Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Supporting the classification of pathology reports: comparing two information retrieval methods.

L M de Bruijn1, A Hasman, J W Arends

  • 1Department of Medical Informatics, University of Maastricht, Maastricht, The Netherlands.

Computer Methods and Programs in Biomedicine
|April 15, 2000
PubMed
Summary

This study compared word-based and N-gram methods for retrieving similar pathology reports using SNOMED codes. The word-based approach demonstrated superior performance in identifying relevant reports, enhancing coding quality.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

IMIA Accreditation of Biomedical and Health Informatics Education: Current State and Future Directions.

Yearbook of medical informatics·2017
Same author

Quality of health care: informatics foundations.

Yearbook of medical informatics·2016
Same author

Image and Signal Processing.

Yearbook of medical informatics·2016
Same author

Education and Research at the Department of Medical Informatics Maastricht.

Yearbook of medical informatics·2016
Same author

Computer-Aided Diagnosis of Abdominal Pain. (Reflections on F.T. De Dombal et al.'s paper: Computer-Aided Diagnosis of Acute Abdominal Pain).

Yearbook of medical informatics·2016
Same author

Computer supported patient management.

Technology and health care : official journal of the European Society for Engineering and Medicine·2014

Area of Science:

  • Medical Informatics
  • Information Retrieval
  • Computational Pathology

Background:

  • Pathology report coding accuracy is crucial for patient care and research.
  • Information retrieval methods can aid in efficiently searching and categorizing large volumes of pathology reports.
  • Standardized terminologies like SNOMED (Systematized Nomenclature of Medicine) are vital for consistent medical coding.

Purpose of the Study:

  • To compare the effectiveness of two information retrieval methods for selecting similar pathology reports.
  • To evaluate the performance of word-based versus N-gram vector representations in this retrieval task.
  • To assess the impact of these methods on improving the quality of SNOMED coding in pathology.

Main Methods:

  • Pathology reports were represented as vectors using either individual words or sequences of characters (N-grams of lengths 4, 5, and 6).

Related Experiment Videos

  • Similarity between reports was measured by comparing their associated SNOMED codes.
  • A comparative analysis was conducted to determine which representation yielded better retrieval results.

    • Main Results:

    • The word-based method consistently outperformed N-gram methods (4-, 5-, and 6-grams) in retrieving similar pathology reports.
    • This suggests that word-level features are more effective than character-level N-grams for this specific information retrieval task.
    • The findings indicate a potential for improved SNOMED coding through optimized report similarity assessment.

    Conclusions:

    • Word-based information retrieval is a more effective strategy than N-gram based retrieval for pathology report similarity assessment.
    • This approach can enhance the accuracy and efficiency of SNOMED coding in clinical pathology.
    • Future research could explore hybrid methods or advanced natural language processing techniques.