A denial tells you two things: this claim didn't get paid, and something in your process produced a claim that wasn't payable. Practices that only respond to the first fact end up running a permanent rework operation. Practices that respond to both eventually run out of that particular denial.
First, separate rejections from denials
The distinction gets blurred constantly, and it matters because the response is different.
A rejection happens before adjudication. The claim failed a format, data or eligibility check at the clearinghouse or at the payer's front door. It is not in the payer's system. It is corrected and submitted as a new claim.
A denial happens after adjudication. The payer processed the claim and refused payment, with a reason code. It is on record, and it generally requires a corrected claim or a formal appeal rather than a straight resubmission.
Reporting that lumps the two together will mislead you about where your problems are.
The recurring categories
Registration and eligibility errors
Wrong member ID, misspelled name, incorrect date of birth, terminated coverage, the wrong payer entity, or an unresolved coordination-of-benefits situation where the primary hasn't adjudicated first. These are consistently among the most common denial categories across the industry, and they are almost entirely preventable at intake and verification.
Missing or invalid prior authorization
No authorization on file, authorization expired, units exhausted, or approval issued for a different code, level of care or provider than the one billed. Note how many of those are tracking failures rather than clinical ones — the authorization existed, it just wasn't reconciled against what was actually delivered.
Coding and modifier issues
Missing modifiers where the payer requires them; modifiers applied where they aren't supported; procedure pairs that are bundled under payer edits; diagnosis codes that don't establish medical necessity for the procedure billed; unit counts that don't match documented time or quantity.
Medical necessity
The payer accepts the service was rendered but concludes the record doesn't meet its published coverage criteria. This is where local and national coverage determinations matter most: they set out explicit expectations about symptom duration, conservative treatment attempted, objective findings and documented response. Medical-necessity denials are the most appealable category and the most documentation-dependent.
Timely filing
The claim arrived after the payer's submission deadline. Almost never recoverable and almost always the downstream symptom of something else — an unmonitored rejection, an unworked A/R bucket, or a backlog nobody flagged.
Non-covered service or benefit exhausted
The service isn't a benefit under that plan, or the patient has used the available visits, sessions or benefit maximum. This is a verification problem, not a billing problem, and it is common in specialties with session or visit limits.
Duplicate claim
Frequently self-inflicted: a claim resubmitted because nobody could tell whether the original was received. Reliable claim-status monitoring removes most of this category.
Credentialing and provider data
The rendering provider isn't credentialed with that plan, isn't linked to the billing entity, or the identifiers on the claim don't match the payer's records. Common in growing practices where a new clinician starts seeing patients before enrolment completes.
The rule that explains most of it
Denials surface in billing. They are usually created before billing. Registration, verification, authorization and documentation produce the conditions; the claim just reports them.
This is why adding capacity to denial rework rarely changes the denial rate. It processes the same volume faster. To change the volume you have to trace denials back to the stage that produced them and change something there.
Building a correction loop
A workable process looks roughly like this:
- Capture every denial with reason code, payer, provider, service, date and dollar value. If denials are being posted as adjustments, this step is already broken.
- Categorise by root cause stage — registration, eligibility, authorization, coding, documentation, filing, credentialing — rather than only by the payer's reason code.
- Rank by frequency and value. A high-frequency, low-value category and a low-frequency, high-value one both deserve attention, for different reasons.
- Work the recoverable ones inside the appeal window, with the documentation the payer's criteria actually cite.
- Take the top two causes back upstream. Not all of them — two. Fix those, then re-measure.
- Re-measure monthly. If a category isn't shrinking, the fix didn't land.
The discipline that matters most is the fifth step, and it is the one most often skipped, because it requires a conversation with the part of the practice that created the problem rather than the part that discovered it.
What about the ones you can't prevent?
Some denials are genuine coverage disagreements, and some payers deny at rates that have little to do with anything your practice did. Those get worked on their merits: appeal with the strongest documentation available, escalate where the policy language supports you, and accept that a proportion won't be overturned.
What you should not accept is not knowing which category you're in. A practice that can't distinguish preventable denials from unavoidable ones can't make a rational decision about where to spend effort.
General information about US healthcare revenue cycle management. Not legal, compliance or coding advice. Payer policies and coverage determinations vary by plan and jurisdiction and change frequently.