Assuring End-to-End Data Quality for Analytics on FHIR
Assuring End-to-End Data Quality for Analytics on FHIR - IOS
A method was presented for ensuring end-to-end data quality when performing analytics on FHIR data.
Summary
This study proposes an approach to ensure data quality when transforming Real-World Data (RWD) from EHRs and registries into analyzable formats. Specifically, it introduces a data completeness assessment pipeline that validates the process across multiple transformation stages involving HL7 FHIR.
Details
The research addresses the substantial potential of accumulating Real-World Data (RWD) from Electronic Health Records (EHRs) and registries for generating Real-World Evidence (RWE). However, realizing robust evidence hinges critically on data quality, especially when heterogeneous data is transformed into standardized, research-ready models. The study introduces a technical solution to assess data completeness across three distinct transformation stages: from the initial data source through Health Level 7 (HL7) Fast Healthcare Interoperability Resources (FHIR), and finally to CSV. Using Trino, a distributed SQL engine, the authors evaluated data completeness by comparing cancer diagnosis counts at these three stages. The modular pipeline design is highlighted for its compatibility with various data sources and its ability to detect errors within ETL processes. Future work aims to expand this system to address additional dimensions of data quality, such as correctness and plausibility, thereby improving the overall robustness of data analytics in federated environments.
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
Original content copyright by respective publishers