From clinical text to a populated EHR

API reference Docs home

The common production task is to take a sentence or document and end up with valid resources written to a FHIR server. Tools let you do it in one call.

One resource

Lang2FHIR maps text to a US Core profile, or can detect the resource type automatically. /tools/lang2fhir-and-create generates the resource and writes it to your provider.

const { fhir_id, fhir_resource } = await client.tools.createFhirResource({
  resource: "condition-encounter-diagnosis",
  text: "Severe persistent asthma with acute exacerbation",
  provider: "550e8400-e29b-41d4-a716-446655440000",
});

A whole note

Multi-resource endpoints detect patients, conditions, medications, observations, and more, link them with references, and submit a FHIR transaction bundle.

US Core identifiers

If required identifiers are missing, Phenoml uses a synthetic identifier namespace so generated resources remain conformant. Real extracted identifiers are preserved.

See Lang2FHIR and FHIR.