How Amazon HealthLake works - Amazon HealthLake

How Amazon HealthLake works

Amazon HealthLake maintains Data Stores of health records in FHIR-compliant format. You can perform the following tasks using the Amazon HealthLake console, AWS Command Line Interface (AWS CLI), or APIs:

  • Create, monitor, and delete a Data Store

  • Import your data from an Amazon Simple Storage Service (Amazon S3) bucket into the Data Store

  • Query data using Create, Read, Update, and Delete functions

  • Use FHIR search functionality

  • Transform your data using integrated medical natural language processing (NLP)

Creating and monitoring Data Stores

With Amazon HealthLake, you can create and manage Data Stores for storing R4 FHIR Resources. Use create-fhir-datastore to create a new Data Store, describe-fhir-datastore to learn more about the properties of a Data Store, and list-fhir-datastore to see all Data Stores associated with your account and their status. When a Data Store is no longer needed, you can delete it using delete-fhir-datastore.

Create, Read, Update, Delete (CRUD) operations

Manage and query data using the CreateResource, ReadResource, UpdateResource, and DeleteResource operations for 71 different FHIR resource types. For a list of FHIR resource types supported by , see Managing FHIR resources. These operations are all handled through an HTTP client.

Integrated medical natural language processing (NLP)

HealthLake has integrated medical natural language processing for the DocumentReference resource type. HealthLake automatically runs the integrated medical NLP on all DocumentReference resources as they are written to the system in an Import operation, during a Create operation, or during an Update operation. The original resource stays unchanged, and the extracted medical information is automatically appended as custom fields. These fields are indexed and queryable after extraction.

FHIR search functionality

You can search health records that are stored in the Data Store either through a specific resource type with supported search parameters or for resource IDs in the server without specifying the resource type. When a FHIR record is created, it is first saved into a FHIR Data Store, where it can be read, updated, queried, or deleted. After the data is ingested, it is indexed using Amazon ES, which makes the data searchable. HealthLake operates with eventual consistency with the Data Store, meaning that there might be a brief latency before the data is searchable within the Data Store.

Import data

Amazon HealthLake enables you to bulk import your files from an Amazon S3 bucket. Use either the console or start-fhir-import-job to begin an import job. Afterwards, you can use describe-fhir-import-job to monitor the status of the job and discover its properties. After the import job is complete, the data can then be added to a Data Store, transformed, or analyzed and used in downstream applications.

Export data

Amazon HealthLake enables you to bulk export your files to an Amazon S3 bucket. Use either the console or start-fhir-export-job to begin an export job. Afterwards, you can use describe-fhir-export-job to monitor the status of the job and discover its properties. After the export job is complete, the data can then be visualized using AWS Quicksight or accessed by other AWS services.