SCAMC features new unified medical language system and standards.
Medical vocabularies, and especially the National Library of Medicine's (NLM) Unified Medical Language System (UMLS), received a great deal of attention during the 14th Annual Symposium on Computers in Medical Care (SCAMC), a conference of the American Medical Informatics Association (AMA). Held in Washington, D.C., November 4-7, 1990, SCAMC drew speakers and an audience from many countries as well as from many disciplines in the health sciences.
Thirty-seven professional organizations, 17 educational institutions, 11 U.S. government agencies - sometimes with several different divisions - six research organizations, five user groups, and four international standards and business organizations were sponsoring and cooperating organizations of this conference.
SCAMC is designed to inform a variety of health care professionals about current and potential applications of computer technology to health care, and to identify areas of research and development that need to be addressed. Among the professionals at this SCAMC were:
* health care administrators
* biomedical scientists
* information scientists
With such a diverse focus and audience it is indeed notable that more than 15 papers - including the opening address of the opening plenary session - were devoted to standards for published literature and their implications for medical informatics.
Linking User and Computer Through
In the opening address, NLM's Betsy L. Humphreys stated that the effective practice of medicine depends on the ability of health professionals to locate relevant information quickly and to interpret it correctly. UMLS is being developed to help make the conceptual link between the user's question and relevant machine-readable information. The paper published in the proceedings was co-authored with Donald A. B. Lindberg, M.D., NLM's Director.
The UMLS project was started in 1986 as a long-term R&D effort designed to ease the retrieval and integration of information from multiple machine-readable biomedical information sources. These sources include:
* descriptions of the biomedical literature
* clinical records
* knowledge-based systems
* directories of people and organizations
The variety of vocabularies and classifications used in these different sources poses a significant barrier to the use of machine-readable information by health professionals and biomedical researchers. In addition, the diversity makes it difficult to develop effective search interfaces to assist these users. Humphreys described a variety of "de facto, derigueur, and even useful" standards for the published literature which, as information professionals know cover descriptive elements and subject content. Players in the field of standards include ALA, NISO, ANSI, ASTM, and NLM; standards accepted or in process are CALS, SGML, ISSN, ISBN, and data exchange protocols.
The UMLS project currently focuses on the development, testing, and evaluation of the first versions of three new knowledge sources:
Information Sources Map
Metathesaurus: The Metathesaurus is a central vocabulary tool of the UMLS. It is a database of information about biomedical terms that appears in several different controlled vocabularies and classifications. Meta-1 was issued for experimental use in Fall 1990. A large database, it encompasses 66,000 concepts and about 100,000 terms, about 200-plus megabytes. Included in the base vocabulary are all terms in the Medical Subject Headings (MeSH), those of several professional associations, and selected terms from the Library of Congress' Subject Headings (LCSH).
Meta-1 preserves the meanings, heirarchical contexts, and relationships about terms present in its source vocabularies; it adds certain basic information about each term and establishes new relationships between terms from different source vocabularies.
Semantic Network: All concepts in Meta-1 were assigned to one or more of 131 basic semantic types or categories. Examples are "Disease or Syndrome," "Virus," "Medical Device," "Hazardous Substance." The Semantic Network defines each of these basic types or categories and it represents a variety of useful potential relationships, among them "Virus" causes "Disease or Syndrome."
Information Sources Map: The Information Sources Map will contain information about databases as a whole, for example, AIDSLINE, DXPLAIN, their content and coverage, location, access conditions, and relationship to other databases. The Map is in the prototype development phase. The goal is that it will assist computer programs to:
* determine which machine-readable sources are likely to be relevant to a particular user inquiry
* supply human-readable information to users about the scope, probable utility, and access conditions of information sources
* make an automatic connection to sources likely to be relevant to a particular user inquiry
* conduct automatically a successful retrieval session on relevant information source(s)
NLM is offering experimental copies of the initial versions in exchange for feedback on improvements. See below for details.
Clinical data is another area hungry for standards. Clement J. McDonald, M.D., Indiana University School of Medicine and Regenstrief Institute for Health Care, is a pioneer in standards for the electronic transfer of clinical data. Collecting the data is the most expensive part of the information process, he explained, and the data is bound to the form in which it is conceived. Standards are needed here because it is difficult to communicate data. Or, as McDonald commented at one point, "Data, data everywhere and not a bit of use." With humor and creativity he emphasized the importance of standards for content, for the data elements in the record, and for exchange of records.
Daniel R. Masys, M.D. described developing standards for electronic images at NLM. He compared the volume and complexity of images with text - text being about 2,000 characters per page - and an image using up to 10 million characters of density, or 1,000 to 10,000 kilobytes more than a page of text.
"Currently few widely accepted standards exist for electronic images," said Dr. Masys, Director of the Lister Hill National Center for Biomedical Communications, "yet such standards will be critical to the feasibility and usefulness of digital image collections in the life sciences." He predicted that "the library of the future will provide electronic access not only to words and numbers, but to pictures, sounds, and other non-textual information."
Today, two-dimensional images, such as those on printed pages, have been supplemented by 3D images which can be rendered, rotated, and "dissected" from any point of view. He described the NLM project to develop a complete volumetric representation of an adult human male and female. Work on this "Visible Human Project" will address the issue of standards for computer representation of biological structure.
While standards have been developed to promote the interchange of pixel-based clinical images such as those produced in CAT scans, and a number of general purpose standards have been developed by the computer graphics community, Dr. Masys indicated that no dominant standard exists for the exchange of volume-based imaging data sets. NLM's efforts will be focused on identifying a small number of sufficiently powerful and flexible formats.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Focus on Medical Information; Symposium on Computers in Medical Care|
|Date:||Mar 1, 1991|
|Previous Article:||National Online Meeting & IOLS '91.|
|Next Article:||ALA midwinter: weather, war, war worries.|