Home | Business News | Browse by Publication | T | The Australian Library Journal

Precision with ease: refining thesaurus support for quality health information searching on HealthInsite.

Article Excerpt
HealthInsite is the Australian Government's Internet gateway to reliable health information online, providing access to over 15,000 information items on the websites of more than 80 approved information partners.

The gateway provides a variety of searching and browsing options to assist users to find information on a wide range of health topics. It relies heavily on subject and other metadata associated with each information resource to provide relevance-ranked search results lists. Subject metadata terms are selected from the Health and Ageing Thesaurus which was developed by the Department of Health and Ageing Library.

All of the HealthInsite search and browse options have been successively improved over the eight years since HealthInsite was launched, but the need for significant enhancements to the search engine and search results displays prompted a major upgrade. The upgrade also presents an opportunity to better integrate the Health and Ageing Thesaurus into the search engine to provide support for topic page authoring and searches on non-preferred terms, synonyms and related terms.

Background

HealthInsite was launched in April 2000 as a gateway to quality health information for Australians. Its main aim is to make it easy for users to find quality-assessed health information.

The Verity Information Server was implemented in mid-2001, with new search interfaces to support Quick Search, Power Search, Thesaurus Search and Refine Search options. These interfaces enable searching of both text and metadata associated with HealthInsite information partners' resources and HealthInsite pages. Search results are sorted according to their relevance scores as calculated by the search engine algorithms, page type, and date of last modification. The topic pages include a comprehensive embedded search that uses more sophisticated metadata searching to retrieve more relevant items than users' own searches.

In late 2002, a 'search for words similar to the one you typed' function was implemented to provide an option for users whose initial search resulted in either no results, or only results with low relevance scores. By clicking on the 'search for similar words' link, users are presented with a search results list for similar spellings.

Following an evaluation of searching on HealthInsite and a review of the literature (Smith et al. 2003), the search interface was redesigned. Enhancements undertaken are listed in the 'Results' section of this paper. A number of search-related recommendations were made in an Operational Review of HealthInsite in 2006 and these are also discussed in the 'Results' section.

Functionality to meet the needs identified through these processes was incorporated into the requirements for the new search engine.

Current thesaurus use in search

HealthInsite relies heavily on metadata elements to support not only searching and results displays, but also the administration and compilation of Web pages. In particular, the subject element in the metadata for each resource accessed through HealthInsite is used in:

* weighting search results

* filtering to limit to subject-related metadata elements

* automatic searches used to create results lists for topic pages

* the Thesaurus search option

* the Thesaurus navigator option

* personalised searches for personal profile users

* displaying subject terms and tree codes for logged-on authors.

The subject metadata element is populated by terms from the Health and Ageing Thesaurus which is managed by the Library of the Department of Health and Ageing. The Thesaurus was originally developed to meet the needs of staff of the Australian government health portfolio agencies but it is also useful for other health-related collections, including consumer resource collections and has been adopted by a number of health organisations in Australia. It is modeled on Medical Subject Headings (MESH) but with a much smaller number of preferred terms (about 3500) and a preference for Australian terminology in common usage.

In HealthInsite the hierarchical structure of the Thesaurus is used in automatic searches in topic pages and in personal profile searches to search for terms within thesaurus trees, and in the Thesaurus Navigator option to display terms for selection and searching. The Thesaurus Search option enables a user to type in a term and get a display of all thesaurus terms that include that term, and then select the term(s) to be searched. However, in the current system, if a user types a term that is not a preferred thesaurus term, a message is displayed simply telling the user that the term is not a preferred term without offering any suggestions for alternative search strategies.

In the weighting of search results, items that have the search term as one of the subject metadata elements are given extra weighting. There is also an Advanced Search filter for 'title, description and subject keywords' that limits results to items that have only the search term in those metadata elements.

Logged-on system authors are also able to see the subject terms and their associated thesaurus tree...

View this article FREE - Now for a Limited Time, try Goliath Business News
Free for 3 Days!