Methodology

Last updated 2026-08-10.

Every figure on this site comes from a public file published by the hospital itself or by a federal agency. This page names each of those files, describes exactly what is done to them, and says where the results are weakest. It is written so that you do not have to take anything here on trust: each number can be checked against its original source.

The short version. Hospitals are required by federal law to publish what they charge. Those files are technically public and practically unreadable — hundreds of thousands of rows, inconsistent formats, and codes rather than names. This site reads them, drops the rows that are not really prices, and lets you look one up. It adds nothing of its own to the numbers.

Where hospital prices come from

Under the CMS Hospital Price Transparency Rule (45 CFR 180.50), most US hospitals must publish a machine-readable file of their standard charges — gross charge, cash price, and the rate negotiated with each insurer — and must update it at least annually.

The rule also requires each hospital to place a file called cms-hpt.txt at the root of its website saying where that price file lives. That is how files are located here: read cms-hpt.txt, follow it to the price file, confirm the hospital's name or tax ID matches. Where that fails — a hospital that never posted the pointer, or posted one aimed at a URL that has since moved — the file is researched by hand, and the hospital is recorded as unresolved until it is found.

Source: each hospital's own published standard-charges file. The exact file used for a hospital, and how old it is, is shown on that hospital's page.

What is removed, and why

Hospital files contain a great many rows that are not usable prices. Three categories are excluded from the figures shown:

Placeholders. A rate published at $1.00 or less alongside a real gross charge is a filler value, not a price. These are dropped. Drug and per-unit rows are exempt from that rule, because a genuine per-vial price can legitimately be under a dollar.
Bundled-payment tokens. Some files express a rate as a percentage, a case rate, or a reference to a contract term rather than as a dollar amount. These are excluded from price statistics rather than shown as if they were dollars.
Codes that do not describe a procedure. Revenue codes name a department ("pharmacy", "room and board") and drug codes name a substance. Both are real billing categories and neither answers "what will this operation cost", so they are not shown as the price of a procedure.

A hospital whose file consists mostly of those categories will show few or no prices here. Its page says so explicitly, and says which case applies, because "no price shown" and "this hospital published nothing" are different statements and only one of them is usually true.

The most important caveat on this site

A hospital's price file does not include the surgeon. It covers the facility's own charges. The surgeon, the anaesthesiologist, the radiologist and the pathologist are usually independent practices that bill you separately, and they appear in no hospital price file at all — not as an omission, but by construction.

So a hospital price is never the whole cost of an operation, and the professional share can be the larger half.

This gets harder still, because some hospitals publish the professional side under the same code with no field marking which is which. Measured on CPT 47562 (laparoscopic gallbladder removal) in Georgia in July 2026: one hospital published $19,238 and another $277. Both are honest numbers. They are prices for different things.

Two guards follow from that. Rows a hospital explicitly declares professional are labelled Physician fee. Rows that sit far below the median of the unlabelled ones are labelled May not be comparable, because they are probably a professional fee that was not declared as one. Neither label is a judgement about the hospital — it is a warning not to compare two numbers that are not measuring the same thing.

For the same reason this site does not rank hospitals against each other on price, and does not show "good value" or "priced above market" badges. The underlying data cannot support that claim honestly, so it is not made.

How the typical price and range are calculated

Where a range is shown, it is the 10th to 90th percentile of the published rates for that code, not the lowest and highest. A single miskeyed row at $2 or $2,000,000 would otherwise define the range, and in files this size there is always at least one. The percentile version describes where prices actually sit.

"Lowest reported" columns, where shown, are the literal minimum and can still be pulled down by a single odd row. They are labelled as reported figures for that reason.

The Medicare comparison

A price of $22,560 means nothing on its own. Shown as a multiple of what Medicare pays the same hospital for the same admission, it becomes interpretable — that comparison is the standard one in health policy research.

The Medicare figure here is computed from CMS's published Inpatient Prospective Payment System files: the operating base payment, which is (labour share × that hospital's wage index + non-labour share) × the DRG weight.

It is a base rate, not a bill. Medicare's actual payment adds adjustments for disproportionate-share, medical education, outlier cases and new technology; none of those are computed here. What is computed is the figure every one of those adjustments is applied to.

Source: CMS Inpatient Prospective Payment System final rule files (MS-DRG weights, wage index). Inpatient only — the outpatient equivalent is behind a licence gate, so no outpatient Medicare comparison is offered rather than a worse one being substituted.

Physician fees, where shown

Because the hospital file omits the professional side entirely, a separate national figure is used to indicate its rough scale: the average Medicare-allowed amount for that code across practitioners.

This is a Medicare average, not what your surgeon charges, and it is never added to a hospital price to produce a single total. Combining a real published rate with a national average would produce a number that looks authoritative and is not true of anywhere. The two are shown separately or not at all.

Source: CMS, Medicare Physician & Other Practitioners — by Provider and Service (average allowed amount).

Quality and outcome measures

Prices without quality invite the wrong conclusion, so each hospital page carries CMS's own published measures: mortality and patient-safety outcomes, 30-day readmissions, healthcare-associated infection ratios, emergency department timeliness, and patient-survey (HCAHPS) star ratings.

These are CMS's measurements and CMS's ratings, reproduced without adjustment. Nothing here is a Clearcharge rating of a hospital, and no measure is combined with another into a composite score.

Source: CMS Provider Data Catalog — Complications and Deaths; Unplanned Hospital Visits; Healthcare Associated Infections; Timely and Effective Care; Patient Survey (HCAHPS).

Financial assistance eligibility

Non-profit hospitals must have a written financial assistance policy to keep their tax exemption, and must report its income thresholds to the IRS on Form 990, Schedule H. Where a hospital's own thresholds are shown, they are read from that filing — so the page states that hospital's actual cut-off rather than a typical one.

Attribution is checked rather than assumed. A single tax filing often covers several hospitals, and a filing found under one hospital's tax ID sometimes describes a different facility entirely. Each filing is confirmed against its Part V Section A facility list, which names the hospitals that organisation actually operates. Thresholds that fail that check are not shown, even though the filing exists. At least one Georgia hospital's apparent filing was confirmed to describe a hospital 180 miles away, and it is excluded.

Where no verified filing exists, the page says what is typical and labels it as typical. It does not present a general figure as that hospital's policy.

Source: IRS Form 990 Schedule H, Part V. Government and for-profit hospitals generally do not file one, so their pages carry the general guidance instead.

Nearby health centers

Hospital pages list federally funded health centers near that hospital. These charge on a sliding scale set by household income rather than being free, and the page says so.

Sites are grouped by operator and address rather than by organisation. A single provider frequently runs a dozen genuinely different locations, and collapsing them by operator would delete the nearest one — which, for someone without a car, is the only thing the list is for. Purely administrative sites are excluded, as no care is delivered at them.

Free and charitable clinics — volunteer-staffed and genuinely free — are a separate class of organisation, are not in this dataset, and are not listed here. The page states that rather than implying the list is complete.

Source: HRSA Health Center Program service delivery sites.

Where this is weakest

Honest limits, rather than a list of features:

Coverage is partial. Not every hospital publishes a usable file, and not every published file can be read. The coverage page lists every hospital and its current status, including the ones with nothing.
Prices go stale. Hospitals must update annually; some files are older than that. Each hospital page shows the age of the file it is using.
Code matching is imperfect. Procedures are identified by billing code, and the same operation can be coded several ways depending on approach and complexity. A code is a good match for a procedure, not a perfect one.
A negotiated rate is not your bill. What you owe depends on your deductible, your out-of-pocket maximum, and how much of each you have already used. This site cannot know any of those and does not estimate them.
Comparisons across hospitals carry the facility/professional problem above. Where two numbers may not be measuring the same thing, the page says so instead of quietly averaging them.

Corrections

If a number here looks wrong, it is usually one of three things: the hospital published it that way, the parser misread their format, or the code is matched to the wrong procedure. The second and third are ours to fix, and reports of them are the single most useful thing anyone sends us.

The hospital name, the procedure or code, and what the page showed you is enough to trace it. Write to [email protected]. Corrections to a displayed figure are made at the source and take effect on the next data load; if a figure is wrong and cannot be immediately fixed, it is removed rather than left up.

Who is responsible for this

Clearcharge is run independently by one person, and is not operated by, funded by, or affiliated with any hospital, hospital system, insurer, or healthcare provider. No hospital or insurer pays for placement here, and no result is ever ordered or ranked by payment.

The operator is not a clinician and does not offer medical or financial advice. That is deliberate rather than a limitation: this site makes no clinical claim and asks for no trust in anyone's professional judgement. Every figure is a public disclosure reproduced with its arithmetic stated, which is why this page exists — the numbers are checkable against their sources by anyone, and that is a stronger guarantee than a credential.

Correspondence, including from journalists and patient-advocacy organisations, reaches a monitored address at [email protected]. See About for how the site is paid for.