Validation of a common data model for active safety surveillance research

Area of Science:

• Pharmacovigilance
• Health Informatics
• Observational Data Analysis

Background:

• Variations in data models and coding systems hinder systematic analysis of observational medical databases for safety surveillance.
• Existing clinical data models can be difficult for analysts to understand and may not support large-scale systematic analysis.
• A Common Data Model (CDM) can facilitate analysis and ensure faithful representation of source data.

Purpose of the Study:

• To validate the suitability of the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) for active safety surveillance.
• To assess the representation of diverse observational healthcare databases within the OMOP CDM.
• To evaluate the performance of analytic methods applied to data within the OMOP CDM.

Main Methods:

• Instantiated the OMOP CDM into a relational database.
• Loaded data from 10 different observational healthcare databases into separate instances.
• Developed and executed a comprehensive array of analytic methods on the instantiated OMOP CDM.

Main Results:

• The OMOP CDM acceptably represented data from 10 distinct observational databases using selected standardized terminologies.
• A range of analytic methods was developed and executed.
• The developed methods demonstrated sufficient performance for active safety surveillance.

Conclusions:

• The OMOP CDM, with standardized terminologies, provides acceptable data representation for active safety surveillance.
• The developed analytic methods are suitable for large-scale systematic analysis using the OMOP CDM.
• The OMOP CDM facilitates efficient and effective active safety surveillance across multiple databases.