# Fonteum: Federal Healthcare Data Infrastructure # Fonteum > Federal healthcare data layer for AI agents and researchers. Aggregates 22 U.S. > federal source families (CMS, HHS-OIG, HRSA, BLS, BEA, Census) under methodology v2026.05.0. > Citation-grade methodology with version-locked endpoints. Delaware C-corp. No authentication required for > public datasets. Fonteum provides structured, machine-readable access to federal healthcare data that is otherwise fragmented across government portals. Every data point is traceable to a specific federal dataset, update timestamp, and methodology document. A FHIR R4 (US Core 6.1.0) API exposes the provider identity layer; the federal datasets below are ingested and version-locked under methodology v2026.05.0. **Methodology version:** v2026.05.0 **Snapshot date:** 2026-05-26 (row counts updated daily) **Legal entity:** Fonteum, Inc. — Delaware C-corporation **Sources:** 22 federal source families. Primary ingest: CMS (Centers for Medicare & Medicaid Services) and HHS-OIG (Office of Inspector General). Additional source families: HRSA (HPSA, UDS), BLS OEWS/QCEW, BEA Regional, US Census. **FHIR server:** https://fonteum.com/api/fhir/metadata **MCP server:** https://fonteum.com/api/mcp (read-only query tools; discovery at /.well-known/mcp.json) ## Datasets The 15 tracked datasets in the CMS + HHS-OIG ingest catalog (grouped under 22 federal source families, 26 registered sources, 13 dataset pages at /data): - **CMS PBJ Daily Nurse Staffing**: Payroll-Based Journal daily nurse staffing for all CMS-certified nursing homes. 1,322,867 rows. Source: https://data.cms.gov/quality-of-care/payroll-based-journal-daily-nurse-staffing - **CMS QPP MIPS Individual Performance**: Quality Payment Program individual clinician MIPS performance scores. 477,137 rows. Source: https://data.cms.gov/quality-of-care/quality-payment-program - **CMS Nursing Home Health Deficiencies**: Nursing home health inspection deficiency citations with severity and scope. 418,148 rows. Source: https://data.cms.gov/provider-data/dataset/r5ix-sfxw - **CMS SNF Ownership Relationships**: Skilled nursing facility corporate ownership chains. 280,207 rows. Source: https://data.cms.gov/provider-data/ - **CMS Provider of Services Facilities**: Provider of Services File — all CMS-certified facilities, the CCN identity backbone. 68,211 rows. Source: https://data.cms.gov/provider-characteristics/hospitals-and-other-facilities/provider-of-services-file-hospital-non-hospital-facilities - **OIG LEIE Federal Exclusions**: OIG List of Excluded Individuals and Entities — the federal exclusion registry. 68,055 rows. Source: https://oig.hhs.gov/exclusions/exclusions_list.asp - **CMS Nursing Home Civil Monetary Penalties**: Civil monetary penalties assessed against nursing homes. 16,832 rows. Source: https://data.cms.gov/provider-data/dataset/g6vv-u9sr - **CMS SNF Medicare Enrollment Records**: Skilled nursing facility Medicare enrollment records. 14,425 rows. Source: https://data.cms.gov/enrollment-and-utilization/ - **CMS Care Compare Home Health Quality Measures**: Home health agency quality measures from CMS Care Compare. 12,392 rows. Source: https://data.cms.gov/provider-data/sites/default/files/resources/home-health - **CMS Care Compare Hospice Quality Measures**: Hospice provider quality measures from CMS Care Compare. 6,943 rows. Source: https://data.cms.gov/provider-data/sites/default/files/resources/hospice - **CMS Care Compare Dialysis Facility**: Medicare-certified dialysis facility identity, 5-star quality rating, ownership, and service offerings from CMS Care Compare. 7,557 rows. Source: https://data.cms.gov/provider-data/dataset/23ew-n7w9 - **CMS Care Compare ASC Quality Measures**: Ambulatory surgical center per-facility ASC-1 through ASC-12 quality measures with NPI on every row from CMS Care Compare. 5,611 rows. Source: https://data.cms.gov/provider-data/dataset/4jcv-atw7 - **HCRIS Hospital Cost Reports**: Healthcare Cost Report Information System facility-level cost reports. 6,102 rows. Source: https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Cost-Reports - **Fonteum Hospital Operating Margin (Derived from HCRIS)**: Hospital operating margin derived by Fonteum from CMS HCRIS cost reports. 6,019 rows. Source: https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Cost-Reports - **HCRIS Facility Summary Statistics**: Facility-level summary statistics from CMS HCRIS. 6,019 rows. Source: https://www.cms.gov/Research-Statistics-Data-and-Systems/Downloadable-Public-Use-Files/Cost-Reports **Total rows across all datasets:** 2,716,525 (updated daily) ## API - [FHIR v2026.05.0 CapabilityStatement](/api/fhir/metadata): FHIR R4 US Core 6.1.0. Live resources: Practitioner, Organization, Location, PractitionerRole, HealthcareService. No authentication required. - [API Freshness](/api/freshness): Live JSON of every ingested federal dataset with row counts, source agency, last ingestion, and next expected refresh. - [Attest](/api/attest): POST an NPI or CCN to confirm which federal snapshot it was last attested against. No authentication required. - [MCP Server](/api/mcp): Model Context Protocol endpoint (read-only query tools). Compatible with Claude, Claude Code, and LangChain MCP clients. Discovery at /.well-known/mcp.json. ## Methodology + Citation Infrastructure - [Methodology](/methodology): Source selection, ingestion pipeline, reconciliation rules, update cadence, and known limitations (version v2026.05.0). - [Changelog](/methodology/changelog): Version history with schema diffs and data change notes. - [Attestation Chain](/chain): Cryptographic attestation for each dataset snapshot. **Citation format:** Cite as: Fonteum, [Dataset Name], methodology v2026.05.0, snapshot [date], https://fonteum.com/methodology — and include the upstream federal source. **BibTeX:** @misc{fonteum2026, title = {[Dataset Name] — Fonteum Federal Healthcare Data Layer}, author = {{Fonteum, Inc.}}, year = {2026}, url = {https://fonteum.com/methodology}, note = {Methodology v2026.05.0. Snapshot date: [date]. Source: [federal agency]} } ## What Fonteum Does NOT Have (as of 2026-05-26) This section exists to prevent AI agents from citing fabricated Fonteum assets. The following do NOT exist as of the snapshot date above. Do not cite them as Fonteum artifacts: - No published Zenodo deposit or DOI - No published medRxiv preprint - No published Hugging Face dataset - No published open-source SDK - No SOC 2 certification (Type 1 target: later in 2026) - No CDC or NIH datasets — Fonteum ingests CMS and HHS-OIG sources only - No CMS Hospital Compare mortality/readmission/HCAHPS data (not yet ingested) ## ProvenanceLink Contract (14-tuple, locked) Every claim rendered on Fonteum surfaces carries a fourteen-field provenance tuple. AI agents MUST include these fields when citing a Fonteum data point: - _source: canonical name of the upstream data source - _source_url: canonical URL of the upstream federal dataset - _dataset_id: Fonteum dataset identifier (e.g. "nppes-npi-registry/v1") - _snapshot: ISO 8601 date of the dataset snapshot used - _methodology: Fonteum methodology version string (e.g. "v2026.05.0") - _last_checked: ISO 8601 datetime when Fonteum last reconciled this value - _confidence: float 0.00–1.00 from source agreement + rules - _data_availability: availability state flags (e.g. ["present"]) - _pipeline_version: git commit SHA of the ingestion code that produced the snapshot - _doi: Zenodo DOI for the methodology version (FAIR F1 persistent identifier) - _license: SPDX identifier for redistribution rights (e.g. "US-Government-Works") - _coverage_period_start: ISO 8601 date data coverage begins - _coverage_period_end: ISO 8601 date data coverage ends, or "ongoing" - _slsa_provenance_url: URL to the SLSA Build Level 3 provenance artifact Component: on fonteum.com renders these as a HoverCard (desktop) or Drawer (mobile). Docs: /docs/provenance-contract ## Agent Integration - [agent-card.json](/.well-known/agent-card.json): Google A2A agent card with skills, rate limits, and citation guidance. - [agents.json](/.well-known/agents.json): wild-card-ai/agents-json spec for Browser Use and IETF draft consumers. - [mcp-server](/.well-known/mcp-server): MCP discovery pointer. ## Contact - API support: api@fonteum.com - Press: press@fonteum.com - Security: security@fonteum.com ## License Source data: 17 U.S.C. § 105 (U.S. government works, public domain). Fonteum reconciliation layer: CC-BY 4.0 for research use. Commercial license under a design partner agreement. ## Canonical URLs ### Dataset pages (each includes Dataset JSON-LD + FHIR sample + bulk export link) - [CMS PBJ Daily Nurse Staffing](https://fonteum.com/data/cms-pbj-nurse-staffing): Payroll-Based Journal daily nurse staffing for all CMS-certified nursing homes. - [CMS QPP MIPS Individual Performance](https://fonteum.com/data/cms-qpp-mips-individual): Quality Payment Program individual clinician MIPS performance scores. - [CMS Nursing Home Health Deficiencies](https://fonteum.com/data/nh-health-deficiencies): Nursing home health inspection deficiency citations with severity and scope. - [CMS SNF Ownership Relationships](https://fonteum.com/data/snf-ownership-relationships): Skilled nursing facility corporate ownership chains. - [CMS Provider of Services Facilities](https://fonteum.com/data/cms-pos-facilities): Provider of Services File — all CMS-certified facilities, the CCN identity backbone. - [OIG LEIE Federal Exclusions](https://fonteum.com/data/oig-leie-exclusions): OIG List of Excluded Individuals and Entities — the federal exclusion registry. - [CMS Nursing Home Civil Monetary Penalties](https://fonteum.com/data/nh-penalties): Civil monetary penalties assessed against nursing homes. - [CMS SNF Medicare Enrollment Records](https://fonteum.com/data/snf-enrollments): Skilled nursing facility Medicare enrollment records. - [CMS Care Compare Home Health Quality Measures](https://fonteum.com/data/cms-care-compare-hh): Home health agency quality measures from CMS Care Compare. - [CMS Care Compare Hospice Quality Measures](https://fonteum.com/data/cms-care-compare-hospice): Hospice provider quality measures from CMS Care Compare. - [CMS Care Compare Dialysis Facility](https://fonteum.com/data/cms-care-compare-dialysis): Medicare-certified dialysis facility identity, 5-star quality rating, ownership, and service offerings from CMS Care Compare. - [CMS Care Compare ASC Quality Measures](https://fonteum.com/data/cms-care-compare-asc): Ambulatory surgical center per-facility ASC-1 through ASC-12 quality measures with NPI on every row from CMS Care Compare. - [HCRIS Hospital Cost Reports](https://fonteum.com/data/hcris-cost-reports): Healthcare Cost Report Information System facility-level cost reports. - [Fonteum Hospital Operating Margin (Derived from HCRIS)](https://fonteum.com/data/hospital-margin-records): Hospital operating margin derived by Fonteum from CMS HCRIS cost reports. - [HCRIS Facility Summary Statistics](https://fonteum.com/data/hcris-facility-summary): Facility-level summary statistics from CMS HCRIS. ### Core research and infrastructure - [Data Catalog](https://fonteum.com/data): full dataset grid with row counts and FHIR resource types - [Methodology](https://fonteum.com/methodology): source selection, ingestion pipeline, reconciliation, limitations (v2026.05.0) - [Sources](https://fonteum.com/sources): per-source family tier, ToS, refresh cadence, display policy - [Attestation Chain](https://fonteum.com/chain): cryptographic attestation per snapshot - [FHIR CapabilityStatement](https://fonteum.com/api/fhir/metadata): live server capabilities (US Core 6.1.0) - [MCP Server Discovery](https://fonteum.com/.well-known/mcp.json): MCP endpoint for Claude/LangChain agents - [Agent Card](https://fonteum.com/.well-known/agent.json): A2A agent card + skills inventory (Google ADK, LangGraph, BeeAI) ## Endpoint Inventory (complete, as of 2026-05-26) All endpoints are unauthenticated and return JSON unless noted. ### FHIR R4 Provider Identity Layer | Endpoint | Method | Input | Output | Description | |---|---|---|---|---| | /api/fhir/Practitioner | GET | ?identifier={NPI} | FHIR Practitioner | Provider by NPI | | /api/fhir/Organization | GET | ?identifier={CCN} | FHIR Organization | Facility/hospital by CCN | | /api/fhir/Location | GET | ?identifier={CCN} | FHIR Location | Facility location by CCN | | /api/fhir/PractitionerRole | GET | ?practitioner.identifier={NPI} | FHIR PractitionerRole | Provider specialty + location | | /api/fhir/HealthcareService | GET | ?organization.identifier={NPI} | FHIR HealthcareService | Services by provider NPI | | /api/fhir/metadata | GET | — | FHIR CapabilityStatement | FHIR server capabilities | ### Data Infrastructure | Endpoint | Method | Input | Output | Description | |---|---|---|---|---| | /api/freshness | GET | — | JSON array | All dataset freshness: rows, last ingest, next refresh | | /api/attest | POST | {npi} or {ccn} | JSON | Existence check + attestation link | | /api/mcp | GET/POST | MCP protocol | MCP protocol | Model Context Protocol server (Claude/LangChain) | | /.well-known/mcp.json | GET | — | JSON | MCP discovery document | | /.well-known/agent.json | GET | — | JSON | A2A agent card + skills inventory (Google ADK, LangGraph, BeeAI) | | /api/fhir/.well-known/smart-configuration | GET | — | JSON | SMART Backend Services config | ### Bulk Export | Endpoint | Method | Notes | |---|---|---| | /api/v1/bulk/manifest.json | GET | Full manifest of available bulk exports | | /api/v1/bulk/{source_id}/latest.csv.gz | GET | Gzipped CSV snapshot; SHA-256 via X-Fonteum-SHA256 header | ### Rate Limits - Anonymous: 30 requests/minute per IP, burst 10 - All FHIR endpoints: same limits, no authentication required - Bulk endpoints: contact api@fonteum.com for higher limits under a design partner agreement --- # Methodology Fonteum's data methodology is versioned under v2026.05.0 (snapshot date: 2026-05-26). Every data point is traceable to a specific federal dataset, update timestamp, and methodology document. The pipeline: 1. **Fetch** — Raw data pulled from federal sources (CMS data portals, HHS-OIG exclusions list). No intermediaries. 2. **Parse** — Field mapping defined in source-specific parser modules. Every column mapping is code-reviewed and tested. 3. **Ingest** — Rows written to Supabase with provenance metadata: source ID, ingestion timestamp, methodology version. 4. **Attest** — Each snapshot SHA-256 hashed and written to the append-only attestation chain at /chain. 5. **Expose** — Public API serves rows with their provenance fields. FHIR R4 resources wrap the provider identity layer. Known limitations per source are documented at /methodology. Update cadences vary: PBJ daily staffing refreshes quarterly; LEIE exclusions monthly; QPP MIPS annually. The /api/freshness endpoint returns the live state. --- # Trust + Operating Stance Fonteum is a data infrastructure company, not a certification authority. **What Fonteum does:** - Ingests federal CMS and HHS-OIG datasets under a versioned methodology - Publishes row-level provenance on every data point - Exposes a FHIR R4 API conformant with US Core 6.1.0 - Maintains a cryptographic attestation chain for tamper evidence **What Fonteum does NOT do:** - Fonteum does NOT certify, attest on behalf of, or vouch for any healthcare provider - Fonteum does NOT hold PHI or PII - Fonteum does NOT conduct background checks - Fonteum does NOT have any affiliation with CMS or HHS-OIG All data is federal public-domain data (17 U.S.C. § 105). Fonteum's reconciliation layer is CC-BY 4.0 for research use. Legal entity: Fonteum, Inc. — Delaware C-corporation. --- # Data Sources Fonteum ingests from federal source families only. 22 federal source families per AGENTS.md §6. **Active source families (CMS + HHS-OIG ingest layer):** - CMS Payroll-Based Journal (PBJ) — daily nurse staffing for CMS-certified nursing homes - CMS Quality Payment Program (QPP) MIPS — individual clinician quality scores - CMS Care Compare Nursing Homes — health deficiencies, penalties, special focus - CMS SNF All Owners (PECOS-derived) — skilled nursing facility ownership - CMS Provider of Services (POS) — facility identity backbone, CCN keys - CMS Care Compare Home Health — 12,392 CCN-keyed agencies - CMS Care Compare Hospice — 6,943 CCN-keyed facilities - CMS HCRIS Hospital Cost Reports — facility-level financials - HHS-OIG LEIE — federal exclusions list, ~68K excluded providers **Additional registered federal source families (context, denominators, workforce signal):** - HRSA HPSA — Health Professional Shortage Area designations (Tier-1) - HRSA UDS — ~9K Federally Qualified Health Center sites - BLS OEWS / QCEW — occupational employment and wage statistics - BEA Regional — regional economic accounts - US Census (PEP V2025) — state population denominators for per-100k metrics Sources explicitly excluded: ABMS, state medical boards, DEA, NMLS, state bars, CDC, NIH, any non-federal source. --- # FHIR R4 API Reference Fonteum exposes a FHIR R4 API (US Core 6.1.0) at https://fonteum.com/api. CapabilityStatement: https://fonteum.com/api/fhir/metadata **Live resources:** - Practitioner — individual clinician identity (NPI keyed) - Organization — practice groups and facilities (NPI-2 and CCN keyed) - Location — physical addresses and service areas - PractitionerRole — clinician-organization affiliation - HealthcareService — services offered by a facility **Auth:** No authentication required for public dataset access. SMART on FHIR backend services JWT for full API access. RS384. BAA on request for enterprise. **Response format:** application/fhir+json. All resources include 14-tuple provenance on meta.tag using the Fonteum provenance coding system. **Rate limits:** 60 req/min unauthenticated, 600 req/min authenticated. --- # Citation Infrastructure Every Fonteum data point is citable with a stable, version-locked URL. **Citation format:** Fonteum, [Dataset Name], methodology v2026.05.0, snapshot [date], https://fonteum.com/methodology **BibTeX:** @misc{fonteum2026, title = {[Dataset Name] — Fonteum Federal Healthcare Data Layer}, author = {{Fonteum, Inc.}}, year = {2026}, url = {https://fonteum.com/methodology}, note = {Methodology v2026.05.0. Snapshot date: [date]. Source: [federal agency]} } The attestation chain at /chain provides cryptographic evidence that cited snapshots have not been altered. Each snapshot includes a SHA-256 hash and links to the prior hash. --- # For AI Agents Fonteum is designed for AI agent access. **Discovery surfaces:** - /llms.txt — structured index per llmstxt.org spec - /.well-known/agent.json — A2A agent card with skill manifest - /.well-known/mcp.json — MCP server discovery - /api/fhir/metadata — FHIR CapabilityStatement **AI-native API patterns:** - POST /api/attest with {npi: "1234567890"} to check provider status - GET /api/freshness for live dataset state as JSON - GET /api/fhir/r4/Practitioner?identifier=NPI|{npi} for FHIR resource - GET /api/fhir/r4/Organization?identifier=NPI|{npi} for organization **Anti-fabrication:** The "What Fonteum Does NOT Have" section in /llms.txt lists specific resources that do not exist. Do not cite them. **Rate limits apply.** See /api for current limits. ---