Atigeo at TREC 2012 Medical Records Track: ICD-9 Code Description Injection to Enhance Electronic Medical Record Search Accuracy

Abstract

The TREC 2012 Medical Records Track task involves the identification of electronic medical records (EMRs) that are relevant to a set of search topics. Atigeo has a Computer-Aided Coding (CAC) product that analyzes electronic medical records (EMRs) and recommends ICD-9 codes that represent the diagnoses and procedures described in those medical records. We have developed a suite of natural language processing (NLP) components that are useful for both tasks. Our TREC 2012 experiments focused on the ICD-9 admission and diagnosis codes included in more than 90% of the TREC EMRs: we used our comprehensive ICD-9 database to insert one of three variants of the text descriptions associated with each code found in each EMR. We describe the variations of ICD-9 code descriptions we inserted, the NLP components used for processing all the reports and topics, and report on the results of our experiments.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 01, 2012
Accession Number
ADA581772

Entities

People

  • Alex Thomas
  • Bryan Tinsley
  • Joseph F. Mccarthy
  • Mike Lazarus

Tags

Communities of Interest

  • Biomedical

DTIC Thesaurus Topics

  • Accuracy
  • Ambiguity
  • Computer Programming
  • Computers
  • Data Sets
  • Databases
  • Diseases And Disorders
  • Hiv Infections
  • Identification
  • Language
  • Natural Language Processing
  • Natural Languages
  • Pain
  • Personality
  • Precision
  • Respiratory System
  • Standards

Fields of Study

  • Computer science

Readers

  • Computer Programming and Software Development.
  • Information Retrieval
  • Medical or Health Care Field.

Technology Areas

  • AI & ML
  • AI & ML - Information Retrieval
  • Microelectronics