Usage Considerations for FHIR Questionnaires

Sharing FHIR Questionnaires

FHIR Questionnaires can be shared between implementations, as their design does not depend on other FHIR resources or external information models. In this way, it is possible to create a repository of FHIR Questionnaires that anyone can access. In the case of Cancer Synoptic Reporting, one group can take the responsibility of transforming the paper-based forms into FHIR Questionnaires and making them available for the community of users.

As an example, the SNOMED International Synoptic Cancer Reporting Clinical Reference Group maintains a GitHub repository with FHIR Questionnaires in: https://github.com/IHTSDO/cancer-synopting-protocol-fhir-questionnaires

This repository records all versions of the questionnaires, with the ability to identify changes and download them for local implementations. Using the collaborative tools in the GitHub repository, it is possible to fork the repository to introduce local customizations, report and discuss bugs and problems in the "Issues" section, and propose questionnaire improvements or changes using the pull request feature.

Rendering Questionnaires in a Clinical Application

It is possible to dynamically create data entry user interfaces based on questionnaire definitions, introducing rich and flexible data capture in clinical applications. Some ready-made open-source libraries can simplify incorporating these functions in clinical software.

This is an example of a Prostate Cancer Form rendered using an open-source library:

SMART on FHIR Integrations

The SMART on FHIR specifications allow the safe incorporation of a questionnaire rendering "plugin" into commercial clinical software. These plugins are executed locally on the client servers and don't share information with external parties.

The National Library of Medicine of the US has published an example SMART Application that can be adapted to render any questionnaire using this standard: https://github.com/lhncbc-fhir/lforms-fhir-app

How do the questionnaires coexist with the rest of the clinical data?

A clinical application records diagnoses, observations, and procedures. Capturing the same information using a FHIR Questionnaire can lead to model duplication, which risks inconsistencies and complicates data retrieval and analytics later. FHIR Standards proposed by the Structured Data Entry group can help minimize these risks and facilitate the integration of the questionnaires with the rest of the information model.

Automatic Population

The use of automatic population helps to reduce the pain of having to fill in the same information 'yet again' by allowing a form to automatically fill in answers already known to the EHR or other data source. The user can then verify that the information is still correct (and revise if necessary) rather than needing to fill out the information all over again (and possibly accidentally omitting or incorrectly entering some data).

Data Extraction

Data extraction procedures allow data captured in a QuestionnaireResponse to be extracted and used to create or update other FHIR resources, enabling the data to be more easily searched, compared, and used by other FHIR systems.

Read more about the Structure Data Capture guides on their website.

Maintaining terminology bindings in FHIR Questionnaires

FHIR Questionnaires contain direct references to SNOMED CT concepts as the codes for questions or responses and references as part of ECL expressions. With each new release of SNOMED CT, it is necessary to validate all referenced content to detect inactivations and make any necessary replacements.

SNOMED International provides a tool for validating terminology bindings, allowing FHIR Questionnaires to be uploaded and inactive codes to be replaced using historical associations published with each release.

https://ihtsdo.github.io/sct-implementation-demonstrator/#/questionnaires