Integrated natural language processing (NLP) for HealthLake - AWS HealthLake

Integrated natural language processing (NLP) for HealthLake

AWS HealthLake provides integrated natural language processing (NLP) libraries for parsing, identifying, and mapping unstructured data stored in FHIR DocumentReference resource types.

Important

Integrated NLP for HealthLake is turned off by default. To have it turned on, submit a support case using AWS Support Center Console. To create your case, log in to your AWS account and choose Create case.

HealthLake integrated NLP works by calling the Amazon Comprehend Medical DetectEntities-V2, InferICD10-CM, and InferRxNorm API actions. The actions append their results as an extension to the DocumentReference resource. When the Amazon Comprehend Medical API actions detect traits that are SIGN, SYMPTOM, and/or DIAGNOSIS, they generate a FHIR Linkage resource. New Condition and Observation resources are created from entities identified with the traits of SIGN, SYMPTOM, or DIAGNOSIS and they are linked to the source document using the Linkage resource.

Note

Although GET requests are supported for FHIR resources generated by HealthLake integrated NLP, FHIR API search functionality is not. To learn more about searching NLP extensions using HealthLake's integration with Athena, see SQL index and query.

Contents