What one source misses, measured against the record.
Fonteum publishes supported public records with available source and observation metadata across healthcare, federal procurement, sanctions and watchlists, federal enforcement, and global company registers. Active production registry status does not establish loaded, complete, or fresh coverage.
Fonteum is a multi-vertical public-records graph spanning healthcare providers, federal procurement entities, sanctions and watchlists, federal enforcement, and global company registers. The same contract applies across the graph: available source, date, and limitation metadata. Snapshot attestations are separate and do not sign individual facts.
The public alternatives — a raw NPPES download, federal -only screening, a single state list — each cover one slice of the provider record. Federal-only screening misses of state Medicaid bars, and no single exclusion list names more than of barred providers. Don’t take the coverage on faith: every number below is lifted from a published study, links back to it, and sits on federal files you can pull and re-run yourself.
Last updated: June 2026 · Sources: published Fonteum June 14, 2026 research snapshot on the OIG LEIE, CMS PECOS, SAM.gov, and ten state Medicaid exclusion lists; current loaded coverage is reported separately
What each screening approach catches
In the June 14, 2026 ten-state study snapshot, no single studied list named every barred provider. Across the OIG LEIE, SAM.gov, and those programs, NPI-identified providers are excluded; the widest single source covers 64.0%. The rows below describe coverage by approach, not by vendor.
full coverage partial gap
| Screening approach | What it catches | What it misses | Field-level provenance |
|---|---|---|---|
| Raw NPPES download | Provider identity only — NPPES carries no exclusion or sanction field | All barred providers — exclusion status is not in the file | File modified date only; no per-field source or limitation |
| Federal OIG LEIE only | of NPI-identified barred providers (64.0%) | (36.0%) — including of state Medicaid bars | Raw federal file; not cited field-by-field |
| Single state exclusion list | One state's bars — e.g. New York names providers absent from the federal list | Every other state and the federal layer; of bars sit on one layer only | One source, one format; no cross-source reconciliation |
| Fonteum source-linked graph | OIG LEIE + SAM.gov + supported loaded state Medicaid lists, joined on NPI; no claim for the other states | Records with no NPI are reported separately, not silently dropped | Available source, date, file-hash, and limitation metadata; individual facts are not signed |
Counts from the excluded-provider landscape study and the federal–state exclusion gap study. A coverage match is a screening flag for a primary-source look, not a Fonteum judgement of any provider.
Four findings a single source cannot produce
of state Medicaid bars are invisible to federal screening
In the June 14, 2026 ten-state study snapshot, of NPI-identified providers excluded across those state Medicaid programs, carry no record on the federal . An employer checking the federal list alone clears nearly two in three state-barred providers as having nothing on file.
is the ceiling for any single exclusion list
In the June 14, 2026 study snapshot across the OIG LEIE, SAM.gov, and ten state Medicaid programs, NPI-identified providers are barred from a public health program — yet the widest single list names only of them. appear on only one of the three layers, so no two-list check reaches them.
barred providers still hold an active Medicare enrollment
Cross-referencing the OIG exclusion list against surfaces 19 of 6,880 in-force federal exclusions that still carry an active enrollment record — two standing over a year, one excluded since 2015. The join is the finding; a single list cannot produce it.
order/refer-eligible providers carry an active exclusion
Of providers cleared to order and refer in Medicare, 170 match an active exclusion or sanction across the OIG, SAM.gov, and the June 14, 2026 study's ten-state ring — 156 on a state list, 18 on the federal LEIE, 6 on SAM.gov. A match is a flag for a primary-source look, not a Fonteum verdict.
Raw NPPES vs the source-linked graph
The NPPES registry is the starting point — active providers; CMS publishes weekly files, while Fonteum's newest observed production system date was June 10, 2026. It is identity, not the rest of the federal record. Fonteum re-ingests the same file and joins it to the sources NPPES does not carry.
| Capability | Raw NPPES download | Fonteum source-linked graph |
|---|---|---|
| Provider identity (NPI, taxonomy) | Yes — the NPPES file is the registry | Same NPPES file; newest serving-table system date observed July 12 was June 10, 2026 |
| Medicare enrollment status | Not in NPPES | CMS PECOS, joined on NPI |
| Federal exclusion flag | Not in NPPES | OIG LEIE, joined on NPI |
| Multi-source exclusion / compromise | Not in NPPES | SAM.gov, supported loaded state Medicaid lists, CMS CMPs, OIG integrity agreements |
| Facility quality, staffing, deficiencies | Not in NPPES | CMS Care Compare, PBJ staffing, NH deficiency citations |
| Per-field source trail | None — a flat file of values | 14-tuple chain: source, date, hash, methodology version |
| Freshness tracking | Whole-file timestamp | Available source and observation dates, scoped by response |
Every joined field carries its source name, snapshot date, and known limitation via the . See the field-level provenance pipeline.
Three dimensions a single download cannot match
Breadth
Fonteum's active production registry includes CMS, OIG, SAM.gov, USASpending.gov, OFAC, SEC, EPA, and company registers. Active status does not establish loaded, complete, or fresh coverage. A dated ten-state study found that 55.6% of matched barred-provider records appeared on only one of LEIE, SAM.gov, or those ten state lists; supported loaded state lists are reported separately.
Freshness
Sources follow different release and ingestion cadences. A daily row-count observation does not mean upstream data refreshed that day. Records expose available release, ingestion, and observation timestamps; source-file SHA-256 coverage also has disclosed gaps.
Provenance
Provenance-aware records expose the available source name, observation date, and limitations. The fourteen-field schema is nullable, and the 26.2M claim ledger currently has no deterministic claim-to-signature linkage.
Coverage, honestly
- What does federal OIG LEIE-only exclusion screening miss?
- In the June 14, 2026 ten-state study snapshot, federal OIG LEIE-only screening missed of state Medicaid bars: of NPI-identified providers excluded across ten state Medicaid programs, carry no record on the federal list. Measured against every barred provider — across the OIG LEIE, SAM.gov, and ten state lists — the federal list names of , so checking it alone clears (36.0%) barred providers as having nothing on file. The gap is the cross-source population a single federal list cannot see.
- How is raw NPPES different from the Fonteum public-records graph?
- Raw NPPES is a weekly flat file of provider identity — NPI, name, taxonomy, address — with no exclusion field, no Medicare enrollment status, no facility quality, and no per-field source trail. The Fonteum serving table had a June 10, 2026 newest system timestamp when checked July 12. Cross-source joins occur only where participating rows publish a shared identifier. Supported loaded state Medicaid coverage is reported separately; the other states are not covered.
- Is there a single list that catches every excluded provider?
- In the June 14, 2026 ten-state study snapshot, no single studied list caught every excluded provider. Across the OIG LEIE, SAM.gov, and those state programs, NPI-identified providers are barred, and the widest single source — the OIG LEIE — names only 64.0% of them. of barred providers appear on only one of the three layers, so even a two-list check leaves a population uncovered. Coverage comes from reconciling the sources against one another on a shared identifier, which is what the compromised-anywhere layer does.
- What does compromised-anywhere coverage add beyond exclusion lists?
- Compromised-anywhere coverage adds the layer between fully clean and formally excluded. Beyond the barred providers, the July 12 production audit counted CMS penalty and payment-denial records across facilities, including 2,553 payment-denial records. Those 2,553 rows are not 2,553 facilities. The records carry no provider NPI, so Fonteum does not force an individual-provider join.
- Are these coverage numbers independently checkable?
- The figures on this page link to named Fonteum studies, which cite the government datasets and dated releases used for their joins. Readers can inspect those source notes and methods. The fourteen-field provenance schema is nullable, and no provenance fact currently links deterministically to a signature.
- Does broader coverage mean Fonteum rates or flags providers itself?
- No. Fonteum does not rate, inspect, endorse, or independently judge any provider, contractor, or listed party. A coverage match means a named public source lists that entity on a given date — the page reports the source record and its date, not a Fonteum verdict. Each match is a screening flag worth a primary-source look, not a Fonteum claim about the entity, and person matches are never made on name alone.
- The excluded-provider landscape → Why no single list names more than 64.0% of barred providers.
- The federal–state exclusion gap → The 64.3% of state Medicaid bars federal screening cannot see.
- Coverage network map → The distribution surfaces that consume the Fonteum graph.
- Source library → Active production registry, plus the smaller public detail-page catalog.
- Field-level provenance → Available source, date, hash coverage, and limitations.