Medical free-text queries often share the same scenario. A scenario represents a repeating task in healthcare. For example, a specific scenario is searching for treatment methods for a specific disease, where "treatment" is a term indicating the scenario. To support scenario-specific retrieval, in this paper we present a new knowledge-based approach to address these problems. In addition, we describe a testbed system developed using the approach. Our specific implementation uses the UMLS Metathesaurus and semantic structure to extract key concepts from a free-text. The approach uses phrase-based indexing to represent similar concepts, and query expansion to improve matching query terms with the terms in the document. The system formulates the query based on the user's input and the selected scenario template such as "disease, treatment" or "disease, diagnosis." Thus, it is able to retrieve documents relevant to the specific scenario.Evaluating the system using the standard OSHMED corpus, our empirical results validate the effectiveness of this new approach over the traditional text retrieval techniques. A. IntroductionThe volume of medical information and clinical data is growing at explosive rates. Ten years ago, medical publications were added to the world's biomedical journal collections at the rate of approximately 3,000 per month. Today, the volume is growing at 1,000 per day in Medline alone [NLM02]. As an artifact of patient care, hospitals generate huge amounts of healthcare data that is digitally available. As has been stated in the Institute of Medicine's report defining a new health system for the 21 st century [IOM01], the delivery of quality healthcare to consumers requires the availability and use of accurate information/knowledge compiled from this large volume of information. The demand by society and professional organizations for the use of evidence-based practices to help improve the quality of care also adds great pressure on healthcare professionals to regularly access the highest quality information during the processes of healthcare planning, decision and delivery. Therefore, computer-assisted information retrieval and processing are necessary today for supporting quality decision-making and helping to overcome human cognitive constraints [Chu02].A Medical Digital Library (MDL) consists of three types of data: structure data such as patient lab data and demographic data; multi-media images such as MRI's; and free-text documents such as patient reports, medical literature, teaching files and news articles. Previous research focused on the effective retrieval of structure data and image data [Chu98a, Chu98b]. However, many medical records are in free-text form and access to these records usually follows well-defined information gathering scenarios. A scenario can be defined as a reoccurring information need where the specific contextual information changes. For example, a physician may pose the following two queries, one for diagnosis and the other for treatment of a disease:• diagno...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.