FHIR Converter

FHIR Converter is an open source project that enables conversion of health data from legacy formats to FHIR.

The first version of the FHIR Converter released to open source on Mar 6th, 2020. It used Handlebars template language and Javascript runtime. A new converter engine was released on Nov 13, 2020 that uses Liquid templating language and .Net runtime.

Both Handlebars and Liquid converters, and corresponding templates/filters, are supported by Microsoft. We recommend using Liquid converter for better alignment with Azure API for FHIR, FHIR Server for Azure, and Microsoft Logic Apps.

The following table compares the two converter engines:

⚠ Rest of this document is about the Liquid converter. For the Handlebars converter, refer to the Handlebars branch.

The Converter makes use of templates that define the mappings between different data formats. The templates are written in Liquid templating language and make use of custom filters, which make it easy with work with HL7 v2 messages.

The converter comes with ready to use templates for HL7v2 to FHIR conversion. These templates are based on the spreadsheet created by the HL7 2-To-FHIR project. If needed, you can create new, or modify existing templates to meet your specific conversion requirements.

FHIR Converter with DotLiquid engine is integrated into the Azure API for FHIR, and FHIR Server for Azure as the $convert-data operation. In addition, it is also available as a command-line tool. The converter transforms the input data into FHIR bundles that can be persisted to a FHIR server.

This project consists of the following components:

1. A command-line tool) for converting data and managing templates.
2. Templates for HL7 v2 to FHIR conversion.
3. Sample data for testing purpose.

Using the FHIR Converter

$convert-data operation in the Azure API for FHIR

FHIR Converter is integrated into Azure API for FHIR, and FHIR Server for Azure to run as part of the service. Refer to the $convert-data documentation for using the FHIR converter in the FHIR Server for Azure.

Command-line tool

Convert Data

The command-line tool can be used to convert a folder containing HL7 v2 messages to FHIR resources. Here are the parameters that the tool accepts:

Example usage to convert HL7 v2 messages to FHIR resources in a folder:

>.\Microsoft.Health.Fhir.Liquid.Converter.Tool.exe convert -d myTemplateDirectory -r ADT_A01 -i myInputDataFolder -o myOutputDataFolder

Manage Templates

The command-line tool also supports managing different versions of templates from Azure Container Registry (ACR). Users can customize templates and store them on ACR if default templates can not meet requirements. After ACR authentication, users can pull and push templates from/to a remote ACR through our tool.

Example command to push a collection of templates to ACR image from a folder:

>.\Microsoft.Health.Fhir.Liquid.Converter.Tool.exe push testacr.azurecr.io/templatetest:default myInputFolder

Example usage of pulling an image of templates in a folder:

>.\Microsoft.Health.Fhir.Liquid.Converter.Tool.exe pull testacr.azurecr.io/templatetest@sha256:412ea84f1bb1a9d98345efb7b427ba89616ec29ac332d543eff9a2161ca12a58 myOutputFolder

More details of usage are given in Template Management CLI tool.

Besides current version of templates given in our project, other versions that released by Microsoft are stored in a public ACR: healthplatformregistry.azurecr.io, users can directly pull templates from healthplatformregistry.azurecr.io/hl7v2defaulttemplates:<version> without authentication.

Note!: Template version is aligned with the version of FHIR Converter.

A note on Resource ID generation

The default templates provided with the Converter computes resource ids using the fields present in the input data. In order to preserve the generated resource ids, the converter generates PUT calls, instead of POST calls.

There are a set of ID generation templates to help generate FHIR resource IDs from HL7 v2 messages.

An ID generation template does 3 things: 1) extract identifiers from input segment or field; 2) combine the identifers with resource type and base ID (optional) as hash seed; 3) compute hash as output ID.

The Converter introduces a concept of "base resource/base ID". Base resources are independent entities, like Patient, Organization, Device, etc, whose IDs are defined as base ID. Base IDs could be used to generate IDs for other resources that relate to them. It helps enrich the input for hash and thus reduce ID collision. For example, a Patient ID is used as part of hash input for an AllergyIntolerance ID, as this resource is closely related with a specific patient.

Below is an example where an AllergyIntolerance ID is generated, using ID/AllergyIntolerance template, AL1 segment and patient ID as its base ID. The syntax is {% evaluate [id] using [template] [variables] -%}.

{% evaluate allergyIntoleranceId using 'ID/AllergyIntolerance' AL1: al1Segment, baseId: patientId -%}

Reference documentation

• Filters summary
• Snippet concept

External resources

• DotLiquid wiki
• HL7 Community 2-To-FHIR-Project

Contributing

This project welcomes contributions and suggestions. Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit the CLA site.

When you submit a pull request, a CLA bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., status check, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.

This project has adopted the Microsoft Open Source Code of Conduct. For more information see the Code of Conduct FAQ or contact [email protected] with any additional questions or comments.