Experimental or Investigational Denials: When the Payer Calls It Unproven

Plans do not pay for every service that exists; they pay for services their coverage terms recognize as established enough to be a benefit. An experimental-or-investigational denial is what the payer returns when it decides the service on the claim has not cleared that bar for the use it was billed for — that the evidence behind it, for this indication, is still developing rather than settled. The care may have been thoughtful and the patient may have benefited; none of that is the question. What is being questioned is whether the service is proven enough, under the plan's definition, to be covered at all.

On a commercial remittance the denial typically arrives as CARC 55, the reason code a payer uses when it deems a procedure, treatment, or drug experimental or investigational. Unlike the general non-covered code, CARC 55 does not require a remittance remark (a RARC) to accompany it, but a remark or the policy identifier the payer can attach is often what names the specific coverage policy behind the decision — and that policy is where the work starts. The bare code says the payer reached the conclusion; it does not say on what basis, and the basis is what an answer has to engage.

There is no single, shared definition of what makes a service experimental — which is exactly why the denial cannot be answered from a general sense of whether a treatment is “new.” The standard that governs is the one written into the plan that denied the claim, and for Medicare it is a standard expressed in different words. Confirming the denial means reading the actual definition the payer applied, not the reputation of the service.

Commercial — the plan's own definition of experimental or investigational

Private plans define “experimental or investigational” in their governing document or medical policy, and the definitions differ from one plan to the next. They commonly turn on factors a claim can be tested against — whether the service has the applicable regulatory approval for the use in question, whether it is supported by peer-reviewed published evidence, whether it is still the subject of ongoing clinical investigation, and whether it has become the recognized standard of care. Each plan maintains its own definition and revises it on its own schedule, so the only definition that matters is the current one for the plan that denied the claim, applied to the indication billed.

Medicare — the reasonable-and-necessary standard, not a separate category

Original Medicare has no statutory “experimental” exclusion for services. An unproven service is denied because it fails the standard behind a medical-necessity denial — that items and services be reasonable and necessary, set by Social Security Act §1862(a)(1)(A) (42 U.S.C. §1395y(a)(1)(A)). So on a Medicare remittance the same denial often reads as not reasonable and necessary rather than as CARC 55. The literal words “experimental or investigational” appear in Medicare regulation for devices: an investigational device the FDA has placed in its experimental category is not covered, while one in the nonexperimental/investigational category may be, along with the routine care furnished in an approved study (42 CFR §411.15(o); 42 CFR Part 405, Subpart B).

Before deciding whether to fight the denial, settle who owns the balance, because the answer is not read from the reason code. As everywhere in this cluster, it is the group code, not the CARC, that assigns responsibility: the same experimental-or-investigational denial can arrive as a contractual-obligation (CO) code, making it the practice's write-off, or as a patient-responsibility (PR) code, making it a patient balance. Which one applies is set by the plan's contract, and a signed advance financial-responsibility acknowledgment is often what a plan requires before an amount it treats as the patient's can be collected.

On Original Medicare the rule follows from the statute the denial rides on. Because an unproven service is denied as not reasonable and necessary, it falls under the limitation-on-liability rule of §1879 (42 U.S.C. §1395pp), which turns on knowledge: the beneficiary can be billed only if they knew, or could reasonably have been expected to know, that Medicare would not pay. In practice that knowledge is established by giving the patient an Advance Beneficiary Notice before the service. With a valid one on file the amount can be billed and generally posts as patient responsibility; without one, the practice absorbs it and refunds anything already collected. The notice as an instrument, and exactly where it applies, is the subject of the Advance Beneficiary Notice.

More of these denials are contestable than the other coverage denials in this cluster, because the thing in dispute — whether a service is established — is a moving judgment rather than a fixed fact. Evidence accumulates, professional guidance changes, and a use that was investigational becomes standard. So the first move is to test the label against the plan's own definition and the indication billed, rather than accept it. The claim resolves into three cases, and the response follows from which is true.

When the case is a genuine dispute — the service is established for this indication and the plan denied it anyway — the appeal is an evidentiary argument, and its shape is set by the plan's own definition rather than by conviction about the treatment. The task is to show, against the criteria the plan actually applied, that the service has cleared the bar the definition describes.

1. 01

   Read the remittance for the code and the policy it points to

   Start from the CARC and any remark or policy identifier on the remittance, and the group-code prefix. On Medicare, expect a not-reasonable-and-necessary denial rather than CARC 55, and read it the same way. The remark or the attached policy is what names the coverage policy the appeal has to engage.
2. 02

   Get the plan's own definition and the specific policy

   The plan's definition of experimental or investigational, and the medical policy it applied, are the document the appeal is measured against. On a group health plan you are entitled to it: the claims-procedure rules require a full and fair review and access, on request and free of charge, to the rule or protocol the plan relied on (29 CFR §2560.503-1). Work from the version in force on the date of service, which is not necessarily the current one.
3. 03

   Build the evidence argument, and attach it

   Assemble the approval status for this indication, the peer-reviewed published evidence, and any professional-society guidance that establishes the service as accepted for this use, and map each to the plan's own criteria. A reviewer decides on what is in front of them, so the supporting evidence goes in with the appeal rather than being described — an appeal that references literature it does not include is one the reviewer cannot act on.
4. 04

   Use the external-review layer where the internal appeal fails

   An experimental-or-investigational determination is one of the medical-judgment decisions that a plan's internal review does not get the last word on. On a non-grandfathered commercial or group plan, federal rules make such a determination eligible for external review by a reviewer the plan does not employ, who decides for themselves whether the service is experimental (45 CFR §147.136). How that layer works — which process applies and why the plan must pay in accordance with the decision — belongs to the levels of appeal, and whether an appeal is the right move at all is in appealing a denial.

That external layer is the reason an experimental-or-investigational denial is worth reading closely before it is written off: it is one of the few coverage denials where an independent decision-maker, not the plan, can settle the question, precisely because the question is one of medical judgment. What to assemble before filing, so the argument goes in complete the first time, is laid out in the denial appeal readiness checklist.

Like the other coverage denials in this cluster, this one is largely knowable before the service, because a plan's view of what it considers experimental is written down in advance. Prevention lives at the front end, where the service is planned and the coverage is checked.

The coverage policy was not checked before the service

The most preventable source. Where a service is one a plan is likely to treat as experimental, the plan's medical policy usually says so, and checking it before the service tells the practice whether it will be covered for this use, whether prior authorization or an advance notice is warranted, and what evidence an eventual appeal would need. Looking it up beats remembering what a plan covered last year.

The indication was not flagged as off-label or emerging

A service established for one condition can be experimental for the one billed, and the difference is invisible unless someone flags the indication. Identifying that a use is off-label or newly established before the claim is built is what lets a practice confirm coverage — or inform the patient — rather than discover the denial on the remittance.

A clinical-trial context was not separated out

When a service is furnished within a clinical trial, the investigational item and the routine costs of participation are adjudicated differently, and billing them as one invites a blanket denial on costs that may in fact be protected. Separating the investigational item from the routine patient costs at the point of billing is what keeps a protected cost from denying with the item.

Coverage — including whether a service is established enough to be covered — is one of the recurring categories mapped in why claims get denied, and the front-end controls that reduce it are the ones set out in preventing denials — here, checking the coverage policy and flagging the indication before the service is ever performed. The rest of this cluster is indexed on the Denials & Appeals pillar.