Skip to main content
This quickstart sends one short note to the FHIR Extraction endpoint and walks through the response. At the end, you have a FHIR Bundle with a Patient, an Encounter, and one resource per clinical fact in the note.
FHIR Extraction is in beta. The request and response schemas can still change. Contact us to share feedback or report issues.

Prerequisites

Send a note

The example note has three facts in it: a diagnosis, a blood pressure reading, and an allergy.
The request sends the note in context, the only required field. It also sends the patient and encounter values that your system already knows: your own IDs, so the response links to your records, and the demographics and visit details, so the endpoint does not have to read them from the note. Do the same in your own requests. Reduce extraction errors explains why.
Request and response lists every request field and its allowed values.

Read the response

The response has two top-level fields: bundle, a FHIR Bundle with all resources, and usageInfo, the credits consumed by the request. For the example note, the Bundle holds six resources: Notice that the endpoint splits the blood pressure into two Observations, one for each number. Every measurement gets its own resource. The full response below shows the real JSON. It is long, mostly because of the extension entries, which the next section explains. It is valid FHIR R4: each clinical resource points to the Patient and Encounter through their fullUrl. See How the response meets FHIR R4.
A coding entry can also carry a version field with the version of the code system the endpoint used. The examples in these docs leave it out.

Check each value against the note

Each clinical resource carries extension entries that point back to the note:
  • Mentions link a field to the words that support it. The systolic Observation has a mention for code (“BP”, characters 59 to 61) and one for valueQuantity (“150”, characters 62 to 65).
  • Alternatives list other codes that could fit. The Condition offers 443694000 Uncontrolled type 2 diabetes mellitus as an alternative to 44054006 Type 2 diabetes mellitus.
The Patient and Encounter carry no mentions, because every value in them comes from the request. Every value points back to the note explains how to read both extensions.

Next steps

How FHIR Extraction works

Learn the rules behind every response, such as positive codes and status fields.

Reduce extraction errors

Send what you know, skip structured data, and review the results.

Filter FHIR resources

Produce only the resources you need and restrict the codes they carry.

FHIR Extraction request and response

Look up every request field, error, and limit.