Controlled Vocabularies for Capturing Clinical Encounters

1
Controlled Vocabularies forCapturing Clinical
Encounters

• James J. Cimino, M.D.
• Department of Medical Informatics
• Columbia University

2
Why Use a Controlled Vocabulary?

• Efficient storage
• Anticipatory data capture
• Record abstraction for billing, reporting,
  research...
• Automated decision support

3
What are the Options?

• ICD-9-CM
• ICD-10
• ICPC
• NANDA
• DSM-IV
• MeSH
• UMLS
• Gabrieli Nomenclature
• Read Clinical Codes
• SNOMED International

4
Why Arent 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

5
Possible SolutionCollaborative Vocabulary
Construction

• Users contribute to design and content
• Content drawn from legacy systems
• Contributions made via the Internet
• Updates in near-real time

6
The InterMed Collaboratory

• HPCC Project from NLM
• Multiple investigator groups
• Columbia University
• Harvard University
• Stanford University
• University of Utah
• Net-based collaborative research

7
The Collaborative Approach

• Agree on a domain of interest
• Model the domain
• Hierarchies
• Semantic attributes
• Contribute content

8
Browsers and Editors

• Ontolingua
• K-Rep
• AccessMED
• Web MED Browser

9
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

10
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