Prior authorization denials: why 88% go unappealed

KFF's analysis of CMS Part C reporting covers 52.8 million prior authorization determinations across Medicare Advantage plans. Two figures from it sit uncomfortably next to each other: 80.7% of appealed prior-authorization denials are overturned, and only 11.5% of denials are ever appealed.

Read together, they describe a system where the appeal almost always works and almost never happens. Nearly nine in ten denied authorizations are simply absorbed — the service is not rendered, is rendered unpaid, or is rescheduled into a later month and forgotten.

This is not a story about payers denying claims they should approve, though that is part of it. It is a story about what happens when the cost of contesting a decision exceeds the staff hours available to contest it.

The AMA's 2025 physician survey puts the administrative load at 13 hours per physician per week on prior authorization alone, across an average of 40 requests. That is most of a full working day, every week, per physician, before a single appeal is filed.

An appeal is strictly additional work on top of that. It requires pulling the clinical documentation, identifying the specific policy criterion the payer applied, writing a letter that addresses that criterion rather than restating the original request, and then tracking the response through a second determination window.

So the decision not to appeal is rarely a clinical judgement that the denial was correct. It is an arithmetic one: the practice has a finite number of staff hours, and the queue is longer than the hours.

The direct cost of contesting a single denied claim averages $57.23 according to Premier's analysis of 280 hospitals. That is the labour to research, document and submit — not the value of the service.

Against that, the expected value of appealing is straightforward to estimate. If the overturn rate on appeal is roughly 80%, the expected recovery on a $400 procedure is about $320 for $57 of work. The economics of appealing are not close.

Optum's index of 124 million hospital claim remits found 84% of denials are potentially avoidable, and that 44% originate at the front end — registration, eligibility and authorization. Authorization and precertification alone account for 12.8% of all denials.

That ordering matters. An appeal recovers revenue after the fact at $57 a claim; a correct authorization on the first submission costs nothing extra. The highest-return work is upstream of the denial, not downstream of it.

An effective appeal answers the specific criterion the payer cited. A letter that restates the original request will usually produce the same determination, because the reviewer is applying a policy test the letter never addresses.

Most payers offer a peer-to-peer review in which the ordering physician speaks directly with the payer’s medical director. It is the fastest available reversal mechanism — frequently resolved on the call rather than in the 30 to 60 days a written appeal takes.

It is also the mechanism practices most often lose by default, because the window is short. Payers typically allow only a few business days from the denial notice to request it, and once that window closes the case drops into the written appeal queue regardless of merit.

Two things make peer-to-peer work. Request it the same day the denial is received, before anyone has decided whether to appeal, because the request can be withdrawn but the window cannot be reopened. Then have the ordering physician cite the payer’s own published medical policy criteria and state which of them the documentation satisfies. A conversation about clinical judgement in general terms tends to fail; a conversation about criterion three of the payer’s own policy tends to succeed.

The most common weakness in a prior authorization appeal is that it argues the patient needs the service. The payer is not disputing that. It is asserting that the documentation submitted did not establish the specific criteria in its coverage policy.

Every commercial payer publishes medical policies stating exactly what must be documented for a given service. An appeal that walks through those criteria in order, quotes the policy, and points to the page and date in the record satisfying each one is answering the question actually asked. An appeal that describes the patient’s condition is answering a different one.

Appeal rights expire, and the windows are set by the payer and the plan type rather than by any single rule. Medicare Advantage, commercial and self-funded ERISA plans all run different clocks, and the clock generally starts at the date on the denial notice rather than the date the practice opened the envelope.

The practical control is a single tracked list with three dates per denial: the peer-to-peer request deadline, the first-level appeal deadline, and the external review deadline where one exists. Denials worked from a general queue without those dates attached are the ones that expire, and an expired appeal right converts a recoverable denial into a write-off with no further options.

Every figure above points at the same conclusion. The denials are largely avoidable, the appeals largely succeed, and the reason neither happens is that nobody has the hours.

That is a staffing and process question rather than a clinical one, which is also why it is solvable. A practice that systematically works its authorization queue — preventing what it can upstream and appealing what it cannot — is not being cleverer than its payers. It is simply doing work that most practices leave undone because they cannot staff it.