What is the denial rate?
A denial is an item of a hospital claim that the health plan (operadora) refuses to pay, in whole or in part. Health plans in Brazil are regulated by ANS (Agência Nacional de Saúde Suplementar, the National Supplementary Health Agency). The denial rate measures how much of what the hospital billed was not accepted when the claim was reviewed. It is the most quoted revenue cycle indicator because it captures rework, cash tied up and permanent loss in one number.
The catch is that “denial rate” can mean different things. The health plan disputes items in the first review (initial denial); the hospital appeals some of them; part is reversed and paid; the rest becomes a loss (final or accepted denial). Each stage has its own number, and each can be divided by a different base. The concept, types and reasons are covered in the pillar guide Claim denials; this guide is about measurement.
The formulas: one indicator, several denominators
The table lists the most common formulas and the public benchmark for each. Use the “Benchmark source” column to know which external number your formula can be compared with.
| Indicator | Numerator | Denominator | Benchmark source |
|---|---|---|---|
| Initial denial (internal) | Amount denied in the health plan's first review | Amount billed in the period | Revenue cycle KPIs guide |
| Managerial initial denial (Anahp) | Amounts denied by health plans still under negotiation | Amounts receivable from health plans (through 2025); amounts received from health plans in the month (from 2026) | Anahp Observatory [1][2][7] |
| Accounting accepted denial (Anahp) | Denied amount that is not paid after negotiation | Gross health plan revenue (until 2022: total net revenue) | Anahp Observatory reports [2][3] |
| Final denial (internal) | Denied amount not reversed after appeals | Gross health plan revenue or amount billed (pick one) | Revenue cycle KPIs guide |
| Initial and final denial (ANS) | Denied amounts (initial or final) in closed claims | Total amount reported in closed claims | ANS Denial Dashboard [5] |
| Denial ratio (ANS, health plan assessment) | 0.75 × (denied ÷ reported amount) + 0.25 × (providers with denial ÷ all providers) | Composite (weighted index) | ANS Health Plan Qualification Program [4] |
| Reversal rate | Amount recovered on appeal | Amount appealed (only denials whose appeal deadline has passed) | Internal |
Three decisions define your formula and should be written down:
- Timing. Initial denial measures rework and delay; final denial measures loss. They are different indicators; the guide Initial and final denials explains the difference.
- Base. Amount billed, amounts receivable or gross health plan revenue. Changing the base changes the result even if nothing changed in operations.
- Cutoff date. Date of service, submission or the health plan's claim review statement. ANS, for example, only counts closed claims, where the amount paid plus the amount denied covers the amount reported [5].
Calculation example (illustrative)
The figures below are hypothetical, only to show the math.
| Item | Amount (BRL) |
|---|---|
| Amount billed to health plans in the month | 10,000,000 |
| Denied in first review | 1,200,000 |
| Amount appealed | 1,000,000 |
| Amount reversed on appeal | 650,000 |
| Denial that stuck (1,200,000 − 650,000) | 550,000 |
| Gross health plan revenue for the month | 10,000,000 |
- Initial denial = 1,200,000 ÷ 10,000,000 = 12.0%.
- Reversal rate = 650,000 ÷ 1,000,000 = 65.0%.
- Final denial = 550,000 ÷ 10,000,000 = 5.5%.
Note that the BRL 200,000 not appealed is part of the final denial: not appealing turns initial denial into loss. The appeal process is covered in the guide Denial appeal.
What is the industry benchmark?
The main public benchmark in Brazil is the Anahp Observatory, which compiles indicators from hospitals that are members of Anahp (Associação Nacional de Hospitais Privados, the National Association of Private Hospitals). It had 194 members at the end of 2025, 199 in Q1 2026 and 203 in Q2 2026, mostly large hospitals [2][3][7]. The guide Anahp Observatory explains the publication and the sample.
| Indicator (Anahp hospitals) | 2022 | 2023 | 2024 | 2025 | Q1 2026 | Q2 2026 |
|---|---|---|---|---|---|---|
| Managerial initial denial (% of amounts receivable from health plans) | — | 11.89% | 15.89% | 15.93% | — | 11.69%** |
| Accounting accepted denial (% of gross health plan revenue) | 0.78%* | 1.17% | 1.96% | 1.72% | 1.61% | 1.35% |
| Accepted denial in Q1 of each year | — | 0.92% | 1.87% | 1.66% | 1.61% | — |
| Initial denial in Q2 of each year | — | 9.27% | 12.45% | 15.63% | — | 11.69%** |
| Accepted denial in Q2 of each year | — | 1.32% | 1.67% | 1.70% | — | 1.35% |
* In 2022 the denominator was total net revenue. Anahp changed the calculation in 2023, so the 2022 figure is not comparable with later years [2]. ** From 2026, initial denial is divided by amounts received from health plans in the month; not comparable with 2023–2025 [7].
How to read these numbers:
- Initial and accepted denial are not divided by each other. One is on amounts receivable; the other, on gross health plan revenue. Dividing 1.72% by 15.93% does not give the sector's reversal rate, because the bases differ.
- Compare quarter with quarter. The 1.61% for Q1 2026 compares with Q1 of prior years (1.66% in 2025), not with the full year [3]. The 1.35% for Q2 2026 compares with 1.70% in Q2 2025 [7].
- The initial denial denominator changed in 2026. Anahp moved from amounts receivable to amounts received in the month. The 11.69% for Q2 2026 does not mean a drop from the 15.63% of Q2 2025 [7].
- Initial denial leaves out what is not yet billed. Anahp notes that the indicator does not include amounts not yet billed, which could make receivables even larger [1].
- The real loss is larger than accepted denial. Anahp notes that hospitals also accept discounts to close payment agreements [2]. In an Anahp survey of 85 hospitals, the allowance for doubtful accounts totaled R$ 1.8 billion in 2024 (R$ 1.4 billion in 2023), and withheld denials, R$ 5.8 billion in 2024[8].
- The sample changes. Anahp grew from 120 hospitals in 2023 to 203 in Q2 2026 and notes, in a footnote on operating indicators, that this growth affects the figures through sample heterogeneity [3].
Another base. The Panorama do Ciclo de Receita 2026 (Revenue Cycle Panorama 2026), by ZG Soluções, based on data from the company's client institutions, reported an average initial denial of 8.86% and a final denial of 6.58% [6]. Its base and definitions differ from Anahp's, so the levels are not directly comparable.
US context. In the US, the closest concepts are the initial denial rate and final denial write-offs. The Brazilian figures above follow Brazilian definitions and should not be compared one-to-one with US benchmarks.
How to compare the denial rate by health plan
The hospital average hides the health plan that denies far more than the others. Breaking the rate down by health plan is the first step in any analysis.
What ANS publishes. ANS maintains a Denial Dashboard (Painel de Glosas), built from TISS data. TISS is the mandatory ANS standard for electronic data exchange between hospitals and health plans. The dashboard shows average payment time, initial denial and final denial by health plan [5]. The formula is the sum of denied amounts in closed claims divided by the sum of reported amounts in those claims, times 100. Reimbursement claims, prepaid arrangements and direct supply of items are excluded. ANS itself notes that it does not receive the denial reason, so the dashboard does not separate administrative from clinical denial [5].
In addition, the ANS Health Plan Qualification Program (Programa de Qualificação de Operadoras, base year 2024) uses a denial ratio indicator. In it, 75% of the weight comes from the ratio of denied to reported amounts and 25% from the share of providers with any denial [4]. Health plans score higher the less they deny: the top score goes to those at or below the 15th percentile of their segment [4]. In negotiations, it helps to know denial feeds into the health plan's public assessment.
What the hospital should build. To compare health plans with each other:
| Precaution | Why | How |
|---|---|---|
| Same formula for all | Different bases distort the comparison | Use amount billed as the initial denial base for every health plan |
| Closed claims only | Claims without a review statement pull the rate down | Count only claims with the claim review statement received |
| Breakdown by reason | The same rate can have opposite causes | Classify using TUSS table 38 codes (TISS standard) |
| Volume next to percentage | 20% at a small health plan weighs less than 8% at the largest | Show denied amount in BRL next to the percentage |
| Rolling window | A single month swings with large batches | Use a 3- or 12-month average |
Reason codes follow the TISS standard, explained in the guide TISS standard.
How to set a realistic target
There is no official denial rate target, from ANS or from Anahp. Copying a market number as a target tends to fail for two reasons: the hospital's formula is rarely the same as the benchmark's, and health plan mix, size and complexity shift the level. The safer path is to derive the target from the hospital's own history.
- Fix the formula. Write down numerator, denominator and cutoff date for initial and final denial.
- Calculate the baseline. Use the last 12 months, by health plan, so seasonality is not mistaken for improvement.
- Isolate the avoidable share. Classify denial by reason and identify how much comes from internal failures (registration, authorization, documentation, coding).
- Set the reduction by reason, not by total. The total target is the sum of the targets by reason.
- Position against the benchmark. Compare final denial with Anahp's accepted denial only if the denominator is gross health plan revenue.
- Review every quarter, comparing quarter with quarter.
Illustrative example (hypothetical figures). A hospital has an average initial denial of 14% of amount billed over the last 12 months. The breakdown by reason shows 5 points come from registration and authorization failures. The team decides to eliminate half of those failures within 12 months: the initial denial target becomes 11.5% (14 − 2.5). For final denial, at 4% of gross health plan revenue and a 50% reversal rate, the target combines less initial denial with more appeals filed on time. The target comes from the diagnosis, not from an external number.
What we see in operations
When we read a hospital's history, the first task is to recalculate the denial rate with a single formula: the same denominator, the same measurement point and closed claims only. Without that, one health plan's initial denial rate and another's final rate end up side by side in the same report. Then we break the rate down by health plan and by reason, which is where a realistic target shows up.
Accrua OS keeps that breakdown current from remittance statements reconciled claim by claim, across every health plan claim. That lets us separate avoidable errors, fixed before submission, from disputes that need an appeal. A target by reason becomes a review rule: the check against the contract, price table and that health plan's denial history flags the error on the next claim.
For leadership, the outcome is a rate that compares month to month and with the industry benchmark under the same formula. Every action that moves the number, by an agent or a person, stays in an audit trail.
Where to start
- Write down the formula for initial and final denial, with denominator and cutoff date, and always use the same one.
- Break the rate down by health plan over the last 12 months and show the BRL amount next to the percentage.
- Classify denial by reason using TUSS table 38 codes.
- Measure the reversal rate only with denials whose appeal deadline has passed.
- Set targets by reason from the hospital's history and review them every quarter.