Hospital price-transparency files, graded

US hospitals must publish machine-readable files of their standard charges. This site grades each published file — never the hospital — on whether it can be retrieved, parsed, and interpreted, with every finding citing the rule or schema requirement it rests on. JSON and CSV publications are graded under separate profiles and never pooled into one distribution.

Files graded under the CMS hospital JSON v3 profile (2026-09-12)

This cohort covers 17 machine-readable files, discovered from hospital cms-hpt.txt documents and collected in one identified run on 2026-09-12. 16 of 17 targets produced a verified body; 16 were graded and 1 recorded as not graded with the reason stated. Subjects were drawn under a stated sampling frame; the counts above describe this cohort, not hospitals as a class. How these subjects were chosen.

Collected 2026-09-12. This page was built 2026-09-19, 7 days later.

13 A1 B2 C1 not graded

What share of the drawn sample this is

the disposition of every facility in this cohort's probability stratum: each published row of that stratum, and each recorded exclusion, as a share of the sample the frame recorded. Shares describe the drawn sample; they are not weighted by bed count, discharge volume, or spend, and they are never pooled across strata or across cohorts. Every share carries its own denominator and a 95% interval computed by the wilson-score method.

DispositionCountShareInterval
published as a row of this cohort11 of 4822.9%13.4% to 36.4%
excluded: format_outside_profile32 of 4866.7%52.7% to 78.2%
excluded: txt_fetch_failed4 of 488.3%3.3% to 19.4%
excluded: txt_published_without_mrf_url1 of 482.1%0.4% to 10.8%

Checked and recorded, not graded

Targets this cohort reviewed but did not grade stay visible, with how far the review went and what it found. An absent or unreachable TXT at one origin is not evidence about the hospital's publication.

Files graded under the CMS hospital CSV v3 (Tall and Wide) profile (2026-09-12)

This cohort covers 25 machine-readable files, discovered from hospital cms-hpt.txt documents and collected in one identified run on 2026-09-12. 20 of 25 targets produced a verified body; 20 were graded and 5 recorded as not graded with the reason stated. Subjects were drawn under a stated sampling frame; the counts above describe this cohort, not hospitals as a class. How these subjects were chosen.

Collected 2026-09-12. This page was built 2026-09-19, 7 days later.

12 A2 B4 C2 D5 not graded

What share of the drawn sample this is

No share is published for this cohort: this cohort accounts for only part of its stratum, so a proportion computed here would have a denominator the frame does not recognize. The accounting this document carries covers 25 facilities.

How grading works, and what it deliberately does not claim · machine-readable comparison (CMS hospital JSON v3) · machine-readable comparison (CMS hospital CSV v3 (Tall and Wide)) · dataset.csv · its Table Schema · JSON API · written-up findings with evidence

A grade describes one published file under one stated policy on one date. It does not rank hospitals, price care, or determine compliance with 45 CFR part 180, and it is not the official CMS validator.