Generating Patient-Specific Summaries of Online Literature

Kathleen R. McKeown, Desmond A. Jordan, and Vasileios Hatzivassiloglou

Medical professionals increasingly use online resources to find journal articles of interest, looking for both the latest news in their specialty, and for articles that discuss results pertaining to patients currently under their care. We present a design for generating summaries of such online articles that are tailored to the characteristics of the patient under consideration. In this model, results of a document search are filtered to highlight articles that are clinically relevant to the patient by matching characteristics of the patients under study against characteristics found in the online patient record. Our summarization algorithm processes each relevant article and retrieves only those pieces of it that are appropriate to the individual patient; these are subsequently combined to produce the summary. We describe a user study which supports our summarization design, and present conclusions from it that indicate the users’ expectations from a good summary. We also outline our approach to several general natural language problems that we have identified as particularly important for the summarization task.


This page is copyrighted by AAAI. All rights reserved. Your use of this site constitutes acceptance of all of AAAI's terms and conditions and privacy policy.