Healthcare Data MCP

Healthcare Data MCP source ledger.

Every useful healthcare AI claim should know where it came from.

19local MCP servers

36catalog source IDs

11workflow plans

v0.2.0release

Named sources, visible limits.

This ledger maps every cataloged source ID to its source system, owning MCP servers, grain, readiness language, limitations, and a practical use. Public data is useful only when the source boundary travels with the answer.

Public and reference data only. No PHI, HIPAA readiness, legal clearance, clinical decision support, reimbursement advice, complete market truth, or current all-payer intelligence is claimed here.

Facility and System Identity Quality and Operations Geography and Access Price Transparency Ownership and Enrollment Public Records State and Finance Workforce and Referrals Research and Web MCP Metadata

Source catalog.

Cadence language follows the local Healthcare Data MCP catalog: CMS public data exposes local freshness through cache status/source metadata, API-backed sources are subject to limits and keys, import-backed sources depend on local imports, and metadata surfaces are release-coupled.

Facility and System Identity

CMS Hospital General Information

Dataset ID: cms_hospital_general_info
Source system: CMS Provider Data Catalog; Medicare-certified hospital identity.
Owning servers: cms-facility, hospital-quality, service-area, drive-time
Grain: One row per Medicare-certified hospital.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Facility identity and current operating facts must remain CCN-scoped and source-period visible.
Example use: Anchor a facility profile or comparison before adding quality, service-area, or access context.

CMS Provider of Services

Dataset ID: cms_provider_of_services
Source system: CMS quarterly POS PUF; certified provider location attributes.
Owning servers: health-system-profiler, cms-facility, public-records
Grain: One row per certified provider location.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Provider category, bed, service, and staffing attributes are source fields, not proof of system affiliation.
Example use: Enrich a CCN with provider location, category, and bed/service attributes.

AHRQ Compendium of U.S. Health Systems

Dataset ID: ahrq_health_system_compendium
Source system: AHRQ Compendium; system and hospital linkage.
Owning servers: health-system-profiler
Grain: System and hospital-linkage files.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: System names and aliases stay source-specific unless exact identifiers support a join.
Example use: Start system reconciliation with AHRQ system IDs before adding facility-level evidence.

NPPES NPI Registry

Dataset ID: nppes_registry
Source system: CMS NPPES API; organization and individual NPI identity.
Owning servers: cms-facility, health-system-profiler, physician-referral-network
Grain: Provider organization or individual NPI.
Cadence/readiness: Live/API-backed; subject to source API limits and key requirements where applicable.
Limitations: NPI identity is not employment verification, affiliation proof, or common-control proof by itself.
Example use: Resolve organization NPIs or physician identifiers for downstream source joins.

Quality and Operations

CMS Hospital Quality Programs

Dataset ID: cms_hospital_quality
Source system: CMS Provider Data Catalog quality programs.
Owning servers: hospital-quality
Grain: Facility-measure rows.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Use exact measure rows; do not substitute HAC totals, HRRP condition rows, or PHC4 reports for exact CMS measures.
Example use: Support a quality fact with a CCN plus exact measure ID receipt.

CMS Hospital Cost Report PUF

Dataset ID: cms_cost_report
Source system: CMS Hospital Cost Report PUF.
Owning servers: hospital-quality, workforce-analytics
Grain: Hospital cost report worksheet rows.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Worksheet facts require fiscal year, worksheet/line/column context, and source period.
Example use: Add public cost-report context to a finance or staffing profile.

CMS Hospital Service Area File

Dataset ID: cms_hsaf
Source system: CMS data.cms.gov service-area files.
Owning servers: service-area
Grain: Hospital-ZIP service area rows.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Medicare service-area context is not all-payer market share or patient leakage.
Example use: Define service-area overlap before a market scan.

CMS Geographic Variation PUF

Dataset ID: cms_geographic_variation
Source system: CMS Geographic Variation PUF.
Owning servers: geo-demographics
Grain: State/county geography by year.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Medicare geography aggregates are context, not facility performance or patient-level evidence.
Example use: Place a county or state in Medicare utilization context.

CMS Medicare Provider Utilization PUFs

Dataset ID: cms_medicare_claims_pufs
Source system: CMS Medicare Provider Utilization and Payment Data.
Owning servers: claims-analytics
Grain: Provider-service rows by discharge or service year.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Public Medicare FFS aggregates are not PHI, all-payer market share, or current operations.
Example use: Support service-line or case-mix context where configured public PUF caches are present.

Geography, Community, Access

Dartmouth Atlas ZIP-HSA-HRR Crosswalk

Dataset ID: dartmouth_hsa_hrr
Source system: Dartmouth Atlas HSA/HRR crosswalk.
Owning servers: service-area, physician-referral-network
Grain: ZIP to HSA/HRR crosswalk.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Healthcare geographies are analytic frames, not legal service areas.
Example use: Translate ZIPs into HSA/HRR context for a market scan.

Census ACS 5-Year API

Dataset ID: census_acs
Source system: U.S. Census Bureau API.
Owning servers: geo-demographics
Grain: Census geography.
Cadence/readiness: Live/API-backed; subject to source API limits and key requirements where applicable.
Limitations: Community/geography values are market context, not facility performance or patient-level facts.
Example use: Add population, income, age, and geography context to a market scan.

CDC PLACES

Dataset ID: cdc_places
Source system: CDC PLACES via Socrata.
Owning servers: community-health
Grain: Measure rows for county, place, census tract, or ZCTA geography.
Cadence/readiness: Live/API-backed; subject to source API limits and key requirements where applicable.
Limitations: Community estimates are not facility-specific outcomes or patient-level evidence.
Example use: Frame chronic disease or access indicators for a defined geography.

Price Transparency

CMS Hospital Price Transparency MRFs

Dataset ID: cms_price_transparency_mrf
Source system: Hospital-hosted CMS standard charge files.
Owning servers: price-transparency
Grain: Hospital item/service payer-plan rate rows.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: MRFs are large, stale or incomplete in places, payer-specific, and not patient out-of-pocket estimates.
Example use: Compare negotiated-rate rows only when payer, plan, code, facility, and source period are preserved.

Ownership and Enrollment

CMS PECOS Medicare FFS Public Provider Enrollment

Dataset ID: cms_pecos_public_provider_enrollment
Source system: CMS Provider Enrollment datasets.
Owning servers: provider-enrollment
Grain: One row per public Medicare enrollment record.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Exact IDs matter; do not merge facilities, owners, or systems by name similarity alone.
Example use: Check enrollment rows by NPI, CCN, or PECOS identifiers.

CMS PECOS Hospital Enrollments

Dataset ID: cms_pecos_hospital_enrollments
Source system: CMS Provider Enrollment datasets.
Owning servers: provider-enrollment
Grain: One row per hospital enrollment record.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Enrollment is a public administrative record, not quality, ownership, or licensure proof by itself.
Example use: Attach enrollment context to a hospital identity profile.

CMS PECOS Hospital All Owners

Dataset ID: cms_pecos_hospital_owners
Source system: CMS Provider Enrollment datasets.
Owning servers: provider-enrollment
Grain: One row per hospital owner or management-control relationship.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Owner IDs and enrollment IDs must travel with the claim.
Example use: Trace source-specific ownership relationships for a CCN.

CMS PECOS Hospital Change of Ownership

Dataset ID: cms_pecos_hospital_chow
Source system: CMS Provider Enrollment datasets.
Owning servers: provider-enrollment
Grain: One row per hospital CHOW event or linked CHOW owner row.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: CHOW context is not legal ownership advice or proof of current control outside the source period.
Example use: Build a CHOW trace with event dates and row receipts.

CMS PECOS SNF Enrollments

Dataset ID: cms_pecos_snf_enrollments
Source system: CMS Provider Enrollment datasets.
Owning servers: provider-enrollment
Grain: One row per skilled nursing facility enrollment record.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Keep SNF enrollment separate from hospital enrollment unless identifiers support the join.
Example use: Review SNF enrollment records by source identifiers.

CMS PECOS SNF All Owners

Dataset ID: cms_pecos_snf_owners
Source system: CMS Provider Enrollment datasets.
Owning servers: provider-enrollment
Grain: One row per SNF owner or management-control relationship.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Owner names are candidate context unless exact owner IDs support the relationship.
Example use: Trace SNF owner relationships with owner IDs preserved.

CMS PECOS SNF Change of Ownership

Dataset ID: cms_pecos_snf_chow
Source system: CMS Provider Enrollment datasets.
Owning servers: provider-enrollment
Grain: One row per SNF CHOW event or linked CHOW owner row.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Do not infer current ownership beyond the source-specific event record.
Example use: Add CHOW event context to a SNF ownership trace.

Public Records and Compliance

Healthcare Public Records

Dataset ID: public_records
Source system: SAM.gov, USAspending, CHPL, HHS OCR, HHS OIG.
Owning servers: public-records
Grain: Entity, certification, contract, or breach records.
Cadence/readiness: Mixed source-published and live/API-backed public records; readiness is source-specific.
Limitations: Screening support only; zero-result screens are not legal clearance.
Example use: Collect public compliance, breach, procurement, and certification context with caveats.

HHS OIG List of Excluded Individuals/Entities

Dataset ID: hhs_oig_leie
Source system: HHS Office of Inspector General LEIE downloadable file.
Owning servers: public-records
Grain: One row per currently excluded individual or entity.
Cadence/readiness: 31-day freshness target for the local LEIE cache.
Limitations: Screening support only; SSN/EIN-level verification is outside the downloadable file.
Example use: Screen an NPI exactly before treating name search as potential-match context.

SAM.gov Exclusions

Dataset ID: sam_gov_exclusions
Source system: SAM.gov Entity Information API.
Owning servers: public-records
Grain: One row per active SAM.gov exclusion record returned by the v4 API.
Cadence/readiness: Live/API-backed; subject to source API limits and key requirements where applicable.
Limitations: Zero-result searches are not legal clearance; verify full SAM.gov records and agency guidance.
Example use: Add UEI/CAGE or name-based SAM context when an API key is configured.

PHC4 Public Reports

Dataset ID: phc4_public_reports
Source system: Pennsylvania Health Care Cost Containment Council public reports.
Owning servers: public-records
Grain: PHC4 public report artifact or extracted table row.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Do not substitute PHC4 public reports for exact CMS quality, cost-report, enrollment, or paid PHC4 discharge facts.
Example use: Use as adjacent Pennsylvania public-report context with report URL and year preserved.

State Health Reporting and Finance

State Public Hospital Data Acquisition Index

Dataset ID: state_health_data
Source system: PA DOH, NJ DOH, Delaware DHSS, PHC4, and AHRQ public artifacts.
Owning servers: public-records, workforce-analytics, financial-intelligence
Grain: Public source artifact, normalized metadata row, or facility-year public metric.
Cadence/readiness: Source-published/import-backed depending on state source; local freshness exposed where cached.
Limitations: State coverage varies; missing state source data is not a negative fact.
Example use: Identify whether a state public report source is available before citing state-level operations or finance facts.

AHRQ Hospital Financial Measures Database

Dataset ID: ahrq_hfmd
Source system: AHRQ HFMD.
Owning servers: financial-intelligence
Grain: Hospital-year financial measure row.
Cadence/readiness: Source-published; local freshness exposed through cache status/source metadata.
Limitations: Do not infer current financial performance from stale public filings or missing fields.
Example use: Add public hospital financial measures to a finance profile with year visible.

Pennsylvania DOH Hospital Reports

Dataset ID: pa_hospital_reports
Source system: Pennsylvania Department of Health.
Owning servers: workforce-analytics
Grain: PA hospital report artifact or facility-year public operations metric.
Cadence/readiness: Source-published/import-backed; availability depends on local source import where applicable.
Limitations: State-specific public reports do not generalize to national coverage.
Example use: Use public PA report context for staffing or throughput only with source period preserved.

New Jersey Hospital Public Data

Dataset ID: nj_hospital_public_data
Source system: New Jersey Department of Health.
Owning servers: financial-intelligence, workforce-analytics
Grain: NJ public hospital financial, charity-care, or utilization artifact row.
Cadence/readiness: Source-published/import-backed; availability depends on local source import where applicable.
Limitations: Missing NJ artifact rows are not proof of no utilization, finance, or charity-care activity.
Example use: Attach NJ public hospital finance or utilization context with artifact identity preserved.

Delaware Hospital Discharge Public Data

Dataset ID: de_hospital_discharge
Source system: Delaware DHSS.
Owning servers: workforce-analytics
Grain: DE public hospital discharge artifact or facility-year utilization row.
Cadence/readiness: Source-published/import-backed; availability depends on local source import where applicable.
Limitations: State discharge artifacts are public aggregate context, not patient-level facts.
Example use: Use Delaware public utilization context only with source year and facility identity visible.

Workforce and Referral Networks

Healthcare Workforce and Labor Datasets

Dataset ID: workforce_labor
Source system: HRSA, CMS, PA DOH, ACGME, NLRB, BLS.
Owning servers: workforce-analytics
Grain: Shortage area, cost report staffing, residency program, or labor action.
Cadence/readiness: Live/API-backed for BLS/HRSA where configured; import-backed for ACGME and local state sources.
Limitations: Workforce values must preserve source period and identity basis; missing local imports are not zero activity.
Example use: Add labor, staffing, shortage-area, or residency context to an operations profile.

Physician Compare and Medicare Utilization

Dataset ID: physician_compare_utilization
Source system: CMS Physician Compare and Provider Summary datasets.
Owning servers: physician-referral-network
Grain: Physician profile and provider-service utilization rows.
Cadence/readiness: Source-published/import-backed; availability depends on local source import.
Limitations: Physician mix and utilization rows are analytic context, not employment verification.
Example use: Build physician specialty mix context when source rows are available.

DocGraph Shared Patient Referral Data

Dataset ID: docgraph_referrals
Source system: CareSet DocGraph.
Owning servers: physician-referral-network
Grain: Directed NPI-to-NPI shared-patient edge.
Cadence/readiness: Import-backed; availability depends on local source import.
Limitations: DocGraph is licensed/import-only and should not imply complete leakage analytics or network adequacy.
Example use: Assess referral/leakage readiness only when licensed import coverage is present.

Research and Web Intelligence

NIH RePORTER Projects

Dataset ID: nih_reporter_projects
Source system: NIH RePORTER API v2.
Owning servers: research-trials
Grain: One row per NIH-funded project result.
Cadence/readiness: Live/API-backed; subject to source API limits and key requirements where applicable.
Limitations: Organization names and aliases need reviewed identifiers before entity-level claims.
Example use: Profile research activity for a system or sponsor with project receipts attached.

ClinicalTrials.gov Studies

Dataset ID: clinicaltrials_gov
Source system: ClinicalTrials.gov API v2.
Owning servers: research-trials
Grain: One row per clinical study.
Cadence/readiness: Live/API-backed; subject to source API limits and key requirements where applicable.
Limitations: Sponsor and site names must not be merged without exact or reviewed identifiers.
Example use: Inventory public trial activity while preserving NCT IDs and unresolved names.

Healthcare Web Intelligence

Dataset ID: web_intelligence
Source system: Google CSE, Google News RSS, Proxycurl, CMS PI, bundled GPO directory.
Owning servers: web-intelligence
Grain: Search result, page, executive, news, or GPO association.
Cadence/readiness: Live/API-backed; subject to source API limits and key requirements where applicable.
Limitations: Public web is untrusted lead evidence; no-result searches are not proof of absence.
Example use: Collect alias, domain, leadership, news, or GPO leads for reviewed follow-up.

MCP Metadata

MCP Metadata and Discovery Surfaces

Dataset ID: mcp_metadata_surfaces
Source system: Healthcare Data MCP server registry and checked-in metadata catalogs.
Owning servers: discovery, gateway
Grain: Server capability, dataset catalog, workflow plan, preset, and gateway document metadata.
Cadence/readiness: Release-coupled registry metadata.
Limitations: The metadata gateway does not proxy live healthcare tools.
Example use: Discover dataset IDs, workflow plans, cache status, presets, and safe gateway metadata.