Executing SQL Queries on FHIR Data: Achieving Tabular Views with FHIRPath

Details

This article discusses 'SQL on FHIR,' a technique introduced by npj Digital Medicine for handling Fast Healthcare Interoperability Resources (FHIR) data. It moves beyond traditional JSON or XML formats, enabling data analysis using a SQL-like query language. The approach treats FHIR resources as table views, allowing developers to extract and manipulate data using functions such as `select()` and `forEach()`. This means that FHIR data can be managed within existing database environments like Spark SQL or PostgreSQL, treating it much like standard tabular data. Validation was performed using large-scale ICU datasets, such as MIMIC-IV. The results demonstrated that this method offers superior performance and flexibility compared to traditional methods. This technology is expected to have significant applications in advanced medical information analysis fields, including Clinical Decision Support Systems (CDSS) and AI development. While the FHIR standard primarily focuses on data provision via REST API or NDJSON, 'SQL on FHIR' integrates this data into a database environment, providing powerful structured query capabilities. This achieves practical data utilization by enhancing the analytical pipeline for healthcare data. It represents a major advancement in making complex medical data actionable.