Billing errors split into two kinds. The first kind stops a claim: the payer denies it, a code lands in the work queue, and someone has to work it. The second kind is quieter. The payer pays something, the claim closes, and the lost dollars sit inside an adjustment code nobody questioned. Both categories are expensive; the second one is harder to see.
Understanding which errors generate the most volume and the most revenue loss is more useful than a comprehensive list of every possible mistake. The goal is to identify the two or three patterns that account for most of a practice's preventable loss, then fix them upstream.
Modifier errors
Missing or mismatched modifiers are the most common fixable denial trigger in most practices. Payer systems check modifier logic automatically, and a single gap stops the claim.
- Missing required modifier. A procedure or E/M code requires a modifier the payer expects (GP on physical therapy services, modifier 25 when billing an evaluation on the same day as a procedure) and the claim goes out without it. CO-4 fires. The fix is a corrected claim with the right modifier, not a reconsideration letter.
- Mismatched modifier. The modifier is present but inconsistent with the procedure code. Modifier 57 on a minor-surgery same-day visit, or modifier 50 with a payer that requires separate LT and RT lines: both produce CO-4 for different reasons.
- Bundling without a modifier override. Two services billed together trigger an NCCI edit and the payer issues a CO-97 denial. If the edit allows a modifier override (indicator 1), adding modifier 59 or 25 to the right line resolves it. If the edit is hard (indicator 0), the services cannot be billed separately regardless of documentation.
Modifier errors are worth addressing systematically rather than claim-by-claim. The same configuration gap typically affects every claim in that workflow, so a single billing rule fix closes more claims than a week of individual corrections.
Diagnosis code errors
ICD-10 specificity requirements catch practices that still code the way they did under ICD-9. Payers reject unspecified or invalid codes, and medical-necessity edits often require a specific diagnosis to trigger coverage.
- Unspecified codes where specificity is available. A code ending in "9" (unspecified) when the visit note clearly documents a more specific finding is both a documentation gap and a medical-necessity risk.
- Diagnosis-procedure mismatch. The procedure code must pair clinically with the diagnosis code on the same claim. When they do not link correctly, the payer denies on medical necessity or applies an LCD restriction.
- Sequencing errors. The primary diagnosis drives medical necessity for most payers. Listing an incidental finding first and the clinical reason for the visit second can reverse the medical-necessity logic and produce a denial the documentation would otherwise support.
Prior authorization gaps
Prior authorization denials (CO-15) are among the most frequently unrecoverable errors in medical billing. Once a service is rendered without an authorization the payer required, a retroactive approval is uncertain and payer-dependent. The appeal goes to a medical director who did not see the patient and is applying criteria written before the visit.
For practices that already have CO-15 volume in the denied queue: group by payer and by CPT code. The pattern will reveal whether auth gaps are tied to a specific payer that changed its requirements, a specific service category, or a scheduling workflow that does not consistently run an auth check.
Timely filing misses
Every payer sets a filing deadline. Most commercial plans require submission within 90 to 365 days of service; the exact window is specified in your contract. A CO-29 timely filing denial is almost always permanent because the appeal window is narrow and the only defense that works is documentation proving the claim arrived before the deadline.
That document is the 277 clearinghouse acknowledgment, the transaction record confirming the payer received and accepted the electronic claim. Without it, a timely filing appeal is difficult to win regardless of the clinical merit. Practices that have a reliable process for capturing and storing 277 acknowledgments by claim number recover far more timely filing appeals than those relying on billing system timestamps alone.
Underpayments treated as routine write-offs
This category does not show up in any denial queue. A payer pays less than your contracted rate, the claim posts as paid, and the difference lands inside a CO-45 adjustment that looks identical to any other contractual write-off. The claim closes. Nobody flags it. The shortfall accumulates.
Catching underpayments requires a separate audit step outside the normal denial workflow: load your contracted rates by CPT code and payer, then compare each allowed amount in your 835 remittance data to what you were owed at contract. Any CO-45 where the allowed amount is below your contracted rate is a recoverable underpayment, not a routine write-off. Dispute it as a contract matter through provider relations, not as a clinical reconsideration.
Underpayments are worth auditing regularly because the pattern is usually systematic: one payer applying the wrong fee schedule to the same CPT code on every claim. A single fee-schedule correction can recover months of accrued shortfall. This is one of the higher-yield items in a full revenue leakage audit.
Duplicate billing
Duplicate claims (the same service submitted more than once for the same patient and date of service) produce CO-18 denials. The most common cause is a manual resubmission after a claim sits in pending status longer than expected, combined with no confirmation that the original actually arrived at the payer.
The practical guard: do not resubmit without a clearinghouse rejection acknowledgment confirming the original claim never reached the payer. If the original was accepted and is in processing, a second submission creates a duplicate that will deny and may complicate the original payment.
Where to start
Sort your denied claims by CARC code and group by payer. CO-4 and CO-97 clusters reveal modifier configuration gaps in a specific workflow. CO-15 clusters point to authorization verification failures at a particular payer or service type. CO-29 concentrations flag a resubmission or clearinghouse timing problem.
Then run the separate underpayment audit on paid claims: 835 allowed amounts compared to contracted rates by CPT code. That data lives outside the denial queue entirely.
Most practices find that two or three error categories account for the majority of preventable loss. Fixing the root cause of each, whether that is a billing rule, a workflow step, or a payer-specific configuration, does more over time than any amount of individual claim corrections.