How we source MRI scan cost figures

There are two kinds of number on this site and we label them differently, because they are not worth the same. Measured figures come from a document we retrieved ourselves: a hospital's own federal price file, a provider's published rate card, a government fee schedule. Each is quoted with the source and the date we read it, and you can check it. Estimated figures are our own modelled bands, used where nobody publishes anything checkable. They carry no source because there is none, and we say so where they appear rather than dressing them in a citation.

If a figure on this site names a publisher, we retrieved that figure from that publisher on the date given. If it does not, it is an estimate. We do not attribute a number to an organisation that did not publish it.

Measured sources

• US hospital price files, parsed by us. Every US hospital must publish a machine-readable list of its standard charges under 45 CFR 180. We download and parse those files ourselves and publish the gross charge, the discounted cash price and the discount side by side, each row stamped with the date of the file it came from. We have now read 1,768 of the 3,107 acute-care hospitals in the United States across ten imaging codes, every state and territory attempted (see New Jersey hospital imaging prices for one state hospital by hospital). Where a hospital publishes a gross charge but no cash price we record it as not published; we never interpolate one. The 1,339 we could not read are counted, not ignored: 522 sit behind a firewall that refuses automated requests, 593 publish no file we could locate, and 189 returned something that was not a machine-readable file. None of them is filled in from a neighbour or a third-party price site.
• Provider rate cards we have retrieved. Published per-scan prices from named providers, each quoted with a retrieval date: Vista Health (UK national list prices), Prenuvo and Ezra by Function (US whole-body screening), RadiologyAssist (US booking-programme rates by city), HonorHealth (published average pricing sheet), Gateway Diagnostic Imaging and Touchstone Medical Imaging (market claims, which we quote as claims, not as prices). Where a provider publishes nothing, we say it quotes on enquiry rather than inventing a figure for it.
• Government schedules and statistics. CMS Medicare Procedure Price Lookup for MRI by CPT code, CMS Part B deductible and the ACA annual limitation on cost sharing, and NHS England Diagnostics Waiting Times and Activity (DM01) for MRI waits. These are what a payer pays or what the state records, not what a cash patient is charged, and we label them that way.
• CMS NPPES registry. The imaging-centre directory is built from the CMS National Plan and Provider Enumeration System NPI file. NPPES is an identifier registry, not a licence register, and its records are self-reported, so we publish the snapshot date and tell readers to confirm before travelling.

Which states are measured, and which are still estimates

The hospital band on this site is no longer a model. In 48 states and territories it is read out of the hospitals' own machine-readable standard-charges files, published under 45 CFR 180. Every page that shows a hospital figure for one of them shows the measured band, says how many hospitals it came from, and carries the gross charge alongside the cash price. The remainder had too few readable hospitals to publish a range from, and keep an estimate that is labelled as one. No page shows both.

Every other state still carries an estimate for the hospital band, and every state carries an estimate for the freestanding imaging-centre band: no imaging operator publishes a per-state cash price and we have not measured one. Those are our own modelled ranges and they carry no source.

We know the limits of that, because we can now see them. In New Jersey the measured cash prices run well outside the band we had modelled in both directions: a brain MRI is $264 at the cheapest hospital and $10,725 at the most expensive, against the modelled range we used to publish, which was $2,500 to $6,800. That is why the estimates are being replaced rather than re-sourced, and why a measured figure always takes precedence over an estimate on the same page.

Three things stay out of the measured bands on purpose. Hospitals that publish their chargemaster unchanged as a “cash price” are excluded and their ceiling is shown separately, because folding them in would hide the difference between a real discount and a restatement. Hospitals whose file we could not read are named and counted rather than filled in from a sibling hospital or a third-party price site. And a measured band is never multiplied by a modelled body-part factor to produce a figure for a scan we did not read, so abdomen, cardiac and whole-body MRI remain estimates everywhere.

What we deliberately do not publish

• Specific named-clinic rate cards. Clinics often quote on enquiry. We publish published-rate bands and aggregator data, not named-clinic specific quote pricing.
• Insurance referral commissions. We do not earn commission on insurance referrals or clinic bookings.
• Personal patient data. Calculators run entirely in your browser. Your scan-type / region inputs are not transmitted, logged, or stored. This site is not a clinical resource and does not provide medical advice.

Editorial position

This site is operated by Digital Signet, an independent AI-development studio. Digital Signet does not sell MRI scans, does not act as a clinic-finder service, does not broker private health insurance, and does not accept paid placements from Bupa, Spire, Nuffield, Circle Health Group or any other healthcare provider. See /about for the operator and the wider network.

Editorial direction is set by the Digital Signet editorial team. Drafts are produced via Digital Signet's autonomous AI development methodology and reviewed against the editorial framework before publication.

Contact

For methodology questions, corrections, or scenarios that don't fit cleanly: oliver@digitalsignet.com.