Characterizing Health-Related Information Needs of Domain Experts

Abstract

In information retrieval literature, understanding the users' intents behind the queries is critically important to gain a better insight of how to select relevant results. While many studies investigated how users in general carry out exploratory health searches in digital environments, a few focused on how are the queries formulated, specifically by domain expert users. This study intends to fill this gap by studying 173 health expert queries issued from 3 medical information retrieval tasks within 2 different evaluation compaigns. A statistical analysis has been carried out to study both variation and correlation of health-query attributes such as length, clarity and specificity of either clinical or non clinical queries. The knowledge gained from the study has an immediate impact on the design of future health information seeking systems.

Analysis of biomedical and health queries: Lessons learned fromTRECandCLEFevaluation benchmarks

“…The basic underlying assumption is that the more specific concepts are involved by the query words, the more specific the query topics are. Hierarchical specificity of a query is computed as the average normalized level of MeSH concepts that map the query words (Znaidi, Tamine, Chouquet, & Latiri, ). Query clarity. Broadly speaking, a clear query triggers a strong relevant meaning of the underlying topics, whereas an ambiguous query triggers a variety of topics meanings that do not correlate necessarily with each other.…”

“…We propose to compute the three following facets of the clarity feature: Clarity through word distribution WTCla ( Q ): from a language modeling viewpoint, a query is clear if it returns a few different topics where a topic can be estimated by the distribution of query words over the result documents (Steve & Croft, ). Thus, the clarity score of a query is computed here at a post retrieval stage, as the Kullback‐Leiber divergence between the query language model and the collection language model. Clarity through topic coverage CTCla ( Q ): a query is assumed to be as much clear as it covers a few general semantic levels of MeSH terminology (Znaidi et al., ). This score is computed at the preretrieval stage. Clarity through concept coverage CCCla ( Q ): a query is assumed to be as much clear as query words match concepts issued from MeSH terminology (Boudin et al., ).…”

“…The basic underlying assumption is that the more specific concepts are involved by the query words, the more specific the query topics are. Hierarchical specificity of a query is computed as the average normalized level of MeSH concepts that map the query words (Znaidi, Tamine, Chouquet, & Latiri, 2013). • Query clarity.…”

“…2. Clarity through topic coverage CTCla(Q): a query is assumed to be as much clear as it covers a few general semantic levels of MeSH terminology (Znaidi et al, 2013). This score is computed at the preretrieval stage.…”