What Timely Filing Actually Costs a Small Practice

The Only Deadline That Is Actually Final

Almost everything in a revenue cycle is recoverable if you get to it.

A coding error can be corrected and resubmitted. A missing modifier can be added. A denied claim can be appealed, sometimes twice. An underpayment can be disputed against the contract. A patient balance can be billed for years. In almost every category, being late costs you time and interest and nothing else.

Timely filing is the exception. Cross the window and the claim is not denied on the merits. It is denied on the calendar, and there is nothing underneath the calendar to argue with.

The service happened. It was medically necessary. It was documented correctly and coded correctly and the patient was eligible. None of that is in dispute and none of it matters. The money stops existing.

That asymmetry is the reason this deserves its own attention rather than being one row in a denial report. Every other line on that report is a claim you can still do something about.

What Actually Happens at the Line

The remittance comes back with CO-29: the time limit for filing has expired.

What follows is usually a short conversation that ends badly. Most payers will accept an appeal only on proof of timely submission, meaning you submitted inside the window and something went wrong in transit. A clearinghouse acknowledgement with a date, a payer-assigned claim number, a rejection you can show you corrected and resubmitted promptly.

If you have that proof, you have a real case. If the claim genuinely sat in a queue for seven months, there is no argument available, because the payer is not wrong. The contract said 90 days and it has been 210.

There is a second-order cost people miss. Some payer contracts prohibit balance-billing the patient for a claim denied as untimely, on the reasoning that the failure was administrative and not the patient's. So the claim does not become a patient balance either. It becomes an adjustment.

The Deadlines, and Where Yours Actually Come From

Filing limits are not one number. They come from four different places and are measured from different starting points, which is most of why practices get them wrong.

Medicare's twelve month limit is set in federal statute and applies uniformly. Every other row on this table is contract-specific or state-specific, which means the only authoritative answer for your practice is your own paperwork.

Three things on that table cost practices real money.

The starting point is not always the date of service. Secondary claims are measured from the primary's remittance date. If the primary took four months to adjudicate, your secondary clock started four months after the visit, and if you assume otherwise you will either file late or panic early.

Resubmission usually does not reset the clock. A claim denied for a correctable reason and resubmitted is still measured against the original date of service at most payers. A claim that bounces three times over five months is burning its window each round, and nothing in the resubmission workflow says so.

Your shortest contract sets your operational deadline. If one payer in your mix gives you 90 days and the rest give you a year, your process has to run at 90 days, because you cannot run two different urgencies through one queue and expect the fast one to be honoured.

Why Claims Sit Long Enough for This to Happen

Nobody decides to let a claim expire. It happens through a specific and predictable set of mechanics.

Queues sort by value, so small claims sink. A worklist ordered by dollar amount is rational for a day and irrational for a quarter. The claims that never reach the top of the list are the ones that eventually age out, and because they are individually small, nobody notices them leaving.

Claims waiting on someone else stall indefinitely. A claim that needs records from a hospital, a corrected referral, a prior authorisation applied retroactively, or a callback from a patient is not in anyone's active queue. It is in a pending state, and pending states have no clock.

Denials that need real work get deferred. A denial requiring a written appeal with supporting documentation takes forty minutes. A denial requiring a resubmitted claim takes four. Given a full day, the four minute items get done. This is correct triage in the short run and it is exactly how the forty minute items reach month seven.

Staff turnover creates a gap. When a biller leaves, the queue they were working continues to age while the role is being filled and the new person learns the systems. Practices routinely lose the oldest, hardest claims across a hiring gap and never connect the two events.

None of these are failures of effort. They are all what happens when work is prioritised sensibly and nothing in the system escalates on the approach of a deadline.

The Aging Report Nobody Runs

Every practice runs an AR aging report. It buckets receivables into 0 to 30 days, 31 to 60, 61 to 90, 91 to 120, and over 120.

That report is genuinely useful and it is answering a different question than the one that matters here.

Aging buckets measure how long money has been outstanding. They were designed for cash flow: how much can I expect this month, how much is slow, how much is stuck. Every bucket is defined by days elapsed since the claim was created.

The question that matters for filing is not how old a claim is. It is how many days it has left. Those are not the same measurement, and the difference is the payer.

A 100 day old claim against a payer with a 12 month window has 265 days left and is merely slow. A 100 day old claim against a payer with a 90 day window is already dead and is still sitting in your 91 to 120 bucket looking like a receivable. Both appear in the same row of the same report. One of them is an asset and one of them is a fiction, and the report cannot tell you which.

That is the whole problem in one sentence. Your aging report shows you how old the money is. It does not show you how close to dead it is.

The report that would tell you is a simple idea and no system produces it, because producing it requires joining two things that live in different places: your open AR, and the filing deadline for each payer on it. The AR is in your billing system. The deadlines are in a filing cabinet or a PDF of a participation agreement. Nothing automatically connects them, so the report does not exist, so nobody sees the number.

How to Find Your Own Number

This is worth doing once by hand, because the result is usually not what people expect and it is a real number rather than a worry.

1. Step 1
   Write down your filing limit for every payer you bill
   Not from memory and not from the portal. From the participation agreement for commercial payers, the state manual for Medicaid, and the plan document where you have self-funded plans. Medicare is twelve months and needs no lookup. This is the tedious step and it is the one that makes everything after it possible.
2. Step 2
   Note what each limit is measured from
   Date of service for most. Date of primary remittance for secondaries. Date of injury for workers' comp in some states. Getting this wrong on secondaries is the single most common error in this exercise.
3. Step 3
   Export your open AR at line level with the date of service and the payer
   You need the actual date of service on each line, not the claim creation date and not the last-touched date. If your system will only give you claim level, take it, but know that it will hide anything unusual happening inside a claim.
4. Step 4
   Calculate days remaining, not days elapsed
   For each open line, subtract days elapsed from that payer's limit. This one subtraction is the entire trick, and it is the number your billing system does not compute for you.
5. Step 5
   Sort ascending and look at the top
   Everything at or below zero is already gone; that is your first number and it is the one that tends to land. Everything between zero and 30 days is what you can still save this month, and it is the only part of this exercise that is urgent.
6. Step 6
   Total the already-expired by payer
   This tells you where the process is failing rather than just how much it cost. One payer dominating the expired list usually means a workflow that does not match that payer's window, which is a fixable thing rather than a diffuse discipline problem.
7. Step 7
   Repeat monthly and watch the zero to 30 band
   The absolute total matters less than whether the band ahead of the cliff is growing. A shrinking band means the process is catching up. A growing one means you are queueing work faster than you are clearing it, and next quarter's number will be worse.

Two hours of a spreadsheet gets you a defensible figure for what you lost last year and a live list of what you are about to lose this month. The second is worth more than the first, because it is the only part you can still act on.

What This Costs, Calculated Honestly

There is no industry benchmark worth quoting here, and any number you see attached to this in an article is almost certainly repeated from another article. Your figure depends on your payer mix, your denial rate, and how your queue is worked. It is genuinely specific to you.

But the shape of the calculation is not complicated. Three components:

• What already expired. Line up your closed adjustments for the last twelve months and pull out anything coded to timely filing. Most practices have never separated this from other adjustments, which is exactly why it has never been a number anyone had to look at.
• What is currently at risk. Your open AR, days remaining, everything under 30. This one is live and shrinking as you read.
• The recurrence rate. The first number divided by twelve, roughly, is what you are losing per month if nothing changes upstream. That is the figure that justifies fixing the process rather than working harder at the queue.

The third number is the one that decides whether this is worth your attention. A practice losing a trivial amount per month to expired claims should go fix something else. A practice discovering it has been quietly adjusting off a meaningful sum every month for two years has found the highest-return thing on its list, because unlike most revenue cycle work, the fix here is a process change rather than an ongoing effort.

What the Fix Actually Looks Like

Not a bigger queue and not more discipline.

A deadline field, populated per payer. If your billing system supports a custom field on the payer record, put the filing limit in it. If it does not, one shared spreadsheet is enough. The point is that the limit lives somewhere a person can see while they work rather than somewhere they would have to go look.

Sort by days remaining, not by age or value. This is the change that does the most and costs the least. It reorders the same work into the order that prevents permanent loss.

An escalation at a fixed threshold. Pick a number, 30 days remaining or 45, and make anything crossing it stop being a queue item and start being an exception that a named person handles. The threshold matters less than the fact that it exists.

A separate adjustment code for timely filing. So that next year the first number in the calculation above takes ten minutes instead of two hours, and so that the trend is visible without anyone running a project to find it.

None of that requires new software. It requires knowing your limits, computing one subtraction, and treating the result as a different kind of work than the rest of the queue.

Common questions

Can a timely filing denial ever be appealed successfully?

Yes, but almost always on procedural grounds rather than clinical ones. The winning argument is proof of timely submission: a clearinghouse acknowledgement with a date, a payer claim number, or documentation that the claim was submitted inside the window and rejected for a reason you then corrected promptly. Appeals arguing that the service was necessary and correctly coded do not succeed, because the payer is not disputing either of those things.

Does resubmitting a denied claim reset the filing clock?

Usually not. Most payers measure the window from the original date of service regardless of how many times the claim has been submitted. That means a claim bouncing repeatedly over several months is consuming its window each round while appearing active in your queue. Some payers allow a separate, shorter window for corrected claims measured from the denial date, which is a different limit rather than an extension, and it is worth confirming per contract.

What is the deadline for a secondary claim?

Typically measured from the date the primary payer adjudicated, not from the date of service. This is the most commonly misunderstood item on the list. If the primary took months to pay, your secondary window may be substantially longer than a date-of-service calculation would suggest, and if you have been assuming the visit date you may be filing later than you think you are.

Our AR looks healthy. Do we still have this problem?

Possibly, because a healthy AR and expired claims are not mutually exclusive. Claims that age out generally get adjusted off, and an adjusted claim leaves the AR entirely. A clean aging report can mean you are collecting well or it can mean the oldest items have already been removed. The way to tell the difference is to look at your adjustments rather than your receivables.

How far back is it worth looking?

Twelve months of adjustments to establish what this has been costing, and your current open AR to see what is at risk now. Going back further is interesting and not actionable, since anything past its window is permanently gone regardless of when you discover it.

Should we just outsource billing to fix this?

It can help and it does not fix it by itself. A billing company works a queue the same way an in-house biller does, and the same prioritisation pressures apply. What matters is whether anyone is computing days remaining rather than days elapsed, and whether anything escalates on approach to a deadline. Ask that question directly before you change vendors, because it is answerable and it predicts the outcome better than the org chart does.