# Data schema — US Hospital Prices

One normalized record per (hospital, code, price type):

```json
{"cpt": "0112U", "type": "cash", "price": 559.20}
```

| Field  | Type   | Description |
|--------|--------|-------------|
| `cpt`  | string | CPT or HCPCS billing code, validated against an OMOP CPT4/HCPCS vocabulary (26,738 codes) |
| `type` | string | `gross` (hospital gross charge) or `cash` (cash / discount price) |
| `price`| number | USD amount for that code at that hospital |

Files are one JSONL per hospital, organized by state: `data/{STATE}/{CCN}.jsonl`,
where CCN is the 6-character Medicare Certification Number.

## Coverage (2026-09-02)

- **3,768 hospitals** across all 50 states
- **30,695,336 price rows**
- Daily snapshots since 2026-08-28 (git history = the time trace)
- Publish metadata for 3,776 hospitals: last file modification date,
  content length, and when we last saw the file live

## Provenance

Source: hospital price-transparency machine-readable files, published under
the federal Hospital Price Transparency Rule (45 CFR §180). Hospital
metadata/URL index from hospitalpricingfiles.org. Normalization code is CC BY 4.0.

## Notes & limitations

- `gross`/`cash` reflect what hospitals publish as standard charges;
  payer-specific negotiated rates are not included in v1.
- Some hospitals' files 404 or move without notice; the snapshot history
  preserves the last good copy.
- Codes are CPT/HCPCS only; DRG and per-day room rates are out of scope in v1.