Phase II-SOF Knowledge Coupler-Based Phase I XML Schema

Abstract

The 2002 digital version of the Special Operations Forces Medical Handbook (SOFMH) includes a comprehensive, searchable database of injuries and illnesses. While it is a complete digital reference source, its utility would be greatly enhanced if a healthcare provider could enter a patient's signs and symptoms into the SOFMH search template and access a list of diagnostic choices in an XML-tagged database. An analysis of the search function indicates that the native search capability of the SOFMH does not inherently contain the requirements to sustain a diagnostic tool. Current search technologies can locate text or indexes, ranked by the frequency a term appears in a document, but not the term's relevance to a set of symptoms. Current search technologies operate by diagnosis category, key words, indices, and content text. The program ranks matches by frequency, index, and content. A medical knowledge coupler requires more sophisticated associations to link a diagnosis to the symptom. XML tagging was selected as the method to identify and assign significance to portions of text information. Initial tagging of the SOFMH did not enable the level of detail required for the diagnostic process. The next step towards a reliable diagnostic tool is establishing the relationships between the symptoms and the diagnosis. A diagnostic tool cannot automatically make these associations; it must be provided the information. Keywords will be used prescribe a certain intuitiveness in the application. A keyword can be significant in many different diagnoses, but will have different weighted values depending on its association with other keywords in a symptom set. Further, the overall weight or ranking of a particular diagnosis in relation to other diagnoses may change or be thrown out completely due to other factors. Collaboration with the Stanford University School of Medicine Department of Medical Information on the Stanford XML Tagging Tool generated a web-based architecture, b7

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Mar 01, 2005
Accession Number
ADA436520

Entities

People

  • Warren L. Whitlock

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Abstracts
  • Biomedical Research
  • Databases
  • Emergency Medicine
  • Environment
  • First Responders
  • Instructions
  • Medical Personnel
  • Medical Specialties
  • Microorganisms
  • Nomenclature
  • Personal Digital Assistants
  • Physicians
  • Special Operations Forces
  • Therapy
  • United States Special Operations Command
  • User Interface

Readers

  • Artificial Intelligence
  • Systems Analysis and Design
  • Trauma or Military Medicine