Representing and utilizing clinical textual data for real world studies: An OHDSI approach

Area of Science:

• Health Informatics
• Natural Language Processing (NLP)
• Real-World Evidence (RWE) Generation

Background:

• Clinical documentation in electronic health records (EHRs) contains rich patient data.
• Natural Language Processing (NLP) is essential for unlocking information within unstructured clinical text.
• The Observational Health Data Sciences and Informatics (OHDSI) consortium focuses on real-world data analysis.

Purpose of the Study:

• To describe a framework for representing and utilizing textual data in RWE generation.
• To detail the development of methods and tools for processing clinical notes using NLP.
• To showcase applications and use cases of the OHDSI NLP solution.

Main Methods:

• Development of a framework for integrating clinical text into the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM).
• Creation of an Extract, Transform, Load (ETL) workflow and associated tools for processing clinical notes.
• Implementation and evaluation of the OHDSI NLP solution across various institutions.

Main Results:

• Successful representation of clinical text information within the OMOP CDM.
• Demonstrated utility of the developed ETL workflow and tools for data extraction.
• Successful applications and use cases in large consortia and individual institutions.

Conclusions:

• The OHDSI NLP framework effectively enables the utilization of textual data in real-world studies.
• The developed tools and workflow facilitate the integration of clinical notes into standardized data models.
• Insights and lessons learned provide guidance for researchers implementing NLP in RWE generation.