Patient Authored FHIR: Recording & Implementation
- Health Level seven (HL7) Fast Healthcare Interoperability Resources (FHIR) offers several ways to distinguish patient contributions, clinician-authored data, and AI contributions within medical records.
- The primary need is to express the provenance of a resource or a specific element within it.
- For instance,tags can indicate that a patient recorded body weights at home or that a patientS partner provided a nickname.
Explore how FHIR and its provenance resources revolutionize managing patient and AI contributions within medical records. We delve into teh methods FHIR offers too differentiate contributions, ensuring data integrity, and proper attribution. Discover how lightweight tags and extensive Provenance resources provide detailed facts,capturing who,what,where,when,and why data is modified. Using tags and the meta.security element allows for the source of the data to be identified, like a clinician, device, patient, or AI. Further, understand the application of DS4P extensions for granular tagging of specific elements.News Directory 3 provides insights into the holistic approach to data management achieved via these methods. Discover what’s next in this ever-evolving landscape.
Tracking Patient and AI Contributions Using FHIR Provenance
Health Level seven (HL7) Fast Healthcare Interoperability Resources (FHIR) offers several ways to distinguish patient contributions, clinician-authored data, and AI contributions within medical records. These methods, also found in CDA and HL7 v2, are essential for maintaining data integrity and ensuring proper attribution.
The primary need is to express the provenance of a resource or a specific element within it. FHIR provides two related but distinct solutions: lightweight tags and more thorough Provenance resources. Tags are suitable for simple indications, while Provenance offers detailed tracking.
For instance,tags can indicate that a patient recorded body weights at home or that a patientS partner provided a nickname. Provenance can detail how AI generated a care plan based on lab results and clinical guidelines.
All FHIR resources have a meta.security coding element that includes provenance codes. Although labeled as security tags, these are also relevant for integrity and availability. The meta.tag element exists, but meta.security is the recommended location for provenance information.
The ValueSet for meta.security covers the security space, including availability and integrity. It contains useful codes for indicating the source of data, such as clinician, device, patient, or artificial intelligence.
Using these codes at the Resource.meta.security level indicates that the entire resource was contributed by a specific source. Such as, an entire observation can be marked as AI-contributed.
However, tagging the entire resource can be too broad. To indicate that only a specific element was contributed differently, such as a patient acquaintance providing a nickname, FHIR uses extensions defined in the Data Segmentation for Privacy (DS4P) implementation guide.
this involves using inline codes at the Resource.meta.security level to show that inline codes are used within the resource. this approach allows specific elements to be tagged with their respective sources.
While tagging provides a focused approach, it may not convey enough information in certain specific cases. Provenance resources offer a more detailed solution,capturing who,what,where,when,and why data was added or modified.
Provenance resources use the same codes as tags for identifying agents but offer more nuance through participation type code systems. They can indicate that an entire observation was contributed by the patient.
The Provenance.target extension can specify that only certain data within a resource was patient-contributed. examples of this can be found in the FHIR documentation.
What’s next
The .meta.security tags and Provenance resources are not exclusively for security purposes. While a Security Labeling Service (SLS) might populate these values, it should not overwrite explicitly set values. Security utilizes many elements in resources that are also valuable for clinical use, ensuring a holistic approach to data management.