No single data source establishes the usual and customary rate for every out-of-network claim. The right source depends on the reimbursement formula in the plan document and on whether federal or state surprise-billing rules apply to the claim. Plans may pay a percentile of claims-based usual, customary, and reasonable (UCR) charges, a stated multiple of Medicare, or another plan-defined amount. Public price transparency data and commercial negotiated rates add market evidence. Each one measures something different, so keep them in separate columns.
This is an analytical guide for payment integrity, repricing, benefits, and advocacy teams. It is not legal advice.
Start with the plan's stated method, then use the evidence that matches it. A UCR-based plan needs UCR data. A Medicare-based plan needs the right Medicare fee schedule. Market data helps you test whether either result is reasonable.
Healthcare.gov defines UCR as the amount paid for a medical service in a geographic area based on what providers in the area usually charge for the same or similar service, and notes that the UCR amount is sometimes used to determine the allowed amount.
Plans and their administrators typically license claims-based charge databases and pay at a chosen percentile. Two plans using different databases or percentiles can reach very different UCR amounts for the same claim.
Claims data has real strengths. It reflects what was actually billed and paid across a large population, and it has a long history behind it. It also has limits worth understanding before you rely on it for a specific claim:
Many plans set out-of-network reimbursement as a percentage of what Medicare would pay. Illustrative example: a plan paying 140% of Medicare, for a service Medicare pays $100, has a reimbursement basis of $140.
Medicare's appeal is that its method is published, geographically adjusted, and updated every year. Use the payment system that matches the claim: the Physician Fee Schedule for professional services, with the correct locality and facility or non-facility rate, and the outpatient, ASC, or inpatient systems for facility claims. Use the fee schedule year that matches the date of service.
The plan document says which method applies. The explanation of benefits or remittance advice shows what the plan actually allowed and paid on this claim. Neither is market data, but no analysis is complete without them. A benchmark that ignores the plan's own formula answers the wrong question.
CMS requires most group health plans and issuers to publish machine-readable files with in-network rates, plus out-of-network allowed amounts and historical billed charges. The out-of-network file is the most direct public evidence of what plans have actually allowed out of network.
It is also thin. Under the current rules, each file covers a 90-day window, and plans must omit any provider and service with fewer than 20 claims. Treat it as a cross-check when it covers your code and market, not as a standalone benchmark.
In-network negotiated rates show what payers and providers in the same market have actually agreed to for the same service. They are the largest and most detailed public pricing source, and they reflect current contracts rather than past charges.
They are market evidence, not a UCR amount, an allowed amount, or a legally required payment. Their value is showing where a billed charge or a proposed payment sits against real contract terms.
They are built differently, and the difference matters most when a specific claim is being challenged. Claims data summarizes what happened in the past across many payers. Transparency in Coverage data shows what each payer has agreed to pay each provider now.
| Question | Claims-based UCR data | Transparency in Coverage data |
|---|---|---|
| How current is it? | Historical claims, which can be one to several years old by the time they are used | Plans must update their files monthly under the current rules |
| What does it show? | What was charged or paid in the past | What payers have contractually agreed to pay |
| How granular is it? | Pooled percentiles across payers and providers | Rate by rate: payer, plan, network, and provider NPI and TIN |
| How deep is coverage for one code and market? | Can be thin, often filled by widening geography or blending services | Deep for in-network rates. Thin for out-of-network allowed amounts |
| Can you trace a number to its source? | Method set by the data provider | Every rate traces back to a public source file |
Neither replaces the plan's own formula. Many teams use both. But when a billed charge has to be defended or challenged, current contract terms at the provider level are harder to argue with than a historical market average.
They answer different questions, so they belong in separate fields in any reimbursement review. Mixing them is the fastest way to lose a dispute.
| Term | Where it comes from | What it measures | When to use it |
|---|---|---|---|
| Billed charge | The provider's claim | What the provider asked for | The number being reviewed, not a benchmark |
| UCR amount | A claims-based charge database chosen by the plan | A percentile of what providers in an area charge | When the plan reimburses on a UCR basis |
| Plan allowed amount | The plan's formula, shown on the EOB | The most the plan recognizes for this claim | Always. It is the plan's actual position |
| Medicare amount | CMS fee schedules | What Medicare pays for the same service, setting, and locality | Medicare-based plans, and as an independent reference |
| Commercial negotiated rate | Payer and hospital price transparency files | What payers have contractually agreed to pay in-network providers | Testing a charge or payment against the local market |
| Published out-of-network allowed amount | Payer Transparency in Coverage files | What a plan allowed out of network, where claim volume is high enough to report | A cross-check when it covers the code and market |
| Qualifying payment amount (QPA) | The plan, under the No Surprises Act methodology | Generally the plan's indexed median contracted rate | Only for claims subject to the No Surprises Act |
State surprise-billing laws can set their own payment standards. Check the state rules separately if the plan is fully insured.
Compare like with like. Hold the service code, billing setting, payer or network, provider type, geography, rate type, and file date constant.
For the commercial-rate step in detail, see how to benchmark commercial negotiated rates. For a worked example on one service, see finding a fair price for an out-of-network MRI.
Transparency in Coverage files provide in-network negotiated rates, out-of-network allowed amounts, and historical billed charges. They need normalization before they support a like-for-like comparison. Raw files mix rate types, settings, provider groupings, and plan structures, and a single payer can publish thousands of them.
Hospital price transparency files add each hospital's payer-specific negotiated charges and discounted cash price, which matters when the claim is from a hospital.
All of it is evidence of published pricing. None of it is an automatic payment rule for a particular claim.
Why file date and rate type change the result: contracts renew and files refresh, so a rate from one file date may not match the date of service. And a file can report a dollar amount, a percentage of billed charges, or a per diem for the same code. Put those in one median and the result describes nothing.
For claims covered by the No Surprises Act, the dispute runs on defined notices, open negotiation, QPA disclosures, and the Federal independent dispute resolution (IDR) process. Preserve every claim-specific document.
The QPA has a narrow legal meaning. Under the CMS methodology, it is generally the plan's median contracted rate as of January 31, 2019 for the same or similar service, in the same insurance market and geographic region, indexed forward for inflation. It sets cost sharing for certain protected services, and certified IDR entities must consider it.
A commercial rate benchmark is not the QPA. It is built from different contracts, a different date, and a different method.
For a claim subject to the Act, the plan's initial payment or notice of denial must include the QPA and contact information for open negotiation. A May 2026 final rule adds further disclosures, including the plan's legal business name and IDR registration number, and claim codes showing whether the claim is subject to the surprise billing provisions.
A general out-of-network claim review and a claim eligible for Federal IDR are different processes. Timelines and eligibility rules change, so check the current requirements at CMS before acting on a specific claim.
A defensible file shows the exact service, setting, geography, plan method, payment notice, and market references used.
Gigasheet gives teams nationwide negotiated-rate data and Medicare reference points, filterable by payer, provider, billing code, location, rate type, and file date. It is healthcare price transparency data and tools built for claim-level work.
Use it for repricing research, claim benchmarking, network comparison, and evidence preparation. Gigasheet supplies market evidence. It does not replace plan terms, calculate a payer's QPA, determine the legally required payment, or guarantee an IDR outcome.
Read the full five-step framework in our post on how to determine fair reimbursement for out-of-network claims, or Request a Demo.
Last reviewed November 20, 2026.
Healthcare.gov defines UCR as the amount paid for a service in a geographic area based on what providers there usually charge for the same or similar service. Plans apply it through their own data source and percentile, so it is not a single national price.
No. Plans and their administrators choose their own sources and percentiles, so two plans can produce very different UCR amounts for the same claim.
Price transparency data. Claims-based benchmarks are built from historical claims that can be one to several years old by the time they are used. Plans must update their Transparency in Coverage files monthly under the current rules, and each rate reflects a contract in effect.
They pool payments across many payers and providers, and a specific code and local market may have too few claims to report. Sources often widen the geography or blend services to fill the gap, which moves the figure away from the claim being reviewed.
Yes. Many plans pay a stated percentage of the Medicare amount for the same service, setting, and locality. The plan document names the percentage.
A negotiated rate is what a plan has contractually agreed to pay an in-network provider. An out-of-network allowed amount is the most a plan will pay a provider it has no contract with.
Yes. Plans publish out-of-network allowed amounts and historical billed charges, but they must omit any provider and service with fewer than 20 claims in the 90-day reporting window, so coverage is thin.
They are strong market evidence, not proof of a required payment. They show where a charge or proposed payment sits against what payers have actually agreed to pay, while plan terms and applicable law set the payment.
CMS describes the QPA as generally the plan's median contracted rate as of January 31, 2019 for the same or similar service in the same market and region, indexed for inflation. It applies only to claims subject to the No Surprises Act.
No. It is a regulatory figure with its own method and base date. Certified IDR entities must consider it, but it is not the same as a current commercial market benchmark.
The claim and itemized bill, EOB or remittance, plan language, codes and place of service, provider identifiers, the plan's reimbursement method, any QPA and open negotiation notice, and your benchmark inputs with source files and dates.
Generally when the claim is subject to the No Surprises Act, such as most emergency services, certain non-emergency services at in-network facilities, and air ambulance. Check current CMS eligibility rules before filing.
Hold the code, setting, provider type, rate type, and file date constant, use one geographic definition, and report the median, percentile range, and record count for each comparison.