Transforming the Medicare Claims Data to the OMOP Common Data Model

Area of Science:

• Health Informatics
• Data Science
• Clinical Research Informatics

Background:

• Integrating diverse healthcare datasets like claims and electronic health records (EHRs) is crucial for real-world data analytics.
• Challenges arise from data structure variations and differing clinical terminologies, hindering interoperability.

Purpose of the Study:

• To develop a systematic Extract, Transform, Load (ETL) pipeline for transforming Medicare claims data.
• To convert Medicare claims into the Observational Medical Outcomes Partnership Common Data Model (OMOP CDM) for standardized analytics.

Main Methods:

• Utilized the ETL-CMS framework and open-source ETL tools, specifically WhiteRabbit, for data transformation.
• Implemented rigorous testing and quality checks using the Data Quality Dashboard to ensure data accuracy and consistency.

Main Results:

• Achieved a 99.2% passing rate in the quality check assessment, confirming data accuracy.
• Demonstrated accurate representation of demographics and clinical comorbidities, with minor variations due to source data syntax.

Conclusions:

• The developed scalable ETL framework promotes healthcare data interoperability and adherence to global standards.
• This approach advances the integration of diverse healthcare datasets, enabling more robust collaborative clinical research and improved analytics.