Design and validation of a FHIR-based EHR-driven phenotyping toolbox

Area of Science:

• Biomedical Informatics
• Clinical Research Informatics
• Health Data Standards

Background:

• Electronic health records (EHRs) contain vast amounts of clinical data crucial for research.
• Standardizing phenotype definitions is essential for reproducible and scalable clinical research.
• Current methods for EHR-based phenotyping can be labor-intensive and prone to error.

Purpose of the Study:

• To develop and validate a standards-based phenotyping tool for authoring EHR-based phenotype definitions.
• To demonstrate the execution of these definitions across heterogeneous clinical research data platforms.
• To facilitate automated and accurate EHR-driven phenotyping.

Main Methods:

• Developed the open-source PhEMA Workbench, utilizing Fast Healthcare Interoperability Resources (FHIR) and Clinical Quality Language (CQL) standards.
• Demonstrated phenotype authoring, execution, and validation using the tool.
• Validated performance by executing a thrombotic event phenotype definition at three clinical sites (Mayo Clinic, Northwestern Medicine, Weill Cornell Medicine).

Main Results:

• The PhEMA Workbench supports phenotype authoring, execution, and publishing to a shared repository.
• A thrombotic event phenotype definition comprised 11 CQL statements and 24 value sets (834 codes).
• Technical validation demonstrated high performance: 100% precision and recall at two sites, and 95% precision/84% recall at a third.

Conclusions:

• The PhEMA Workbench facilitates EHR-driven phenotype definition, execution, and sharing in diverse research environments.
• Integrating phenotype definitions with standards-compliant systems promotes automation and reduces human error.
• This tool enhances the reliability and efficiency of clinical research using EHR data.