Every practice knows what a denied claim looks like on a remittance. Less visible is the second cost attached to it: the labor required to recover money the practice has already earned.
That cost rarely appears as a line item. It sits inside payroll, in the hours a billing team spends on hold with payers, interpreting reason codes in web portals, and drafting appeals one at a time. Across thousands of claims a month, it becomes one of the larger uncounted expenses on the income statement.
This paper refers to that cost as the Administrative Tax: the operational cost of managing claim denials by hand. It has two components. The first is the labor spent reworking the denials a team can reach. The second, usually larger, is the revenue abandoned on the denials a team cannot reach: claims that were recoverable but were written off because pursuing them manually cost more than they would return.
The benchmarks make the case directly. Initial denial rates have risen for four consecutive years and now sit near 11.8% nationally. Reworking a single claim costs an ambulatory practice roughly $25, and several times that for complex clinical denials. More than half of denied claims are never reworked, and fewer than one percent of appealable denials are ever appealed, even though roughly seven in ten denials are overturned when someone pursues them.
The implication is direct. For most practices, the problem is not that payers deny too much. It is that manual capacity limits how much a practice can recover, and everything above that limit becomes margin forfeited by default.
The 2026 denial landscape
Denials are no longer occasional friction in the revenue cycle. They are a structural cost, and it is growing.
Kodiak Solutions, drawing on data from more than 2,100 hospitals and 300,000 physicians, found that the average initial denial rate reached 11.8% in 2024, up from 10.2% in 2020. That was the fourth consecutive annual increase. Experian Health's State of Claims survey reports the same trend from the provider's side: 41% of organizations now report denial rates of 10% or higher, up from 30% just three years earlier. A majority also report that claim errors are rising rather than falling.
The expense compounds where it meets staffing. The roles responsible for denials are among the hardest to keep filled. Experian's staffing research found that nine in ten providers report double-digit turnover in administrative roles, and 43% report administrative-role turnover above 25%, several times the national workforce average.
These forces reinforce each other. Rising denials require more follow-up labor at the moment that labor is scarcest and most expensive. When a team is understaffed, triage wins: new claims that must go out to keep cash flowing take priority over aging denials that require investigation. The denials do not disappear. They age past filing deadlines and convert into permanent write-offs. This is how a staffing problem becomes a margin problem.
Deconstructing the Administrative Tax
Sizing the tax starts with the unit cost of a single denial. Reworking one denied claim costs an ambulatory practice about $25. That figure rises sharply with complexity. For hospitals and health systems, Premier puts the average administrative cost of fighting a denial at about $57; other industry analyses put it closer to $118; and the most complex clinical denials, which require chart review, clinical documentation, and several rounds of appeal, can run as high as $181 per claim. By comparison, submitting a clean claim in the first place costs a small fraction of that. Reworking a claim costs three to five times more than getting it right the first time.
The reason this expense stays invisible is that nearly all of it is labor. Roughly 90% of claims-processing cost is staff time, and most denied claims require a person to handle them before they can be resubmitted. That labor does not appear in a budget as a "denial expense." It appears as headcount.
Breaking a single denial into steps shows where the time goes. A biller first determines why the claim was denied, usually by logging into a payer portal and interpreting a reason code that may not reflect the underlying problem. If the portal is unclear, the next step is the telephone, where hold times routinely exceed 45 minutes and payers often limit how many claims can be discussed per call, forcing repeat attempts across separate days. The correction follows: re-verifying eligibility, a step that now takes more than ten minutes for over half of providers; assembling supporting documentation; and drafting an appeal letter to the payer's specific requirements. A simple denial may take ten to fifteen minutes to resolve. A typical one takes twenty-five to forty-five. A complex clinical appeal, with records to gather and several review cycles, can take hours.
No individual step is difficult. The difficulty is their volume. Each step is manual, each repeats hundreds or thousands of times a month, and the total registers as a large, unlabeled cost that no one deliberately chose to incur.
The abandonment crisis
This is the point at which the tax turns from expensive to damaging.
Because reworking denials is slow and labor is finite, teams cannot possibly work every denial they receive. Industry estimates suggest that between 50% and 65% of denied claims are never reworked at all. On the appeals side the picture is even starker. KFF's analysis of ACA marketplace plans found that in 2024, insurers denied roughly 19% of in-network claims, yet consumers appealed fewer than one percent of them. In some earlier years, that appeal rate was as low as two-tenths of one percent.
At the level of a single claim, abandonment is rational. When a biller weighs a low-dollar denial, the arithmetic often argues against pursuing it: if the labor to recover a claim exceeds what the claim is worth, an hour and several payer calls cannot be justified. So the claim is written off. It was often recoverable; the manual cost of recovery simply exceeded its value. Repeated across a month of small-balance denials, a practice can surrender tens of thousands of dollars in collectible revenue while each individual decision remains defensible.
What turns this into a crisis rather than a rounding error is one fact: most of that abandoned revenue was recoverable. Premier found that roughly 70% of denials are ultimately overturned when providers actually pursue them. KFF found that even among the rare marketplace appeals that do get filed, insurers reverse their original decision in close to half of cases. The denials were not final. They were only treated as final because no one had the capacity to fight them.
That reframes the problem. Revenue lost to denials is, for the most part, not lost to payers who refuse to pay. It is lost to a manual process that does not scale, one in which the cost of recovery is high enough, and the capacity to pursue it low enough, that a majority of collectible claims are surrendered by default. Because the loss accumulates one write-off at a time, it rarely draws the scrutiny a comparable six-figure loss would otherwise attract.
A worked example: the compounding cost
A concrete example makes the figures easier to weigh. The following model describes a mid-sized practice. The full calculation appears in the box below, and every input can be replaced with a practice's own figures.
Illustrative model. Replace every input with your own.
A mid-sized practice submits 5,000 claims a month at a 10% denial rate, conservative against the national average. That is 500 denied claims a month, or 6,000 a year.
Part A, rework labor spent. At 30 minutes per denial, working all 500 consumes about 250 staff hours a month, roughly 3,000 hours a year, or nearly one and a half full-time employees doing nothing but rework. At the $25 benchmark, that is $12,500 a month, or $150,000 a year.
Part B, recoverable revenue abandoned. Applying the abandonment range, about 65% of those denials, roughly 325 a month, are never reworked. At an average claim value of $150, that is about $48,750 in denied revenue set aside each month. Because roughly 70% of pursued denials are recoverable, the practice forgoes about $34,000 a month, or roughly $409,500 a year, in revenue it could have collected.
Total annual Administrative Tax, about $559,500, the rework labor spent plus the recoverable revenue abandoned.
Combined, the Administrative Tax for a single mid-sized practice approaches roughly half a million dollars a year: a six-figure rework-labor bill, plus several times that amount in recoverable revenue abandoned. Neither figure appears as a line on the income statement. The first sits in payroll; the second appears only as revenue that never arrived. The pattern matters most for organizations that are growing, because both figures rise with claim volume and denial rate. The tax grows with the practice rather than shrinking as it scales.
The solution: flipping the math
Every figure in this paper traces back to one cause. The cost of recovery is set by manual labor, and manual labor does not scale. Changing that variable inverts the equation.
This is the function of technology-enabled automation. It does not work harder than a human team; it removes the manual steps that make each denial expensive. Three capabilities do most of the work. Automated tracking replaces portal-by-portal status checks: a system ingests remittance data, sorts each denial by reason and payer, and presents the full inventory in one place, so nothing ages out unseen. Algorithmic prioritization ranks denials by recoverability and dollar value, directing staff time to the claims most likely to pay, while the low-dollar claims a person would have written off can be pursued automatically at little marginal cost. Automated appeal generation assembles payer-specific letters and their supporting documentation in seconds rather than the twenty-five to forty-five minutes a manual draft requires.
The effect on the arithmetic is direct. When the per-claim cost of pursuit approaches zero, the calculation that forces abandonment no longer holds. A low-dollar claim becomes worth recovering. The denials a manual team surrenders become revenue the practice can capture, because capturing it no longer costs more than it returns. More consequential than the lower cost is what automation removes: the abandonment rate itself, which is the single largest source of denial-related loss.
The gains upstream are larger still. The most valuable denial is the one that never occurs. Automated eligibility verification and pre-submission claim scrubbing catch the eligibility errors, coding mismatches, and missing-information problems that account for most preventable denials, and they catch them before a claim leaves the building. The results are measurable: in one industry-wide evaluation, 83% of organizations using AI-driven RCM reduced denials by at least 10% within six months, and pre-bill scrubbing routinely moves denial rates out of the double digits and into the mid-single digits.
The most telling figure may be how few organizations have moved. Only about 14% of providers currently use AI in their claims process, though two-thirds say they believe it would help. That gap is the opportunity. Practices that automate denial management now recover revenue that competitors still write off by hand.
Calculate your own Administrative Tax
The following framework estimates an organization's annual exposure. It has two parts, the labor spent and the revenue abandoned, and the sum is the total Administrative Tax.
Part A. Rework labor spent
Monthly claims × denial rate × rework cost per claim × 12
Part B. Recoverable revenue abandoned
(Monthly claims × denial rate) × share of denials never reworked × average claim value × share recoverable × 12
The Administrative Tax is Part A plus Part B. Applied to the mid-sized practice above:
Worked example, illustrative only.
Component Calculation Result Part A: rework labor spent 5,000 × 0.10 × $25 × 12 $150,000 Part B: recoverable revenue abandoned (5,000 × 0.10) × 0.65 × $150 × 0.70 × 12 ~$409,500 Total annual Administrative Tax ~$559,500
With that figure in hand, the comparison is straightforward: weigh the annual Administrative Tax against the fully loaded cost of automating denial management. For most practices, the labor line alone justifies the investment. The revenue recovered by ending abandonment is what turns the decision from a cost saving into a source of growth.
The denials arriving in this month's remittances are, for the most part, recoverable. The question is how much of that revenue the current process allows a practice to keep.
Methodology
The financial models in this paper combine national median benchmarks from established revenue-cycle sources: MGMA, KFF, Premier, the AMA, the Journal of AHIMA, Kodiak Solutions, and Experian Health. Where a single authoritative figure was unavailable, the model uses the midpoint of published industry ranges and states the assumption in the relevant calculation.
These benchmarks describe national medians, and results at any individual organization will differ. Denial rates and rework costs vary by specialty, geography, and payer mix. Behavioral health, for instance, typically runs higher than general medical practice, and denial rates differ widely across payers, with traditional Medicare well below Medicaid managed care. The worked example is illustrative and uses deliberately conservative inputs; readers should substitute their own claim volume, denial rate, average claim value, and labor cost to model their exposure. The worksheet above is built for that purpose.
About the data
Figures in this paper are drawn from: Kodiak Solutions Revenue Cycle Analytics (2024 initial denial rate); Experian Health, State of Claims (2025); Premier, Inc. provider denial-cost analysis (2023 data); MGMA benchmarks on per-claim rework cost and unworked denials; KFF, Claims Denials and Appeals in ACA Marketplace Plans (2024 data); Poland & Harihara, "Claims Denials: A Step-by-Step Approach to Resolution," Journal of AHIMA (2022); the AMA Prior Authorization Physician Survey (2024); and Black Book Research's industry-wide evaluation of AI-driven RCM solutions (2025).
Related: Preventing prior-authorization and administrative denials · Navigating AI and algorithmic claim denials · All guides