A study of extracting knowledge from guideline documents

Abstract

Guideline documents offer a rich repository of information on clinical decisions, actions and prescriptions. However, clinicians do not use them as much as expected since health care organisations started to develop them. One alternative to promote the use of guidelines is to automatically select the relevant information at the point of care. But, extracting knowledge from a guideline document is an arduous and complex task. In this paper, we propose to apply the methodology CommonKADS in the analysis phase of a clinical practice guideline, with the aim of systematizing knowledge acquisition, providing a methodological support that helps to detect and document all the transformations from natural language to the structured representation of a knowledge model. When forcing to the knowledge engineer to keep these transformations, the knowledge modelling becomes more gradual. © 2009 Springer-Verlag Berlin Heidelberg.