Nursing Home Deficiency Citations (CMS) Scraper

Pulls every nursing home health deficiency citation CMS publishes, about 418,000 citations across the last three inspection cycles at every certified nursing home in the US, into clean JSON records. One item per citation: facility, survey date, F-tag with CMS's plain-English definition, scope and severity already decoded, correction status, and a link to the facility's Care Compare page.

CMS puts this data behind a browser table and a monthly CSV. This Actor filters it at the source by state, CCN, facility name, survey date, scope/severity, tag and inspection cycle, so a nursing home inspection history for one facility or every immediate-jeopardy citation in one state is a single cheap run.

Who uses this data

• Elder-law and nursing home abuse attorneys screening a defendant facility or building its citation history. Filter scopeSeverityCodes to J, K, L and you have every immediate-jeopardy finding in your state, newest survey first, each with CMS's definition of the tag cited.
• Long-term care liability underwriters and reinsurers pricing a facility or a chain. Pull a CCN or a provider-name pattern and you get the full three-cycle citation record, including scope, severity and whether the deficiency was corrected.
• Nursing home brokers, lenders and REIT analysts running diligence on a facility before a sale or refinance.
• Senior placement services and care advisors vetting homes for a family. providerUrl on every record links to the facility's Care Compare page.
• Researchers, journalists and state agencies studying enforcement patterns, tag frequency, or how quickly deficiencies are corrected.

What you get

One dataset item per citation, most recent survey date first. A real record:

{
  "source": "https://data.cms.gov/provider-data/dataset/r5ix-sfxw",
  "attribution": "Centers for Medicare & Medicaid Services (CMS)",
  "citationId": "505154|2026-06-22|F-0656|D|1|S|",
  "ccn": "505154",
  "providerName": "TACOMA NURSING AND REHABILITATION CENTER",
  "providerAddress": "2102 SOUTH 96TH STREET",
  "city": "TACOMA",
  "state": "WA",
  "zipCode": "98444",
  "providerLocation": "2102 SOUTH 96TH STREET,TACOMA,WA,98444",
  "providerUrl": "https://www.medicare.gov/care-compare/details/nursing-home/505154/?state=WA",
  "surveyDate": "2026-06-22",
  "surveyType": "Health",
  "inspectionCycle": 1,
  "deficiencyTag": "F-0656",
  "deficiencyPrefix": "F",
  "deficiencyTagNumber": "0656",
  "deficiencyCategory": "Resident Assessment and Care Planning Deficiencies",
  "deficiencyDescription": "Develop and implement a complete care plan that meets all the resident's needs, with timetables and actions that can be measured.",
  "scopeSeverityCode": "D",
  "severity": "No actual harm with potential for more than minimal harm that is not immediate jeopardy",
  "scope": "Isolated",
  "isImmediateJeopardy": false,
  "deficiencyCorrected": "Deficient, Provider has no plan of correction",
  "correctionDate": null,
  "isStandardSurvey": true,
  "isComplaintSurvey": false,
  "isInfectionControlSurvey": false,
  "isUnderIdr": false,
  "isUnderIidr": false,
  "processingDate": "2026-07-01",
  "fetchedAt": "2026-08-23T11:40:03.834Z"
}

severity and scope are the decoded meaning of scopeSeverityCode, taken from CMS's own scope/severity grid (Table 1 of the Five-Star Technical Users' Guide): severity runs from "No actual harm with potential for minimal harm" up to "Immediate jeopardy to resident health or safety", and scope is Isolated, Pattern or Widespread. A code outside that grid is passed through with severity and scope left null rather than guessed.

What deficiencyDescription is, and is not. It is CMS's plain-English definition of the F-tag, the same wording for every citation of that tag. It is not the surveyor's facility-specific statement of deficiencies (the CMS-2567 narrative), which CMS does not publish in this dataset. Verified: across 126 distinct tags in a 1,500-row sample, no tag carried more than one distinct description. Use it to know what was cited; get the 2567 from the state survey agency.

Dates are ISO 8601 calendar dates exactly as CMS publishes them; a value that is not a real date becomes null and is counted in the run summary, never repaired into something plausible. Missing values are null, never an empty string. Y/N columns become real booleans. Records are deduplicated on citationId, a composite of facility, survey date, tag, scope/severity, cycle and survey kind, because CMS publishes no row id of its own.

A run summary is written to the OUTPUT key: the filters actually sent to CMS, how many citations matched, how many HTTP requests it took, and whether a charge limit stopped the run.

Input

Everything is optional. Run it with no input and you get the 100 most recently surveyed citations nationwide.

Every filter below is applied by the CMS API itself, not after download, so a narrow search is also a fast and cheap one.

Field	Type	Default	Notes
states	string[]	-	Two-letter postal codes, exact. ["NY","NJ"].
ccns	string[]	-	CMS Certification Numbers, exact. Keep the leading zero: ["015009"].
providerNameContains	string	-	Case-insensitive substring on the facility name, e.g. MANOR.
surveyDateFrom	ISO date	-	Surveys on or after this date.
surveyDateTo	ISO date	-	Surveys on or before this date.
scopeSeverityCodes	string[]	-	Letters A to L, exact. ["J","K","L"] is immediate jeopardy.
tagPrefix	string	-	Single letter. Every citation here is an F tag; see Limitations.
tagNumbers	string[]	-	Four-digit tag numbers without the prefix, e.g. ["0689"].
inspectionCycles	string[]	-	1, 2 or 3, where 1 is the most recent cycle.
surveyKind	enum	any	any, standard, complaint or infectionControl.
maxResults	integer	100	1 to 50000. This is your cost cap: you pay per delivered record.
includeRaw	boolean	false	Attach a raw object with the CMS columns under their original names.

Example inputs

Immediate jeopardy citations in one state (plaintiff screening, underwriting):

{
  "states": ["NY"],
  "scopeSeverityCodes": ["J", "K", "L"],
  "maxResults": 500
}

Full citation history for one facility (case file, diligence on a sale):

{
  "ccns": ["505154"],
  "maxResults": 1000
}

Recent abuse and neglect citations nationwide (F-0600 is the free-from-abuse tag; F-0689 is accidents and supervision):

{
  "tagNumbers": ["0600", "0689"],
  "surveyDateFrom": "2026-01-01",
  "surveyKind": "complaint",
  "maxResults": 2000
}

Pricing

$0.003 per record; you pay only for records delivered. One charge per dataset item, nothing else. Set maxResults and you have set your bill: 1,000 records is $3.00. Filters are applied at the source, so a narrow search costs less than a broad one, and a filter that matches nothing costs nothing.

Data source, refresh and licence

Data comes from the Centers for Medicare & Medicaid Services (CMS) Provider Data Catalog, dataset "Health Deficiencies":

• Landing page: https://data.cms.gov/provider-data/dataset/r5ix-sfxw
• Read through the datastore query API: https://data.cms.gov/provider-data/api/1/datastore/query/154d0bc8-8214-5dd3-a552-f3aa2e034177
• Field definitions: Nursing Home Data Dictionary, Table 7

Refresh: CMS refreshes this dataset monthly. At the time of writing the published file was NH_HealthCitations_Jul2026.csv, last modified 2026-07-01, with CMS's catalog giving 2026-08-26 as the next expected update. Every record carries CMS's own processingDate so you can see which monthly cut you received. A monthly schedule tracks the source; running it more often will return the same data.

Licence: This is CMS provider data, public domain as a US government work. CMS's catalog at https://data.cms.gov/data.json states the licence for its datasets as https://www.usa.gov/government-works; the Provider Data Catalog sub-catalog (https://data.cms.gov/provider-data/data.json) publishes no per-dataset licence field, so that parent-catalog licence is what this listing relies on.

Attribution: Centers for Medicare & Medicaid Services (CMS), carried on every record in the attribution field. This Actor is not affiliated with or endorsed by the Centers for Medicare & Medicaid Services. No login, API key or account is involved in reading this data.

https://data.cms.gov/robots.txt declares crawl-delay: 10 and its User-agent: * rules do not disallow /provider-data/api/. The Actor makes strictly sequential requests with a 10 second gap between them, sends a descriptive User-Agent, and caches nothing between runs. Pages are requested at the API's own maximum of 1,500 rows, so a 100-record run is a single request and even a 50,000-record run is 34.

Limitations

• Health deficiencies only. Every citation here carries the F prefix, confirmed against the live data: a query for any other prefix returns zero rows. Fire safety and life safety citations (K and E tags) are a separate CMS dataset and are not included.
• Three inspection cycles, not three calendar years. CMS's catalog blurb says "the last three years", but the data dictionary says "the last three cycles", and the data agrees with the dictionary: survey dates observed on 2026-08-23 ran from 2017-03-23 to 2026-06-22. A facility surveyed infrequently can carry an old citation.
• Not every citation is used for the star rating. CMS notes that this dataset includes citations excluded from the health inspection rating, such as those under Informal Dispute Resolution (isUnderIdr / isUnderIidr) and cycle 3 standard surveys. Two tags, F731 and F884, are not reported on Care Compare; F884 does not appear in this dataset at all.
• A citation is an allegation of noncompliance as recorded by a state surveyor, not a finding of liability. deficiencyCorrected and correctionDate tell you the compliance status CMS recorded, and correctionDate is blank on roughly 2,400 rows.
• No full statement of deficiencies. As described above, deficiencyDescription is CMS's per-tag definition, not the surveyor's CMS-2567 narrative for a specific survey.
• Monthly, not live. A survey conducted after the last monthly cut will not appear until the next one.

FAQ

How do I look up nursing home inspection results and violations by state?

Set states to one or more two-letter codes. The filter runs on the CMS API itself, so a single state is one or two requests. Add surveyDateFrom to keep only recent surveys, or scopeSeverityCodes to keep only the serious ones.

What do nursing home scope and severity codes mean?

CMS grades every deficiency with a letter from A to L. Severity rises down the grid (A to C no actual harm with potential for minimal harm; D to F potential for more than minimal harm; G to I actual harm; J to L immediate jeopardy) and within each band the scope is Isolated, Pattern or Widespread. Every record carries the letter plus the decoded severity and scope text and an isImmediateJeopardy boolean.

Is this the CMS-2567 statement of deficiencies?

No. CMS does not publish the surveyor's facility-specific narrative in this dataset. What you get is the tag cited and CMS's standard definition of that tag. Request the 2567 from the state survey agency.

Does this include fire safety (life safety code) citations?

No. Every citation in this dataset is an F-tag health deficiency. Life safety citations (K and E tags) are a separate CMS dataset.

How far back does nursing home deficiency data go?

Three inspection cycles per facility. On 2026-08-23 the oldest survey date in the file was 2017-03-23 and the newest 2026-06-22. CMS does not publish older citations in this dataset, and this Actor does not archive them.

How is this different from Nursing Home Compare star ratings?

Star ratings are a summary score. This is the underlying citation-level record: every deficiency, its tag, its severity and its correction status, including citations CMS excludes from the rating calculation.

Related datasets

Other datasets by the same author, all pay-per-record with source, refresh cadence and licence stated on each listing:

Nursing home and long-term care data

• Nursing Home Fines & Civil Money Penalties (CMS) Scraper
• Nursing Home Directory & Ownership (CMS Form 671) Scraper
• Nursing Home Daily Nurse Staffing (CMS PBJ) Scraper
• Nursing Home Change of Ownership (CMS SNF CHOW) Scraper
• New York Nursing Home Bed Availability (NYSDOH) Scraper

Medicare provider enrollment data

• Medicare Revoked Providers List (CMS Revocations) Scraper
• Medicare Opt-Out Providers List (CMS Affidavits) Scraper
• Medicare Revalidation Due Date List (CMS) Scraper

Public procurement and RFP feeds

• E-Rate Form 470 RFP Scraper (USAC School & Library Bids)
• NYC PASSPort Procurement Scraper (RFPs, Bids & Contracts)

Developer and AI test data

• Synthetic Invoice OCR Ground Truth Dataset Generator