Translating from OMOP to FHIR to bridge the gap between electronic health record formats

Industry Sector: Health Care

Business Function: Research & Development

Flexibly convert between Observational Medical Outcomes Partnership (OMOP) and Fast Healthcare Interoperability Resources (FHIR) formats for healthcare data to allow researchers and clinicians to access the data in the format that’s most valuable and familiar to them.

Challenge

Having flexibility between Clinical Data Management (CDM) systems for Electronic Health Record (EHR) data is essential to healthcare organizations as it bridges the gap between analytical Real World Evidence (RWE) workflows and operational clinical ones. Of the many CDMs out there, OMOP and FHIR are widely used, yet have very different formats and serve fairly different purposes. Users who are familiar interacting with FHIR may not be familiar with OMOP, and vice versa.

Solution

The OMOP to FHIR mapping allows us to convert between the two out of the box, via a series of templatized Foundry transforms.

Translating from OMOP to FHIR to Bridge the Gap Between Electronic Health Record Formats

Stakeholders and user groups

Researchers
Clinical Practitioners

Impact

A large medical research agency was able to collect EHR data from a large number of hospitals and safely convert it to OMOP and then to FHIR using this approach, from which it was used in a wide range of RWE and clinical workflows. It’s available as a Foundry Template, meaning that it can be deployed in a matter of days instead of months of development.

How it’s made

The OMOP-to-FHIR connector is a series of transforms that map the OMOP specification into the FHIR specification (links below). There’s some loss of fidelity between the two formats, but it provides a good basis for understanding the OMOP data and can be further refined as necessary. It’s been packaged in a Foundry Template, meaning that it’s redeployable in hours or days rather than needing to be developed over months.

OMOP - Maintained by OHDSI (https://www.ohdsi.org/data-standardization/the-common-data-model/)
FHIR - Maintained by HL7 (https://www.hl7.org/fhir/overview.html)

Want more information on this use case? Looking to implement something similar? Get started with Palantir. ↗