“Aboutness” and other Problems of Text Retrieval in the Pharmaceutical Industry

In-house text databases of biomedical publications, as used by medical information departments in the pharmaceutical industry, are described in terms of their general features and styles of use. Problems in their upkeep include loss of precision in searching as the volume of publications grows, inconsistency in contents and in indexing philosophy, thesaurus maintenance, and the handling of newly emergent concepts or terms. The need to capture implicit information and the “aboutness” of documents impose heavy requirements for intellectual effort, but are essential for effective retrieval. Another conflict is between the need to identify specific documents and the need to create syntheses of state-of-the-art information in the form of knowledge bases. The extraction of meaning as opposed to the identification of possibly relevant documents remains a central task for information retrieval in the pharmaceutical industry.