Biomedical Engineering Reference
In-Depth Information
medical billing coder may code the event in the ICD-9 [9] vocabulary as
410.02 'Acute myocardial infarction of anterolateral wall subsequent
episode of care', whereas a cardiologist may classify the infarction as an
'acute Q wave infarction - anterolateral' which has a SNOMED CT [10]
code of 233827001. Obviously, the cardiologist needs a more detailed
and descriptive vocabulary than the billing coder. HL7 refers to these
different sets of vocabularies as code systems. Where possible, HL7 does
not duplicate these code sets but instead relies on the code sets already
developed by other organizations. In HL7 V3.x, when using a code it is
necessary to refer to the code system from which it originated. For
example, take the following XML snippet:
<code code='233827001' codeSystem='2.16.840.1.113883.6.96'
codeSystemName='SNOMED CT' displayName='acute Q wave infarction -
anterolateral'>
<code code='233827001' codeSystem='2.16.840.1.113883.6.96'
codeSystemName='SNOMED CT' displayName='acute Q wave infarction -
anterolateral'>
Most of this is self-explanatory, except for the code system. This is an
object identifi er (OID), which is a code used to refer to the SNOMED CT
vocabulary itself. Another dimension relevant to analytics is that systems
are not fl at but hierarchical in nature. An acute Q wave infarction is a
type of acute myocardial infarction, which is a type of acute heart disease,
and so forth. To be useful, an analytical query on heart disease will need
to contain all records that contain references to any type of heart disease.
The Unifi ed Medical Language System (UMLS) is a project from the
US National Library of Medicine (NLM) to bring together various
medical vocabularies to enable interoperability [11]. UMLS is organized
into a concept hierarchy, starting with a unique medical concept. The
unique concept is then related to all known lexical and string variants of
the concept. These are then related to individual codes from various
clinical vocabularies. Although UMLS is not strictly open source, there is
no charge associated with licensing UMLS and the actual data provided
can be queried without special tools. However, there may be separate
license fees associated with the use of specifi c terminology sets such as
SNOMED CT. The NLM does require that users give a brief report
annually of the usefulness of the Metathesaurus.
The UMLS data set is broken up into three areas:
1. Metathesaurus, which contains the vocabularies and relationships;
2. Semantic Network, which contains a set of semantic types (ex.
Anatomical Structure, Substance, Finding) and semantic relationships
(ex. disrupts, causes, manifestation of), which are used to provide
additional meaning to the Metathesaurus;
Search WWH ::
Custom Search