CO-97 reads: “The benefit for this service is included in the payment/allowance for another service or procedure that has already been adjudicated.”

In plain terms: the payer decided you already got paid for this. Not that the service was unnecessary. Not that documentation was missing. That it was part of something else on the claim.

That distinction matters, because CO-97 is the denial most practices write off automatically — and a meaningful share of CO-97s are wrong.

What actually triggers CO-97

1. A same-day E/M billed alongside a procedure, without modifier 25. The most common trigger by a wide margin. Every procedure code carries a built-in evaluation component. Bill an office visit the same day as a procedure and the payer bundles the visit into the procedure unless modifier 25 tells it not to.

2. Modifier 25 was appended, but the documentation doesn’t carry it. This is the expensive version. The modifier passed the edit, the payer paid, and a post-payment audit reversed it. Modifier 25 requires a significant, separately identifiable E/M service. Two notes about the same problem, one of which happens to precede a procedure, is not separately identifiable — regardless of how much was written.

3. NCCI procedure-to-procedure edits. Two procedure codes where one is a component of the other. Some pairs allow a modifier to override; some are hard edits that never unbundle. Appealing a hard-edit pair wastes the appeal.

4. Global surgical period. Post-operative visits inside the global window are already paid inside the surgical fee. These CO-97s are usually correct — unless the visit was unrelated to the surgery, which is what modifier 24 exists for.

The part most practices get backwards

CO-97 is not automatically a write-off. Two of the four scenarios above are legitimately appealable: modifier 25 omitted in error where documentation does support a separately identifiable service (a corrected claim, not an appeal), and NCCI pairs that permit a modifier override where the clinical circumstance justifies it.

The other two — hard NCCI edits and true global-period visits — are correct denials and should be written off promptly rather than worked repeatedly.

The failure pattern we see most often is a practice that treats all four the same way. Either everything gets written off, and the appealable share is donated to the payer, or everything gets appealed, and staff time burns on edits that will never pay.

Why the volume is invisible

CO-97 shows in the denial report, so it looks like a managed problem. But the related loss usually isn’t in the denial data at all.

When a practice takes enough CO-97 hits on same-day E/M claims, the common response is to quietly stop billing the E/M. Staff learn that “those get denied” and stop creating the claim. From that point the revenue never appears anywhere: not as a denial, not as a write-off, not as a variance. It becomes an absence.

A claim nobody submits cannot appear in any report you run. That is the actual size of the CO-97 problem, and it is invariably larger than the denial line.

Run this on your own remits

Three checks. About 30 minutes with a report writer.

1. Sort 12 months of CO-97 by CPT pair. Group by the two codes involved rather than by payer or date. A handful of pairs will account for most of the volume, and each one has a single root cause — a coding habit, a template, or a scheduling pattern.

2. Pull the CO-97s where modifier 25 was present and the claim still denied. Read five of those notes. If the E/M documentation restates the procedure’s indication rather than addressing a separate concern, the modifier is being applied as a billing reflex instead of a clinical fact. That is an audit exposure, not just a denial.

3. Count same-day E/M claims per provider, per month, over 24 months. Look for a provider whose count drops and never recovers. That is the invisible loss above — the point where someone stopped billing rather than stopped being denied. Compare against a peer with a similar panel; the gap is the annualized number.

Find something and want the fix? Reply and say so. Find nothing? That is a genuinely good outcome, and worth knowing for certain.

When CO-97 is not the real problem

Check the accompanying RARC before working any CO-97. CARC codes carry a remittance advice remark code that names the specific edit. A CO-97 paired with an N-series remark pointing at an NCCI edit is a different problem, with a different fix, than a CO-97 with no remark at all. Working the CARC without reading the RARC is the most common reason a correctly appealable denial gets abandoned.