Output schema¶

The pipeline emits dashboard_data.json -- a single JSON object containing the full processed bundle. This page documents its structure.

Top-level shape¶

{
  "metadata": { "generated_at": "...", "total_patients": 96 },
  "patients": [...],
  "documents": [...],
  "encounters": [...],
  "problems": [...],
  "medications": [...],
  "labs": [...],
  "vitals": [...],
  "labs_vitals": [...],
  "procedures": [...],
  "allergies": [...],
  "immunizations": [...],
  "careplans": [...],
  "diagnostic_reports": [...],
  "goals": [...],
  "notes": [...],
  "document_references": [...]
}

Per-record schemas¶

Most clinical records share a common base. Per-tab specifics extend it.

Common fields (every clinical record)¶

Field	Type	Description
`patient_id`	string	Foreign key to `patients[].patient_id`
`code`	string	Primary clinical code (e.g. SNOMED, RxNorm, LOINC, CPT)
`code_system`	string	System name (`SNOMED-CT`, `RxNorm`, `LOINC`, etc.)
`display_name`	string	Human-readable name for the code
`all_codings`	array	All codings present on the source CodeableConcept
`effective_date`	string (ISO date)	Primary date for the record
`source`	string	`"fhir"` or `"ccda"` -- which input produced this record

Patient¶

Field	Type
`patient_id`	string
`first_name`, `last_name`	string
`mrn`	string
`dob`	string (`YYYY-MM-DD`)
`gender`	string
`num_documents`	int -- count of documents linked to this patient

Document¶

Field	Type
`patient_id`	string (nullable)
`document_uuid`	string
`source_file`	string
`source_format`	string -- `"ccda_xml"`, `"html_fragment"`, `"rtf"`, `"pdf"`, `"unknown"`
`source_type`	string -- `"rtf_note"` for RTF, otherwise same as `source_format`
`plain_text`	string -- extracted text, capped at 30,000 chars

Encounter¶

source, display_name (class display), type (visit type), code (class code), effective_date, end_date, status. Aliases: start_date, visit_type.

Problem¶

source, code, code_system, display_name, effective_date, clinical_status. Aliases: onset_datetime.

Medication¶

source, code, code_system, display_name, effective_date, status, fhir_resource_type (MedicationRequest / MedicationStatement / MedicationAdministration). Aliases: authored_on, start_date, medication_subtype.

Observation (lab or vital)¶

source, code, code_system, display_name, value, unit, category, effective_date. Aliases: effective_datetime.

The category field distinguishes vitals from labs. The labs_vitals top-level key is a convenience union.

Allergy¶

source, code, code_system, display_name, effective_date, severity, clinical_status. Aliases: allergen, allergen_code, allergen_system, recorded_date.

Immunization¶

source, code, code_system, display_name, effective_date, status. Aliases: vaccine, vaccine_code, vaccine_system, occurrence_datetime, recorded_date.

Procedure¶

source, code, code_system, display_name, effective_date, status. Aliases: performed_datetime.

CarePlan¶

source, title, description, status, effective_date.

DiagnosticReport¶

source, code, code_system, display_name, status, effective_date, conclusion.

Goal¶

source, description, status, effective_date.

Section narrative (CCDA)¶

Field	Type
`patient_id`	string
`section_title`	string -- e.g. `"Problems"`, `"Medications"`, `"Vital Signs"`
`narrative_text`	string -- full text content of the CCDA `<text>` element
`character_count`	int
`source_file`	string
`source`	string -- always `"ccda"`

DocumentReference (FHIR)¶

patient_id, type_display, status, effective_date, source.

Field aliasing¶

Many records carry both a canonical name and one or more aliases. This is intentional -- it lets downstream code use whichever schema is most convenient. Aliases are added in Stage 4.

Example record¶

{
  "patient_id": "demo-patient-001",
  "code": "9322",
  "code_system": "RxNorm",
  "display_name": "Riluzole 50 MG Oral Tablet",
  "all_codings": [
    { "code": "9322", "system_name": "RxNorm",
      "display": "Riluzole 50 MG Oral Tablet" }
  ],
  "effective_date": "2023-06-01",
  "effective_datetime": "2023-06-01",
  "authored_on": "2023-06-01",
  "start_date": "2023-06-01",
  "status": "active",
  "source": "fhir",
  "fhir_resource_type": "MedicationRequest",
  "medication_subtype": "MedicationRequest"
}