By: Dr. Uri Lerner

Healthcare organizations are increasingly transitioning to FHIR to improve interoperability, enhance data sharing, and comply with evolving regulations. Mandates like the United States Core Data for Interoperability (USCDI) and the Israeli Medical Information Mobilization Law, 5784-2024, are driving this shift, necessitating the migration of legacy systems to standardized formats that support seamless data exchange. This transition, while challenging, is essential for achieving a more connected, efficient and patient-centric healthcare ecosystem.

The process of converting legacy pathology data to FHIR is intricate, requiring a nuanced understanding of both clinical data and interoperability standards. The following ten recommendations are designed to facilitate an efficient and accurate conversion. While these principles are applicable to all FHIR-related initiatives, the pathology domain has unique characteristics that pose specific challenges, as elaborated below. 

1. Comprehensively Analyse Legacy Data Sources and Dataflow

A deep understanding of the structure, content, and context of legacy pathology data is fundamental. Determine whether the data reside in flat files, relational databases, transferred via HL7 v2 messages, or use outdated terminologies such as SNOMED M and SNOMED T, or other proprietary formats. This foundational knowledge is crucial for effective mapping during the transition to FHIR. Understand the dataflow and decide whether to convert the complete business scenario (e.g., from test ordering to result delivery) or just specific components.

2. Select Appropriate FHIR Resources

Legacy pathology data should be accurately mapped to relevant FHIR resources such as DiagnosticReport, Observation, and Specimen. The choice of resource must align with the data's clinical semantics to preserve its context and intended use. For example, DiagnosticReport should represent the entire pathology report, whereas Observation can encapsulate individual measurements. If converting the complete workflow, understand which ServiceRequests are created during the process.

3. Leverage FHIR Profiles and Extensions

Employ FHIR profiles to enforce adherence to specific constraints and maintain alignment with established standards. Where unique data elements exist that are not covered by base FHIR resources or local CORE profiles, organizational profiles can be created and extensions should be utilized judiciously to accommodate these without compromising interoperability. Check out the Israel Core FHIR project here.

4. Engage Multidisciplinary Expertise

Successful data conversion hinges on collaboration between technical and clinical experts. Work with pathologists, informaticists, and terminology experts to ensure that both clinical relevance and technical rigor are maintained throughout the conversion process. This section is crucial for constructing pathology results, especially when encoding in standard terminologies like SNOMED CT, which must represent multiple parameters such as body site, tissue type, collection method, morphological abnormalities, and malignancy indications.

5. Develop a Detailed Data Mapping Strategy

Construct a robust mapping plan that translates legacy data fields to their corresponding FHIR elements. This plan should be meticulously documented to maintain transparency and to facilitate ongoing quality assurance. Special attention should be paid to maintaining semantic equivalence, which means preserving the meaning and clinical context of the data as it is mapped to new structures, ensuring that no critical clinical information is lost. Given the inherent complexity of pathology data—where multiple samples may be obtained during a single procedure, each specimen can yield multiple results, and results are often developed through several collaborative steps involving lab technicians and pathologists, incorporating both quantitative and qualitative aspects—ensuring semantic equivalence is a critical component for maintaining data quality.

6. Preserve Complex Data Relationships

Pathology data often include intricate relationships—such as between specimens, observations, and diagnostic reports—that must be preserved. Use FHIR references to maintain these connections accurately, ensuring the integrity and clinical validity of the data.

7. Standardize and Normalize Data

Legacy data are frequently heterogeneous, containing inconsistencies, missing elements, and mixed terminologies. Invest in comprehensive data normalization and standardization efforts, ensuring that units of measurement, terminologies, and date formats are harmonized to match FHIR requirements.

8. Align Terminology with Standard Codes

Terminological consistency is paramount for interoperability. Employ standard coding systems like SNOMED CT or LOINC for representing pathology concepts and results. This facilitates semantic interoperability and supports secondary uses of the data, such as research and quality improvement. Avoid retaining outdated organizational terminologies, as they can introduce data quality assurance (DQA) issues.

9. Implement Rigorous Testing Protocols

Testing must be integrated throughout the conversion. Use FHIR validators to ensure compliance and develop test cases that reflect real-world clinical scenarios to evaluate the robustness of the mapping. Iterative testing identifies issues early and ensures data quality. Form a dedicated testing team comprising both FHIR experts and pathology professionals to comprehensively validate all aspects of the results. Develop detailed test scenarios, focusing particularly on outlier cases and clinically critical situations, such as terminal cancer, advanced malignancies, or urgent but treatable conditions. This multidisciplinary approach ensures the robustness of the mapping and the clinical reliability of the data.

10. Maintain Data Provenance and Traceability

Ensure that data provenance is meticulously maintained throughout the conversion. Include metadata detailing the origin, version history, privacy considerations, and any modifications applied to the data. This provenance information is essential for ensuring transparency, establishing trust in the converted data, and complying with audit requirements.

Conclusion

Converting legacy pathology data to FHIR is inherently complex, demanding meticulous planning, deep clinical understanding, and stringent adherence to data integrity. By implementing these ten strategic recommendations, healthcare organizations can significantly improve interoperability, safeguard the clinical utility of pathology data, and enable more effective data sharing across diverse systems, ensuring both data accuracy and regulatory compliance.

Converting medical data of any type, not just pathology, can be arduous work. We at Outburn make it easy with FUME, our FHIR converter. You will definitely want to check it out!