The separately dated platform snapshot reports active source-registry rows as of August 11, 2026. Registry status does not establish loaded, complete, or current coverage. The serving-table ribbon below reports only counts returned by its own query and keeps unavailable values absent.
- Provenance register
- v1.1
- Last reviewed
- 2026-07-12
- Serving counts checked
- 2026-08-12
The data graph · one second
Public records → cited fields, in five gated stages.
48 registered source pages are organized through five reproducibility stages into four public surface groups.
Sources
48 registered · grouped for display
Provider identity
Federal identifiers + Medicare enrollment
Care quality
CMS facility quality data
Enforcement & integrity
Exclusions, penalties, and agreements
Federal context
Tier-1 aggregates (research only)
Pipeline
5 reproducibility stages
- 01
Source registry
Authority, rights, cadence, limitations
- 02
Ingestion run
Dated execution · row count · errors
- 03
Normalize
Typed parsing · null preservation
- 04
Source identifiers
NPI, CCN, UEI, or source-native key
- 05
Serving surface
Source and observation metadata where available
48
Sources registered
1,322,946
Nurse-staffing records (CMS PBJ)
69
Research snapshots
4
Displayed source groups
Legend
- Tier-2 · live
- Tier-1 · research-only
- Pending records request
Source nodes link to /sources · pipeline detail at seven-stage breakdown.
What is in the graph today
Serving counts checked 2026-08-12- 111integrated sourcesActive data_sources rows in the platform-stats snapshot as of August 11, 2026
- 1,322,867nurse-staffing recordsCMS Payroll-Based Journal (PBJ) daily nurse-staffing records across Medicare-certified nursing homes
- 93profile-enrichment sourcesActive Tier-2 profile-enrichment rows in the source registry
- 65ingestion runsRecorded source-pipeline runs in the ingestion audit table
- 9research snapshotsDated aggregate artifacts registered for research surfaces
These are serving-table counts observed when this ribbon was queried. They do not establish that an upstream source refreshed that day; source dates and limitations remain source-specific.
Seven stages from public source to served record.
Fonteum keeps source families separate through ingestion and serving. Public responses expose the source, observation date, and limitation fields that exist for that response; absent metadata stays absent.
- STAGE 01
Public source
An official, public-record dataset published by a federal or state agency: NPPES, CMS PECOS, CMS Care Compare, BLS, HRSA, US Census.
data.cms.gov · npiregistry.cms.hhs.gov · hrsa.gov - STAGE 02
Source registry
Each source page records its publisher, rights posture, expected cadence, limitations, and current serving status.
data_sources · src/lib/brand/sources-registry.ts - STAGE 03
Ingestion run
Tracked ingestion runs record a dated script execution plus its fetch method, record count, and run notes. Not every active source-registry row has a current run.
ingestion_runs - STAGE 04
Normalize
Source-specific parsers preserve published nulls, normalize documented fields, and record parse failures without inventing values.
src/lib/sources/parsers/ - STAGE 05
Source identifier
Records retain the identifier supplied by the source—such as NPI, CCN, UEI, CAGE code, or a source-native record key—so consumers can reproduce the lookup.
NPI · CCN · UEI/CAGE · source-native keys - STAGE 06
Source-specific serving table
Normalized records remain in source-specific tables rather than being collapsed into an inferred cross-source record.
cms_* · oig_* · sam_* · global_* tables - STAGE 07
Public response
Provider, facility, enforcement, research, and export surfaces expose source and observation metadata when the underlying response supplies it.
/providers · /data · /research · API/MCP
The pipeline lives in code. Source definitions at src/lib/brand/sources-registry.ts; source parsers at src/lib/sources/parsers/; ingestion and snapshot scripts at scripts/research/.
Three source clusters. One provenance contract.
We organize Fonteum by the public datasets that back it — not by a flat list of provider category pages. Every cluster ships with a registered source, an ingestion runner, and a published research asset.
- Cluster
Healthcare graph
Loaded NPPES and CMS PECOS records can link by NPI where both sources publish it. NPPES supplies enumeration fields; PECOS supplies Medicare enrollment fields where present.
Source:CMS NPPESChecked Newest system date Jun 10, 2026Source:CMS PECOSChecked Source release Jun 18, 2026- NPI and taxonomy where present in the loaded NPPES record
- Medicare fee-for-service enrollment fields where present
- Provider identity and specialty-supply research from federal source records
- Cluster
Care graph
CMS Care Compare publishes facility-quality data for nursing homes, home health, hospice, and dialysis. Four research artifacts live across the supported modules.
Source:CMS Care CompareSnapshot Source release May 7, 2026- CMS overall star ratings (1.0–5.0) where CMS publishes them
- 4 research snapshots: nursing-home, dialysis, home-health, hospice
- Home-health availability and quality records with source limitations
- Cluster
Research graph
BLS OEWS, BEA Regional, HRSA HPSA, and US Census state population. Tier-1 research-only sources that add employment, income, shortage-area, and per-capita context.
Source:BLS OEWSSnapshot May 2026Source:HRSA HPSASnapshot May 2026Source:US CensusSnapshot May 2026- Per-capita density, wage context, shortage-area context
- Cited inside every Sprint-1 study and the Care Compare snapshots
- Never attached to individual provider profiles
Three source-linked examples.
Each card shows the field as it would render on the relevant Fonteum surface, sourced verbatim from the public dataset, with the “what it means” and “what it does not mean” framing CMS itself documents.
When a field is in the database but not on the page.
A parsed value does not automatically become a customer-facing claim. Four rules govern which fields render, where they render, and what context accompanies them.
- Rule 1
Why don't BLS / BEA / HRSA / Census numbers appear on individual provider profiles?
Aggregate datasets (BLS Occupational Employment & Wage Statistics, BEA Regional Income, HRSA Health Professional Shortage Areas, US Census state population) describe markets, not individuals. Attaching a state-mean wage figure to one practitioner would imply an individual signal the data doesn't carry. They render only on /research aggregate pages.
How it's enforcedSource-pack manifest tier='tier1-research-only' + the data:bls-bea-not-in-profile-components launch gate. - Rule 2
Why does Fonteum prefer source-native identifiers?
An NPI, CCN, UEI, CAGE code, or source-native record key gives consumers a reproducible way to locate the upstream record. When a deterministic identifier bridge is unavailable, Fonteum labels the limitation instead of presenting a name-only inference as certain.
How it's enforcedSource-specific parsers and API contracts retain published identifiers; response metadata can be null when no reproducible bridge exists. - Rule 3
Why are some published fields absent from customer-facing pages?
A value's presence in a federal file does not make every presentation accurate or useful. High-risk enforcement and quality fields require source-specific context, dated limitations, and an approved rendering before they appear on a customer-facing page.
How it's enforcedSource registry status, typed response schemas, page-level methodology, and customer-facing claim tests form the display boundary. - Rule 4
Why do we show "missing" or "unknown" instead of a best guess?
If the source dataset doesn't carry a value, we don't infer one. CMS NPPES doesn't publish a phone number for every NPI. CMS Care Compare leaves 35.8% of home-health agencies unrated. CMS Hospice General Information has no overall star rating. In every case the public surface shows the explicit absence — not a model-imputed guess — and the methodology page documents why.
How it's enforcedManifest field-level rules + per-page methodology disclaimers + the 'unrated ≠ low-quality' doctrine sentence in every CMS-cited UI.
Three classes the dataset can fall into.
These three classes describe the source and rights model used where a field has registered provenance metadata. Not every public field has a populated class, date, or attribution value.
- PUBLIC_RECORD
Public record (federal / state)
Federal and state public-record datasets can fall in this class when their source terms and display rules permit it. Rights and attribution remain source-specific.
Examples: CMS NPPES NPI registry, CMS PECOS Medicare enrollment, CMS Care Compare star ratings, CMS Nursing Home deficiency citations, OIG LEIE exclusions.
- RESEARCH_AGGREGATE
Research aggregate (Tier-1)
Sourced from federal aggregate datasets that describe markets, not individuals. Renders only on /research surfaces — never attached to per-provider profiles.
Examples: BLS Occupational Employment & Wage Statistics (OEWS), BLS QCEW, BEA Regional, HRSA HPSA, US Census state population.
- OWNED
Owned by Fonteum
Generated or assigned by Fonteum's own systems. Full rights to store and display indefinitely.
Examples: Record identifiers, slugs, snapshot timestamps, methodology-version tags.
Provider and facility records use source-published identifiers such as NPI, CCN, and PAC ID where those identifiers are available; coverage differs by source.
| Field | Source | Retention | Observed serving date |
|---|---|---|---|
| NPI | CMS NPPES | Stored serving row; history varies | System date 2026-06-10 |
| CCN (CMS Certification Number) | CMS Care Compare | Stored serving row; history varies | Source date 2026-05-07 |
| PAC ID | CMS PECOS | Stored serving row; history varies | Source date 2026-06-18 |
Per-provider descriptive fields can expose source, observation-date, and limitation metadata when the serving response supplies it; metadata can be unavailable.
| Field | Source | Retention | Observed serving date |
|---|---|---|---|
| Taxonomy code | CMS NPPES | Stored serving row; history varies | System date 2026-06-10 |
| Medicare enrollment | CMS PECOS | Stored serving row; history varies | Source date 2026-06-18 |
| MIPS score | CMS QPP | Stored when present | Not included in July 12 freshness query |
CMS-published quality signals for Medicare-certified facilities. Fonteum republishes the CMS value with attribution; it does not independently rate.
| Field | Source | Retention | Observed serving date |
|---|---|---|---|
| Overall star rating | CMS Care Compare | Stored serving row; history varies | Source date 2026-05-07 for checked modules |
| Deficiency citations | CMS Nursing Home Deficiency file | Stored serving row; history varies | Not included in July 12 freshness query |
Federal program-exclusion records. Surfaced as a sourced compliance flag, never as a Fonteum judgement about a provider.
| Field | Source | Retention | Observed serving date |
|---|---|---|---|
| Exclusion flag | OIG LEIE | Stored serving row; history varies | Source date 2026-05-08 |
| Exclusion type | OIG LEIE | Stored serving row; history varies | Source date 2026-05-08 |
| Statutory authority | OIG LEIE | Stored serving row; history varies | Source date 2026-05-08 |
We cite source-backed facts. We do not pay-to-rank, issue trust badges, invent ratings, award providers, or claim license credentials we cannot trace.
The boundary described in the homepage “What Fonteum does NOT claim” block applies here: Fonteum does not independently rate, inspect, verify, endorse, or guarantee any provider. Supported responses cite CMS, NPPES, OIG LEIE, and other public-record sources where linked source metadata is populated.
- Sources → The generated source-catalog subset — documented entries with tier, publisher cadence, fields used where specified, and limitations.
- Methodology → Documented sourcing, publisher cadence, observed freshness, and corrections policy.
- Home Health methodology → The source-specific ingestion methodology for CMS Care Compare home-health data.
- Research → Tier-1 dated aggregates published from the source graph.
- Editorial policy → Independence, sourcing, conflicts, corrections, retractions.
- Corrections log → The log dates accepted corrections and names the documented cause.
- Data & press kit → Cite our data, request a custom export, or reach press.