It usually surfaces the same way. The denial report lands on the manager's desk, and one line keeps repeating: a lab test the practice ordered, performed and documented has been denied as a duplicate. Someone on the team says the test really was done twice. Someone else says that is exactly the problem. Both can be right — and until the practice can tell which situation it is looking at, the same denial comes back next month wearing a different patient's name.
On October 1, 2026, CMS used its MLN Connects newsletter to remind practices of the rule behind those denials: Medicare doesn't pay for duplicate laboratory tests unless they're medically necessary. The reminder was short. The confusion it points at isn't. In professional coding forums, billers have argued for years about when a repeated test earns its own claim line and which modifier proves it — and the honest answer is that three different problems hide inside the single word “duplicate.” Sorting them out is the whole job.
First, the words — because the denial uses them loosely
A duplicate test, in Medicare's usage here, is a second test that adds nothing the first one didn't already provide: the same test again without a medical reason, or a test that was already included inside a panel ordered the same day. A panel is a single code that bundles a defined group of tests — order the panel, and by definition you've already ordered everything inside it.
A repeat test is different. It is the same test performed again later the same day because the clinician needs a new result to manage the patient — a second measurement, taken at a second moment, that can change a decision. Think of it the way you think about vital signs. Reprinting the morning's blood pressure reading isn't a new measurement. Taking the pressure again after treatment, to see whether it moved, is. Medicare's repeat-test rule works on the same logic: a new result, at a new time, for a clinical reason.
Modifier 91 is the flag that tells Medicare a claim line is that second kind of event. Appended to a laboratory code, it identifies a repeat clinical diagnostic laboratory test — the same test, repeated the same day, to obtain multiple results in the course of treatment. It's a statement, not a key. It says “this was a genuine repeat,” and the record has to back that up.
Problem one: the test was already inside the panel
The most mechanical of the three problems needs no modifier at all, because there's nothing to defend. CMS points practices to section 90.5 of the Medicare Claims Processing Manual, Chapter 16: when you order a mix of panels and individual tests, review what each panel already contains, and don't order individual tests that duplicate tests inside the panel.
Clinically, this is ordering a basic metabolic panel and then ordering a potassium from the same draw as if it were a separate event. The potassium was never missing. It was inside the panel from the start. When both lines reach Medicare, the individual test reads as a duplicate — because, to the panel's definition, it is one. The fix lives in the order set, not in the appeal: compare standing orders and favorite lists against panel contents, and remove the individual tests that the panel already covers.
Problem two: the rerun
The second problem is a test repeated because the first attempt did not produce a usable result — a specimen problem, an equipment problem, or repeated simply to confirm what the first result already showed. The work may genuinely have happened twice. It's still not a billable repeat.
The Medicare Administrative Contractor billing and coding article that CMS directs practices to — on repeat or duplicate services on the same day — is explicit about modifier 91's limits: not for a rerun performed to confirm the initial results, not for a problem with the specimen or the equipment, and not when another code already describes a series of results. A rerun fixes the laboratory's process. It doesn't give the clinician new information, and Medicare doesn't pay for it as a second test. Appending modifier 91 to push a rerun past a duplicate edit doesn't cure the problem. It converts a denial into a documentation risk, because the claim now asserts a medically necessary repeat that the record doesn't describe.
Problem three: the real repeat — the one case modifier 91 exists for
The third situation is the legitimate one. A patient's condition is evolving during the day, the clinician repeats the same test hours later to obtain a subsequent result, and that result feeds a treatment decision. CMS's instruction is direct: if repeat tests are medically necessary, report them with CPT modifier 91.
Notice what carries the weight in that sentence. Not the modifier — the medical necessity, and the record that shows it. A defensible repeat shows why the test was needed again, and when each test was performed, so a reviewer can see two clinical moments rather than one test billed twice. CMS also gives the boundary in the other direction: the Medicare Administrative Contractor will deny duplicate tests that are not medically necessary (Manual Chapter 16, section 90.3), and its example is deliberately ordinary. Medicare will not pay for a second vitamin D test unless it is medically necessary. A test with a long clinical memory, repeated without a documented reason, is where that denial concentrates.
Why CMS is raising this now
The October reminder did not arrive alone. CMS also noted that its calendar year 2027 Physician Fee Schedule proposed rule included a request for information on duplicate laboratory testing, imaging, and how results are shared between systems — and that the agency is exploring options to address these issues in future rules. Read that at its exact weight: it is a request for input and a signal of attention, not a new rule, and nothing in it changes how a claim should be billed today. What it tells a practice manager is simpler. Duplicate testing is on the agency's list. Order sets and repeat-test habits that have never been audited are worth auditing while the standard is still the familiar one.
What to check on Monday
- Pull your duplicate denials before anything else. Take the last 90 days of laboratory denials and sort them by patient, date of service and code. The pattern matters more than any single denial: the same code pairs appearing again and again point to an order-set problem, while scattered repeats point to documentation.
- Audit order sets against panel contents. For every standing order set or favorite list your clinicians use, list the panel codes and remove individual tests that duplicate a component of a panel ordered with it (Manual Chapter 16, section 90.5).
- Separate your reruns from your repeats. If a second run happened to confirm a result or because of a specimen or equipment failure, it's not a modifier 91 case. Make sure nobody on the team is treating the modifier as the tool that gets a duplicate past an edit.
- Check the record behind every billed repeat. For each same-day repeat billed with modifier 91, the documentation should show the clinical reason for repeating the test and the time of each performance. If the note cannot show two distinct clinical moments, the claim is exposed even if the care was real.
- Look at your repeats of slow-moving tests. Before a test such as vitamin D is reordered, check whether a recent result already exists and document why a new one is needed. That is CMS's own example of where medical necessity gets questioned.
- Keep other payers separate. This article describes Medicare's framework. Commercial and Medicaid plans run their own duplicate and repeat rules; verify each major payer's policy before applying this workflow to its claims.
The bottom line
A denied second test is rarely one problem. Sometimes the test was already inside the panel. Sometimes it was a rerun that should never have been billed as a repeat. And sometimes it was a genuine, necessary repeat whose record simply never said so. Each of those has a different fix, in a different place — the order set, the laboratory workflow, the note. Our reading at Healthcode RCM is that practices lose this money quietly because all three arrive under one denial label and get worked the same way. They should not be.
Healthcode RCM works with small, independent practices on exactly this layer: denial-pattern reviews that sort repeat denials by cause, and claim-workflow checks that compare what was ordered, what was documented and what was billed. If your duplicate lab denials have never been sorted, that sorting is the right first step.
Tell a colleague this: a repeat test earns its own line only when it produces new clinical information — if it only repeats the first result, Medicare sees a duplicate, not a service.
Sources
• Centers for Medicare & Medicaid Services. Medicare Claims Processing Manual, Chapter 16 — Laboratory Services (sections 90.3 and 90.5, as cited by CMS in the October 1, 2026 MLN Connects newsletter).
• Medicare Administrative Contractor billing and coding article, “Repeat or Duplicate Services on Same Day” — linked by CMS from the October 1, 2026 MLN Connects newsletter; states the limits on modifier 91 summarized in this article.
* Note on status: the CY 2027 Physician Fee Schedule rule is a proposed rule. Its request for information on duplicate testing is an exploration for possible future rulemaking, not a current billing requirement.
Loading approved comments…