Claim denials, explained

A denial is a claim the payer adjudicated and refused to pay, in whole or in part, with a stated reason. It is different from a rejection (never adjudicated) and from an underpayment (paid, but less than the contract requires). Distinguishing the three is the first thing a revenue operation has to get right, because each has a different fix and different appeal rights.

​

The three failure modes

Underpayments are the invisible one. A rejection and a denial both announce themselves. An underpayment looks exactly like a payment — the money arrives, the claim closes, and nobody notices that the allowed amount was below your contracted rate. Detecting it requires loading your fee schedules into your PM system and comparing. See Underpayments, fee schedules, and payer contracts.

​

The denial families

Six families produce the overwhelming majority of denials.

​

1. Eligibility

Typical CARCs: 26 (expenses incurred prior to coverage), 27 (expenses after coverage terminated), 31 (patient cannot be identified as our insured), 32 (dependent not an eligible dependent). Root cause: no eligibility check, a check run too early, or a data-entry error in the member ID. Fix: verify current coverage, correct the insurance record, rebill to the correct payer. Often the patient has coverage, just not the coverage you billed. Prevention: run eligibility at scheduling and at check-in, and save the 271 response.

​

2. Authorization

Typical CARCs: 197 (precertification/authorization absent), 198 (precertification exceeded). Root cause: authorization not obtained, obtained but mismatched to the service billed, or, very commonly, obtained and never captured on the claim. Fix: if an authorization exists, submit a corrected claim with the number. If not, request retro-authorization where the payer allows it. If neither, appeal on medical necessity or absorb it. Prevention: build a per-payer auth-required list into scheduling, and a scrubber edit that blocks submission when a listed code has no auth number. CO-197 is a contractual denial. You generally cannot bill the patient for it — the patient did nothing wrong. Writing it off or billing the patient are both wrong answers.

​

3. Coding and medical necessity

Typical CARCs: 11 (diagnosis inconsistent with the procedure), 16 (lacks information), 50 (not deemed a medical necessity), 167 (diagnosis not covered), 97 (bundled into another service). Root cause: code-diagnosis mismatch, missing modifier, documentation that doesn’t support the code, or a service excluded by a local or national coverage determination. Fix: review documentation against the code. If the code is right, appeal with the note and the applicable coverage policy. If the code is wrong, correct it. Prevention: scrubber edits for code-pair and modifier rules; clinician education on documentation; awareness of your MAC’s LCDs.

​

4. Timely filing

Typical CARC: 29 (time limit for filing has expired). Root cause: charge entry lag, or a claim that bounced in a rejection loop and was never resubmitted. Fix: appeal with proof of timely submission — clearinghouse acceptance reports are the standard evidence. If there was genuinely no timely submission, this is usually unrecoverable. Prevention: same-day charge entry, same-day rejection resolution, and a report of claims with no acceptance confirmation. See Beat timely filing limits.

​

5. Coordination of benefits

Typical CARCs: 22 (may be covered by another payer), 23 (impact of prior payer adjudication). Root cause: wrong payer billed first, or the payer’s COB record is stale. Fix: determine the correct order, bill the primary, then submit to the secondary with the primary’s 835 detail. Often the patient must call the payer to update COB — you cannot do it for them. Prevention: ask about other coverage at every registration, and verify it in the 271.

​

6. Credentialing and enrollment

Typical CARCs: 8 (procedure inconsistent with provider type), 185 (rendering provider not eligible), 206 (NPI missing or invalid). Root cause: provider not credentialed, not linked to the group contract, taxonomy mismatch, or a service date before the effective date. Fix: resolve with the payer’s enrollment department, then rebill. If the service predates the effective date, pursue a retro-effective date — this is where your documented follow-up log earns its keep. Prevention: don’t schedule insured patients with a clinician before their effective date. See Handle credentialing delays.

​

Plus: not actually errors

CARC 96 (non-covered) and 204 (not covered under the plan) are usually benefit design, not mistakes. The service isn’t covered. Route to patient responsibility if your contract and an advance beneficiary notice permit, or write it off. Chasing these as appeals wastes the queue’s time.

​

Why denials are mostly upstream

Look at where the six families originate: Not one of them originates in the billing department. Billers are the people who discover and remediate failures that happened somewhere else. A denial-management strategy that consists only of working the queue harder is treating symptoms. This is why the root-cause tagging loop matters more than the queue itself: it is the only mechanism that pushes fixes back to where the failure happens. See Build the billing rhythm.

​

Prevention economics

The arithmetic that should drive your priorities:

• Reworking a denied claim costs staff time, research, correction, resubmission, and often a follow-up call. Industry estimates commonly place the cost of reworking a claim in the range of tens of dollars per claim, and appeals cost multiples of that.
• Preventing the same denial costs close to nothing once the control exists, a scrubber edit, a required field, a checklist item.
• A meaningful share of denied claims are never reworked at all. Those are pure loss.
• Appeal success rates are far from 100%, so even a worked denial has an expected value below its face amount.

The practical conclusion: a denial prevented is worth substantially more than a denial worked, and roughly an order of magnitude cheaper to achieve. A team that spends all its time on the queue and none on prevention is optimizing the expensive half.

​

Turning denials into management information

The transformation that separates practices that improve from practices that just cope:

​

Benchmarks

Directional, and highly specialty-dependent: Track your own trend rather than chasing an external benchmark. A denial rate falling from 12% to 7% matters more than whether 7% is “good.”