Modern Methods of Handling the World's Biomedical Literature, 1965
Page  1

modern methods of handling the worlds biomedical literature martin m cummings md i am pleased to have this opportunity to participate in the inaugural scientific program which celebrates the opening of the new st paul-ramsey hospital i am honored also to do this under the sponsorship of the ramsey county tuberculosis and health association which has done so much to assist in the eradication of tuberculosis in this community i note also with historical interest that the ancker hospital first opened its doors in 1873 the same year in which the national library of medicine then the surgeon generals library issued volume i of its catologue which was americas first effort co control and provide access to the worlds medical literature it should be pointed out also that it was a physician-surgeon dr john shaw billings who developed this catalogue as well as initiating the publication of index medicus in 1879 which has since served as the most comprehensive index to the worlds serial publications in the field of medicine it is not well known that the same dr billings wrote the specifications for the invention of an electronic tabulator which used punched cards for sorting and collating data in the census of 1880 and 1890 and which is considered to be the fore runner of the modern electronic computer the recording of medical knowledge dates back to the 3rd millennium when prescriptions of various drugs and other methods of dealing presented at the inaugural scientific program st paul-ramsey hospital st paul minnesota november 12 1965

Page  2 2 with diseases were written on sumerian tablets and egyptian papyri among the earliest evidence of logical classification of such records is the catalog of hieratical books graven on one of the walls of the ancient egyptian library at edfu centuries elapsed before private collecting of books led to the establishment of public libraries and before the ever-increasing volume of literature led to the establishment of specialized scientific and medical libraries 2 price has stated that both the number of scientific periodicals and the number of papers in them have increased tenfold every half century beginning with 1750 when there were about ten scientific journals in the world today there are more than 50000 the biomedical literature has followed this trend with more than 15000 serial titles currently published abstracting services were introduced about 1850 but their rapid growth often has served to further clog the literature rather than reduce it a basic distrust of the accuracy of abstracts compels many physicians and particularly scientists to insist on access to the primary source of information the scientific article bearing on this growth is the rapid increase throughout the world of persons involved in medical and health related activities indeed in the united states alone the number of people engaged in health professions multiplied by a factor of 5.8 between 1900 and 1960 while the total 4 population multiplied by a factor of only 2.4 the phenomenal increase in the volume of biomedical literature and

Page  3 3 in the number of contributors and users has not however been paralleled bu equivalent growth of medical libraries this lack of parallel growth of medical literature and of information storage and retrieval methods and facilities adequate for todays needs has led to serious inadequacies in the capability of libraries to catalog index store and retrieve published literature for the use of scientists physisicts and others the computer has been seized upon as a new means of assisting and supplementing the traditional library approch to management of the published literature. at the national library of medicine the successfully used for retrieval or reference an exciting future for the entire field of information retrieval manual techniques for preparing our bibliographies listing of worlds medical literature could not meet demands for quick retrieval of citations from the rapidly expanding biomedical literature to provide better control of this literature the national library information storage and retrieval system which is called medlars medical literature analysis and retrieval system

Page  4 4 1 to improve the quality of index medicus while increasing the number of journals indexed in it 2 to reduce the time required for preparing the monthly editions of index medicus 3 to provide for increased depth of indexing 4 to provide for the production of compilations similar to index medicus in form and content but devoted to specialized subject fields 5 to provide for prompt and efficient searching of a larger computer store of information for citations to biomedical and biomedically related literature 6 to reduce duplication of literature screening activities carried on by academic and other institutions and organizations in the united states 7 to provide for the incorporation of monographs and other nonjournal 9 literature into medlars the system design was completed in january 1962 the preparation of computer programs installation of equipment training of personnel and detailed system testing took approximately 2 years the new system went into operation in january 1964 medlars has three major subdivisions i an input subsystem in which the skills of professional indexers are used in connection with the capabilities of a large-scale digital computer ii a retrieval subsystem in which the capabilities of professional literature searchers are used in connection with computer manipulations and iii a publication subsystem that converts

Page  5 5 retrieved citations and descriptors characterizing these citations into photopositive film slide 1 through the computers input programs the information on the paper tape is recorded on reels of magnetic tape edited and incorporated into two major data files the compressed citation file and the processed citation file the compressed citation file contains highly coded citations that can be retrieved as demand bibliographies that is bibliographies intended not for publication but for individual use of the requester and providing references on highly specific subjects the processed citation file on the other hand contains citations used in publishing index medicus and other recurring bibliographies - that is bibliographies intended for serial publication and wide dissemination retrieval subsystem through the retrieval system citations stored in the computer are recovered on demand requests for bibliographic citations from physicians scientists librarians and others are forwarded to a staff of specialists with extensive training in indexing and in the logic of computer searching these search specialists analyze the requests enumerate the relevant search elements and formulate search statements linking these elements logically formulated statements de signed to retrieve demand bibliographies are punched into paper tape and fed into the computer the retrieved citations machine-edited and sorted are recorded on magnetic tape and decoded into natural language prior to being printed either on cards or on continuous tabulating paper by the computer printer

Page  6 6 slide 2 computer 3 tape store publication subsystem the publication subsystem see fig 2 pro vides recurring bibliographies by processing citations for eventual printing from photopositive film the magnetic-tape files of retrieved citations are used for preparation of print copy by a computer driven phototypesetter called grace and acronym for graphic arts composing equipment grace is also used occasionally for preparing demand bibliographies 4 grace operating at a speed of about 300 characters per second grace can use 226 different characters in preparing 23-centimeter-wide positive photographic film or paper 5 page because medlars is a man-machine system its success is directly dependent on the intellectual input such input is provided by several kinds of specialists specialists in medical subject headings are responsible for formu lating the controlled vocabulary basic to consistent efficient and accurate analysis indexing and searching of biomedical literature this controlled vocabulary known as medical subject headings is the dic tionary to index medicus and its related publications it is the key to retrieval of information for recurring and for demand bibliographies and the base for most of the computer programs not only does it provide descriptors for indexing and for retrieving citations to the biomedical literature stored in the computer it also provides descriptors for cata loging books and for preparing the public card catalog used at the national library of medicine and elsewhere

Page  7 7 the controlled-vocabulary list is developed from recommendations made by i trained professionals engaged in indexing and in searching the biomedical literature ii users of index medicus and medlars and iii advisory panels composed of physicians biomedical scientists and other health specialists in 1965 edition of medical subject headings published as part 2 of the january 1965 issue of index medicus contains over 6300 subject headings arranged in two sections alphabetically with cross references and in subject categories the magnitude and significance of the indexers task may be seen from the fact that 152030 articles were indexed in 1964 and that the librarys indexing section plans to index 300000 articles annually by 1970 worldwide surveys have shown that some 15000 biomedical and related serial publications are published annually of these 6000 serial publications containing over 300000 articles annually are regarded as worthy of being indexed by the national library of medicine the magnitude and significance of the indexers task may be further seen from the various publications that are dependent on indexing - index medicus bibliography of medical reviews and a number of specialized bibliographies we view the specialized bibliography as a most important contribution to physicians and scientists who wish to keep up with their medical specialties without having to plow through the enormous acreage of literature which is generated preparation of the basic publication index medicus is a demanding

Page  8 8 and difficult task in 1961 an average monthly issue of index medicus had 450 pages and contained references to more than 10000 articles the january 1965 issue of index medicus had 691 pages cited 14665 articles and in addition contained two special features - i the latest revision of medical subject headings and ii the list of journals indexed in index medicus giving the titles of the 2472 journals then indexed by the library in 1961 the annual total of articles indexed was 132154 in 1964 it was 152030 in 1961 12661 journal issues were indexed in 1964 the figure rose to 15497 beginning with the march 1965 issue index medicus contains a monthly bibliography of medical reviews this publication is important to those who wish to become familiar with a special field of medicine in composite form the annual bibliography of medical reviews now being assembled will contain references selected from over 2400 journals regularly indexed in index medicus and from 539 journals not regularly indexed still other specialists whose skills are used in connection with the computer are the searchers after journal articles have been indexed and the citations have been fed into the computer trained searchers formulate subject statements necessary for retrieving those citations that satisfy inquiries received from clinicians scientists teachers and librarians more than 4000 such demard bibliographies have been produced by medlars since it became operational references retrieved by computer search may be organized in various formats and printed if desired all subject headings assigned by an

Page  9 9 indexer to a given citation can be printed with the citation thus providing a profile that aids the requester in judging the appropriateness of the citations retrieved slide 6 c-v bib representative of recurring bibliographies are index keriicus and index to the dantal literature both published by the library the cerebrovascular bibliography published by the national institute of neurological diseases and blindness and the national heart insan be printed with the citation thus providing a profile that aids the requester in judging the appropriateness of the citations retrieved slide 6 c-v bib representative of recurring bibliographies are index keriicus and index to the dantal literature both published by the library the cerebrovascular bibliography published by the national institute of neurological diseases and blindness and the national heart institute and index of rheumatology published by thean be printed with the citation thus providing a profile that aids the requester in judging the appropriateness of the citations retrieved slide 6 c-v bib representative of recurring bibliographies are index keriicus and index to the dantal literature both published by the library the cerebrovascular bibliography published by the national institute of neurological diseases and blindness and the national heart institute and index of rheumatology published by the american rheumatism association recurring bibliographies in various formative stages are those concerned with fibrinolysis and thrombolysis sudden death in infants and children smoking and health drug-induced abnormalities and venereal diseases slide 7 index rheumat of special interest is the index of rheumatology as a result of cooperation between the american rheumatism association and the national library of medicine this index is now being produced semi-monthly by the library and published by the association expected to provide over 6000 citations annually from the periodicals now indexed by the library the index is the first regular subscription bibliography aside from the librarys own publications to be fully attributable to medlars first year experience of particular significance is the practical experience gained in the first year of medlars operation despite im perfections in the controlled-vocabulary list need for improvements in indexing practices and techniques and some retrieval difficulties imposed by these deficiencies the results have been most heartening experience gained from more than 4000 searches made for physicians scientists teachers librarians and others has established the practicability of medlars beyond question slides 8-9 - demand searches

Page  10 10 appreciating that there is as yet no wholly satisfactory method of objectively evaluating the effectiveness of information storage and retrieval systems the library has relied heavily on consumer reaction and appraisal evaluation of critical reports indicates that the percentage of missed entries is minimal furthermore the relevance of retrieved citations as determined by the individual requesters evalua tion of demand bibliographies appears to be satisfactory new and more precise measurements of relevance are under study for the library to try to index all of the worlds substantive biomedical literature in the depth required by specialized information centers seems impracticable it is the librarys intent therefore to avoid unnecessary duplication by supplying specialized information centers with relevant bibliographies on magnetic tapes it is hoped that these centers will then analyze and refine the material in the light of their specialized requirements already cooperative arrangements involving the library and several universities have been made to test this concept slide nlm specialized information centers because medlars was designed to serve three purposes - to provide demand bibliographies recurring bibliographies and high-speed printing - it is likely to prove more economical and more utilitarian than a system designed to serve a single function also its location in a research library increases the efficiency of the system since the library can provide original documents or photocopies of the articles required by the user and traditional reference services are available to supplement the services of the computer system

Page  11 11 research and development the library realizes that medlars is only an initial response to the need for improved documentation technology directed toward better biomedical communication and modernization of library techniques it is in this spirit that the following activities have been planned for implementation within the near future ; decentralization in accordance with the librarys legal responsibility for wide dissemination of scientific and technical information in medicine a three-phase program to decentralize the medlars search and retrieval capability was begun in late 1964 this program calls for production of compatible duplicate tapes which would be made available to medical schools research institutions governmental organizations and industry for use in their own computer facilities these groups would then have the same search and retrieval capabilities as medlars in the first phase of decentralization the library has initiated two pilot studies one at the university of california los angeles with computer equipment not compatible with that installed at the library and the second at the university of colorado with compatible equipment the university of california is reprogramming the librarys honeywell tapes for use on ibm computers and will test the concept of regional ization of services it will also use the tapes to support the activities of a specialized brain research center the contract with the university of colorado provides for experimental use of medlars as in studies of selective dissemination of information a medlars center will be establisned in the united kingdom and sweden next year

Page  12 12 the second phase will be the formation of a network of additional search centers in areas with large concentrations of scientific manpower interest in this part of the program is so high that requests for sharing in the medlars searching capability have already been received from more than 40 university medical centers private corporations and government agencies as an aid in the selection of additional search centers the library has developed certain criteria the service potential of the institution its computer resources and consequent responsiveness to bibliographic demands generated by the automated system its interest in conducting further research and development based on use of medlars tapes and its ability to provide services on a regional basis such as distribution of specialized monthly bibliographies to physicians and scientists at local institutions the third phase of the decentralization program will be the provision of data tapes and program tapes at cost to interested insti tutions within the united states and abroad and experimentation with the linkage of established centers in the united states through the use of data transmission equipment automated acquisitions and cataloging system the acquisition of new books by a library is comparable to the purchase-order and accounts payable function of a private firm this library function and the re lated cataloging of acquisitions are obvious candidates for automation a systems analysis of these functions has been made by the national library of medicine and an automated system has been designed to

Page  13 13 improve the capacity and efficiency of these processes through the use of available computer equipment computer programming to implement the first step of this new system is now under way mechanization of the cataloging operation will produce i a rapid cataloging service that can be used by other libraries for their acquisitions and cataloging activities ii a printed book catalog listing all new acquisitions in the national library of medicine each year with a cumulative list published every 5 years and iii incorporation into medlars of citations to selected monographs graphic storage and retrieval system plans are now under way for development of a graphic-image storage and retrieval system that will permit rapid photocopy retrieval of the full text of documents in the librarys collection the library envisions a system that will tie the bibliographic capabilities of medlars to the graphic capabilities of this new photocopy system conclusion by its development of improved media for dissemination of information the national library of medicine is fostering a greater awareness and a better understanding of research and development efforts in behalf of public health and clinical medicine and a more rapid translation of research into clinical application the librarys transformation from a passive repository of information to an active ally of the researcher teacher and clinician has led to increased use of medical library facilities and has stimulated clinical application of research findings through better communication of published information the enactment of new legislation

Page  14 14 to support the upgrading of our nations medical libraries will permit the development of an efficient and orderly information system designed to meet pressing needs which have existed for more than a decade we expect to utilize the skills and technologies available in universities and industry to achieve this goal however no degree of mechanization can replace or substitute for reading and learning our job is to provide access to information your task is to evaluate and use it for improved health of people everywhere