Abstract
Inability to find answers to clinical questions is a major obstacle to obtaining just-in-time information during patient encounters. We have introduced a new model for describing the content of documents in domain-specific collections, using document classes and semantic components, that may supplement existing indexing and searching techniques and improve information retrieval. In this paper we describe the model and present the results of our investigations into using the model to represent clinical questions in the medical domain. We manually mapped generic questions from a clinical question taxonomy to two web-based document collections using the document classes and semantic components we identified for each collection. We successfully mapped 36 of 50 question categories in one resource, and 34 of 50 in the other. Based on the frequency of the question types in the taxonomy, over 92% of questions were covered by the mappings in both resources. We conclude that our model is capable of usefully representing information needs and we propose ways to use the model for retrieval.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have