Phenomics Exchange for Research and Diagnostics (Jira PSS-2385)

This project aims to address the ability to record, query for, and exchange structured and coded patient-level phenotypic information in an interoperable fashion. Such information is increasingly utilized in a variety of clinical and research applications such as rare disease diagnostics, public health surveillance, adverse events, real world data analytics, and cancer treatment selection. The project’s goal is closely related to but still distinct from other phenotyping efforts where the focus is on developing search algorithms and criteria to facilitate information retrieval and cohort identification given the current state of the underlying clinical information. Our project is more focused on improving the encoding and exchange of the underlying clinical information in order to improve higher level use cases that can benefit from an improved representation of phenotypic information.

Patient level phenotypic information is frequently spread across different EHR systems and different types of medical record artifacts. For example, phenotypes can be described in free text clinical notes and various other types of reports such as radiology, pathology, and laboratory reports. The majority of this information is unstructured or non-standard and this leads to significant difficulties in extracting and sharing such information to support the various research and diagnostic use cases that can benefit from such information. For example, our working use case is focused on defining the encoding and exchange of such information through FHIR to a laboratory system as part of a request to perform genomic testing for a rare disease diagnostic use case. The laboratory and medical geneticist need such information in order to provide a report of the genomics findings in the context of the phenotypic picture of the patient and their family. Currently, this information is not directly accessible by the laboratory system, and is frequently sent in unstructured formats and without the use of standard terminologies.

Ongoing efforts exist to record variants and genetic diagnoses within the EHR (see Genomics Reporting IG by the Clinical Genomics WG). However, efforts to implement similar solutions for phenotypic information is still nascent and this project will attempt to address these needs through the development of a FHIR IG to define a limited set of resource profiles and related conformance resources in order to facilitate the capture and exchange of a rich and structured phenotypic picture of a patient for both research and diagnostic purposes.

On the implementation side, our project aims to provide mappings and tooling to facilitate interoperability with the Global Alliance for Genomics & Health (GA4GH) Phenopackets Schema standard. This is a well established and ISO approved standard that is actively used in the community and providing FHIR interoperability will be a valuable resource. The goal is to enable the exchange of information through both FHIR and Phenopackets to address the needs of our research and diagnostics communities.

The initial scope will be limited to a small set of resource profiles and related conformance resources to exchange the coded phenotypes of a patient. One important scoping and use-case constraint is the ability to integrate and interoperate with any existing workflows that use the Phenopackets Schema standard. The goal is to allow such workflows to adopt FHIR as an exchange standard even if the native workflows will continue to use the Phenopackets Schema as their internal data model. However, other use cases and communities are welcome to join our project to contribute to additional development as needed.

The development of the IG started as a National Library of Medicine funded effort to model a subset of the data elements from the Phenopackets Schema in FHIR . The initial IG content was influenced by the team's well established understanding of the Phenopackets Schema while learning how to develop a FHIR IG, and the content was written from that perspective.

We are revising a significant portion of the current IG content as part of migrating the IG to be a HL7 WG IG in order to facilitate its use by the wider FHIR community without needing a deep understanding of the Phenopackets Schema. This issue is being addressed by rewriting the narrative descriptions, renaming artifacts, and adopting other FHIR IG best practices. Guidance related to interoperability with the Phenopackets Schema will still be included in the IG in separate pages and mapping artifacts accessible to the communities needing such information. We also aim to simplify and generalize the IG’s artifacts to facilitate their use in other contexts, one of which is the interoperability with the Phenopackets Schema. We will also be reaching out to other phenotyping efforts occurring in communities such OHDSI and eMERGE that are more focused in information retrieval to benefit from learned lessons and in order to provide solutions that support their use cases.