Searching for biomedical information on the World Wide Web

The rapid growth of biomedical information available via the Internet and its most popular retrieval system, the World Wide Web, has fostered active research and development directed toward locating resources that are appropriate for answering specific queries. The goal is to create tools that optimize information retrieval (as measured by two quantities, precision and recall) while minimizing the effort required by the user. Existing Web retrieval tools can be divided into the following groups: manually maintained topical lists; automatically generated word-based indices; software agents and multi-index searching aids; network cataloging methods; and miscellaneous hybrid and newer approaches. Improvements in current methods should arise from further research into: methods of describing objects on the Web; improved ways of searching for (and within) collections of documents as opposed to single documents; the ability to search for fielded documents; and ways to describe resources that span intra- and interdisciplinary as well as cross-cultural linguistic differences. For this last problem, the U.S. National Library of Medicine's Unified Medical Language System (UMLS) will be of great help. As online information retrieval improves, efforts are underway to improve the online information itself; quality control over content is being addressed as the peer-review systems of traditional printed journals migrate into the realm of electronic publication.