Development of a Corpus for Evidence Based Medicine Summarisation

Development of a Corpus for Evidence Based Medicine Summarisation Diego Molla Department of Computing Macquarie University Sydney, Australia diego.molla-aliod@mq.edu.au Maria Elena Santiago-Martinez Department of Computing Macquarie University Sydney, Australia maria.santiago-martinez@mq.edu.au Abstract In this paper we introduce some of the key NLP-related problems related to the practice of Evidence Based Medicine and propose the task of multi-document query-focused sum- marisation as a key approach to solve these problems. We have completed a corpus for the development of such multi-document query- focused summarisation task. The process to build the corpus combined the use of auto- mated extraction of text, manual annotation, and crowdsourcing to find the reference IDs. We perform a statistical analysis of the corpus for the particular use of single-document sum- marisation and show that there is still a lot of room for improvement from the current base- lines. 1 Introduction An important form of medical practice is based on Evidence Based Medicine (EBM). (Sackett et al., 1996; Sackett et al., 2000). Within the EBM paradigm, the physician is urged to consider the best available evidence that is relevant to the patient at point of care. However, the physician is currently overwhelmed with the large volumes of published text available. For example, the US National Library of Medicine offers PubMed1, a database of medical publications that comprises more than 19 million ab- stracts. The median time spent to conduct a clinical systematic review is 1,139 hours (Allen and Olkin, 1999). In contrast, the average time that a physician spends searching for a topic is two minutes (Ely et 1http://www.ncbi.nlm.nih.gov/pubmed al., 1999). In practice, the physician would typically try to keep up to date by reading systematic reviews. However, systematic reviews are generic studies that may or may not be applicable to the particular case that the physician is concerned with. When there are no appropriate systematic reviews, the physician will need to search over the research literature, find the relevant information, and appraise it in terms of quality of the results and applicability to the patient (Sackett et al., 2000). There is a range of NLP tasks that have been attempted on this area, but so far not much work has been done on multi-document query-based sum- marisation. We argue that this task would greatly help the physician but the lack of appropriate cor- pora has hindered the development and testing of such query-based summarisers for this domain. In this paper we present such a corpus, show some characteristics of the corpus, and advance some spe- cific tasks that the corpus is suited for. Section 2 introduces EBM and its connection with tasks related to multi-document query-based sum- marisation. Section 3 describes the corpus. Sec- tion 4 details how the corpus was built. Section 5 gives an indication of the use of the corpus for the specific task of single-document summarisation. Fi- nally, Section 6 concludes the paper. 2 Evidence Based Medicine and Summarisation In this section we introduce EBM and present work related to the use of NLP for …

Content truncated to 500 words