Unbundling: Reporting the Parts Instead of the Whole
Unbundling is reporting the parts of a service separately when the code set already has a code for the whole. The reason it persists in practices with clean edit reports is not subtlety — it is that the rule and the machinery that enforces the rule are not the same size, and CMS says so directly: the edits do not cover every type of unbundling that exists, and the obligation to code correctly applies whether an edit is watching or not.
Updated 12 min read
On this page
Key takeaways
- The governing rule is short: report the most comprehensive code that describes the services performed, and do not report the components of a service separately from it.
- The correct-coding manual states that its edits and policies do not include all possible combinations or types of unbundling, and that providers are obligated to code correctly even where no edit exists to stop them.
- So a claim that cleared every edit has not been validated. It has been unopposed, and those are different findings.
- The manual names four forms, and only two of them resemble the component-of-a-comprehensive-service pattern most people picture.
- One of the four is not a code pair at all — splitting a bilateral code into two unilateral reports — which a procedure-to-procedure edit is structurally unable to catch, because there is no pair to test.
- The manual rules out both disguises of that split at once: two units of the unilateral code, and the unilateral code on two lines with the side modifiers.
- The same passage names the route for a practice that finds its own error — contact the contractor about potential payment adjustments — which is worth knowing before the question becomes urgent.
The rule is about the code set, not about the edits
The correct-coding manual states the obligation in two sentences and does not qualify them: procedures are to be reported with the most comprehensive code that describes the services performed, and the services a code describes must not be unbundled. Everything else in this subject is application.
Notice what the rule refers to. It refers to what the code describes — a fact about the code set — and not to whether any particular combination appears in an edit table. The edit tables are an enforcement mechanism built on top of the rule, and like any enforcement mechanism they are narrower than the thing they enforce. NCCI procedure-to-procedure edits covers what an individual edit asserts and how to read one; this article is about the obligation that exists whether an edit does or not.
Two questions that are easy to merge and should not be
Four forms, and two of them do not look like unbundling
The manual enumerates rather than defines, which is more useful. Its examples fall into four shapes, and the last two are the ones that survive review because they do not match the mental image of a component billed separately from its whole.
| Form | What was reported | Why it survives |
|---|---|---|
| Reporting parts where a combined code exists | Two or more codes describing services that one combined code already describes in full. | It requires knowing the combined code exists. A coder confident in each individual code can be wrong about the set. |
| Fragmenting one procedure into components | The base version of a procedure plus a separate code for a step performed within it. | Each code is individually accurate about something that happened, which is exactly why the claim reads as well documented. |
| Splitting a bilateral code into two unilateral reports | The unilateral code with two units, or the unilateral code on two lines carrying the side modifiers, where a bilateral code exists. | There is no code pair, so a procedure-to-procedure edit has nothing to test. It presents as a units question or as nothing at all. |
| Reporting a service integral to a larger one | A code for something the larger procedure necessarily required — the manual's example is surgical access. | Integral steps are real work, genuinely performed, and often separately documented. The test is not whether it happened. |
The manual also carries a specific rule for biopsies, which sits alongside these: a biopsy is separately reportable only where the pathologic examination results in a decision to immediately proceed with a more extensive procedure on the same lesion, or where it was performed on a separate lesion.
The bilateral split is ruled out in both of its disguises
The absence of an edit is not permission
This is the part of the subject with the clearest statement behind it and the least attention paid to it. The manual's own introduction says that its edits and policies do not include all possible combinations of correct coding edits or types of unbundling that exist, and that providers are obligated to code correctly even if edits do not exist to prevent use of an inappropriate code combination.
Read that against how a billing operation actually works and the consequence is sharp. A claim moves through a scrubber, then through a payer's edits, and arrives paid. Every checkpoint it passed was testing a subset, and none of them was ever advertised as complete. The claim has been unopposed. Treating that as validation is a reasonable inference from a system that gives no other feedback, and it is wrong for a documented reason.
What a passed edit actually tells you
That this combination is not in the tested set, or is in it with an indicator that permitted the modifier that was used. It is a statement about the tables, not about the codes.Where the standard actually lives
In what the codes describe. Which is why a practice with a recurring pattern is better served reading the code descriptors for the combination than reading the edit file for it — the file may simply not contain the pair.What the manual says to do about a discovered error
It names a route rather than leaving it open: a provider who determines they have been coding incorrectly should contact their Medicare Administrative Contractor about potential payment adjustments. Whether a particular situation calls for that, for a different disclosure route, or for advice first is a legal question and not a coding one.
Where this stops being a coding question
Three things unbundling is not
- Not an edit firing
- An edit is a control, and controls produce false positives. A pair can be flagged and the reporting still be correct, which is what the exception grounds and the correct coding modifier indicator exist for. The edit is evidence that a question exists, not an answer to it.
- Not the same as too many units
- Reporting more units of one code than the encounter supports is a different error with a different control — a medically unlikely edit tests the count on a line. Unbundling is about which codes were used. The bilateral split is the case where the two overlap, which is exactly why it is the hardest of the four to see.
- Not solved by a modifier
- A modifier asserts that the ordinary relationship between two codes does not apply on this claim — a separate encounter, a separate anatomic site. Where the services genuinely were components of one procedure, no modifier makes the report correct, and appending one converts a coding error into an unsupported assertion on a submitted claim. How a modifier changes adjudication covers why that is worse rather than better.
The check that finds the forms edits cannot
What this article deliberately does not tell you
It names no code numbers and reproduces no code descriptions. The manual's examples are built out of paired code numbers and their full published descriptions, and those descriptions are licensed by the bodies that maintain them. Each example here is therefore given structurally — the shape of the error rather than the instance — and the source is cited so the worked examples can be read where they are published.
It also makes no estimate of how much of correct coding the edit tables cover. The manual states that the coverage is incomplete and does not quantify it, and a figure invented here would be exactly the kind of confident number this subject is already full of.
Finally, it does not advise on whether to disclose a discovered pattern, or to whom. The manual names a contractor route for payment adjustments; other routes exist, they carry different consequences, and which one fits is a question for counsel with the facts in front of them rather than a rule that can be stated in advance.
Common questions
Our claims pass every edit. Doesn't that mean our coding is correct?
No, and the correct-coding manual says so in its own introduction: the edits and policies do not include all possible combinations of correct coding edits or types of unbundling that exist, and providers are obligated to code correctly even where no edit exists to prevent an inappropriate combination. A claim that cleared the scrubber and the payer's edits was unopposed, not validated — every checkpoint it passed was testing a subset that has never been advertised as complete. The standard lives in what the codes describe, so a recurring combination is better checked against the code descriptions than against the edit file, which may simply not contain the pair.
What actually counts as unbundling?
The manual enumerates four situations rather than giving a definition. Reporting two or more codes where a single comprehensive code already describes the services in full. Fragmenting one procedure into component parts — the base procedure plus a separate code for a step performed inside it. Splitting a bilateral code into two unilateral reports. And reporting a service that was integral to a more comprehensive procedure, with surgical access as the named example. It adds a specific rule for biopsies: separately reportable only where the pathologic examination leads to a decision to immediately proceed with a more extensive procedure on the same lesion, or where the biopsy was on a separate lesion.
We report the unilateral code twice with left and right modifiers instead of the bilateral code. Is that acceptable?
No — the manual rules out that exact form. Where a bilateral code exists, it forbids reporting the unilateral code with two units and forbids reporting it on two lines with the side modifiers, naming both in the same instruction. This is worth flagging because the two-lines-with-side-modifiers version is sometimes adopted as the fix for the two-units version, and it is the same error in a second costume. It is also the form least likely to be caught, because there is no code pair for a procedure-to-procedure edit to test, so it surfaces as a units question or not at all.
Is unbundling fraud?
Not by itself, and the two questions have different tests. In the correct-coding manual, unbundling is stated purely as a coding obligation — there is no intent element and no penalty attached to it there. Liability under the False Claims Act turns on a knowledge standard that reaches beyond intent, which is why the distinction matters in practice: an isolated coding error is a coding error, while a pattern that continues after somebody identified it is being evaluated under a different rule. The safe operating position is that the coding question is settled by what the codes describe, and everything downstream of that is a legal question for counsel.
Can a modifier make a bundled combination reportable?
Only where the ordinary relationship genuinely does not apply on this claim — a separate patient encounter or a separate anatomic site — and where the edit's own indicator permits a modifier at all. Where the services really were components of one procedure, no modifier makes the report correct. Appending one anyway does not neutralize the error; it adds a second one, because a modifier is an assertion about the record and an unsupported assertion on a submitted claim is a worse position than a plain coding mistake.
We found a pattern in our own coding. What does the manual say to do?
It names a route, in the same passage that says the edits are incomplete: a provider or supplier who determines that they have been coding incorrectly should contact their Medicare Administrative Contractor about potential payment adjustments. That is the manual's own instruction and it is useful to know it exists before the question becomes urgent. Whether a specific situation calls for that route, for a different disclosure route with different consequences, or for legal advice first is not a coding decision and is not answered here — it depends on scope, duration and what was known, and it is worth taking to counsel with the facts assembled.
How do we find unbundling if the edits will not show it to us?
Change what is being reviewed. A report of edits that fired can only surface the forms an edit can test, which is two of the four. The review that finds the others starts from the practice's own highest-volume code combinations and asks, for each, whether the code set already contains a code describing them together — and separately, whether any service being reported alongside a procedure was necessarily required by it. That is a coding review against the descriptors rather than a claims review against a file, and it belongs in the internal audit cycle rather than in daily claim work.
Key terms in this article
Defined once, on their own pages.
Continue learning
The control that catches some of this, the one that catches the units, and where the coding question becomes a legal one.
NCCI Procedure-to-Procedure Edits
What an individual edit asserts, the comprehensive/component misnomer, and the indicator that says whether any modifier can move it.
Medically Unlikely Edits
The control on how many units of one code a line may carry — the test the bilateral split lands in front of, if it lands anywhere.
The Bilateral Procedure Modifier
How a bilateral service is correctly reported, including when the code's own terminology already makes it bilateral.
The False Claims Act
Where the coding question becomes a legal one, and the knowledge standard that reaches conduct short of intent.
Primary sources
- CMS National Correct Coding Initiative Policy Manual for Medicare Services — Introduction (opens in a new tab)
States that the NCCI edits and policies do not include all possible combinations of correct coding edits or types of unbundling that exist; that providers and suppliers are obligated to code correctly even if edits do not exist to prevent use of an inappropriate code combination; and that a provider or supplier who determines they have been coding incorrectly should contact their Medicare Administrative Contractor about potential payment adjustments. Also records that the edits are built for services billed by a single provider or supplier for a single patient on the same date of service.
- CMS National Correct Coding Initiative Policy Manual for Medicare Services, Chapter I — General Correct Coding Policies (opens in a new tab)
Requires procedures to be reported with the most comprehensive code that describes the services performed, and prohibits unbundling the services a code describes. Enumerates the forms: reporting multiple codes where one comprehensive code describes the services; fragmenting a procedure into component parts; unbundling a bilateral procedure code into two unilateral procedure codes, ruling out both the two-units form and the two-lines-with-side-modifiers form; and unbundling services integral to a more comprehensive procedure, with surgical access as the example. Sets the separate rule for biopsies performed alongside a more extensive procedure.