Information systems to support medical practice and scientific discovery.

In many educational applications of the computer, systems are designed to contain information or knowledge that is to be conveyed to the student or user via an interactive learning scheme. In contrast, most NLM systems strive merely to increase the user's access to the broad range of published scientific literature, or to link the user to appropriate factual databases. Providing or enhancing access is, of course, made more difficult than in didactic systems because of the relatively large size of the files that make up the bibliographic and factual databases. For example, the MEDLARS files contain over 12 million records, made up of roughly thirty billion online characters. Users send electronic queries to these files on-line via commercial value-added networks (24 hrs/day, 365 days/year), in total about 4 million queries per year. Of the 4 million, approximately half of the queries were (in the opinion of the users polled) prompted by questions that arose in caring for a particular patient. The remaining 2 million queries concerned research, teaching, and public policy.