718 US nursing homes have been barred from new Medicare admissions; the average ban lasted 29 days

Where a denial sits in the enforcement ladder

A DPNA is one rung on a graded federal enforcement ladder, and reading the rung correctly is the key to interpreting this dataset. When a state survey agency cites a nursing home, the severity and scope of the finding determine what CMS can do next. Minor, isolated findings draw a plan of correction and a follow-up visit. More serious or repeated findings draw money: a civil money penalty, assessed either per day the violation persists or per instance. When money alone has not produced a fix, or when the finding is grave enough on its own, CMS turns to remedies that touch the facility’s ability to operate normally — and the Denial of Payment for New Admissions is the first of those.

Above the denial sit the most severe steps: a state-appointed temporary manager, and ultimately involuntary termination from the Medicare and Medicaid programs, which ends federal payment entirely and usually forces the home to close or sell. A DPNA is deliberately short of that. It is meant to be reversible: serve the denial, fix the deficiency, pass a revisit survey, and resume admissions. The most urgent findings — those CMS labels immediate jeopardy, meaning a situation likely to cause serious injury or death — compress this whole timeline, because the law requires a fast remedy or termination. Many of the longest denials in this dataset trace back to findings that started at, or escalated to, that immediate-jeopardy level and then resisted a quick correction.

That ladder is why the counts here should be read as a measure of enforcement, not a leaderboard of the “worst” homes. A facility lands in this dataset because a serious finding was followed by a slow or incomplete fix — a sequence that depends on the citation, the operator’s response, and the rhythm of the state survey agency, all at once. Two homes with similar underlying problems can end up on different rungs depending on how quickly each returned to compliance. The denial marks the escalation; it does not, by itself, rank the care.

Why no one had counted this

The raw material has been public for years, yet no running count or named roster of payment-denied nursing homes existed anywhere on the open web. The reason is structural. CMS publishes denials inside a penalties file built primarily around monetary fines, where a payment denial is just one of two penalty types and carries a length in days rather than a dollar figure. There is no “banned facilities” feed, no press release per action, and no aggregate. The information is open, but it is shaped for compliance recordkeeping, not for a reader who wants to know how many homes were affected and which ones.

Two extra problems keep the number out of reach. First, the penalty file’s facility names are inconsistent and sometimes truncated, so a naive list produces unreliable identities; resolving each provider number against the official Provider of Services file is what turns a messy export into a clean, nameable roster. Second, the file is a rolling window with survey-event dating, which means a casual reader cannot tell a current ban from a long-closed one without computing the implied end date. We did both — identity resolution and date math — so the result is a roster that is both nameable and honestly framed. The point of publishing it is not to shame any home; it is to make a real, federally documented enforcement record legible for the first time.

Key findings

The headline count is 718 distinct nursing homes. Because 195 of them were placed under a denial more than once — one facility drew 6 separate denials — the total number of enforcement actions is higher, at 1,000. The median denial ran 20 days and the mean 29 days; the gap between the two reflects a long right tail of unusually extended bans. The single longest denial in the window — THE REHABILITATION CENTER OF NORTH HILLS in NORTH HILLS, CA — ran 404 days, from 2024-08-09 to 2025-09-17. 718 of the 718 banned facilities resolved to an official CMS Provider of Services record (718/718), so the named lists below carry no unattributed identities.

State-by-state breakdown

DPNA actions are not evenly spread. IL alone accounts for 299 banned facilities across 500 actions — far more than any other jurisdiction — while several states recorded only a single facility, and at least one recorded none at all in the window. Concentration reflects two things at once: the raw size of a state’s nursing-home sector, and how aggressively its state survey agency cites and escalates. Filter or sort the table below; counts are denial actions, not quality judgments.

Map: banned facilities by state

Each tile is a state, positioned roughly geographically; darker tiles have more distinct facilities banned. This cartogram gives every state equal visual weight so small, dense states are not lost — a deliberate alternative to an area choropleth, where geographically large but sparsely affected states would dominate.

Tiles show distinct facilities banned. Territories outside the 50-state grid are listed in the table above. SD recorded no DPNA action in the window.

Denial actions per year

The penalties file is a three-year rolling window, so the first and last calendar years are partial — the count tapers at both ends not because enforcement slowed, but because the window’s edges clip them. The fuller middle years carry the bulk of the actions.

Read the middle of the series rather than the ends. The fullest year in the window carried more than a thousand denial actions across roughly the same number of distinct facilities, which tells you most homes drew a single action that year while a minority cycled through more than one. Year-to-year movement in a rolling file like this should be read cautiously: a dip at the leading edge is an artifact of data not yet landing, not a real decline, and the same caution applies to the trailing edge. The durable signal is the scale — well over two thousand denial actions in three years — not the shape of the tails.

How long the bans run

Most denials are short. The largest single group is brief windows of a few weeks, consistent with CMS’s aim of forcing a fast return to compliance rather than closing a home. But a meaningful tail extends for months — and that tail is where the most serious, slowest-to-fix failures sit. The distribution below counts denial actions by length.

The longest bans, named

Below are the 25 facilities that served the longest single denial in the window, each resolved to its official CMS identity by provider number (CCN) before listing. Names, cities, certified bed counts, and chain affiliations come from the CMS Provider of Services file; denial dates and lengths come from the penalties file. Where a chain is shown, it is the ownership affiliate CMS records for that facility — not a judgment about the chain.

“Implied end” = denial start date + length in days. The penalties file carries no explicit end-date column. Download the full action-level dataset as CSV.

Reading the top of the list

The single longest denial in the window belongs to THE REHABILITATION CENTER OF NORTH HILLS in NORTH HILLS, CA, a 138-bed facility. Its denial ran 404 days — from 2024-08-09 to 2025-09-17 — more than a full year without paid new Medicare admissions. A window that long almost always reflects a serious finding that proved slow to correct across multiple revisit surveys, because CMS lifts a denial the moment a facility passes a revisit; the clock only keeps running while the home remains out of compliance.

The next two longest tell the same story in different settings. ELEVATE CARE NILES in NILES, IL served 155 days, and SOUTH ELGIN LIVING & REHAB CENTER in SOUTH ELGIN, IL served 152. None of these are small or obscure buildings — several of the longest-banned facilities carry some of the highest certified bed counts in the dataset, which means a single extended denial can sideline a large share of a local market’s skilled-nursing capacity for months. That is part of why these actions matter beyond the individual home: in a tight market, a long denial can ripple into hospital discharge planning and bed availability for an entire area.

A practical caution for readers scanning the table: the presence of a facility here describes a federal enforcement action with a specific start and end, not a permanent label. Most of these denials have already closed, and a facility that served a denial and returned to compliance is, by the program’s own design, back in good standing for new admissions. The value of the list is historical and pattern-level — it shows where and how often the federal government has had to pull this particular lever.

Repeat denials and the chain question

The gap between 718 facilities and 1,000 actions is the most telling number in the study. 195 facilities — roughly one in five of those banned — were placed under a denial more than once inside this single three-year window. One home reached 6 separate denials. A repeat denial is a strong signal: it means the corrective remedy worked well enough to lift the first ban, but the underlying conditions returned, and CMS had to escalate again. For families, a facility with multiple denials in a short span is a different proposition than one with a single, long-closed action.

Ownership context matters too, which is why the named table carries the chain affiliation CMS records for each facility. Several multi-state operators appear more than once across the longest bans — a pattern worth watching, because operating decisions about staffing, budgeting, and corrective response are frequently made above the level of the individual building. We surface the affiliate name as a fact from the Provider of Services file and stop there: this study does not compute chain-level ban rates or attribute cause to any operator. Readers who want to trace ownership further can follow the SNF ownership records, where the same facilities resolve to their disclosed owners and management companies.

Geography compounds the picture. IL’s 299 banned facilities are not only the most of any state; they also include some of the largest buildings in the dataset by certified bed count, so a single denial there can affect hundreds of beds’ worth of admissions capacity at once. A handful of states with large nursing-home sectors — concentrated in the Midwest and on the West Coast — account for a disproportionate share of the national total, while many states contribute only a few facilities each. That spread is visible in both the filterable table and the tile-grid map above.

Methodology & reproducible SQL

A DPNA action is a single row in the CMS Nursing Home Penalties dataset (CMS Provider Data Catalog, dataset g6vv-u9sr) whose penalty type is Payment Denial. Each carries a payment_denial_start_date and a payment_denial_length_days. We count distinct facilities by provider number (CCN), count actions by row, and compute ban length directly from the length field. Facility identity is resolved against the CMS Provider of Services (POS) file by CCN; POS supplies the canonical name, city, certified bed count, and chain affiliation. Method version dpna-bans/v1. The headline counts are reproduced by:

-- Headline: distinct banned facilities + total denial actions
SELECT
  COUNT(*)                       AS denial_actions,
  COUNT(DISTINCT ccn)            AS facilities_banned,
  ROUND(AVG(payment_denial_length_days)) AS avg_ban_days,
  MAX(payment_denial_length_days)        AS longest_ban_days
FROM nh_penalties
WHERE penalty_type = 'Payment Denial';

-- State breakdown (distinct facilities + actions per state)
SELECT state,
  COUNT(DISTINCT ccn) AS facilities,
  COUNT(*)            AS actions,
  ROUND(AVG(payment_denial_length_days)) AS avg_days,
  MAX(payment_denial_length_days)        AS max_days
FROM nh_penalties
WHERE penalty_type = 'Payment Denial'
GROUP BY state
ORDER BY facilities DESC;

-- Named facilities, identity resolved against CMS POS
SELECT p.ccn,
  COALESCE(pos.facility_name, p.provider_name) AS facility_name,
  p.state, pos.beds_certified, pos.chain_name,
  p.payment_denial_start_date AS ban_start,
  p.payment_denial_length_days AS ban_days,
  (p.payment_denial_start_date
     + (p.payment_denial_length_days || ' days')::interval)::date AS implied_end
FROM nh_penalties p
LEFT JOIN cms_pos_facilities pos ON pos.ccn = p.ccn
WHERE p.penalty_type = 'Payment Denial'
ORDER BY p.payment_denial_length_days DESC;

A few field-level decisions shape the counts. We treat each penalties row whose type is Payment Denial as one action, and we count a facility once per provider number regardless of how many denials it drew. The named “longest bans” table deduplicates to distinct facilities, keeping each home’s single longest denial, so no facility appears twice. Average and median ban length are computed across actions, not facilities, because the denial — not the building — is the unit of enforcement. Where the penalties file and the Provider of Services file disagree on a facility’s name, we display the Provider of Services name as canonical and fall back to the penalty-file name only when no Provider of Services record exists; across this window every banned provider number resolved against the Provider of Services file, and a single truncated source name was reconciled to its canonical form.

The page reads these counts live from the database on a daily refresh and falls back to the 2026-06-04 snapshot for static rendering, so the figures you see are always traceable to the SQL above. Federal data is public domain under 17 U.S.C. §105; we publish it free under a CC BY 4.0 license with full method disclosure. Read how every Fonteum figure is sourced on the sources page.

Key terms

• DPNA (Denial of Payment for New Admissions). The federal remedy at 42 CFR §488.417 under which Medicare and Medicaid stop paying for residents admitted after a set date, while existing residents stay covered. In effect, a ban on taking new paid patients.
• CCN (CMS Certification Number). The provider number that uniquely identifies a certified facility. We count distinct facilities by CCN and use it to resolve each home’s official identity against the Provider of Services file.
• Substantial compliance. The standard a facility must return to for a denial to lift — broadly, no deficiency posing more than minimal potential for harm. A revisit survey confirms it.
• Revisit (follow-up) survey. The on-site check after a cited facility reports it has corrected its problems. Passing the revisit is what ends an active denial; failing it keeps the clock running, which is how the longest bans accumulate.
• Civil money penalty (CMP). A fine assessed per day or per instance of non-compliance — a separate penalty type in the same CMS file. This study counts denials, not fines; the two remedies often apply to the same facility but answer different questions.

Limitations

• No end-date field. The penalties file records a denial start date and a length in days, not an explicit end date. We compute the implied end as start + length. Many of these windows have already closed; some near the snapshot edge may still be open. For a single home’s current standing, check Medicare Care Compare.
• Rolling three-year window. CMS publishes a moving window, so the first and last calendar years are partial and earlier denials age out. Counts describe this window, not the full history of US nursing-home enforcement.
• Survey-event dating. Penalty dates in the source are tied to the survey event that triggered the action, not the date a denial was formally issued or collected. We use the denial start date for the window, which is the operative date for the ban.
• No causal or quality claim. A count of DPNA actions is a count of enforcement events. Fonteum does not rate, rank, or judge facilities, and a ban is not a statement that a home is the “worst” — only that CMS applied a specific payment remedy. We do not join these counts to star ratings here; that is a separate study.
• Identity edge cases. Every banned CCN in this window resolved against CMS POS, and a single source-file name truncation was reconciled to the POS canonical name. Had any identity been unresolved, that facility would have been kept in the aggregate count but withheld from the named table.

Frequently asked questions

How to use this data

For families and patient advocates, the practical workflow is two steps. Use the named table and the state breakdown to see whether a facility — or the operators active in your area — has a denial on record in this window, then confirm a specific home’s current standing on Medicare Care Compare, which reflects active enforcement in close to real time. A past, closed denial is context for a conversation with an admissions coordinator, not a verdict; a recent or repeat denial deserves a direct question about what changed and what was fixed.

For reporters and researchers, this study is a starting index across 23 jurisdictions, not the end of the inquiry. The action-level CSV gives every denial with its facility identity, start date, and length, which you can join to ownership records, staffing data, and the underlying health-deficiency citations to reconstruct why a given home ended up on this rung. Because every figure traces to the published SQL against official CMS files, the counts are independently reproducible — re-run the queries against the same dataset and you will land on the same numbers. We ask only that re-publication preserve the framing: these are facilities that have been banned over a defined window, dated and bounded, not a live blacklist.

For operators and their counsel, the value is benchmarking. The ban-length distribution and the state-level averages put any single action in context: a brief denial that closed on schedule looks very different against a backdrop where most denials run only a few weeks than it does in isolation. Nothing here assigns fault or predicts outcomes; it simply makes a previously scattered public record legible in one place.

Related research

• Care Compare: Nursing Homes — quality, staffing, and ownership for every certified US facility.
• Nursing-home health deficiencies — the survey citations that precede most enforcement.
• How long facilities take to correct deficiencies — the compliance clock a DPNA is meant to speed up.
• SNF ownership & chain affiliation — who operates the facilities in the named table.
• OIG exclusions (LEIE) — a separate federal sanctions track.
• NPI lookup — resolve any provider NPI to its NPPES, PECOS, and OIG record.
• Sources & methodology index.

Sources

1. U.S. Centers for Medicare & Medicaid Services, Nursing Home Penalties (Provider Data Catalog, dataset g6vv-u9sr). data.cms.gov/provider-data/dataset/g6vv-u9sr
2. 42 CFR §488.417 — Denial of payment for new admissions. ecfr.gov · §488.417
3. U.S. Federal Register — CMS enforcement rulemaking and notices. federalregister.gov
4. Medicare Care Compare — current facility standing. medicare.gov/care-compare

Fonteum Research · Published 2026-06-04 · Data source: CMS, public domain under 17 U.S.C. §105 · Published free under CC BY 4.0 · Method dpna-bans/v1.