Modern Processing of Information in the NLM, 1969
Page  1

[Seldo - place in my boxes of reprints- WC] [Reprinted by the U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE National Institutes of Health] Reprinted from the BULLETIN OF THE AMERICAN COLLEGE OF PHYSICIANS July 1969, Vol. 10, No. 7, pp 326-330 Copyright 1969, by American College of Physicians Modern Processing of Information in the National Library of Medicine MARTIN M. CUMMINGS, M.D. Director, National Library of Medicine The explosive growth of the world's medical literature makes it virtually impossible for physicians and scientists to satisfy their information needs without efficient indexing and abstracting services which lead them to relevant portions of the literature. American medicine has been favored by the availability of a specialized library which has for more than a century concerned itself with the informational needs of the community related to health. The National Library of Medicine, one of the nation's three national libraries, is the world's largest research library in a single and professional field. The Library has a statutory mandate from Congress to apply its resources broadly to the advance of the medical and health-related sciences. It collects, organizes and makes available biomedical information to investigators, educators and practitioners and carries out programs designed to strengthen existing and develop new medical library services in the United States. It is the central resource for the existing national biomedical information system. The collection was started in 1836 as the "Library of the Surgeon General's Office" (Army) and developed as a national resource under Dr. John Shaw Billings, Librarian from 1865 to 1895. Named "Army Medical Library" in 1922 and "Armed Forces Medical Library" in 1952, it became part of the Public Health Service, U. S. Department of Health, Education, and Welfare in 1956 under legislation introduced by Senator Lister Hill and the late President John F. Kennedy. At this time it was designated the National Library of Medicine. In 1962 the NLM moved from its old quarters in downtown Washington, D. C., where it was associated with the Armed Forces Medical Museum, to a new building on the campus of the National Institutes of Health in Bethesda, Maryland. In 1968 the NLM became part of the National Institutes of Health. The Library's holdings now total nearly 1,500,000 books, journals, theses, pamphlets, prints and microfilms. More than 70 languages are represented in the collection. New material is acquired at the rate of approximately 100,000 items a year. NLM provides interlibrary loans upon request from other libraries. No periodicals leave the library. Instead, single photocopies of requested individual articles are substituted. NLM developed a unique mobile camera whereby the camera van and operator go to the bound volumes rather than the books being carried to the camera. Through the use of specially designed duplicating equipment, the microfilms are reproduced in finished copy for mailing to the requesting library. Although much of the bibliographic material a physician needs to read is readily available through his own medical library, that which is not within his collection can be obtained through this mechanism. More than 160,000 such loans are provided each year. The Library pioneered in the first large scale computer-based operation of biomedical reference storage and retrieval techniques with the development and implementation of MEDical Literature Analysis and Retrieval System. MEDLARS became operational in January, 1964. It was designed to store, analyze and retrieve bibliographic citations to biomedical literature automatically. This sys-

Page  2 tem uses a computer driven phototypesetter to produce the monthly Index Medicus and Current Catalog, the latter a biweekly listing (cumulated quarterly) of citations to publications cataloged by the Library, or one of its "cooperative cataloging" partners. Of the many publications published by the National Library of Medicine the most important is Index Medicus, published monthly and cumulated annually. This is a bibliographic list of references to current articles from 2,260 of the world's biomedical periodicals. Monthly issues of Index Medicus are ar ranged in four sections. In the Bibliography of Medical Reviews subject section and the Bibliography of Medical Reviews author section, the articles cited represent surveys of recent biomedical literature. In the Index Medicus subject section each article cited appears printed under several subject headings, which represent the most important concepts of that article. The entry appears in full under each subject heading. Each entry contains the following elements, in order: original English title of article or English translation of title; senior author (if more than one) or sole author; journal title abbreviations; volume number; inclusive pagination; date of issue; and an abbreviation indicating the language of the article if other than English. References in the author section of Index Medicus, cite the first three authors' names. The title of the article in the vernacular appears next, followed by: the journal title abbreviation, volume number, inclusive pagination, date of issue, and language origin. Index Medicus coverage is limited to periodical literature. Proceedings of congresses, symposia and similar materials are not indexed unless published in periodicals. The ever increasing volume of biomedical literature has made it desirable to have judgments and recommendations from consultants as to the number and quality of journals to be included in Index Medicus. The National Library of Medicine is advised by a group of distinguished physicians, medical editors, and medical librarians. The Library indexes the periodical literature that is judged most important by this advisory group. Unfortunately it is not possible to include every journal that might contain useful articles and more journals are usually recommended for inclusion than can be actually indexed. The current members of the Journal Selection Committee are: Dr. William B. Bean (Chairman), Professor and Chairman,Department of Internal Medicine, State University of Iowa; Dr. Morris Fishbein of Chicago, former Editor of the Journal of the American Medical Association; Dr. Franz J. Ingelfinger, Editor of the New England Journal of Medicine; Dr. John H. Talbott, Editor of the Journal of the American Medical Association; Mr. William K. Beatty, Librarian of the Archibald Church Medical Library, Northwestern University; Miss Myrl L. Ebert, Librarian, Division of Health Affairs, University of North Carolina, and Mr. Harold L. Bloomquist, Librarian, The Francis A. Countway Library of Medicine, Boston, Massachusetts. When articles are indexed for Index Medicus, the most specific term available in MeSH is used. NLM's controlled vocabulary is a vital part of the Library's computer-based Medical Literature Analysis and Retrieval System (MEDLARS). MeSH is the list of descriptors used by NLM for indexing articles from the world's leading biomedical journals. MeSH descriptors are subject oriented terms which characterize an article's content. By co-ordinating MeSH terms, the computer is able to store and retrieve bibliographic citations for publication in the monthly Index Medicus and in response to demand search requests from scientists, practitioners and educators. MeSH is also used for cataloging the Library's books and other documents. The Medical Subject Heading Staff is continually revising and updating the vocabulary. Subject specialists are responsible for areas of the health sciences in which they have knowledge and experience. The specialists collect new terms as they appear in the scientific literature, define them, establish their relationships with other conceptual terms in the vocabulary, and recommend their addition to the Subject Heading list. The Library staff meet frequently with physicians and scientists working in specialized biomedical disciplines and undertake a revi-

Page  3 sion of the terminology relating to their areas of competence. Acting as consultants to the Library, these groups provide the additional specialized knowledge required to keep the vocabulary current with progress being made in the sciences. The MeSH vocabulary published annually as Part 2 of the January Index Medicus consists of an alphabetical listing, with cross references, and a categorical arrangement that places the terms in relationship to other terms representing similar areas and concepts. For example, anatomical terms are listed together, disease terms in a second category and drugs and chemicals in another. Thus, if an article is concerned with erythromycin, it would be indexed under that term, rather than the broader "antibiotics." Under the latter term, one would find only: (a) articles dealing with antibiotics generally, (b) those that deal in a very general way with a large number of named antibiotics, or (c) those that deal with specific antibiotics which have not been separately listed in MeSH. The user who is interested in one concept may find that the broader term is also worth examining for other relevant articles. Although one may be interested primarily in penicillin sensitivity, articles on the general subject of antibiotics sensitivity may also be useful. The 1969 revision of MeSH contained 7,450 terms, 60 subheadings, and 11,000 cross references. For computer searches more than 100,000 useful combinations of terms and subjects may be obtained. Using the same data source as Index Medicus the Library prepares recurring bibliographies in specialized subject areas for publication by cooperating professional organizations. During the past year 15 recurring bibliographies have been prepared and are printed and distributed by nonprofit professional organizations and government agencies with whom the Library cooperates. They include: (1) The Bibliography on Medical Education, published monthly in the Journal of Medical Education; (2) The quarterly Cerebrovascular Bibliography, prepared under the auspices of the joint Council Subcommittee on Cerebrovascular Disease, National Heart Institute; (3) The monthly bibliography, Fibrinolysis, Thrombolysis and Blood Clotting, distributed by the National Heart Institute; (4) The monthly Index of Rheumatology, available from the American Rheumatism Association Section of the Arthritis Foundation; (5) The quarterly Index to Dental Literature, sold by the American Dental Association; (6) The quarterly international Nursing Index, sold by the American Journal of Nursing Company; (7) The quarterly Artificial Kidney Bibliography, published by the National Institute of Arthritis and Metabolic Diseases; (8) The bimonthly Endocrinology Index, published by the National Institute of Arthritis and Metabolic Diseases; (9) The Bibliography of Surgery of the Hand, a quarterly, published and distributed by the American Society for Surgery of the Hand; (10) The Anesthesiology Bibliography, a bimonthly, published and distributed by the American Society of Anesthesiologists; (11) The quarterly Toxicity Bibliography, published by the National Library of Medicine; (12) The monthly Current Bibliography of Epidemiology (CUBE), published by the American Publice Health Association; (13) The Neurosurgical Biblio-Index, a quarterly, published by the American Association of Neurological Surgeons; (14) The Cranio-Facial-Cleft Palate Bibliography, published quarterly by the American Cleft Palate Association; (15) The Index of Investigative Dermatopathology and Dermatology, published monthly by the Universities Associated for Research and Education in Pathology, Inc. Another product of MEDLARS is the "demand search," resulting from machine searches of the computer file (covering the period January, 1964, to date) in response to complex questions which cannot be handled effectively by the use of traditional printed indexes or catalogs. Almost 1 million augmented bibliographic citations are now stored on magnetic tape. This means of access to the MEDLARS store of information is through a specific individualized search of the file. Specific bibliputer file (covering the period January, 1964, to date) in response puter file (covering the period January, 1964, to date) in response to complex questions which cannot be handled effectively by the use of traditional printed indexes or catalogs. Almost 1 million augmented bibliographic citations are now stored on magnetic tape. This means of access to the MEDLARS store of information is through a specific individualized search of the file. Specific bibliputer file (covering the period January, 1964, to date) in response to complex questions which cannot be handled effectively by the use of traditional printed indexes or catalogs. Almost 1 million augmented bibliographic citations are now stored on magnetic tape. This means of access to the MEDLARS store of information is through a specific individualized search of the file. Specific bibliographic references are prepared by MEDLARS in response to requests from individual health professionals through their local or regional medical libraries. An article may be indexed under as many

Page  4 as 10 categories in Index Medicus. MEDLARS however, may carry 25 or more descriptors of the same references. The use of the magnetic tapes that contain the data base of bibliographic citations and the larger list of descriptors permits the electronic computer to select from the larger for those most likely to be relevant. The efficiency of the system is not determined as much by its hardware and software as by its "liveware," human professional indexers. The physician or scientist will find his inquiries best answered when it has been formulated after he has reached a good understanding with his librarian or with a search analyst as to what he really wants and what he doesn't want. We are attempting to link a number of medical libraries with one another and with the NLM to increase access to MEDLARS and other library services. This represents a deliberate attempt to make medical libraries in general and the NLM in particular more responsive to the needs and desires of its users. A more recent innovation in Library operations and one which has been well received are the NLM Selected Literature Searches. These are bibliographies originally produced as "demand searches" in response to many individual requests for the same item. When there are repeated requests for a search and it is of current topical interest, it is reformulated and printed for immediate widespread distribution to interested health professionals. Currently some 69 titles are available. During the past year more than 18,000 copies were mailed out on request. The list of latest available Literature Searches appears in the NLM News, JAMA, Public Health Reports, and other medical publications. From the beginning the National Library of Medicine has recognized that the more uses that are made of the centrally prepared store of citations on magnetic tape, the greater would be the return to the national health effort. The Library, therefore, has assisted in the establishment of six decentralized MEDLARS stations to afford health professionals greater access to the store of journal article citations and thus to provide immediate implementation of improved library services. MEDLARS centers are located at: the Biomedical Library of the University of California at Los Angeles, the University of Colorado Medical Center in Denver, the University of Alabama Medical Center in Birmingham, the University of Michigan in Ann Arbor, Harvard Medical School in Boston, and the Ohio State University Medical College, Columbus. In addition, MEDLARS search stations are in operation in Sweden and the United Kingdom. Based on the experience gained at these search stations, the Library plans further decentralization to other institutions to develop a network of MEDLARS stations providing rapid access to the world's biomedical literature on a regional basis. This is a network in the true sense of the word. Access to this system is best obtained through a local or regional library, although individual requests from individuals who do not have access to a library are honored by NLM. This service is important to physicians in rural areas of the country. The Library's Remote Information Systems Center (RISC) is capable of searching information files distant from Bethesda. Beginning in March, 1968, electronic linkage was established between two of the stations (Harvard and Colorado) and the National Library of Medicine. Searches formulated at these stations are transmitted via IBM 1050 equipment which produces punched cards at the Library. These punched cards are then used to query the computer. RISC is an excellent example of the type of modern facility which permits the sharing of resources in the development and use of information and communications system. RISC permits the National Library of Medicine, for the cost of a terminal device and smaller user charges, to access large information and computational systems which other organizations have developed at great expense over long periods of time. Through RISC, the following on-line, timesharing computer systems are accessed on a regular basis: System Development Corporation's Time-Sharing System (TSS) in Santa Monica, California; System Development Corporation's Centralized Information Reference and Control On-Line (CIRCOL) in Dayton,

Page  5 Ohio; State University of New York (SUNY) Biomedical Communications Network computer system at the Upstate Medical Center, Syracuse, New York; Massachusetts Institute of Technology's MAC/Technical Information Program (TIP) in Cambridge, Massachusetts; Massachusetts Institute of Technology's Lincoln Laboratory System in Lexington, Massachusetts; IBM's QUIKTRAN; TYMSHARE's computer system in Englewood Cliffs, New Jersey, and Stanford Research Institute's Question-Answering System (QAS) in Palo Alto, California. The list of systems available through RISC is growing steadily as new on-line systems of interest are identified and the necessary accessing arrangements are made with the operators of the systems. During the past three years the National Library of Medicine has been developing the framework for a biomedical communications network. In August, 1968, the President signed into law legislation designating the planned National Center for Biomedical Communications "which is to be constructed and located as part of this library," as the Lister Hill National Center for Biomedical Communications. The first responsibility assigned to the Center was the development of the Biomedical Communications Network. Three other responsibilities delegated to the Center by the Secretary of DHEW are also important: (1) to apply existing and advanced technology to improve biomedical communications; (2) to serve as the focal point in the Department of Health, Education, and Welfare for the technological aspects of biomedical communications, information systems, and network projects; (3) to represent the Department in those activities of the President's Office of Science and Technology, other Federal agencies, and interagency committees related to information and communications. Important too for the Lister Hill Center is the application of the latest technology in such areas as graphic image storage and retrieval, and the facsimile transmission of documents. New technology is already being applied at the Library's National Medical Audiovisual Center in Atlanta where attention is centered on the audiovisual services required for the network. NMAC operates a central facility for the development, production, distribution, and utilization of motion pictures, videotapes, and other audiovisual forms. It coordinates a comprehensive audiovisual program to assure maximum responsiveness and economy of funds and manpower, provides consultation and assistance in the development of specialized audiovisual activities, and encourages the production, dissemination, and utilization of medical films and other audiovisuals in the schools of health professions. It operates a national clearinghouse and archival program and acts as a national/international film and videotape center for the distribution and exchange of biomedical audiovisuals. The National Library of Medicine is also rich in historical material. NLM"s Incunabula Room, with some 500 books published before 1501, is as much the medical historian's laboratory as it is a showpiece. The oldest piece in the collection is an Arabic manuscript, dated 1094 by the physician Rhazes. There are more than 4,800 works printed between 1500 and 1600 and almost 500,000 books and journals printed before 1871 which makes it the largest collection of this type in the United States. Our strength and dedication to service rests on the close collaboration between the medical community and its National Library sustained over a century. With a combination of traditional library services and new, forward looking programs, the National Library of Medicine continues to be much more than a repository of documents. It is a dynamic information center, actively engaged in tying the past to the present while planning for, adapting or applying the latest in communications technology for the benefit of the entire medical community in the years ahead.