generated: '2026-09-05'
method: derived
source: >-
  openapi/centers-for-disease-control-and-prevention-dibbs-ecr-refiner-openapi.json
  (85 component schemas, $ref graph and *_id reference fields),
  openapi/centers-for-disease-control-and-prevention-dibbs-query-connector-openapi.yaml,
  openapi/centers-for-disease-control-and-prevention-soda-v2-1-api-openapi.yml, and the DCAT-US
  catalog at well-known/centers-for-disease-control-and-prevention-data-json-catalog.json.
description: >-
  CDC has no single data model — it has four unrelated ones, one per surface. The only surface
  with a rich, resolvable entity graph is the DIBBs eCR Refiner; the public data APIs are
  schemaless by design (a dataset row is whatever that dataset's columns are).
domains:
  - name: eCR Refiner configuration
    surface: DIBBs eCR Refiner
    entity_count: 85
    root: Configuration
    entities:
      - name: Configuration
        schemas: [GetConfigurationResponse, GetConfigurationsResponse, CreateConfigurationResponse, GetConfigurationResponseVersion]
        id_field: id
        relationships:
          - {type: belongs_to, target: Condition, via: condition_id}
          - {type: has_one, target: DbConfigurationStatus, via: $ref}
          - {type: has_many, target: CustomCodeResponse, via: $ref}
          - {type: has_many, target: DbCode, via: $ref}
          - {type: has_many, target: DbConfigurationSectionProcessing, via: $ref}
          - {type: has_many, target: IncludedCondition, via: $ref}
          - {type: has_one, target: LockedByUser, via: $ref}
          - {type: has_one, target: DbTotalConditionCodeCount, via: $ref}
          - {type: has_one, target: Configuration, via: draft_id, note: a configuration points at its own draft}
          - {type: has_one, target: Configuration, via: active_configuration_id, note: and at the active version it supersedes}
      - name: Condition
        schemas: [Condition, GetConditionResponse, GetConditionsResponse, IncludedCondition]
        id_field: id
        relationships:
          - {type: has_many, target: DbCode, via: $ref}
          - {type: has_one, target: CompletenessStatus, via: $ref}
          - {type: has_many, target: CodeSystemsReponse, via: $ref}
      - name: Code
        schemas: [DbCode, CodeResponse, CustomCodeResponse, UpdateCustomCodeInput]
        id_field: id
        relationships:
          - {type: belongs_to, target: DbCodeSystem, via: system_id}
          - {type: belongs_to, target: Condition, via: condition_id}
      - name: DbCodeSystem
        id_field: id
        note: The terminology system a code belongs to (LOINC, SNOMED, ICD-10 and so on).
      - name: AuditEvent
        id_field: id
        relationships:
          - {type: belongs_to, target: Condition, via: condition_id}
        note: Surfaced through EventsResponse; the Refiner keeps an audit trail per configuration change.
      - name: UserResponse
        id_field: id
        relationships:
          - {type: belongs_to, target: Jurisdiction, via: jurisdiction_id}
          - {type: has_one, target: NotificationsToRender, via: $ref}
        note: >-
          jurisdiction_id is the multi-tenancy boundary — a Refiner deployment is scoped to a
          public health jurisdiction.
      - name: DiscoveredConfigurationSet
        relationships:
          - {type: belongs_to, target: Condition, via: condition_id}
          - {type: has_many, target: DiscoveredConfigurationVersion, via: $ref}
        note: Produced by the simulator's discoverConfigurations operation.
      - name: Release
        relationships:
          - {type: has_one, target: ReleaseNotes, via: $ref}
  - name: FHIR query
    surface: DIBBs Query Connector
    entity_count: 0
    note: >-
      The spec models no entities of its own — request and response bodies are HL7 FHIR R4
      resources (US Core Patient in the published example), so the data model is FHIR's, not
      CDC's. The queryable identifiers are given, family, dob, mrn, phone, gender, race and
      ethnicity, with race/ethnicity drawn from the CDC Race & Ethnicity code system
      urn:oid:2.16.840.1.113883.6.238.
  - name: Open data row
    surface: SODA (data.cdc.gov, chronicdata.cdc.gov)
    entity_count: 1
    note: >-
      A single generic entity: the dataset row, addressed by an eight-character four-by-four
      identifier and shaped entirely by the dataset. There is no cross-dataset schema and no
      foreign key between datasets. Per-dataset column metadata is available at
      /api/views/{four-by-four}.json.
  - name: Dataset catalog
    surface: data.cdc.gov/data.json
    entity_count: 1385
    schema: DCAT-US 1.1 (Project Open Data Metadata Schema)
    entities:
      - name: dataset
        id_field: identifier
        relationships:
          - {type: has_many, target: distribution, via: "distribution[]"}
          - {type: belongs_to, target: publisher, via: publisher}
          - {type: has_one, target: contactPoint, via: contactPoint}
        note: >-
          Verified against the harvested catalog: 1,385 dataset entries, each identified by a
          https://data.cdc.gov/api/views/{four-by-four} URI, which is the join key back to the
          SODA row model above.