloted in many others, and I think it will be useful and helpful for repeat CPT codes. I have a different view on the inpatient side, though — I have yet to see any AI that can reliably and consistently evaluate inpatient documentation and assign the correct codes without a coder doing the critical thinking. That’s where I stand.

Q: Can you provide more information about how you get claims coded so quickly? Is there technology used for that?

Laura Legg: I’m not using a lot of technology for that at this time — I wish I was. That’s something I want to grow in, and I think we all can. We are going to see increased use of AI for flagging claims for us going forward.

Q: What’s one signal you’re seeing today that doesn’t match your reported coding accuracy?

Laura Legg: I’ll sometimes see a coding accuracy report showing something like 97%, but when I start digging into which DRGs, whether it was a diagnosis issue, a procedure issue, or a sequencing issue, I keep asking questions. Those signals are the answers to those questions — are we performing well on PCS? If not, is it body part, is it root operation? The signals are there if you dig deep enough to find them. That’s a good question, thank you for that.

Ella Jeffries: Well, thank you — that was all the time we have for today. I want to thank Laura for an excellent presentation, as well as Coronis Health for sponsoring today’s webinar. Thank you for joining us today, and we hope you have a wonderful rest of your day. Thank you, everyone.