Delivery

Where the tables land, and how a trial works.

Free California sample on every marketplace. 30-day full-table evaluation on request.

Rails

Five ways to receive the same tables.

Same schema on every rail: one schema per product, one table per exported file, three meta columns (_dataset, _vintage, _exported_at) on every table.

Trial policy

Two stages. No card, no call required for the first.

Stage 1: free California sample

Every Snowflake, BigQuery and Databricks listing serves a one-state slice (California) of each table, free, with the full column set and the dictionary. Take it today without talking to us.

Stage 2: 30-day full-table evaluation

On request, the full tables for 30 days under evaluation terms: internal use, no redistribution, no derived product. Delivered on the rail you name. Have an analyst reconcile the HCA capex roll-up against the 10-K, or the California quarterlies against the annual cost report.

A trial does not discount the first contract. One licence per firm, not per pod.

Compliance

Public filings. No patient data.

Every table is built from public federal and state filings: CMS cost reports, enrollment and ownership rosters, Care Compare, hospital price-transparency files, NADAC, IRS Form 990, BLS OEWS, and the OIG, GSA, FDA, Treasury and state exclusion lists. There is no patient-level data anywhere in the catalogue and no scraped private panel. Personal information is limited to what those statutory registries publish about licensed providers and excluded individuals: names, NPIs, practice addresses, sanction actions and officer compensation from Form 990.

Healthparse is not a consumer reporting agency and the data is not a consumer report. It is not for FCRA-regulated decisions about an individual. Per-dataset attribution and redistribution terms are published at api.healthparse.io/v1/license.

Dictionary standard

Written against the export, not the source spec.

Each product ships a dictionary listing every exported column with its type, a description of what the field is in the source system, and notes on nullability and coverage measured on the exported data. Where a source mixes encodings or drops fields, the dictionary says so with counts. Row counts, vintages and known limits are stated per table and re-measured at each refresh.

Refresh cadences

By product.

ProductDelivery cadenceLag
NADAC weekly drug pricesWeekly, the Monday after CMS postsOne to two weeks
Medicare Advantage enrollmentMonthly, loaded on the 2nd of the following monthTwo to four weeks
Provider change eventsTwice weekly (Tuesday and Friday), weekly and quarterly by sourceDays for roster sources; quarters for cost-report sources
Price-transparency negotiated ratesMonthly snapshotsCurrent
HCRIS hospital financialsQuarterly, on the HCRIS release cycleSeveral quarters after fiscal year end
State quarterly hospital financialsQuarterly, when the state postsAbout four months
Hospital capital purchases panel (A-7)Quarterly, on the HCRIS release cycleSeveral quarters after fiscal year end
Hospital equipment purchase signalsQuarterly, on the HCRIS release cycleSeveral quarters after fiscal year end
Nonprofit hospital finances (Form 990)Monthly, as new filings postA year or more after tax year end
Facility quality and star ratingsQuarterlyWeeks after the CMS release
Exclusion and sanction screeningWeeklyDays
Medicare Part B utilizationAnnualOne to two years
Wage benchmarks (BLS OEWS)AnnualAbout a year
Provider-to-operator crosswalkMonthly, on the CMS ownership releaseWeeks