Denial reports are long, but denial causes are not. In most practices a small group of Claim Adjustment Reason Codes accounts for the large majority of denied dollars, and almost every one of them originates before the claim is ever built: at scheduling, at registration, at the point of documentation. That matters, because it means the fix rarely lives in the billing office.
Below are the seven that show up most often, what each one is really telling you, and the workflow change that stops it recurring.
The seven, at a glance
| Code | What the payer is saying | Where it is actually caused |
|---|---|---|
| CO-16 | Claim lacks information needed for adjudication | Registration and claim scrubbing |
| CO-197 | Precertification or authorization absent | Scheduling and pre-visit workflow |
| CO-97 | Payment is bundled into another adjudicated service | Coding and modifier selection |
| CO-50 | Not deemed a medical necessity | Clinical documentation and diagnosis linkage |
| CO-29 | Time limit for filing has expired | AR follow-up cadence |
| CO-18 | Exact duplicate claim or service | Resubmission discipline |
| PR-204 | Not covered under the patient's current plan | Eligibility verification |
CO-16, the one that hides everything else
CO-16 is not a diagnosis, it is a category. On its own it tells you almost nothing, which is why practices that track it as a single bucket never make progress against it. The useful information sits in the Remittance Advice Remark Codes attached to it, which name the specific missing element: a subscriber identifier that does not match, an absent referring provider NPI, a taxonomy mismatch, a missing accident date.
The corrective action is to stop reporting CO-16 as one line. Split it by remark code, and you will usually find two or three concrete registration defects driving most of the volume. Those are fixable at the front desk in a week.
CO-197, the most expensive to fix late
Authorization denials are unusual because the window to fix them cheaply closes fastest. Once the service has been rendered without authorization, your options narrow to a retroactive authorization request, which many payers grant only in defined circumstances and within a short window, or a medical necessity appeal with clinical documentation attached.
- Maintain a live list of which CPT codes require authorization by payer and plan type, and check it at scheduling rather than at check-in.
- Track authorization numbers, unit counts and valid date ranges as structured data, not as a note in the chart.
- Watch for expiry: a valid authorization that has run out of units or dates denies identically to no authorization at all.
- Escalate quickly. Retroactive windows are typically measured in days, not weeks.
CO-97 and the bundling question
CO-97 means the payer considers the service already paid within another line on the same claim or episode. Sometimes it is correct and the write-off is legitimate. Often it is not, and the difference is usually a modifier that was warranted but not applied, or two services that were genuinely distinct but not documented as such.
Before appealing, check the code pair against the National Correct Coding Initiative edits and confirm whether a modifier is even permitted for that pair. If the edit carries a modifier indicator of 0, no modifier will override it and the appeal will fail regardless of documentation. If it is 1, the question becomes whether your note supports the separate service.
CO-50, where documentation meets coding
Medical necessity denials are a mismatch between the diagnosis submitted and the payer's coverage policy for that procedure. For Medicare, the governing documents are the National and Local Coverage Determinations, which state explicitly which diagnosis codes support which services. Commercial payers publish their own medical policies, and they do not always agree with CMS.
The failure is usually specificity. An unspecified diagnosis code frequently falls outside a coverage policy that the specific version of the same condition would have satisfied. This is a documentation and query problem before it is a billing problem.
CO-29, the avoidable one
Timely filing denials are pure lost revenue: the service was rendered, the claim was payable, and nobody worked it in time. Filing limits vary widely by payer and contract, and the clock does not always start where you assume, particularly on secondary claims where it may run from the primary remittance date.
- Record each payer's filing limit in your practice management system and surface claims approaching it as an aging alert, not a monthly report.
- Keep proof of timely submission. Clearinghouse acknowledgement reports will often overturn a CO-29 on appeal when the original submission was in fact on time.
- Treat rejections differently from denials. A clearinghouse rejection never reached the payer, so the filing clock is still running while it sits unworked.
CO-18 and PR-204
Duplicate denials usually mean somebody resubmitted rather than corrected. If a claim needs changing after submission, it generally requires a corrected claim with the appropriate frequency code and the original claim reference, not a fresh copy. Sending the same claim again produces a duplicate denial and adds nothing.
PR-204 shifts the balance to the patient because the service is not a covered benefit. The question worth asking is whether eligibility was verified on the date of service, and whether the patient was told before the appointment. Discovering non-coverage at the remittance stage means an awkward conversation and a low probability of collection.
What to do on Monday
- Pull ninety days of denials and rank them by dollars, not by count. The most frequent denial is rarely the most expensive one.
- Split CO-16 by remark code so it stops masking its own causes.
- For your top three categories, identify the step in the workflow where the defect is introduced, and change that step rather than the appeal process.
- Convert each resolved root cause into a pre-submission edit so it cannot recur.
- Re-measure after one full billing cycle. Denial work without measurement tends to move volume between categories rather than reduce it.