Skip to content

Advertisement

You're viewing the new version of our site. Please leave us feedback.

Learn more

BMC Proceedings

Open Access

KART, a knowledge authoring and refinement tool for clinical guidelines development

  • E Pasche1,
  • D Teodoro1,
  • J Gobeill2,
  • D Vishnyakova1,
  • P Ruch2 and
  • C Lovis1
BMC Proceedings20115(Suppl 6):O49

https://doi.org/10.1186/1753-6561-5-S6-O49

Published: 29 June 2011

Introduction / objectives

Optimal antibiotic prescriptions rely on evidence-based clinical guidelines, but creating such guidelines requires a time-consuming systematic review of the literature. We aim at facilitating this process by proposing an innovative tool to extract antibiotic treatments from the literature.

Methods

We develop a web application, embedding a question-answering (QA) module based on EAGLi (Engine for Question-Answering in Genomics Literature), which has been specifically tuned for antibiotherapy. The users ask questions (i.e. what antibiotic is used to treat cystitis caused by E. coli?) to which the system answers by retrieving a set of MEDLINE records from which the most frequently associated antibiotics are extracted and returned in a relevance-ranked list. The users can then access the annotated abstracts of the publications supporting the antibiotic as being a potential treatment, thus allowing them to use their expert judgment to accept or reject the assumption.

Results

The tool is accessible at http://eagl.unige.ch/KART. The QA engine was able to answer correctly to more than half of the queries (top-precision=0.56). In addition, infectious disease specialists from the University Hospitals of Geneva evaluated KART with several clinical scenarios. Despite an overall appreciation of the system and the recognition of its usefulness, improvements are required to use it when generating or updating clinical practice guidelines.

Conclusion

KART seeks to facilitate medical knowledge building by providing an advanced retrieval engine. It provides a novel approach to cope with high volumes of literature generated over systematic reviews by facilitating access to pertinent information on antibiotic-related treatments.

Disclosure of interest

None declared.

Authors’ Affiliations

(1)
Division of Medical Information Sciences, University Hospitals and University of Geneva
(2)
Information Science Department, University of Applied Sciences

Copyright

© Pasche et al; licensee BioMed Central Ltd. 2011

This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Advertisement