Survey of Medical Informatics
Download
Report
Transcript Survey of Medical Informatics
Survey of Medical
Informatics
CS 493 – Fall 2004
September 27, 2004
Health Care Data Standards
Chapter 4: Patient Safety - Achieving a
New Standard of Care.
IOM Report
Terminology Standards
Why terminology standards?
Standardized terminologies facilitate:
Electronic data collection
Retrieval of relevant data or knowledge
Allows data to be reused for multiple purposes
such as syndromatic surveillance, clinical decision
support and quality and cost monitoring
Criteria for terminologies
Technical Criteria used by NCVHS for
evaluating and selecting terminologies
Page 145 – Table 4-2
CHI focus areas
Page 146 – Table 4-3
Terminology Standards
November 13, 2003 – NCVHS recommends:
the adoption of SNOMED CT as the terminology of choice
for representing clinical terms
LOINC is endorsed as the standard to use for Laboratory
concepts coding
RxNORM
National Drug File Clinical Drug Reference Terminology
(NDF RT)
Also have recommended the adoption of ICD-10
January 29, 2004 – CHI endorses SNOMED CT for
anatomy, nursing, diagnosis and problems, and nonlab interventions and procedures
Overview of Core and Supplemental
Terminologies
Box 4-1
Pages 150-151
Figure 4-4
Page 157
Unified Medical Language
System
UMLS
Presentation Source Material: Oliver Bodenreider: “The Unified Medical
Language System (UMLS) integrating biomedical terminologies,” Nucleic Acids
Research, 2004, Vol. 32, Database issue D267-D270
UMLS
Repository of Biomedical Vocabularies
It integrates
2 million names
900,000 concepts
60 families of biomedical vocabularies
12 million relations among concepts
UMLS Metathesaurus
Sample vocabularies integrated here:
NCBI taxonomy
Gene Ontology
The Medical Subject Headings (MeSH)
OMIM
Digital Anatomist Symbolic Knowledge Base
ICD-9
CPT
UMLS Metathesaurus
Repository of biomedical concepts drawn
from all the different terminologies integrated
Online Mendelian
Inheritance in Man
National Center
For Biotechnology
Information
Gene Ontology
University of Washington
Digital
Anatomist
Figure Source Material: Oliver Bodenreider: “The Unified Medical
Language System (UMLS) integrating biomedical terminologies,” Nucleic Acids
Research, 2004, Vol. 32, Database issue D267-D270
Terminology Integration Principles
Knowledge organized by concept/meaning
Synonymous terms associated with concepts
Concepts can be linked to other concepts
Hierarchical, part-of, inheritance, associative
links and statistical relations between
concepts are represented in the
Metathesauraus.
Each Metathesauraus concept is categorized
currently into one of 135 high-level categories
Structure of Metathesaurus
Allows
Collecting terms associated with a particular
concept
Exploring relationship between concepts
Browsing the concepts associated with a
particular category
Section of the UMLS Semantic Network
Source: http://www.nlm.nih.gov/research/umls/META3_Figure_1.html
Section of the UMLS Semantic Network
UMLS Tools
MetamorphoSys – helps users customize
Metathesaurus
lvg : a program to generate lexical variants
MetaMap : extracts Metathesaurus concepts
from text
LOINC
Source: “LOINC, a Universal Standard for
Identifying Laboratory Observations: A 5-Year
Update” Clement J. McDonald, Stanley M. Huff,
et. al. Clinical Chemistry 49:4, 624-633 (2003)
http://www.clinchem.org/cgi/reprint/49/4/624
Logical Observation Identifiers Names
and Codes
Started by a group of researchers who met in
1994 in Regenstrief Institute
1995 first release of 6000 laboratory test
results
17 releases since then
Regenstrief LOINC Mapping Assistant
(RELMA)
Examples
LOINC codes for partial pressure of arterial
blood oxygen (Po2) and percentage
lymphocytes
EKG measurements
Vital signs
LOINC only codes for observation types not
the values that are associated with a
particular test
Knowledge Representation
Clinical Guideline Representation Model
Clinical Guidelines
National Guideline Clearinghouse contains
1,000 publicly accessible guidelines
Box 4-2 pg. 159