CohortLoom

Sample evaluation licence

Release sample-v1. Approved 2026-10-17. The same text is published at /legal/sample-license/ on the CohortLoom site.

What this is

This free sample is synthetic by construction: simulated patients, with documents written by language models from the simulated charts. It contains no real patient data and no PHI.

You may

  • use the sample to evaluate CohortLoom and for non-commercial research;
  • share the sample with credit to CohortLoom.

You may not

  • present the sample as real patient data, de-identified data, or clinically validated data;
  • use it to make decisions about the care of real people.

Third-party material

  • ICD-10-CM: CDC/NCHS, public domain.
  • LOINC codes are used under the LOINC licence. LOINC is copyright Regenstrief Institute, Inc.
  • RxNorm names are from the U.S. National Library of Medicine. CVX: CDC.
  • The code-system notices for the FHIR export are in fhir/README.md.
  • Synthea (the patient simulator) is Apache-2.0. The scoring tool in scoring/ is Apache-2.0 (see its LICENSE).

No warranty

The sample is for testing, development and research and is provided "as is". Machine-written documents can contain clinical errors; that is expected in test data and is why each file ships with an answer key.

Questions: [email protected]

This licence covers the free sample only. Paid datasets have their own licence: see the dataset license summary.