CMS publishes overall hospital star ratings and facility attributes through Care Compare / the Provider Data Catalog. The primary machine-readable source for facility-level general information (including overall rating) is the Hospital General Information dataset on data.cms.gov/provider-data (dataset id xubh-q36u).

There is no separate “secret” star-ratings product — the stars are fields on the public Provider Data Catalog tables.

Official API pattern (high level)

1. Call the Provider Data Catalog datastore query endpoint for xubh-q36u.
2. Filter with conditions such as state, citytown, hospital_type, or hospital_ownership.
3. Read hospital_overall_rating (and related facility fields) from each result row.
4. Optionally filter client-side for minimum rating or emergency-services flags when the catalog does not support that operator server-side.

Ready-made JSON options

• Apify Actor (free): CMS Hospital Quality / Care Compare Scraper
• HTTP API: CMS Healthcare Data API on Zyla — GET /v1/hospital-quality

Example Apify input

{
  "state": "RI",
  "minOverallRating": 3,
  "emergencyServicesOnly": true,
  "maxItems": 50
}

Example Zyla-style query

GET /v1/hospital-quality?state=RI&minOverallRating=3&limit=25

What you get back

Typical cleaned fields include CCN / facility id, facility name, address, city, state, hospital type, ownership, emergency services flag, and overall star rating.

Use cases

• Payer network adequacy and quality dashboards
• Journalists comparing hospitals by state
• PE / diligence screening of acute-care facilities
• AI agents that need structured Care Compare JSON

Related

• Nursing homes have a separate Care Compare provider-info table — see CMS Nursing Home Care Compare Scraper.
• Physician payments: CMS Open Payments Scraper.

Disclaimer

Public CMS data only. Ratings methodologies change; always cite the CMS release you used. Not medical or legal advice.

CMS Open Payments (the Sunshine Act transparency program) publishes payments and transfers of value from drug and device manufacturers to physicians, non-physician practitioners, and teaching hospitals. Data is released as large public files and also exposed through the official Open Payments datastore query API on openpaymentsdata.cms.gov.

If you only need filtered rows — for example consulting fees in one state, or payments tied to a specific manufacturer — you usually do not need to download the entire yearly CSV.

Official API pattern (high level)

1. Resolve the dataset UUID for the program year and payment type (general, research, or ownership).
2. POST (or GET) to the datastore query endpoint for that dataset with limit, offset, and conditions (state, NPI, nature of payment, etc.).
3. Map the returned rows into the fields your pipeline needs.

CMS documents the metastore and datastore query interfaces on the Open Payments site. Always prefer the federal source of truth for schema changes.

Faster path: ready-made JSON Actors / APIs

If you want filtered JSON without writing the client yourself:

• Apify Actor (free): CMS Open Payments Scraper — wraps the official API; filter by state, NPI, manufacturer, nature, amount; export JSON/CSV.
• HTTP API: CMS Healthcare Data API on Zyla — GET /v1/open-payments with plan key.

Example Apify input

{
  "paymentType": "general",
  "year": 2024,
  "recipientState": "RI",
  "natureOfPayment": "Consulting Fee",
  "maxItems": 100
}

Example Zyla-style query

GET /v1/open-payments?recipientState=RI&year=2024&natureOfPayment=Consulting%20Fee&limit=25

Common pitfalls

• Unfiltered dumps of a full program year are huge. Always pass at least one selective filter (state, NPI, manufacturer, or nature).
• Payment amounts are often stored as text in the source files; normalize to numbers in your client.
• Research and ownership files share many columns with general payments but not all filters apply equally.

Related datasets

• CMS Hospital Quality / Care Compare Scraper
• CMS Medicare Physician Scraper

Disclaimer

This article describes public federal open data. It is not legal advice. Review CMS terms and applicable privacy law for your use case.