This document describes how to map Unified Medical Language System (UMLS) codes to International Classification of Diseases 10th Revision (ICD10) codes. It involves building an ontology of the ICD10 code hierarchy and using SPARQL queries over the ontology and UMLS concepts to find the most specific ICD10 code for a given set of UMLS codes describing a patient's diagnosis. An application was created to demonstrate mapping UMLS codes from an electronic medical record to the appropriate ICD10 code for billing and other purposes.
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Mapping UMLS to ICD10 codes
1. UMLS TO ICD10
MAPPING.(OR: GIVEN A SET OF UMLS CODES HOW TO MAP IT TO A ICD10 CODE?)
PRESENTED BY
PRAMOD C KONERU , WEB 3.0 - DEMO
Friday, April 26, 13
3. WHAT!!ARE THESE CODES
UMLS
(Unified
Medical
language
system),
is
a
set
of
files
and
so9ware
that
brings
together
many
health
and
biomedical
vocabularies.
ICD10
(InternaEonal
ClassificaEon
of
Diseases),
are
designed
to
classify
healthcare,
providing
a
system
of
diagnosEc
codes
for
classifying
diseases,
including
nuanced
classificaEon
of
wide
variety
of
signs,
symptoms,
abnormal
findings,
complaints
etc
etc.,
Friday, April 26, 13
4. WHY?
DO WE NEED ICD10
1.
In
the
healthcare
industry
these
codes
are
vital
for
billing
and
health
insurance
reimbursement,
automated
decision
support
and
government
sta:s:cal
reports
on
morbidity
and
mortality.
2.
It
is
mandatory
for
healthcare
services
to
encode
pa:ent
diagnosis
informa:on
in
these
codes
3.
Deadline
Friday, April 26, 13
5. HOW??
SYSTEM TO ANNOTATE EMR
AGAINST
UMLS CODES.
ONTOLOGY OF UMLS
HIERARCHY.
CAN WE BUILD OUR APPLICATION ON TOP OF THESE?
Friday, April 26, 13
6. ARCHITECTURE
EMR DOCUMENT CONVERTED INTO
TRIPLES
WITH CHIEF COMPLAINT
EACH
CLASS
IS
EMBEDDED
WITH
ITS
SPARQL
QUERY
PREFIX icd: <http://www.knoesis.org/icd10mapping/>
PREFIX ezdi: <http://www.ezdi.us/cardio.owl#>
PREFIX rdfs: <http://www.w3.org/2000/01/rdf-schema#>
ASK FROM <http://knoesis.wright.edu/umls2icd10/sample>
WHERE
{
{{SELECT ?x FROM <http://knoesis.wright.edu/healthcare> WHERE { ?x rdfs:subClassOf ezdi:C0259752. }}
{icd:patient icd:suffering_with ?x.}}
UNION
{icd:patient icd:suffering_with ezdi:C0259752.}
}
ALGO
ASK QUERY
TRUE
GET SUBCLASSES
Friday, April 26, 13
13. THINGS THAT I HAVE
LEARNED
ONTOLOGY BUILDING,
in-‐fact
I
have
built
a
small
ontology
which
holds
the
the
disease
hierarchy.✓
ADVANCED SPARQL,
which
includes
the
use
of
SPARQL
inferencing
and
sub
querying.
✓
Friday, April 26, 13