This abstract explores how integrating Intelligent Medical Objects (IMO) Precision Normalize™ API into the OMOP ETL process can significantly improve the fidelity of mapping EHR data to SNOMED-CT® concepts. Results show a dramatic increase in patient recall for over 4,400 ICD-10-CM codes, strengthening the accuracy of observational research using the OHDSI Common Data Model.

Key Highlights:

• Identifies major data quality gaps when mapping EHR interface terminology solely via ICD-10-CM.

• Compares legacy ETL vs. enhanced ETL pipelines leveraging IMO’s terminology service.

• Reveals 4,473 ICD-10 codes with improved patient recall using the enhanced approach.

• Demonstrates significant gains in retrieving patients for under-represented diagnoses (e.g., 6,000+ added for Z40.00).

• Underscores the importance of faithfully capturing clinical intent to support high-quality real-world evidence.