Every denied claim has a reason code. Most of them are fiction — or close enough that the distinction does not matter to the hospital waiting 90 days for payment. Understanding why claims actually get denied, as opposed to why the payer says they were denied, is the difference between a revenue cycle that leaks and one that holds.
The uncomfortable truth first: the majority of denials are preventable, and the majority of those are front-end failures. A mistyped policy number. A missed prior authorization. An eligibility check skipped at registration. The industry's own data keeps confirming it — the errors that cost the most happen before the claim is ever submitted.
The Four Horsemen
Prior authorization is the largest single category and the most strategically deployed. Kodiak's 2024 data showed authorization-related initial denials falling 7.7% as providers built better front-end processes — and payers promptly shifted to medical-necessity and information-request denials instead. The authorization regime keeps expanding regardless: more services require it, the criteria keep changing, and the administrative burden of obtaining it falls entirely on the provider. When authorization is denied retroactively — after the care was delivered — the hospital holds the bag.
Medical necessity is the denial reason with the most discretion and the least accountability. The payer's reviewer, who has never seen the patient, overrules the treating physician's judgment. These are the denials providers find hardest to overturn, and they are growing: up 5% in 2024. They cluster in high-cost service lines — inpatient stays, advanced imaging, specialty procedures — which is to say, exactly where the dollars are.
Coding errors are the quiet killer. Wrong diagnosis code, missing modifier, procedure code that does not match the documented diagnosis — each one is an automatic denial, and each one was preventable. With rework costs running $47.77 per Medicare Advantage denial and $63.76 per commercial denial according to HFMA research, a coding department's error rate is a direct P&L line.
Timely filing is the most unforgiving. Miss the payer's filing deadline — often 90 to 180 days, varying by contract — and the claim is dead with no appeal. It is the denial category that turns administrative delay into permanent revenue loss.
The Appeals Playbook
Here is what the industry does not advertise: appealed denials get overturned at remarkable rates. Premier's data suggests roughly 70% of denied claims are ultimately paid. Kodiak's 2025 benchmarking put the average overturn rate at 42.1% across all denials — lower, because it includes the unwinnable ones — but well-worked clinical appeals routinely reverse more than half.
The playbook is unglamorous and it works. First, triage by value and winnability: not every denial deserves an appeal, and the art is knowing which ones do. High-dollar clinical denials with strong documentation get the full treatment; low-dollar technical denials get fixed and resubmitted fast. Second, build the clinical record before you need it — the appeal is won or lost in the documentation that existed at the time of service, not in the letter written afterward. Physician-to-physician peer reviews move medical-necessity denials when paper appeals do not. Third, track everything by payer, by reason code, by service line. Denial patterns are payer behavior patterns, and a hospital that can show a payer its own data negotiates from strength.
The deeper fix is structural: eligibility verification and prior-auth clearance before service, coding audits that catch errors pre-bill, and denial-prevention analytics that flag at-risk claims before submission. AI-driven denial prediction is gaining real traction — early adopters report roughly 20% declines in rejected submissions. The technology exists. The question is whether providers deploy it faster than payers deploy new denial logic.
Because that is the actual dynamic. It is an arms race, and the payer has the bigger R&D budget.
The Structural Fixes Gaining Ground
Technology helps at the margin, but the only durable fixes are structural — and several are moving. Gold-carding laws, now enacted or proposed in multiple states, exempt providers with high approval rates from prior-authorization requirements entirely: prove you order appropriately, and the payer stops second-guessing you. It is the rare reform that rewards good behavior instead of punishing bad.
CMS has tightened prior-authorization rules for Medicare Advantage, requiring plans to follow traditional Medicare coverage rules and limiting their ability to impose additional clinical criteria. Enforcement remains the open question — rules without penalties are suggestions — but the regulatory direction is unmistakable.
The transparency lever may matter most. Payers are not currently required to publish meaningful denial-rate data by category, which means the behavior Kodiak documents in private datasets stays invisible to regulators and the public. Mandatory denial-rate reporting — by payer, by service line, by reason code — would change the incentive structure overnight. Sunlight does not fix everything. But an industry that publishes its denial rates starts managing them.
Until those reforms bite, the playbook stands: prevent what you can, appeal what you must, and track everything. The hospitals winning this fight treat denials as a data problem, not a billing problem — because that is what the payers turned it into.
"The appeal you win in 90 days was lost the day the claim went out the door. Prevention is the only strategy that scales."