Related Experiment Video
Updated: Jun 7, 2026
07:50
A Metadata Extraction Approach for Clinical Case Reports to Enable Advanced Understanding of Biomedical Concepts
Published on: September 20, 2018
Determining correspondences between high-frequency MedDRA concepts and SNOMED: a case study
Prakash M Nadkarni1, Jonathan D Darer
1Geisinger Health Systems, Danville, PA, USA. Prakash.Nadkarni@yale.edu
BMC Medical Informatics and Decision Making
|October 30, 2010
Summary
Improving MedDRA to SNOMED CT mappings enhances pharmacovigilance. Most MedDRA terms can be mapped compositionally to SNOMED CT, with errors often due to missed correspondences.
Area of Science:
- Medical Informatics
- Pharmacovigilance
- Clinical Terminology
Background:
- Systematic Nomenclature of Medicine Clinical Terms (SNOMED CT) is proposed for clinical documentation.
- Current pharmacovigilance relies on Medical Dictionary of Regulatory Activities (MedDRA) for adverse event classification.
- High-quality MedDRA-to-SNOMED CT mappings are crucial for advancing pharmacovigilance.
Purpose of the Study:
- Identify challenges in mapping MedDRA to SNOMED CT.
- Analyze patterns in unmapped, high-frequency MedDRA concepts.
- Detect integration errors in MedDRA-to-UMLS mappings.
Main Methods:
- Analyzed one year of US FDA Adverse Event Reporting System data.
- Identified MedDRA preferred terms covering 95% of adverse events and indications.
- Attempted mapping of unmapped terms to SNOMED CT with software assistance.
Main Results:
- Most MedDRA terms (645 Adverse-Event, 141 Therapeutic-Indications) could be composed using SNOMED CT.
- Few unmapped terms required more than three SNOMED CT concepts.
- 30% of terms showed missed one-to-one correspondences, causing duplication in UMLS.
Conclusions:
- Composite mapping patterns and error identification can refine MedDRA-to-SNOMED CT mappings.
- Improved mappings support the development of an adverse-event ontology.
- Focusing on high-frequency terms and common errors optimizes mapping efforts.
