Denials Are Born at the Signature Line, Not the Claim Line
Why billing companies are moving validation upstream into their clients' intake, and the three books of business where it pays back first.
Every billing company owner I meet has a version of the same story. A denial lands on a Tuesday. Someone on the team pulls the chart, finds an order that was signed without a diagnosis supporting the item, and starts an appeal that will take three weeks. Nobody bills the client for those three weeks. You absorb them.
That is the part of the outsourced billing model nobody wrote into the contract. You are paid on collections, usually 4% to 9% of them in 2026, but a large share of your labor goes to repairing errors created before the encounter ever reached your queue. The claim was dead at the referral. You just found out about it at adjudication.
"You are not losing money on the claims you work. You are losing it on the ones that were never winnable and still had to be worked."
The margin math got worse, and the fee did not move
Tebra's 2026 State of the Medical Billing Industry report found that 54% of billing companies expect gross margins of 10% or below, and 46% saw denial rates rise during the year. In the same market, 98% now sell services beyond claim submission. Read those three numbers together and the strategy writes itself: the percentage fee is capped by competition, the work per claim keeps going up, so the only thing left with real slope is preventing the denial in the first place.
CMS has already told everyone where denials come from
This is the part that gets skipped in most denial-prevention pitches, because the answer is public and unglamorous. Every year the Comprehensive Error Rate Testing program reviews a random sample of Medicare fee-for-service claims and publishes why they failed. In the 2025 reporting period, roughly 53% of improper payments were attributed to insufficient documentation, another 12% to no documentation at all, and 15.3% to medical necessity that the record did not support. MedPAC's June 2026 report to Congress put documentation deficiencies at 68% of all improper payments across Medicare FFS.
Incorrect coding is real, but it is nowhere near the top of that list. Which means the tooling most billing companies have invested in, the scrubber and the clearinghouse edit set, is aimed at a minority of the problem. A scrubber is excellent at catching an invalid modifier or a bad payer ID. It cannot tell you that the face to face note is missing, that the written order has no length of need, or that the physician signed on a date outside the payer's window. Those failures live in documents that arrive weeks before the charge is created, and they arrive in your client's fax queue, not yours.
The commercial side tells the same story. In KFF's August 2026 analysis of the first year of mandated payer prior authorization metrics, Medicare Advantage plans denied 12% to 13% of standard prior authorization requests, Medicaid managed care 14%, and ACA Marketplace plans 18%. Denials were rarely appealed. When they were, 67% of Medicare Advantage denials were overturned. KFF's own reading of that overturn rate is worth quoting the substance of: either the request should have been approved initially, or the initial submission was missing the documentation needed to justify the service.
"A denial that gets overturned on appeal was usually not a coverage dispute. It was an incomplete packet that eventually got completed."
Three books where this shows up fastest
This does not apply evenly across a client roster. In a straightforward primary care book, most denials really are eligibility and coding, and a good scrubber earns its keep. The documentation problem concentrates in books where a payer requires proof of medical necessity, assembled from several signed documents, before the service is delivered. Three segments stand out.
DME and HME books
MedPAC's June 2026 report put the DMEPOS improper payment rate at 24.2%, close to four times the 6.6% Medicare fee-for-service average and the worst of any claim type. CMS is specific about the cause. In its DMEPOS supplier guidance, continuous glucose monitors carried a 25.2% improper payment rate with the majority of errors related to missing or insufficient documentation, and orthotic braces ran between 35.2% and 54.4%.
The denial reason is almost never a code. It is a Certificate of Medical Necessity with a blank field, a written order missing length of need, an oxygen qualifying test outside the allowed window, or a face to face encounter note that does not mention the item ordered.
Resupply makes it compounding. One incomplete file does not fail once. It fails every month it recurs, and the supplier usually finds out three cycles later. Add the fact that Medicare DME MAC jurisdictions apply different local coverage rules, and a book spread across states is carrying several rule sets at once, usually inside one or two people's heads.
Home health and hospice books
CMS's own provider compliance data for the 2024 reporting period attributes 51.4% of home health improper payments to insufficient documentation and 33.7% to medical necessity. Incorrect coding accounts for 3.4%. If your denial strategy for a home health client is built around coding accuracy, it is aimed at roughly a thirtieth of the problem.
Look at the face to face denial reason codes Medicare publishes and the pattern is unmistakable. The certification is invalid because the encounter document was missing, or untimely, or performed by a practitioner who does not qualify, or not related to the primary reason for home health. Four different codes, none of them about the care delivered, all of them about the paperwork proving it.
These are conditions of payment, not billing details. When they fail there is often nothing to appeal, and the agency also loses the referral source, because the physician's office remembers who chased them for a corrected order three weeks after discharge.
Prior authorization heavy specialty books
Published specialty benchmark compilations drawing on MGMA, Experian, HFMA and MDaudit data put behavioral health denial rates at roughly 20% to 30%, orthopedics at 14% to 22%, and physical therapy and chiropractic at 15% to 20%. These are the books where authorization packets fail on conservative care history, functional scores, and treatment plan detail. All of it exists somewhere in the chart. None of it is assembled at submission.
The clock makes it worse. Since January 2026, Medicare Advantage plans must decide standard requests within 7 days and expedited requests within 72 hours. A shorter window only helps the party whose packet was complete when the window opened.
What DocuFindr does
DocuFindr is a validation layer that sits at intake. It reads the inbound packet, whether that arrives by fax, PDF, or portal, checks it against the payer's requirements for that item and jurisdiction, and returns a completeness score with the specific gap named: which field is blank, which date is out of window, whose signature is missing. The coordinator sees it before anything is submitted, while the referring physician is still reachable and still remembers the patient.
It is not a scrubber, which works after the charge exists. It is not a denial management tool, which works after the money is gone. It works at the signature, not the appeal.
Why billing companies deploy it into the client's intake
Three patterns, in the order most partners adopt them.
On pricing, one thing worth stating plainly because it comes up in every legal review: DocuFindr is a flat monthly fee by document volume. We do not price on a percentage of collections or a share of recovered claim value, in either direction. That keeps the arrangement clean under anti kickback analysis and keeps your cost predictable in a month when a client's volume spikes.
"It also changes the sales conversation. You stop competing at 5.5% against someone at 6% and start competing on first pass rate."
The 2027 clock is the part people are underestimating
CMS-0057-F requires impacted payers to run FHIR based prior authorization APIs from January 1, 2027. Most of the commentary treats it as a speed story, and it is. But faster, structured, machine read submission does not forgive an unsigned order. It removes the excuses that currently absorb the blame. When a payer can answer in 72 hours, a denial is unambiguously about what you sent, and every client will be able to see exactly how long the packet sat with their billing partner. Whoever is validating at intake by then looks like a different category of vendor. Whoever is still working appeals looks like a cost center.
A 90 day test that does not require faith
- Pick one book. DME, home health, or your highest denial specialty. Do not start with the whole roster.
- Baseline it. 90 days of denials by CARC, split documentation origin versus coding origin. If the documentation share is under a quarter, buy nothing and go work on something else.
- Route intake through validation for one client in that book. Keep everything else identical so the comparison holds.
- Measure four things: share of packets with a gap caught before submission, first pass rate for that cohort, days from referral to clean submission, and rework hours per 100 claims.
That last metric is the one that shows up in your margin. The others are how you sell it to the client.
Want to see it against your own denial file?
Send an anonymized 90 day denial extract and we will show you which of those claims were decided before submission. Fifteen minute walkthrough, no deck required.
DocuFindr is a pre-submission documentation validation platform for DME and HME suppliers, home health agencies, billing companies, and specialty clinics. HIPAA compliant, BAA available, patent pending, NVIDIA Inception member, Microsoft Azure ISV Success, MATTER Select member.
Sources
- CMS, Medicare Fee-for-Service Supplemental Improper Payment Data, 2025 reporting period (claims July 1, 2023 – June 30, 2024): $28.8B in improper payments at a 6.6% rate; approximately 53.0% attributed to insufficient documentation, 15.3% medical necessity, 12.0% no documentation.
- MedPAC, June 2026 Report to Congress: DMEPOS improper payment rate of 24.2% against a 6.6% Medicare FFS average; documentation deficiencies accounting for 68% of all improper payments.
- CMS, Provider Compliance Hot Spot: DMEPOS Suppliers: FY2024 continuous glucose monitor improper payment rate of 25.2% with the majority of errors related to missing or insufficient documentation; orthotic brace rates of 35.2% to 54.4%.
- CMS Medicare Learning Network, Provider Compliance Tips: Home Health Services, 2024 reporting period: insufficient documentation 51.4%, medical necessity 33.7%, incorrect coding 3.4%.
- CMS, Home Health Services Review Reason Codes (HH01A–HH01E): face to face encounter document missing, untimely, performed by an unapproved practitioner, or unrelated to the primary reason for home health.
- KFF, Prior Authorization Metrics Provide New Insights into Insurer Practices, but Gaps Remain, August 2026: Medicare Advantage denied 12–13% of standard prior authorization requests, Medicaid managed care 14%, ACA Marketplace 18%; 67% of appealed Medicare Advantage denials overturned.
- Tebra, 2026 State of the Medical Billing Industry Report: 54% of billing companies expecting gross margins at or below 10%, 46% reporting rising denial rates, 98% offering services beyond claim submission.
- MGMA DataDive and Experian Health State of Claims: average initial denial rate of approximately 11.8%; 41% of providers reporting denial rates above 10%. Specialty ranges drawn from published compilations of MGMA, Experian, HFMA and MDaudit data.
- CMS, Interoperability and Prior Authorization Final Rule (CMS-0057-F): FHIR prior authorization APIs required from January 1, 2027; Medicare Advantage decision windows of 7 days standard and 72 hours expedited effective January 1, 2026.