On A&P-merged templates, the Core API can return Assessment & Plan as structured subsections: one per clinical problem, each with a stable id. Use those IDs to attach ICD-10 codes per problem and to export the note problem-by-problem into an EHR.

When structured A&P is returned

Structured subsections appear when the note uses an A&P-merged template (keys ending in _AP_MERGED, for example GENERIC_SOAP_AP_MERGED). Custom organization templates with a merged Assessment & Plan section qualify too.

Discover templates and sections for the authoritative list available to your organization.

Read the generated note

Every section carries a required content union: TEXT or SUBSECTIONS.

Text sections (default, including A&P on non-merged templates):

{
  "key": "HISTORY_OF_PRESENT_ILLNESS",
  "title": "History of present illness",
  "content": {
    "type": "TEXT",
    "text": "The patient reports..."
  }
}

Structured Assessment & Plan (A&P-merged templates only):

{
  "key": "ASSESSMENT_AND_PLAN",
  "title": "Assessment & Plan",
  "content": {
    "type": "SUBSECTIONS",
    "subsections": [
      {
        "id": "550e8400-e29b-41d4-a716-446655440000",
        "title": "Hypertension",
        "text": "Continue lisinopril 10 mg daily."
      },
      {
        "id": "6ba7b810-9dad-11d1-80b4-00c04fd430c8",
        "title": "Type 2 diabetes",
        "text": "Maintain metformin; recheck A1c in 3 months."
      }
    ]
  }
}

A single note mixes both variants: A&P uses SUBSECTIONS; all other sections use TEXT.

Note request bodies accept content.type: "TEXT" or content.type: "SUBSECTIONS" on the Assessment & Plan section. To receive subsection_id on normalized conditions, the input note must use content.type: "SUBSECTIONS" with stable subsections[].id values; do not send content.type: "TEXT" when generation returned SUBSECTIONS.

Generate ICD-10 codes by subsection

Paths below use the Server API; User API equivalents exist at /v1/core/user/... with the same shapes.

1. Generate or edit the note (POST /v1/core/server/generate-note or POST /v1/core/server/edit-note-with-instructions).
2. Call POST /v1/core/server/generate-normalized-data with the full note. Preserve subsection id values on the A&P section when you resend the note:

{
  "note": {
    "locale": "ENGLISH_US",
    "template_key": "GENERIC_MULTIPLE_SECTIONS_AP_MERGED",
    "sections": [
      {
        "key": "ASSESSMENT_AND_PLAN",
        "title": "Assessment & Plan",
        "content": {
          "type": "SUBSECTIONS",
          "subsections": [
            {
              "id": "550e8400-e29b-41d4-a716-446655440000",
              "title": "Hypertension",
              "text": "Continue lisinopril 10 mg daily."
            },
            {
              "id": "6ba7b810-9dad-11d1-80b4-00c04fd430c8",
              "title": "Type 2 diabetes",
              "text": "Maintain metformin; recheck A1c in 3 months."
            }
          ]
        }
      }
    ]
  }
}

Each returned condition includes subsection_id: a UUID matching one content.subsections[].id from the input note's A&P section when the condition is tied to structured A&P, or null otherwise:

{
  "conditions": [
    {
      "coding": {
        "system": "http://hl7.org/fhir/sid/icd-10-cm",
        "code": "I10",
        "display": "Essential (primary) hypertension"
      },
      "categories": ["ENCOUNTER_DIAGNOSIS"],
      "subsection_id": "550e8400-e29b-41d4-a716-446655440000"
    },
    {
      "coding": {
        "system": "http://hl7.org/fhir/sid/icd-10-cm",
        "code": "E11.9",
        "display": "Type 2 diabetes mellitus without complications"
      },
      "categories": ["ENCOUNTER_DIAGNOSIS"],
      "subsection_id": "6ba7b810-9dad-11d1-80b4-00c04fd430c8"
    }
  ]
}

• One condition is one code; a subsection may appear on multiple conditions.
• subsection_id is null when the condition is not tied to a subsection (for example problem-list items).

Preserve subsection IDs

When editing a note before normalization or export, round-trip SUBSECTIONS verbatim so id values stay stable. POST /v1/core/server/edit-note-with-instructions accepts the same content union on each section.

Async responses and webhooks

POST /v1/core/server/generate-note-async poll responses use the X-Nabla-Api-Version header on the poll request. They follow the same content union as synchronous endpoints.

Webhook payloads (generate_note_async.succeeded, generate_normalized_data_async.succeeded) use your organization's pinned API version, not the version header on the original async POST. Until you upgrade the pinned version, webhook note sections keep the legacy { key, title, text } shape even if you call /next endpoints.

Normalized codes are not included in note generation responses. Call POST /v1/core/server/generate-normalized-data (or the async variant) after you have the final note.

See Setup Webhooks for delivery details.

Validation

• content.type: "SUBSECTIONS" on a non-A&P section returns 400 BAD_REQUEST.

See Note customization for how split_by_problem differs from structured subsections.