The Denial Funnel: Where Your Recoverable Dollars Actually Sit
Why the denial rate is the wrong headline metric — and why the money is in the claims that are never reworked and the payments made below contract that nobody examines.
Ask a revenue cycle leader how denials are performing and the answer is usually a percentage: 6%, 9%, 12%. It is useful. It is also incomplete.
A denial rate tells you how often a payer made an adverse first decision. It does not tell you how much of that denied value was actually recoverable, how much your team touched, how much was appealed on time, how much was overturned, how much was written off — or how much revenue never appeared in the denial report because the payer sent a payment that was simply lower than the contract required.
The denial rate is not the wrong number. It is the wrong headline. For finance leaders, the more important question is: of the dollars we earned, where did they stop moving — and which of those dollars can still be recovered?
HFMA's claim-integrity framework makes this point explicitly. It does not stop at one denial rate. Its recommended measures include initial denials by both volume and dollars, denial write-offs as a percentage of net patient service revenue, time from denial to appeal, time from denial to resolution, and the percentage of denials overturned. Those are not redundant measures. They answer different questions about whether a denial became a temporary delay, a recoverable asset, or a permanent loss.
A 7% denial rate with disciplined recovery can be financially healthier than a 5% denial rate with weak follow-up, missed appeal windows and silent underpayments.
That distinction matters more as denials become harder to manage. Experian Health's 2025 State of Claims survey found that 41% of surveyed providers said more than 10% of their claims were being denied. In a separate 2026 survey of 210 revenue cycle leaders, 25% reported that denial rates had continued to rise over the prior year. The operational response cannot be to stare harder at the top-line percentage. The response has to be to understand the funnel beneath it.
This is the familiar layer: initial denials segmented by payer, reason, location, provider, service line and dollar value. It tells you where friction enters the revenue cycle. It is essential for root-cause prevention — but it still does not tell you what the denial is worth after the payer's first decision.
A $40 denial and a $4,000 denial each count as one denied claim if you manage only by volume. Even a dollar-based denial rate can mislead if it uses gross charges rather than expected reimbursement. Finance needs the expected allowed amount — the revenue the contract says the claim should produce — because that is the value at risk.
Not every denial deserves an appeal. Some claims are non-covered by policy, unsupported by the record or otherwise not payable. Others are administrative failures that can be corrected, medical-necessity decisions that can be supported, authorization mismatches that can be challenged, or payer processing errors that can be reversed. The financial exposure is not the total denied balance. It is the portion of the denied balance for which the evidence supports payment.
This is where a denial report can look better than the cash reality. A denial may be correctly identified, categorized and placed into a work queue — and still receive no meaningful action before the appeal or timely-filing window closes. If the organization cannot separate denials received from recoverable denials worked, it cannot measure the amount of revenue being lost through capacity, prioritization or workflow failure.
That gap is not theoretical. In June 2026, HHS-OIG reported on a specific set of Medicare Advantage prior-authorization denials for skilled nursing facility admissions. The 19 organizations reviewed denied 12% of requests. Only 18% of those denials were appealed. But when they were appealed, 95% were overturned in the enrollee's favor. This is not a benchmark for physician-practice claims and should not be treated as one. It is a powerful demonstration of the measurement problem: the first denial rate alone said almost nothing about the value sitting behind the appeals that were — and were not — pursued.
An overturn rate without dollars can flatter performance. A recovery total without labor can do the same. Mature denial management asks both: what percentage of recoverable dollars came back, and what did it take to get them? That allows leaders to distinguish a payer that requires one corrected claim from a payer that routinely forces two appeals, a medical-record submission and 60 days of follow-up before paying the same contractual obligation.
That difference belongs in payer strategy, staffing decisions and contract negotiations. A payer's nominal rate is not its economic rate when collection requires disproportionate administrative work.
The most dangerous assumption in revenue cycle reporting is that a claim marked paid is finished. CMS describes the 835 electronic remittance advice as the transaction that reports the payer's adjudication, payment and adjustments at the claim or service-line level. That tells you what the payer did. It does not independently prove that the payer calculated the contractual allowed amount correctly.
A contractual adjustment can be legitimate. So can bundling, multiple-procedure reductions, modifier logic, site-of-service differences, unit limits and payer-specific reimbursement rules. But unless actual payment is compared to an expected-payment model built from the contract and applicable reimbursement logic, a short payment can close as a zero-balance account with no denial and no work queue. The transaction looks resolved. The revenue is not.
If you are trying to prevent denials, study initial-denial incidence and root cause. If you are trying to size staffing, study recoverable inventory, touch rate and time to resolution. If you are trying to evaluate payer behavior, study overturn rates, administrative effort, underpayment variance and days to cash. If you are reporting financial leakage to the CFO, use dollars — not just claim counts — and reconcile final write-offs and contract variance back to expected reimbursement.
For an executive dashboard, five measures are more decision-useful than a denial percentage by itself: denied dollars as a percentage of expected reimbursement; recoverable-denial work rate; recovery yield; final denial write-offs as a percentage of net patient service revenue; and paid-to-expected contract variance. Add time to appeal and time to resolution when cash velocity matters — which, for most practices, is always.
It is: How many dollars that should have been paid are still sitting somewhere between first adjudication and final resolution? Some are obvious denials. Some are denials nobody had capacity to touch. Some are appeals never filed. Some have already been written off. And some are sitting in accounts your system calls paid.
That is why Emphanex starts with the remittance data rather than a denial-rate benchmark. The goal is not to tell you whether your percentage looks normal. It is to quantify where the recoverable dollars actually sit, which losses are operational, which are payer-driven, which payments are below expectation, and which problems are large enough to deserve management attention first.
A benchmark can tell you whether you look like everyone else. A denial funnel can tell you where your money went.
This article is for general informational purposes only and does not constitute legal, coding, reimbursement or compliance advice. Contract interpretation and recoverability depend on the specific payer agreement, plan terms, claim facts and applicable law.