Open research datasets harmonized to the OMOP CDM can be used to evaluate COPD AI models

This work presents the harmonization of three open research datasets, Kaggle COPD datasets, MIMIC3, and UK Biobank, to the OMOP Common Data Model to support the development and evaluation of AI models for COPD diagnosis and management. Through structural and semantic mapping and implementation of ETL processes, COPD related clinical variables were standardized to ensure consistency and interoperability across databases. The study demonstrates how publicly accessible datasets mapped to OMOP can facilitate cross database analyses, simplify data access, and foster collaborative development and assessment of AI solutions for respiratory disease research.