Methodology — FairRate

Sources

Insurer in-network files: The federal Transparency-in-Coverage disclosures. For each insurer we take the flagship commercial network and read the professional negotiated rates by tax ID, NPI, and billing code. These are multi-gigabyte compressed files; we stream them rather than loading them, and refresh monthly as insurers republish.
Hospital price transparency: Hospital machine-readable files under 45 CFR 180. The 2026 format adds actual-paid medians and percentiles derived from remittance data, plus claim counts — used to validate negotiated rates and confirm real billing volume behind them.
Medicare Physician Fee Schedule: The final CY2026 PFS, computed at each metro's own Medicare locality (carrier 04412) from the national relative value file and that locality's geographic indices. This is the reference every commercial contract is quoted against.
NPPES registry: The national NPI registry, used to restrict every benchmark to physicians carrying the relevant specialty taxonomy and practicing in the target metro's counties.

From raw file to benchmark

The filters and transforms applied to every rate, in order.

Professional fees only. Institutional and facility-side line items are excluded; a practice negotiates its professional fee schedule, so that is what we benchmark.
Place of service is respected. Procedure codes are benchmarked in the setting they are actually performed (ambulatory surgery / hospital outpatient); office visits are benchmarked at POS 11. The Medicare comparison is matched to the same setting.
Specialty and geography. Rates are kept only for NPIs with the relevant taxonomy in the target counties, filtering out unrelated providers who happen to appear in a group's filing.
Deduplication. Identical tax-ID × rate observations are collapsed so a single contract published across many plan files counts once, not hundreds of times.
Percentiles, not averages. Each code × insurer is summarized as the 10th / 50th / 90th percentile of surviving rates — robust to outliers in a way a mean is not.

The corrections that matter

Real disclosure files are messy in specific, repeatable ways. Reading them naively produces confident, wrong numbers. Three we explicitly correct for:

Out-of-market contamination

National insurer files list a group's rate if any one of its NPIs matches the target — which pulls in out-of-state contracts held by large multi-market groups. We require meaningful local overlap before a rate counts toward a metro benchmark, so a Texas number reflects Texas practices.

Per-modifier fractional rows

At least one major insurer publishes the same procedure as separate rows for each billing modifier — discontinued, reduced, or split-care versions, each a fraction of the full fee. Left in, they collapse a median toward zero. We benchmark the full unmodified professional fee.

Ghost rates

A file may list a rate for a tax ID that never bills the code. We cross-reference the specialty registry and, where available, hospital claim-count disclosures to keep the benchmark grounded in rates that reflect real billing.

Scale and freshness

Because the files run to gigabytes and change monthly, the pipeline is built to re-run cleanly on each refresh — so a benchmark reflects the current published contracts, not a stale one-time snapshot.

Questions we get

Is any of this private or patient data?

No. Every input is a public disclosure an insurer or hospital is legally required to publish. There is no claims data, no remittance detail tied to individuals, and nothing that touches protected health information.

How current are the figures?

Insurers republish their in-network files monthly. Benchmarks are rebuilt against the latest files each cycle, so a packet reflects the contracts in force at the time it is produced.

Which specialties and metros do you cover?

Today: gastroenterology, ophthalmology, dermatology, and otolaryngology (ENT), across the Dallas–Fort Worth, Houston, San Antonio, and Austin metros — with gastroenterology the most complete. Procedure-heavy specialties come first because they concentrate revenue in a small set of codes, so a compact benchmark covers most of the book, and their procedures occur in settings where hospital disclosures provide an independent cross-check.

Can you benchmark my specialty or metro?

The pipeline generalizes to any specialty and market where the disclosures exist, and we are actively adding both. Tell us what you need.