Your billing team hands you a denial report every month. You scan the denial rate, note whether it moved, and approve the rework queue. That report feels like the scoreboard for billing accuracy in your practice. It isn’t. It’s one line item in a much larger bill, and it happens to be the smallest one.
The problem with billing errors isn’t that practices ignore them. It’s that the errors a practice can see are structurally different from the errors that cost the most. A denied claim announces itself. It generates a remittance code, enters a work queue, and shows up in a report. An undercoded claim that gets paid on the first pass generates nothing at all — no alert, no queue, no line on any dashboard you currently review.
This article gives you a four-layer cost model and a calculation you can run this month using data already sitting in your practice management system. By the end, you’ll have a defensible annual dollar figure for what billing errors cost your specific practice — not a national average, and not a vendor’s estimate.
The Denial Report Shows Maybe a Third of the Damage
A three-physician primary care group came to us convinced their billing was in good shape. Clean claim rate sat at 94%. Denial rate hovered near 9%, comfortably inside what their EHR vendor called normal. Their office manager reworked denials within four days on average, which is genuinely good performance.
Then we sampled thirty charts against submitted claims.
Eleven of them documented a higher level of service than what was billed. Not fraud, not aggressive coding — documentation that plainly supported a 99214 submitted as a 99213, mostly because the provider’s note landed after the claim went out and nobody reconciled the two. Those eleven claims were paid. Cleanly. On the first submission. They appeared in every report as successes.
That’s the shape of the problem. Practices measure billing accuracy using denial data, and denial data is blind to the single largest category of billing error by dollar value.
What Medical Billing Errors Actually Cost a Practice: The Four Layers
Billing error cost splits into four layers. The first is visible in your existing reports. The other three are not, which is precisely why they persist.
Layer 1 — Rework Labor
Every denial that re-enters your workflow consumes staff time for investigation, correction, payer contact, and resubmission. MGMA data places the administrative cost of reworking a single denied claim between $25 and $118 depending on complexity. This is the only layer most practices can quote a number for.
Layer 2 — Permanent Write-Off
A meaningful share of denials never get reworked at all. Industry reporting puts the never-resubmitted rate somewhere between 35% and 60%, driven by timely filing windows closing, low-dollar claims judged not worth the labor, or denials that simply age out of the queue. Every one of those is earned revenue converted to zero — and according to AHIMA analysis of denial prevention, the overwhelming majority of denials trace back to administrative or process causes rather than genuine clinical disputes, which means most of this layer was avoidable.
Layer 3 — Silent Undercoding
Services performed and documented but billed at a lower level than the record supports. These claims pay cleanly, so they never generate a denial, never enter a work queue, and never appear in a denial report. The revenue difference vanishes without leaving evidence.
Layer 4 — Carrying Cost of Delayed Cash
Error-driven denials extend A/R days. Money you’ve earned sits with a payer instead of in your operating account, and that delay has a real cost — whether you’re carrying a line of credit or simply forgoing what that cash could do inside the practice.
Why a Healthy Clean Claim Rate Can Sit on Top of a Real Leak
Clean claim rate answers one question: did the payer accept this claim without kicking it back? That’s a formatting and eligibility test. It says nothing about whether the claim captured the full value of the service delivered.
A claim can be perfectly clean and still be wrong by fifty dollars.
This is where a lot of practice managers get misled, and it isn’t their fault. Every tool in the standard billing stack is built to catch the errors that cause rejection, because rejection is what payers penalize. Claim scrubbers check field completeness and code compatibility. Clearinghouse edits flag mismatches. None of that machinery examines whether the code you submitted matches the note the provider wrote. That comparison requires someone to read the documentation, and no scrubber does that.
Most practices are running a quality system optimized for payer acceptance while assuming it’s optimized for revenue capture. Those are different objectives.
The Layer Almost Nobody Measures
Here’s what we’ve found consistently across independent practices that ask us to look at their billing: undercoding produces more annual revenue loss than denials do, and it’s usually the layer with zero assigned ownership.
The reason is structural rather than behavioral. Denials come with a feedback loop built in — the payer tells you something is wrong, someone gets assigned to fix it, and the fix is measurable. Undercoding has no feedback loop whatsoever. Nobody tells you. The claim pays. The report looks fine. There is no signal to respond to, so no process forms around it.
Two patterns drive most of it. The first is documentation timing: the claim goes out before the provider closes the note, so the biller codes from an incomplete record and never revisits it after the note is finalized. The second is defensive downcoding, where a biller who has been burned by an audit develops a habit of coding one level conservative on anything ambiguous. That habit is rational for the individual and expensive for the practice. It also compounds quietly, because nothing ever corrects it.
Neither pattern shows up in accuracy metrics. Both show up in the annual number. If your practice is working toward the 95% coding accuracy benchmark, it’s worth knowing that a practice can hit that benchmark and still be losing five figures a year to conservative coding, because accuracy measures whether a code is defensible — not whether it’s complete.
How to Calculate Your Practice’s Annual Billing Error Cost
Set aside one afternoon. You need your last twelve months of claims data and a thirty-chart sample. Every figure below is illustrative — replace each one with your own.
Step 1: Pull Your Denial Volume and Rework Cost
Take total claims submitted per month and your initial denial rate. Multiply. Then multiply by a blended rework cost — start at $48 unless you’ve timed your own staff, in which case use your real loaded hourly rate times average minutes per denial.
Worked example: 950 claims monthly at an 11.8% denial rate gives 112 denials. At $48 each, Layer 1 costs $5,376 per month, or roughly $64,500 annually.
Step 2: Isolate What Never Gets Reworked
Run an aging report on denials older than ninety days with no resubmission. Divide by total denials to get your true abandonment rate — this is almost always higher than what your team estimates. Multiply abandoned claims by your average claim value.
Worked example: 35% of 112 monthly denials abandoned equals 39 claims. At an average claim value of $142, Layer 2 costs about $66,500 annually.
Step 3: Audit Thirty Charts for Undercoding
Pull thirty recent paid encounters at random. Have a certified coder compare the documentation against the code submitted. Count how many were billed below what the note supports and calculate the average dollar differential.
Worked example: 12% of charts undercoded at an average $38 differential, across 950 monthly claims, puts Layer 3 near $52,000 annually.
Step 4: Price Your Delayed Cash
Subtract a 32-day A/R benchmark from your actual A/R days. Multiply the excess by your average daily collections, then by your cost of capital.
Worked example: 15 excess A/R days against $4,932 daily collections ties up roughly $74,000; at 8%, Layer 4 costs about $5,900 annually.
Step 5: Read the Distribution, Not Just the Total
Add the layers. For the example practice, the total lands near $188,900 against $1.8 million in collections — about 10.5% of revenue.
The total matters less than the split. Layer 1, the only layer this practice was actively managing, accounts for roughly a third of the loss. Two-thirds of the money was leaking through channels that generated no report, no alert, and no work queue. That distribution is what should drive your next decision.
Four Expensive Mistakes Practices Make Trying to Fix This
Buying a better claim scrubber and considering it handled. Scrubbers reduce Layer 1 and Layer 2. They have no visibility into Layer 3 whatsoever, because they never read the clinical note. Practices routinely spend on scrubbing technology while the largest layer goes untouched.
Fixing the most frequent error instead of the most expensive one. Demographic entry errors are usually the highest-volume denial category and among the cheapest to rework. Prior authorization and medical necessity denials are far less frequent and carry much higher claim values. Sort your denial log by dollars, never by count.
Treating this as a performance problem. When a practice reads its error cost as evidence that the billing staff is underperforming, the usual response is pressure — and pressure produces more defensive downcoding, which increases Layer 3. Billing error cost is a workflow measurement, not a personnel evaluation. It’s also worth noting that turnover in front-office and billing roles is itself a driver of error rates, which is one reason the true cost of front desk staff turnover and billing error cost tend to move together.
Running the audit once. A thirty-chart sample gives you a snapshot. Undercoding patterns drift with staff changes, payer policy updates, and new service lines. Practices that audit annually catch the drift a year late.
Turning Error Cost Into a Standing Operating Metric
Once you have a baseline, the version of this that actually changes outcomes is a recurring measurement rather than a one-time project.
Track error cost per encounter — total four-layer cost divided by total encounters — and review it quarterly alongside your collections figures. It’s a single number that moves when your workflow changes, and it’s specific enough that you can attribute movement to a cause.
Segment it by payer and by service line. Error cost is almost never distributed evenly. One payer’s authorization requirements or one procedure’s modifier rules will typically account for a disproportionate share, and that concentration tells you exactly where to put attention first. Practices that already track billing analytics and RCM KPIs can fold error cost per encounter into an existing review rhythm without building anything new.
Then close the documentation timing gap. Setting a rule that claims don’t transmit until the note is closed removes the largest single input to Layer 3, and it costs nothing beyond workflow discipline.
For practices where the audit reveals the loss is concentrated in undercoding and unworked denials rather than in claim formatting, the capacity problem is usually specific: nobody owns chart-to-claim reconciliation, because it never fits between everything else the billing staff already handles. That’s the work a dedicated medical billing virtual assistant is structured around — consistent claim review against documentation, denial follow-through past the ninety-day mark, and pattern reporting on recurring causes.
What Your Number Should Change
The point of running this calculation isn’t the total. It’s the distribution across the four layers, because that distribution tells you what kind of problem you actually have.
If most of your loss sits in Layer 1, you have a claim quality problem and better front-end scrubbing will move it. If it sits in Layer 2, you have a follow-through capacity problem — denials are being worked until the queue gets busy, then abandoned. If it sits in Layer 3, you have a reconciliation gap, and no amount of denial management will touch it. Practices that skip the calculation almost always assume they have the first problem, and most of them have the third.
Run the thirty-chart audit before you buy anything, hire anyone, or change a workflow. It takes an afternoon and it will tell you more about your revenue cycle than a year of denial reports.
Frequently Asked Questions
These are the questions practice managers raise most often once they start sizing their own error cost.
How much do billing errors typically cost a medical practice each year?
Published estimates of revenue lost to billing errors and leakage generally range from 4% to 15% of collections, with the variation driven by specialty, payer mix, and how the loss is defined. That range is too wide to plan against, which is why running the four-layer calculation on your own claims data produces a far more useful number than any published benchmark.
What does it cost to rework a single denied claim?
MGMA data places the administrative cost between $25 and $118 per claim, with complexity being the main variable — a demographic correction sits at the low end while a medical necessity appeal requiring documentation sits at the high end. Timing your own staff on a sample of denials gives you a practice-specific figure worth substituting in.
Is undercoding actually more expensive than claim denials?
For many independent practices, yes. Denials are recoverable in large part and generate a visible correction process, while undercoded claims pay cleanly and leave no trace. The only reliable way to size it is a documentation-versus-code audit on a random chart sample, since it appears in no standard report.
Can a claim scrubber prevent most billing errors?
Scrubbers are effective against format errors, code compatibility issues, and eligibility mismatches — the categories that cause outright rejection. They cannot detect undercoding, because that requires comparing the submitted code against the clinical note. Expect a scrubber to reduce two of the four layers and leave the largest one untouched.
How often should a practice audit for billing errors?
A thirty-chart sample each quarter is enough to detect drift without creating a burden. Practices that audit only once a year tend to discover coding pattern changes long after the revenue is unrecoverable, particularly following staff turnover or a payer policy update.
Should we fix denials or undercoding first?
Sort by dollars rather than by volume and address whichever layer is larger in your own numbers. If your abandonment rate on denials is above 30%, that’s usually the faster recovery, since those claims are already identified and some remain inside appeal windows. Undercoding takes longer to correct but produces a permanent structural gain once the reconciliation step exists.
Ready to Find Your Number?
Most practices we work with are surprised less by the size of their billing error cost than by where it’s concentrated. The denial report was never going to show them, and that’s not a failure of their team — it’s a limitation of the instrument.
If you’d like a second set of eyes on your claims data and a chart sample reviewed by a certified coder, book a free 15-minute consultation with the Care VMA Health team. We’ll walk through the four-layer calculation on your actual numbers and show you which layer is costing you most.
