Distributed Cognition and Knowledge-based Controlled Medical Terminologies

From Clinfowiki
Revision as of 16:21, 3 March 2015 by Hgarcia2 (Talk | contribs)

Jump to: navigation, search

Introduction

In his 1998 article entitled Distributed Cognition and Knowledge –Based Controlled Medical Terminologies, Cimino opens with the observation that the standard of medical care now demands a scope of knowledge, expertise, and responsiveness that exceeds the retention and processing capacity of the individual practitioner. [1] He then examines the evolution of several terminologies into knowledge bases capable of representing medical concepts and describes these knowledge bases as supporting two forms of distributed cognition. The first of these is the aggregation of people and applications for complex medical decisions. The second is the distributed work to coordinate and maintain the terminologies themselves.

Key Concepts

Distributed Cognition - Cimino introduces the concept of distributed cognition through the example of consultative support and specialization of medical practice. He then offers the definition of distributed cognition as that which happens "when intellectual processes are shared among multiple participants, especially in order to solve a particular task in a particular context." [1] Controlled medical terminologies (CMT) - Whether between humans or computer systems, efficient and accurate distribution of this work depends on concise, unambiguous, accurate, and sufficiently descriptive messages. Turning his focus on computer systems in distributed cognition, Cimino acknowledges the requirement of CMTs.

Distributed Development of CMTs

  • Traditionally developed in isolation, constricted to particular application and domain
  • Growth increased utility within a domain, but left the CMT unusable across domains
  • US National Library of Medicine's (NLM)MeSH was created to provide application agnostic, trans-domain system to join many domain specific terminologies (e.g. SNOMED-CT, ICD-9-CM, etc.) [2]
  • To reduce redundancy and ambiguity in CMTs, developers are working to create named computer manipulable inter-relationships among concepts.
  • This creates a set of conceptual frames with names slots that refer to other frames.
  • This referencing is the conduit for the distributed cognition central to this discussion.

Distributed Knowledge for Maintenance

  • Addition of new terms includes addressing synonyms, redundancies, how, and where the new term should be added.
  • Manual process tedious, slow, error prone
  • Positive indications of feasibility of automated classification and terminology maintenance.
  • > Columbia University's Medical Entities Dictionary (MED)
  • > InterMEd's Ontolingua
  • > Intermountain Healthcare's-Intermountain Healthcare VOSER

Knowledge-based CMTs to Support Distributed Cognition in Patient Care

  • Potency and maintainability of Clinical decision support systems (CDSS) depends on CMTs
  • Effective triggering depends on ability to accurately reference presenting case to reference cases and reasoning processes.

Discussion

Historically, the limited quality of terminologies limited their ability to "participate" in distributed cognition. Cimino identifies the evolution of CMTs to knowledge-bases as contributing to improved quality. A degree of inferencing is now available with draws these terminologies into the realm of automated processing, and the creation of valid interrelationships. Cimino acknowledges that we are far from intelligent CMTs with automated integration, but that the field is moving in that direction. He identifies a major challenge to the utilization of computers for this maintenance is the need explicit definitions for terms that stable across contexts. While difficult, this article reflects Cimino's gratification with the trend.

Commentary

In order to reach the potential benefits from broad deployment of electronic health information systems, the effective adoption of CMTs must occur. It is difficult to read an article about interoperability that doesn't at least allude to the challenges related to standardized terminologies across disparate systems and domains. The concept and measured optimism in this article provide both reason for hope and potential guidance for career selection. This was an assessment of the state of the art for a critical area of health informatics in 1998. Searching the literature for related developments since then would certainly be of interest to anyone interested in this work.

References

  1. 1.0 1.1 Cimino, J. J. (1998). Distributed cognition and knowledge-based controlled medical terminologies. Artificial Intelligence in Medicine, 12(2), 153-168. http://www.sciencedirect.com.ezproxyhost.library.tmc.edu/science/article/pii/S0933365797000481
  2. Humphreys, B. L., Lindberg, D. A., Schoolman, H. M., & Barnett, G. O. (1998). The Unified Medical Language System An Informatics Research Collaboration. Journal of the American Medical Informatics Association, 5(1), 1-11. http://jamia.oxfordjournals.org/content/5/1/1.short