Login

Publications  •  Project Statistics

Glossary  •  Schools  •  Disciplines
People Search: 
   
Title/Abstract Search: 

Dissertation Information for Debbie Travers

NAME:
- Debbie Travers

DEGREE:
- Ph.D.

DISCIPLINE:
- Library and Information Science

SCHOOL:
- University of North Carolina, Chapel Hill (USA) (2003)

ADVISORS:
- Stephanie Haas

COMMITTEE MEMBERS:
- Gary Marchionini
- Barbara Marie Wildemuth
- Olivier Bodenreider
- Joanne Gard Marshall

MPACT Status: Fully Complete

Title: Identification of concepts from emergency department text using natural language processing techniques and the Unified Medical Language System RTM

Abstract: This research is part of a larger project to develop a thesaurus for emergency department (ED) chief complaint (CC) information. The CC is the patient's reason for visiting the ED, and is determined by the triage nurse in the initial minutes of the visit. The nature and severity of the CC directly influence many aspects of the patient's ED visit. CC data are also vital for public health surveillance activities. Despite the significance of the CC, there is no standard terminology for CC.

The goals of this research were to identify the concepts that comprise the domain of ED CC, and to develop a modular natural language processing (NLP) system for use in processing clinical text. The resulting Emergency Medical Text Processor (EMT-P) system is a series of modules that extracts standardized terms from clinical text using NLP and the Unified Medical Language SystemŽ . After applying EMT-P to a corpus of CC data representing all visits to three EDs during a one-year period of time, 83% of the original CC entries matched a UMLS concept. Samples of text/UMLS concept matches and non-matches were evaluated to determine the accuracy of EMT-P. 96% of the matches were rated equivalent or related, and 38% of the non-matches were found to match UMLS concepts. The results show that EMT-P Version 1 is relatively accurate; areas needing improvement in future versions of EMT-P were identified.

In the course of this study, a modular NLP system called EMT-P was developed and used to process a corpus of clinical text and extract standardized terms from the majority of entries. 3898 ED CC concepts were identified for possible inclusion in an ED CC Thesaurus, and produced a model of the domain of ED CC. A set of recommendations for developing the ED CC Thesaurus was also compiled, and included further validation of EMT-P, some specific areas of content to be included, the need to address CC-related data, and operational issues regarding design and implementation. Future plans include application of EMT-P to other types of clinical text, including triage nurses' notes, and clinical reports.

MPACT Scores for Debbie Travers

A = 0
C = 0
A+C = 0
T = 0
G = 0
W = 0
TD = 0
TA = 0
calculated 2008-01-31 06:20:51

Advisors and Advisees Graph

Directed Graph

Students under Debbie Travers

ADVISEES:
- None

COMMITTEESHIPS:
- None