Unbilled and Held Claims: The Receivable No Report Shows
Every A/R report a practice runs is built from claims that were submitted. The work that has been done and not yet billed is therefore invisible in it — not shown as old, not shown as at risk, not shown at all. That would be a reporting inconvenience if the clock waited for the claim. It does not: the filing period runs from the date of service, so this population is spending its window from the moment the patient leaves.
Updated 9 min read
On this page
Key takeaways
- Unbilled work is receivable. The obligation exists from the date of service; the absence of a claim does not pause it.
- The filing clock starts at the date of service, not at the date the claim became billable. A claim held for missing information is spending the same window a submitted claim spends.
- Aging reports are built from billed claims, so an account never billed has no bucket. The largest avoidable loss in a revenue cycle is the one its main risk report structurally cannot display.
- The exceptions to the filing limit cover the program's errors and the patient's circumstances — administrative error, retroactive entitlement, another payer recovering months later. None of them covers the practice being behind.
- Two different populations get conflated: work not yet billed because it is not ready, and claims deliberately held pending something. They have different owners and different fixes.
- A hold with no owner, no reason and no expiry date is a write-off with a delay. The discipline is that every held claim carries all three.
- The measure to keep is the unbilled inventory by age and by reason, alongside charge lag. A single total tells you nothing about which of these it is.
The clock started when the patient was seen
The rule that makes this population urgent rather than merely untidy is a simple one, and it is widely misread. Under 42 CFR 424.44 (opens in a new tab), the Medicare filing period runs from the date of service. Not from the date the encounter was closed, not from the date coding finished, not from the date the claim cleared the scrubber. Commercial windows are set by contract and are frequently shorter, but they are almost always anchored the same way.
So a claim sitting in a hold queue is not waiting outside the process. It is inside it, consuming the only resource that cannot be replaced. Two claims from the same date of service — one submitted the next morning, one held six weeks for a missing referral number — face the same deadline, and the second has spent six weeks of it doing nothing.
And the exceptions are not about you
Why the report cannot show it
An aging report sorts open balances into buckets by how long they have been outstanding, and the event it measures from is submission. That is a reasonable design for what the report was built to do — but it means an account with no claim has no age, and an account with no age has no row. A/R aging buckets covers what that report does show and how to read it; the point here is the shape of what it omits.
The structural consequence, stated plainly
This is a specific instance of a general rule the corpus states elsewhere: a queue and a report answer different questions, and building one from the other hides whatever the source excludes. The unbilled population needs its own inventory, produced from the encounters rather than from the claims, because those are the only records that exist for it.
Two populations that get counted as one
“Unbilled” is routinely used for two different things with different owners and different remedies. Separating them is the first thing a practice can do about the number, because a single total gives no indication of which problem it has.
| Population | What it is waiting on | Where the fix lives |
|---|---|---|
| Not yet billable | The claim cannot be produced yet: a charge not captured, an encounter not closed, coding not complete, documentation not signed. | Upstream of billing, in clinical and front-office workflow. The billing team can report it and escalate it; it cannot resolve it. |
| Billable but held | The claim exists and is being withheld: a scrubber edit unresolved, missing registration or insurance data, an authorization number not yet received, a secondary claim waiting on the primary remittance. | Inside the billing operation, mostly. Each hold has a specific missing item and someone who can obtain it — which is what makes an expiry date reasonable to demand. |
A third category hides between them and is worth naming: claims held for a reason that was resolved and never removed from the hold. Nothing surfaces those, because the condition that put them there is gone and nobody is looking for a queue with no current cause.
What a hold has to carry
Holding a claim is often correct. Submitting one that is known to be wrong produces a denial, a correction and a second submission, all of which cost more than waiting a day for the right information. The problem is not that claims are held; it is that holding is usually a state with no attributes, so nothing distinguishes a claim held deliberately this morning from one held eight weeks ago for a reason nobody remembers.
A reason, from a defined list
Not free text. A hold reason that can be counted is what turns the queue into a diagnosis — the same argument the corpus makes about posting a write-off with a reason code. If one reason accounts for most of the queue, that reason is a process to fix rather than a batch to work.An owner who can actually clear it
The person who can obtain the missing item, not the person who noticed it was missing. A hold assigned to the billing team for a document only the clinician can sign is a hold assigned to nobody.An expiry date, set against the filing window
Every hold gets a date by which it either clears or escalates, and that date is derived from the deadline the claim is already spending — not from a general sense of how long is reasonable. A hold without an expiry is a write-off with a delay, and the delay is the part that makes it invisible.A decision at the expiry, not a renewal
When the date arrives, the choice is to submit with what exists, to escalate to someone who can supply what is missing, or to record a deliberate decision not to bill. All three are outcomes. Quietly re-dating the hold is not one, and it is what most systems make easiest.
The two measures worth keeping
Common questions
Is unbilled work really part of A/R?
It is a receivable in every sense that matters operationally: the service was furnished, the money is owed by someone, and the window in which it can be claimed is already running. Whether it appears in the accounting A/R balance depends on when the practice recognizes the charge, which is a bookkeeping question. The management question is not, and the answer to it is unambiguous — this is money at risk, on a clock, and excluding it from the receivable view is what makes it dangerous.
Does the filing clock start when we submit or when we treat?
When the service was furnished. The Medicare period runs from the date of service, and commercial windows — which are contractual, and frequently shorter — are almost always anchored the same way. This is the single most consequential fact about held claims, because it means a claim waiting on a missing authorization number is spending exactly the same window as one already with the payer, while producing none of the signals that would let anyone notice.
If we could not bill because information was missing, is there any relief?
Not on that basis. The exceptions that extend the Medicare filing period cover error or misrepresentation by a Medicare employee, contractor or agent; a beneficiary notified of retroactive entitlement after the date of service; and a State Medicaid agency or Medicare Advantage plan recovering payment months after the service. They are about the program's mistakes and the patient's circumstances. There is no exception for a coding backlog, an unsigned note, or a missing referral number, and planning as though there might be is how a held queue becomes a loss.
Should we ever submit a claim we know is incomplete?
It is a real trade-off rather than a rule, and the expiry date is what forces it to be made deliberately. Submitting a claim known to be wrong produces a denial and a rework cycle, which is why holding is often correct in the first days. As the filing window closes, the calculus inverts: a submitted claim that is denied can usually be corrected and resubmitted within the window, while a claim never submitted cannot be anything. The point of setting the expiry against the deadline is that somebody makes that call while both options still exist.
How do we find claims held for a reason that no longer applies?
By reviewing the queue against its own reasons rather than working it by age. A hold is created by a condition; nothing in most systems watches for that condition being resolved, so the claim stays put after the missing document arrived or the eligibility question was answered. Periodically re-testing each hold reason against current data — rather than waiting for someone to open the account — is the only thing that surfaces them, and it tends to clear a meaningful part of the queue the first time it is done.
Key terms in this article
Defined once, on their own pages.
Continue learning
The report this population is missing from, and the deadline it is spending.
A/R Aging Buckets
What the aging report shows, and why an account never billed has no row in it.
Designing an A/R Follow-Up Process
The workable-state test that expressly excludes this population — and what to do with what it excludes.
Preserving Timely Filing Evidence
What proves a claim was submitted in time, once it finally is.
Timely Filing Denials
What happens when the window closes, and the narrow grounds for an exception.
Pre-Submission Claim Validation
The scrubbing layer that creates many of these holds, and how to keep it from creating them twice.
Charge Lag Calculator
Work out how long your own work takes to become a claim, from your own figures.
Authoritative sources
- 42 CFR § 424.44 — Time limits for filing claims (opens in a new tab)
Sets the period for filing a Medicare claim by reference to the date of service, so the window is measured from when the service was furnished rather than from when the claim was prepared or submitted. Its exceptions permitting an extended period cover error or misrepresentation by a Medicare employee, contractor or agent; a beneficiary notified of retroactive Medicare entitlement after the date of service; and a State Medicaid agency or Medicare Advantage plan that recovered payment from the provider months after the service — with a further extension where the last day of the period falls on a Federal nonworkday.