Making the Conceptual Connections

The UMLS Section in this issue of JAMIA is dedicated to the memory of Marsden Scott Blois, Jr., a pioneer in medical concept representation. Dr. Blois, an internationally recognized physician-investigator and medical informatician, made important contributions to melanoma research and to biomedical informatics during his distinguished and varied career. His keen insight in the early days of the UMLS project contributed significantly to the future direction and success of the project.

This issue of JAMIA helps mark the tenth anniversary of the Unified Medical Language System (UMLS) project. From the beginning, the project has focused on overcoming the barriers users face when attempting to interact with computerized health information systems. The UMLS developers and their collaborators envisioned and then created a set of knowledge sources designed to support the development of sophisticated and accessible information systems. The UMLS knowledge sources have grown to now encompass the Metathesaurus (the largest of the knowledge sources), the Semantic Network, the Information Sources Map, and the SPECIALIST lexicon with its accompanying lexical programs. The Metathesaurus integrates more than 30 biomedical thesauri. The most recent release contains 331,756 concepts named by 739,439 different terms, including translations of some of the terminology into several other languages. Metathesaurus concepts are assigned semantic types from the Semantic Network, which, through its 135 semantic types and 51 relationships, provides a unifying semantic structure for the Metathesaurus terminology. The Information Sources Map is designed to serve as a resource for identifying databases and other information resources that are relevant to users' particular queries. The SPECIALIST lexicon is designed for use in natural language processing applications. The accompanying lexical programs work together with the lexicon (currently containing some 90,000 lexical records) in recognizing lexical variation in biomedical terminologies and texts. The UMLS data are distributed on CD-ROM, but, increasingly, researchers are gaining access by using the UMLS Knowledge Source Server, which provides flexible Internet-based access to all the knowledge sources.¹

The papers included here necessarily represent only a small portion of the active research that has been conducted on the UMLS over the past ten years. For more detail, readers should examine the recently compiled bibliography of some 280 selected citations on the UMLS project.² The current issue reflects a spectrum of publications that characterize UMLS research and development efforts. The contributions include a historical perspective on the UMLS project by its leaders³; a viewpoint paper by a prominent group not directly involved in core UMLS efforts on the role the ULMS now plays in addressing basic issues related to terminology development and use⁴; a paper on anatomic knowledge representation that proposes certain extensions to the UMLS Semantic Network and Metathesaurus⁵; a paper that describes semantically based methods to improve UMLS content quality and maintenance techniques⁶; and three reports on experiments applying the UMLS to specific clinical objectives.^7,8,9 Two papers^7,9 represent international work on using and extending UMLS knowledge in health information systems.

A decade of UMLS research has seen the development and testing of a rich set of continuously evolving knowledge sources. These knowledge sources are distributed regularly to the research community, a decision that was motivated by the belief that systems and tools improve through actual testing and use. Since early 1997, when the eighth annual release of the knowledge sources was announced, over 500 individuals and organizations worldwide have requested and received access to this eight edition. Since the first release of the knowledge sources in 1990, investigators have experimented with the UMLS data, and particularly its Metathesaurus, in a variety of application areas.^2,3 The UMLS—which incorporates multiple terminologies, each designed for its own purposes and users—has, through its rich set of interrelationships, resulted in an evolving system that is greater than the sum of its component parts.⁴ As the coverage of the UMLS grows and tools to improve its update and maintenance flourish,⁶ it should find increased applicability in supporting research and education in the basic clinical sciences (e.g., Rosse et al.⁵), in bibliographic and information science research and products (e.g., the Internet Grateful Med Fact Sheet¹⁰), and in the clinical arena (e.g., Joubert et al.,⁷ Cooper and Miller,⁸ and Bodenreider et al.⁹)