clinical-data-extraction-schema.json

{
  "$schema": "http://json-schema.org/draft-07/schema#",
  "type": "object",
  "properties": {
    "header_fields": {
      "properties": {
        "Patient_Name": {
          "$comment": "Patient's full name as it appears across documents.",
          "description": "Full name of the patient.",
          "examples": ["GEORGE VETTICADEN"],
          "type": "string"
        },
        "Date_Of_Birth": {
          "$comment": "Date of birth is crucial for patient identification.",
          "description": "Patient's date of birth.",
          "examples": ["1980-03-26"],
          "format": "date",
          "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
          "type": "string"
        },
        "Report_Date": {
          "$comment": "The date the health record was generated.",
          "description": "Date the health report was generated.",
          "examples": ["2025-05-16"],
          "format": "date",
          "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
          "type": "string"
        },
        "Patient_Age": {
          "$comment": "Age calculated from date of birth to report date.",
          "description": "Patient's age at the time of the report.",
          "examples": [45],
          "type": "integer"
        }
      },
      "required": ["Patient_Name", "Date_Of_Birth", "Report_Date", "Patient_Age"],
      "type": "object"
    },
    "Provider_Information": {
      "description": "Information about healthcare providers.",
      "items": {
        "properties": {
          "Provider_Name": {
            "$comment": "Provider names must be exact as they appear in documents.",
            "description": "Name of the healthcare provider.",
            "examples": ["Duly Health and Care", "Northwestern Medicine", "Edward-Elmhurst Health"],
            "type": "string"
          },
          "Provider_Type": {
            "$comment": "Type categorizes the provider's specialty or role.",
            "description": "Type or specialty of the provider.",
            "examples": ["Primary Care", "Specialist", "Hospital"],
            "type": "string"
          },
          "Provider_ID": {
            "$comment": "Optional identifier for the provider.",
            "description": "Unique identifier for the provider if available.",
            "type": "string"
          }
        },
        "required": ["Provider_Name"],
        "type": "object"
      },
      "type": "array"
    },
    "Allergies": {
      "description": "Patient allergy information.",
      "items": {
        "properties": {
          "Allergy": {
            "$comment": "Allergy information is crucial for medication safety.",
            "description": "Description of the allergy.",
            "examples": ["No Known Allergies", "Penicillin"],
            "type": "string"
          },
          "Record_Date": {
            "$comment": "Record date helps track when allergies were documented.",
            "description": "Date when the allergy was recorded.",
            "examples": ["2025-04-09"],
            "format": "date",
            "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
            "type": "string"
          },
          "Provider": {
            "$comment": "Provider who documented the allergy.",
            "description": "Healthcare provider who recorded the allergy.",
            "examples": ["Duly Health and Care"],
            "type": "string"
          }
        },
        "required": ["Allergy", "Record_Date", "Provider"],
        "type": "object"
      },
      "type": "array"
    },
    "Conditions": {
      "description": "Medical conditions and diagnoses.",
      "items": {
        "properties": {
          "Diagnosis_Date": {
            "$comment": "Diagnosis dates enable chronological analysis.",
            "description": "Date when the condition was diagnosed.",
            "examples": ["2023-09-28"],
            "format": "date",
            "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
            "type": "string"
          },
          "Provider": {
            "$comment": "Provider information helps track diagnosis sources.",
            "description": "Healthcare provider who made the diagnosis.",
            "examples": ["Duly Health and Care"],
            "type": "string"
          },
          "Condition_Name": {
            "$comment": "Condition names must be standardized for analysis.",
            "description": "Name or description of the medical condition.",
            "examples": ["Generalized anxiety disorder", "Attention deficit hyperactivity disorder", "Rheumatoid arthritis"],
            "type": "string"
          },
          "Status": {
            "$comment": "Status indicates if condition is active or resolved.",
            "description": "Current status of the condition.",
            "examples": ["Active", "Resolved", "Chronic"],
            "type": "string"
          },
          "Related_Symptoms": {
            "$comment": "Symptoms help characterize the condition presentation.",
            "description": "Symptoms associated with the condition.",
            "type": "string"
          }
        },
        "required": ["Diagnosis_Date", "Provider", "Condition_Name"],
        "type": "object"
      },
      "type": "array"
    },
    "Immunizations": {
      "description": "Vaccination history.",
      "items": {
        "properties": {
          "Immunization_Date": {
            "$comment": "Immunization dates track vaccination history.",
            "description": "Date when the immunization was administered.",
            "examples": ["2022-10-22"],
            "format": "date",
            "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
            "type": "string"
          },
          "Provider": {
            "$comment": "Provider information tracks vaccination sources.",
            "description": "Healthcare provider who administered the vaccine.",
            "examples": ["Duly Health and Care"],
            "type": "string"
          },
          "Vaccine_Name": {
            "$comment": "Vaccine names must be standardized for analysis.",
            "description": "Name of the vaccine.",
            "examples": ["influenza, injectable, quadrivalent", "Pfizer-BioNTech COVID-19 Vaccine"],
            "type": "string"
          },
          "Vaccine_Type": {
            "$comment": "Type categorizes the vaccine's method or purpose.",
            "description": "Type or category of the vaccine.",
            "examples": ["Influenza", "COVID-19", "Tdap"],
            "type": "string"
          }
        },
        "required": ["Immunization_Date", "Provider", "Vaccine_Name"],
        "type": "object"
      },
      "type": "array"
    },
    "Procedures": {
      "description": "Medical procedures performed.",
      "items": {
        "properties": {
          "Procedure_Date": {
            "$comment": "Procedure dates enable chronological analysis.",
            "description": "Date when the procedure was performed.",
            "examples": ["2020-11-24"],
            "format": "date",
            "pattern": "^\\d{4}-\\d{2}-\\d{2}$",
            "type": "string"
          },
          "Provider": {
            "$comment": "Provider information helps track procedure sources.",
            "description": "Healthcare provider who performed the procedure.",
            "examples": ["Duly Health and Care"],
            "type": "string"
          },
          "Procedure_Name": {
            "$comment": "Procedure names must be standardized for analysis.",
            "description": "Name or description of the procedure.",
            "examples": ["NJX AA&/STRD TFRML EPI LUMBAR/SACRAL 1 LEVEL", "Neurosurgery Procedure"],
            "type": "string"
          },
          "Procedure_Type": {
            "$comment": "Type categorizes the procedure's method or purpose.",
            "description": "Type or category of the procedure.",
            "examples": ["Surgical", "Diagnostic", "Therapeutic"],
            "type": "string"
          },
          "Procedure_Result": {
            "$comment": "Results describe the procedure outcome if available.",
            "description": "Outcome or result of the procedure.",
            "type": "string"
          }
        },
        "required": ["Procedure_Date", "Provider", "Procedure_Name"],
        "type": "object"
      },
      "type": "array"
    }
  },
  "claude:options": {
    "extractionHints": {
      "global": "Extract all dates in ISO 8601 format (YYYY-MM-DD) for database compatibility. Numbers should be extracted as numeric values without text modifiers.",
      "header_fields": "Extract patient name exactly as it appears in the document. Calculate age accurately from Date_Of_Birth to Report_Date.",
      "Provider_Information": "Standardize provider names to match exactly one of the three main providers (Duly Health and Care, Northwestern Medicine, Edward-Elmhurst Health) or their sub-entities.",
      "Allergies": "Extract allergies exactly as listed in the document. If 'No Known Allergies' is indicated, record it as such.",
      "Conditions": "Extract condition names exactly as they appear in the document. Standardize status values to 'Active', 'Resolved', or 'Chronic' where possible.",
      "Immunizations": "Extract vaccine names exactly as they appear in the document. Categorize vaccines into standardized types where possible.",
      "Procedures": "Extract procedure names exactly as they appear in the document. Categorize procedures into standardized types (Surgical, Diagnostic, Therapeutic) where possible."
    },
    "validationHints": {
      "Dates": "Verify all dates are in valid ISO 8601 format (YYYY-MM-DD) and chronologically possible.",
      "PatientConsistency": "Ensure patient name is consistent across all documents for the same patient.",
      "ProviderConsistency": "Standardize provider names to match exactly one of the three main providers or their sub-entities."
    },
    "globalFormatting": {
      "dateFormat": "All dates must be extracted in ISO 8601 format (YYYY-MM-DD) for proper database compatibility and date calculations.",
      "numberFormat": "Extract numeric values without text modifiers, preserving decimal precision as shown in the document."
    }
  }
}