Skip to main content

​
Processor Output

The Processor Output object represents the structured data extracted by an Enzo document processor. It contains fields, subfields, insights, confidence scores, and references to the original document.

​
Structure

A processor output contains one or more field objects, each representing a specific type of extracted information.

​
Properties

  • fields (array)
    An array of field objects extracted from the document.
  • processorId (string)
    A unique identifier for the document processor that generated this output.

​
Field Object

Each field in the processor output represents a specific type of information extracted from the document.
  • id (string)
    A unique identifier for the field.
  • name (string)
    The name of the field, describing the type of information it contains.
  • type (string)
    The data type of the field (e.g., “object”, “array”, “string”, “date”).
  • description (string)
    A description of what the field represents.
  • value (object)
    The extracted value for this field, which may contain nested subfields.
    • subfields (array)
      An array of subfield objects that provide more detailed information about the field.
    • insights (array)
      Additional insights or analysis derived from the field data.
    • confidence (number)
      A confidence score between 0 and 1 indicating the processor’s confidence in the extracted information.
    • references (array)
      References to the original document where this information was extracted from.

​
Subfield Object

Subfields provide more granular information about a field.
  • id (string)
    A unique identifier for the subfield.
  • name (string)
    The name of the subfield.
  • type (string)
    The data type of the subfield (e.g., “string”, “date”, “number”).
  • description (string)
    A description of what the subfield represents.

​
Example

{
  "fields": [
    {
      "id": "field_Tz9",
      "name": "activity_restrictions",
      "type": "object",
      "description": "Any restrictions on the patient's activities",
      "value": {
        "subfields": [
          {
            "id": "subfield_Ii1",
            "name": "text",
            "type": "string",
            "description": "the verbatim quote from the medical record."
          },
          {
            "id": "subfield_XGH",
            "name": "date",
            "type": "date",
            "description": "when was the restriction given?"
          }
        ],
        "insights": [],
        "confidence": 0,
        "references": []
      }
    }
  ],
  "processorId": "dp_k4taf8H1eZVDKB0YQUda_"
}

​
Common Field Types

Depending on the document processor, various field types may be extracted:
  • Patient Demographics: Information about the patient such as name, date of birth, gender, etc.
  • Medical History: Patient’s past medical conditions and treatments.
  • Diagnoses: Current medical conditions with associated ICD codes.
  • Medications: Current medications with dosage, frequency, and route.
  • Vital Signs: Measurements like blood pressure, heart rate, temperature, etc.
  • Lab Results: Laboratory test results with reference ranges.
  • Activity Restrictions: Limitations on patient activities as shown in the example.
  • Treatment Plans: Recommended treatments and follow-up care.

​
Field Confidence

The confidence score indicates how certain the processor is about the extracted information:
  • 0.9-1.0: Very high confidence, likely accurate
  • 0.7-0.9: High confidence
  • 0.5-0.7: Moderate confidence
  • 0.3-0.5: Low confidence
  • 0.0-0.3: Very low confidence, may require manual review
Fields with low confidence scores may need human verification before being used for clinical decision-making.