Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Natural language processing systems for capturing and standardizing unstructured clinical information
Kreimeyer K., Foster M., Pandey A., Arya N., Halford G., Jones S., Forshee R., Walderhaug M., Botsis T. Journal of Biomedical Informatics73 (C):14-29,2017.Type:Article
Date Reviewed: Jun 11 2018

Clinical information could be the poster child for big data and artificial intelligence (AI). It encompasses large amounts of unstructured text, which has to be harnessed to be effectively used (1) by medical personnel treating a patient on the spot, and (2) as a valuable data repository to run statistics on treatment options, efficacy of prescription drugs, and overall diagnostic support.

The authors’ goal is to present a systematic approach to investigate clinical natural language processing (NLP) systems in use today. They use a set of criteria to evaluate 5,000 articles with references to such systems, and eventually cull 71 NLP systems from the articles to analyze further. By outlining their methodology, their selection process becomes transparent.

The authors provide a listing of systems/projects, which is a useful starting point for practitioners and researchers in the field of clinical NLP systems. They provide two tables with basic (for example, NLP type, framework, open source, references) and detailed (for example, descriptions, evaluation, performance, usage) assessments of the 71 NLP systems, with some pointers to the underlying algorithms and methodologies. But the shortcomings of their methodology become obvious when many of their assessments of individual projects are left incomplete--for example, framework “not found”; evaluation and performance “unknown” (pp. 20-27).

The authors indicate that their research is a work in progress and that their ultimate goal is “making quality NLP applications available for specific use cases” (p. 22). Rather than presenting a laundry list of projects and systems, it would have been more useful to provide a smaller set of systems with tighter criteria, such as “in actual use” or “applied to industrial strength problems.” Moreover, issues like the use of ontologies, the processing of acronyms, and handling temporal information in clinical texts need to be addressed in a more comprehensive framework. A tighter coupling of actual use and chosen framework would lead to more expressive taxonomies of such systems.

For the researcher or practitioner in the field, the lists provided are at best pointers to the more research necessary to judge any applicability of the presented systems/projects to clinical information processing at large. This paper is a shortcut to find listings of clinical NLP projects/systems, but a long shot to determining if most of them are suitable for industrial use in clinical domains at large. It also points to the need for a more stringent approach when it comes to establishing frameworks and criteria for determining what makes NLP systems viable in clinical information processing.

Reviewer:  Klaus K. Obermeier Review #: CR146076 (1808-0449)
Bookmark and Share
  Reviewer Selected
Featured Reviewer
 
 
Natural Language Processing (I.2.7 )
 
 
Medical Information Systems (J.3 ... )
 
Would you recommend this review?
yes
no
Other reviews under "Natural Language Processing": Date
Current research in natural language generation
Dale R. (ed), Mellish C. (ed), Zock M., Academic Press Prof., Inc., San Diego, CA, 1990. Type: Book (9780122007354)
Nov 1 1992
Incremental interpretation
Pereira F., Pollack M. Artificial Intelligence 50(1): 37-82, 1991. Type: Article
Aug 1 1992
Natural language and computational linguistics
Beardon C., Lumsden D., Holmes G., Ellis Horwood, Upper Saddle River, NJ, 1991. Type: Book (9780136128137)
Jul 1 1992
more...

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy