Unified Medical Language System UMLS

1
Unified Medical Language System (UMLS)

• Yildiray Kabak, METU-SRDC

2
Outline

• Introduction
• Knowledge Sources
• Metathesaurus
• Semantic Network
• SPECIALIST Lexicon
• UMLS Knowledge Source Server
• UMLS Applications
• EHR perspective

3
Introduction (Purpose)

• lack of a standard language in medicine
• main purpose is to build required vocabulary
• to facilitate the development of computer systems
  that behave as if they "understand" the meaning
  of the language of biomedicine and health

4
Introduction (Purpose)

• for system developers in building or enhancing
• electronic information systems that create,
  process, retrieve, integrate, and/or aggregate
  biomedical and health data and information

5
Introduction

• The UMLS Knowledge Sources are multi-purpose
• not optimized for particular applications
• The associated UMLS software tools assist
  developers in customizing or using the UMLS
  Knowledge Sources for particular purposes

6
Introduction

• consists of three Knowledge Sources
• Metathesaurus
• concepts that include the various names
  representing the same meaning from different
  source vocabularies
• Semantic Network
• 135 semantic types, 54 semantic relations
• SPECIALIST Lexicon
• dictionary of biomedical terms and common words,
  lexical tools and records used in natural
  language processing

7
Outline

• Introduction
• Knowledge Sources
• Metathesaurus
• Introduction
• Concepts, Terms, Strings, Atoms
• Relations
• Metathesaurus Files
• Semantic Network
• SPECIALIST Lexicon
• UMLS Knowledge Source Server
• UMLS Applications
• EHR perspective

8
Introduction

• very large concept-oriented database
• holds concepts, their various names and
  relationship among them
• Links alternative names and views of the same
  concept together and identify useful relations
  between different concepts
• 2004AA 1,020,866 concepts and 2.8 million terms

9
2002AC

• 870,000 concepts (Eye, Oculus 1)
• 1,756,000 terms (Eye, Eyes, eye 1)
• 2,083,103 strings/concept names (Eye, Eyes,
  eye 3)
• 11,479,000 relationships between concepts
• 7 million of relationships between concepts and
  English words
• 113 source vocabularies
• 15 different languages

10
Metathesaurus

• It is built from the electronic versions of many
  different thesauri, classifications, code sets,
  and lists of controlled terms used in
• patient care,
• health services billing,
• public health statistics,
• indexing and cataloging biomedical literature,
• and/or basic, clinical, and health services
  research.

11
Types of Metathesaurus sources

• Thesauri, e.g., MeSH
• Statistical Classifications, e.g., ICD-9, ICD-10,
  ICPC
• Billing Codes, e.g., CPT, CPT Spanish version,
  HCPCS
• Clinical Coding Systems, e.g., SNOMED, Read
• Nursing Vocabularies, e.g., NIC, NOC, OMAHA
• Alternative/Complementary Medicine ALTLINK
• Drug Sources Multum, Micromedex, VANDF
• Drug Regulatory, e.g., MedDRA
• Lists of controlled terms, e.g., COSTAR, HL7
  values

12
Introduction

• All concepts are assigned to at least one
  semantic type
• consistent categorization of all concepts at the
  relatively general level
• Metathesaurus must be customized to be used
  effectively

13
Metathesaurus

• do not provide a structure
• provide unified meaning for a concept from
  different vocabularies
• provide the data itself

14
Concerning message ontology

• Assume that all the concepts are defined in OWL
• CUI for ALLERGY
• /

15
Concerning message ontology

• An instance for example
• Anaphylaxis
• C0002792
• ....
• Metathesaurus source vocabularies include
  terminologies designed for use in patient-record
  systems

16
Metathesaurus structure

• Concepts (CUI)
• Terms (LUI)
• Strings (SUI)
• Atoms (AUI)
• Relations

17
Concept

• A concept is meaning
• A meaning can have many different names
• link all the names from all of the source
  vocabularies that mean the same thing
• each concept (meaning) has a concept unique
  identifier (CUI)

18
(No Transcript)
19
Concept Names and String identifiers

• Each string in the concept names has a unique
  identifier (SUI)
• Any variation in character set, upper-lower case,
  punctuation is a separate string with a separate
  SUI
• The same string in different languages have
  different SUI

20
Atoms

• Each and every occurrence of a string in each
  source vocabulary is an atom
• every atom has an atom identifier (AUI)
• In other words, Atoms are the entries in the
  source vocabularies

21
Terms

• All the variants of a string is grouped into a
  term
• a term is the group of all strings that are
  lexical variants of each other
• Each term has a lexical identifier (LUI)

22
Example
23
(No Transcript)
24
(No Transcript)
25
Metathesaurus
CONCEPTs
TERMs
STRING
CUIs
SUIs
LUIs
STRING
STRING
Is organized by concept or meaning its purpose
is to link alternative names and views of the
same concept together and to identify useful
relationships between different concepts.
STRING
STRING
STRING
26
Metathesaurus relationships

• Apart from the synonymy
• Intra-source relationships between concepts from
  the same vocabulary
• Inter-source relationships between concepts in
  different vocabularies

27
Intra-source

• Hierarchical
• immediate-parent
• immediate-child
• immediate-sibling
• Broader (RB) Has a meaning which includes that
  of the concept.
• Narrower (RN) Has a meaning which is included in
  that of the concept
• Statistical
• if two concepts co-occurred as key topics within
  the same articles

28
Inter-source

• (Note some of the below may be statistical)
• Other related (RO) Has a relationship other than
  synonymous, narrower, or broader
• Like (RL) The two concepts are similar or
  "alike".
• RQ related and possibly synonymous
• SY source asserted synonymy

29
Relationships

• In 2002AC version
• 5M Inter-source Hierarchical relations
• 6.5M statistical relations

30
Metathesaurus Files

• when installed, a set of files is created. That
  is Metathesaurus is just a set of files
• you are responsible for reading from that files
• you can also customize the files according to
  your needs
• For example, MRCONSO.RRF
• CUI,LAT,TS,LUI,STT,SUI,ISPREF,AUI,SAUI,SCUI,SDUI,S
  AB,TTY,CODE,STR,SRL,SUPPRESS, CVT
• There is exactly one row for each atom

31
(No Transcript)
32
Outline

• Introduction
• Knowledge Sources
• Metathesaurus
• Semantic Network
• SPECIALIST Lexicon
• UMLS Knowledge Source Server
• UMLS Applications
• EHR perspective

33
Semantic Network

• provides consistent categorization of all
  concepts and the relations between the types
• Note that these relations are between types not
  concepts
• They are different from the relations in the
  Metathesaurus

34
Semantic Network

• Broad subject categories
• Represent the biomedical domain
• 2 main categories
• Entity
• Event
• Semantic type is assigned to Metathesaurus
  concepts at the most specific level

35
UMLS Semantic Net
Entity
Event
36
Relations

• Primary link isa ? establishes the hierarchy
• Five group other than isa
• physically related to
• spatially related to
• functionally related to
• temporally related to
• conceptually related to
• inheritance supported

37
Semantic Net 54 Links
38
Example
39
In addition.. Semantic Groups

• 15 Semantic Groups
• Smaller set of categories (135 15)
• Broader, coarser groupings
• Partition 99.5 of UMLS Metathesaurus concepts
• Used for
• Word sense disambiguation
• Profiling, analyzing vocabularies
• Display in Semantic Navigator

40
Semantic Groups

• Activities and Behavior
• Anatomy
• Chemicals Drugs
• Concepts Ideas
• Devices
• Disorders
• Genes Molecular Sequences
• Geographic Areas
• Living Beings
• Objects
• Occupations
• Organizations
• Phenomena
• Physiology
• Procedures

41
(No Transcript)
42
Outline

• Introduction
• Knowledge Sources
• Metathesaurus
• Semantic Network
• SPECIALIST Lexicon
• UMLS Knowledge Source Server
• UMLS Applications
• EHR perspective

43
SPECIALIST Lexicon and Lexical Tools

• Lexicon a database of syntactic, morphological
  and orthographic information for commonly
  occurring English language words and biomedical
  vocabulary for natural language processing
• Tools assist in detecting and abstracting away
  from the inflectional, case and word order
  variations

44
Lexicon

• Many of the words and multi-word terms that
  appear in concept names also appear in the
  SPECIALIST lexicon
• The lexical tools are used to generate the word,
  normalized word and normalized string indexes
  (connect each word to all related string, term,
  concept) to the Metathesaurus

45
Lexicon

• English words
• 20,000 (initial) test set from MEDLINE abstracts
• 10,000 American Heritage Dictionary frequency
  list
• 2,000 Longman's Dictionary of Contemporary
  English
• verbs and adjectives identified by heuristics
• Biomedical terms in the Metathesaurus

46
Example Lexical Variant Generator

• 3 primary programs
• Normalizer(norm)
• Word index generator (wordInd)
• Lexical variant generator (LVG)

47
Normalization

• Abstracts away
• Case
• Punctuation
• word order
• possessive forms
• inflectional variation
• Generates strings in Metathesaurus normalized
  string index (MRXNS)

48
Example

• Hodgkin Disease
• HODGKINS DISEASE
• Hodgkin's Disease
• Disease, Hodgkin's
• HODGKIN'S DISEASE
• Hodgkin's disease
• HodgkinsDisease
• Hodgkin's disease NOS
• Hodgkin's disease, NOS
• Disease, Hodgkins
• Diseases, Hodgkins
• HodgkinsDiseases
• Hodgkinsdisease
• hodgkin'sdisease
• DiseaseHodgkins
• Disease, Hodgkin

Disease hodgkin
49
Lexical Tools

• For example
• MMTX designed to map arbitrary terms to concept
  names or to discover concepts within free text

50
Outline

• Introduction
• Knowledge Sources
• Metathesaurus
• Semantic Network
• SPECIALIST Lexicon
• UMLS Knowledge Source Server
• UMLS Applications
• EHR perspective

51
UMLS Knowledge Source Server

• Internet access to the Knowledge Sources
• Browser
• API

52
Outline

• Introduction
• Knowledge Sources
• Metathesaurus
• Semantic Network
• SPECIALIST Lexicon
• UMLS Knowledge Source Server
• UMLS Applications
• EHR perspective

53
Application Example
They developed their own database from the files
54
PubMed
Bronzed disease
UMLS tools
Addisons disease(MesH term)
Search MeSH indexed database
55
Outline

• Introduction
• Knowledge Sources
• Metathesaurus
• Semantic Network
• SPECIALIST Lexicon
• UMLS Knowledge Source Server
• UMLS Applications
• EHR perspective
• GEHR
• openEHR

56
GEHR
You define the structure on your own
57
GEHR

• Wherever PLAIN_TEXT or TERM_TEXT appears in the
  GOM, the expansion of a termset code may appear.
• GEHR uses the CUI (Concept Unique Identifier) of
  the UMLS to specify concept codes for any
  attribute in the model

58
GEHR
59
openEHR
60
openEHR (Data Package)
61
Future Work

• Investigate the API, for example
• Find the concepts of a semantic type
• Find the concepts of a vocabulary
• Find the concepts of the semantic type, which is
  functionally related to a semantic type of a
  given concept