Hospital Standard-Charges Price Benchmark

An independent tool. It is not affiliated with, endorsed by, or operated by any hospital, health system or payer whose files it reads. All data is read live from each hospital's own federally mandated standard-charges machine-readable file, never scraped from any patient-facing site.

Give it a batch of hospitals (each a name you choose and the direct link to that hospital's own published CMS standard-charges file) and a list of procedure codes; get back, for each hospital and code, where that hospital's own negotiated rate sits against every other hospital in your batch -- a percentile, the batch median/min/max, and the underlying raw payer-specific charges the comparison was built from.

Why this is hard, and why that is the point

There is no central directory of the roughly 6,000 US hospitals' file locations -- CMS does not host one, so you supply the direct link per hospital. The file itself comes in at least two column shapes hospitals actually use (verified against CMS's own published reference examples): TALL, one row per code/payer/plan, and WIDE, one row per code with the payer and plan baked directly into the column name, repeated once per contracted plan a hospital has forty or more of. This Actor reads either shape automatically, streams a file that can run past a gigabyte within a disclosed byte budget (max_scan_mb), and tolerates the malformed rows a known minority of hospitals ship without losing the rest of the file.

What it returns

For each hospital and procedure code, one row per underlying payer/plan charge: hospital_label/hospital_ccn/hospital_city/hospital_state, procedure_code/code_type, payer_label/plan_label, the raw standard_charge_negotiated_dollar/standard_charge_gross/ standard_charge_cash/standard_charge_min/standard_charge_max, and the shared benchmark figures for that hospital+code: batch_percentile_rank (0 = cheapest in your batch, 100 = most expensive), batch_median, batch_min, batch_max, and batch_sample_size (how many hospitals in your batch actually had data for that code).

What it does not do, on purpose

It never charges for a single hospital's price as if it were a comparison. A procedure code with data from only one hospital in your batch is delivered free, with the reason -- add more hospitals to that code's comparison to turn it into a real benchmark. It also never guesses at a hospital's file location: you supply mrf_url directly, because 45 CFR 180.50 requires the hospital -- not a search engine -- to publish that link, and a wrong guess would silently compare the wrong file.

Finding a hospital's standard-charges file

Every hospital is required to post one on its own public website, usually linked from a page titled "Price Transparency" or "Standard Charges" (search "<hospital name> standard charges" or "<hospital name> price transparency"). File names commonly follow the pattern <EIN>_<hospital-name>_standardcharges.[csv|json].