Unified Medical Language System

Unified Medical Language System
Name	Unified Medical Language System
Founded	1986
Founder	National Library of Medicine
Type	National Institutes of Health program
Headquarters	Bethesda, Maryland

Unified Medical Language System The Unified Medical Language System provides a compendium of biomedical terminologies and standards designed to enable interoperability among National Library of Medicine, World Health Organization, Centers for Disease Control and Prevention, Food and Drug Administration, and European Medicines Agency systems. It supports mapping among vocabularies used by institutions such as Mayo Clinic, Johns Hopkins Hospital, Massachusetts General Hospital, Cleveland Clinic, and Kaiser Permanente while facilitating research at organizations like National Institutes of Health, Wellcome Trust, Howard Hughes Medical Institute, and Bill & Melinda Gates Foundation. The project informs standards work at Health Level Seven International, International Organization for Standardization, SNOMED International, and LOINC.

Overview

The system aggregates lexical and semantic resources from sources including SNOMED CT, LOINC, ICD-10, MeSH, RxNorm, Disease Ontology, Gene Ontology, and UMLS Metathesaurus to create concept mappings used by stakeholders such as World Health Organization, Centers for Medicare & Medicaid Services, Veterans Health Administration, Department of Veterans Affairs, and European Commission. It links clinical terminologies used in Epic Systems Corporation, Cerner Corporation, Allscripts, and MEDITECH with research vocabularies favored by PubMed, ClinicalTrials.gov, Cochrane Collaboration, and EBSCO Information Services. The resource supports data exchange across projects like All of Us Research Program, Human Genome Project, ENCODE Project, and UK Biobank.

History and development

Development began under the auspices of the National Library of Medicine in the mid-1980s with collaborations involving Harvard University, University of California, San Francisco, Columbia University, Stanford University School of Medicine, and Yale School of Medicine. Early efforts tied to initiatives at National Institutes of Health and funding from National Science Foundation aligned with policy work at Office of the National Coordinator for Health Information Technology and international dialogues at World Health Assembly. Major milestones included incorporation of ICD-9 to ICD-10 mappings, integration of MeSH descriptors, and alignment with SNOMED CT after negotiations with International Health Terminology Standards Development Organisation partners. The program evolved alongside influential projects such as Human Genome Project and regulatory developments like the Health Insurance Portability and Accountability Act.

Components

Core deliverables include the UMLS Metathesaurus, the UMLS Semantic Network, and the UMLS SPECIALIST Lexicon. The Metathesaurus unifies source vocabularies from SNOMED CT, LOINC, ICD-10, MeSH, RxNorm, Gene Ontology, and proprietary terminologies used by Mayo Clinic and Cleveland Clinic. The Semantic Network provides high-level categories referenced by scholars at Harvard Medical School, Johns Hopkins Bloomberg School of Public Health, Columbia University Mailman School, and University of Michigan Medical School. The SPECIALIST Lexicon supports natural language processing used in systems developed at Massachusetts Institute of Technology, Carnegie Mellon University, SRI International, and IBM Research.

Applications and uses

Researchers at National Institutes of Health, Wellcome Trust, European Research Council, and Bill & Melinda Gates Foundation leverage the resource for data harmonization in projects like All of Us Research Program and UK Biobank. Hospitals such as Mayo Clinic and vendors including Epic Systems Corporation use mappings for electronic health record interoperability required by Centers for Medicare & Medicaid Services quality reporting and by Office of the National Coordinator for Health Information Technology certification. Public health agencies like Centers for Disease Control and Prevention and World Health Organization employ the resource for surveillance, while pharmacists reference RxNorm alignments in systems tied to American Society of Health-System Pharmacists workflows. Academic groups at Stanford University, University of California, San Francisco, Johns Hopkins University, and University of Pennsylvania use it for natural language processing, machine learning, and secondary use of clinical data integrated with PubMed and ClinicalTrials.gov.

Evaluation and challenges

Independent assessments by teams at National Library of Medicine, MITRE Corporation, RAND Corporation, Agency for Healthcare Research and Quality, and university laboratories have highlighted strengths and limitations in coverage, concept granularity, and mapping precision between ICD-10, SNOMED CT, and LOINC. Challenges include licensing tensions involving SNOMED International and national release policies of ICD-10 adopted by World Health Organization member states, technical integration issues reported by vendors like Epic Systems Corporation and Cerner Corporation, and the need for continuous updates to reflect discoveries from Human Genome Project derivatives and initiatives such as ENCODE Project. Evaluators from American Medical Informatics Association and Institute of Medicine have recommended governance and tooling improvements.

Governance and access

Governance is administered by the National Library of Medicine within the National Institutes of Health with community engagement from American Medical Informatics Association, Health Level Seven International, SNOMED International, and international partners including World Health Organization and European Commission. Access policies balance open research use by Harvard University, Stanford University, and University of California groups with licensing obligations required by contributors such as SNOMED International and proprietary vendors like Elsevier and Thomson Reuters. Training and outreach are provided through collaborations with Centers for Disease Control and Prevention, Office of the National Coordinator for Health Information Technology, and academic training programs at Johns Hopkins Bloomberg School of Public Health and Harvard Medical School.

Category:Biomedical informatics