Potential of FHIR Standard for Data Extraction: Lessons Learned from PCORnet® Pilots

Key Players

Details

A study published by Duke University investigated the utility of the Fast Health Care Interoperability Resources (FHIR) standard in supporting data activities across the PCORnet® infrastructure. The background noted that while participating institutions transform local EHR data into the proprietary PCORnet® Common Data Model (CDM), non-participating sites might find the cost prohibitive. Therefore, FHIR was tested as an alternative method for obtaining EHR data. The project aimed to determine if FHIR-based data extraction could be a viable study solution for sites not participating in PCORnet®. The methodology included surveying sites on FHIR availability and comparing FHIR extracts (using REDCap) against PCORnet® CDM extracts across 3 sites, focusing on 40 patients. Results showed that 73% of responding organizations reported deploying FHIR in a production capacity. However, the concordance measures varied significantly by data type: Cohen kappa ranged from 0.01 to 0.76 for diagnoses, 0.24 to 0.84 for lab results, and 0.1 to 0.87 for medications. In conclusion, while differences in the extracted data were observed, the study concluded that FHIR-based extracts may be well-suited for certain types of studies. This highlights FHIR's potential role as a critical interoperability standard for large-scale epidemiological research and data sharing infrastructure.

Technology Note

FHIR（Fast Healthcare Interoperability Resources）は医療データ交換の国際標準。このエントリの関連技術: FHIR