Transcript Setting for Vocabulary Test - People | Columbia University

Controlled Vocabularies for
Capturing Clinical Encounters
James J. Cimino, M.D.
Department of Medical Informatics
Columbia University
Why Use a Controlled Vocabulary?

Efficient storage

Anticipatory data capture

Record abstraction for billing, reporting,
research...

Automated decision support
What are the Options?
ICD-9-CM
 ICD-10
 ICPC
 NANDA
 DSM-IV
 MeSH
 UMLS
 Gabrieli Nomenclature
 Read Clinical Codes
 SNOMED International

Why Aren’t We Using Them?

Abstracting systems provide insufficient detail

Comprehensive vocabularies may be too
expressive

Organizational mismatch between developers
and users

Adaptation of legacy systems to standard
vocabularies
Possible Solution:
Collaborative Vocabulary Construction

Users contribute to design and content

Content drawn from legacy systems

Contributions made via the Internet

Updates in near-real time
The InterMed Collaboratory

HPCC Project from NLM

Multiple investigator groups
» Columbia University
» Harvard University
» Stanford University
» University of Utah

Net-based collaborative research
The Collaborative Approach

Agree on a domain of interest

Model the domain
» Hierarchies
» Semantic attributes

Contribute content
Browsers and Editors

Ontolingua

K-Rep

AccessMED

Web MED Browser
Contributions to the InterMed MED

Find a class for your term in the Web browser

Click on the “Add” button

Enter the name

Add other parents and children

Refine semantic attributes

Click on “UMLS” button
Outcomes

Exploration of Internet-based vocabulary
collaboration

Expansion of domain coverage

Vocabulary that fits user needs

Contributions to the UMLS
http://www.cpmc.columbia.edu/intermed_proj.html