Long Range Plan National Library of Medicine Locating and Gaining Access to Medical and Scientific Literature iyi@i ^tefe^S • • in "5 JWT 1 '"*■ 194 1 ; .t | ^gp- —^a| U.S. Department of Health and Human Services Public Health Service National Institutes of Health Report of Panel Long Range Plan National Library of Medicine Locating and Gaining Access to Medical and Scientific Literature U.S. Department of Health and Human Services December 1986 Public Health Service National Institutes of Health Members and Staff of Panel 2 Locating and Gaining Access to Medical and Scientific Literature Chairperson Nancy M. Lorenzi, Ph.D. Associate Senior Vice President University of Cincinnati Medical Center Cincinnati, Ohio Members Douglas Brutlag, Ph.D. Associate Professor, Biochemistry Stanford University Medical Center Stanford, California Cyril Feng, M.S.L.S. Director Health Science Library University of Maryland Baltimore, Maryland Hugh Harroff, Jr., D.V.M. Veterinarian Battelle Memorial Institute Columbus, Ohio Mary M. Horres, M.S.L.S. Biomedical Librarian University of California, San Diego La Jolla, California James Isbister, M.A. President Consolidated Technologies Washington, D.C. Allan M. Kulakow, Ph.D. Director, African Programs Academy for Educational Development, Inc. Washington, D.C. Gertrude Lamb, Ph.D. Director Health Science Libraries Hartford Hospital Hartford, Connecticut George Lundberg, M.D. Editor Journal of the American Medical Association Chicago, Illinois Richard Reitemeier, M.D. Professor Mayo Clinic and Mayo Medical School Rochester, Minnesota Patricia Schwirian, Ph.D., R.N. Director, Office of Information Management Services, College of Nursing The Ohio State University Columbus, Ohio Robert Wedgeworth, M.L.S. Dean School of Library Science Columbia University New York, New York Martha Williams, M.A. Professor of Information Science University of Illinois Urbana, Illinois NLM Staff Susan Buyer Slater, M.A. Executive Secretary Lois Ann Colaianni, M.L.S. Resource Person John Anderson, M.S. Resource Person Consultants to Panel 2 Naomi C. Broering, M.L.S., M.A. Director Dahlgren Memorial Library Georgetown University Medical Center Washington, D.C. Holly Shipp Buchanan, M.Ln., M.B.A. Director, Corporate Information Resources NKC Hospitals, Inc. Louisville, Kentucky Stanley Foster, M.D. Assistant Director International Health Program Office Centers for Disease Control Atlanta, Georgia Lillian Haddock, M.D. Dean for Academic Affairs & Professor in Medicine School of Medicine University of Puerto Rico, San Juan, Puerto Rico R. Brian Haynes, M.D. Professor Department of Clinical Epidemiology and Biostatistics McMaster University Hamilton, Ontario, Canada Richard Janeway, M.D. Vice President for Health Affairs and Dean Bowman Gray School of Medicine Wake Forest University Winston-Salem, North Carolina Carol Jenkins, M.L.S. Director Health Sciences Library University of North Carolina at Chapel Hill Chapel Hill, North Carolina Jean K. Miller Director Health Science Center at Dallas Library University of Texas Dallas, Texas Acknowledgments This report is the culmination of one year of work on the part of the Panel Two members, consultants, and staff. The Panel wishes to thank all those involved for their generous contribution of time and talent to this effort. In particular, the Panel wishes to thank Carol Jenkins, special consultant to the panel, for sharing her considerable expertise and writing skills in developing the Panel report. Contents Long Range Plan 1 Background and Context 6 Report of 2 NLM Programs and Recent Accomplishments 8 3 A Vision of the Future 10 Research 10 Education 10 Medical Practice 12 Administration 13 Library and Information Science 13 4 Major Issues and Future Directions 14 Maintaining and Improving Access Through the Infrastructure of Organizational Units 14 Linkages and Networks Standards Technology Supporting Access to the Infrastructure by Users of Health Information 22 Users: A Model Opportunities To Improve Access for Users International Issues 5 Observations and Recommendations 28 Support for the Infrastructure 29 Support for User Access 30 Summary 31 References 31 Appendix A The Medical Library Assistance Act 32 Appendix B NLM Planning Process 50 Background and Context The timely transfer of scientific and biomedi- cal information is the fundamental purpose of the NLM (National Library of Medicine). The Nation's investment in medical research, education, health care, and development can be fully realized only when the results of research are applied by health professionals. For this to occur, the biomedical literature that reports the results of medical research and development must be acquired, organized, and made available to researchers, practitioners, educators, administrators, and students in the health professions. Panel Two, whose topic was "locating and gaining access to medical and scientific liter- ature," had as its general focus two ques- tions: (1) who are and who should be the users of NLM, and (2) what is the infrastruc- ture necessary to facilitate access to informa- tion. Those questions were posed against a background of rapid and massive changes in the amount of biomedical information being generated and in the means available for its transfer and distribution. The AAMC (Association of American Medi- cal Colleges) report, Academic Information in the Academic Health Sciences Center: Roles for the Library in Information Management, succinctly summarizes the state of the art of information access and dissemination: The generation of new knowledge has continued its unrelenting course, dou- bling at least every 10 years. The world's knowledge base is shifting inexorably from a paper to an electronic base. New ways of synthesizing, compacting, and representing knowledge are develop- ing .. .Transfer of quantities of data and information can be accomplished rela- tively inexpensively and nearly instan- taneously with the use of phone lines, satellites, and microwaves over any dis- tance. The societal view of the impor- tance of individualized access to and control of information is changing; in a highly competitive technologically based world, individuals and organizations with better quality information services are more productive and effective.1 Rapid technological developments continue to expand the possibilities for information access and dissemination. While technology development in and of itself is important, it must be wed to key integrating concepts, most notably those emerging from NLM's IAIMS (Integrated Academic Information Management Systems),2 so that information can be successfully updated, verified, and easily accessed. In the following report on access to medical information, Panel Two urges the NLM leadership to recognize and use the strengths of the existing infrastructure to achieve the scenarios envisaged and described by this Panel. The strongest components of the NLM infrastructure should be closely studied to determine how they can provide base sup- port for our vision of the future. This should include: (1) an examination and possible redefinition of the RML (Regional Medical Library) Network; (2) improvement of direct services to the user of information sources and services; and (3) support for programs of graduate education in library and informa- tion sciences. In addition, the Panel strongly supports NLM's efforts to develop technolo- gies related to knowledge-based systems and recommends further improvement in dis- seminating biomedical information. Panel Two encourages NLM to explore the further expansion of its services to include consumer health information, improved organization and availability of research data, and appen- diceal files. Two recent NLM initiatives are strongly supported by this Panel as well: first, the development of a Unified Medical Lan- guage is critical to several key NLM direc- tions and would serve also to significantly improve the infrastructure. Second, the IAIMS projects have created effective models that provide a process, framework, and clear set of goals for the year 2006. Several appendices to this report amplify the text. A 20-year review of the MLA (Medical Library Assistance) Act, prepared by several members of Panel One, provides insight into the information needs of health professionals and how the medical library network can satisfy these needs. A brief description of the NLM planning process is also appended. 7 NLM Programs and Recent Accomplishments MEDLARS Online Network 8,000 73 U.S. Online Users Millions 1986 Searches Online & Offline Since 1836, NLM has given health profes- sionals vital access to biomedical informa- tion. Starting as a small collection of books by the Surgeon General of the Army, NLM, then called the Armed Forces Medical Library, is today the largest medical library in the world. The Act establishing NLM as a national library under the Public Health Service in 1956 (PL84-941) formally affirmed NLM's purpose: "to assist the advancement of medical and related sciences, and to aid the dissemination and exchange of scientific and other information important to the pro- gress of medicine and to the public health." The NLM Act specifies that the Library shall disseminate bibliographic information describing the materials in the NLM collec- tion and shall make these materials available through loans or copying procedures. Together these activities provide access to both the physical collection and information about content (e.g. journal citations and abstracts) that may identify relevant portions of the collection for a particular user. NLM enhances access to the literature by distributing in a variety of formats indexed citations to articles in journals and catalog- ing records for books, serials, and audiovisual programs. MEDLARS, NLM's automated Medical Literature Analysis and Retrieval System, was first developed in 1964 to provide automated support to the production of Index Medicus and to improve its currency.3 4 5 NLM also publishes printed catalogs, the NLM Current Catalog and the NLM Audi- ovisual^ Catalog; lists of biomedical serial titles; and many indexes to articles on special subjects in biomedical journals. NLM cooperates with other organizations and publications such as the American Hospital Association, the American Journal of Nursing, and the American Dental Association to produce Hospital Literature Index, the International Nursing Index, and the Index to Dental Literature as well as smaller recurring bibliographies and publish- ed literature searches. The Library has recently issued an experimental video disk with an accompanying catalog that includes images of 1,000 pictures from the Library's prints and photographs collection. Using the NLM computer system, over 3,800 institutions and close to 1,000 individual health professionals each year conduct over 3,000,000 searches of the Library's data bases. NLM has developed a microcomputer search interface to MEDLINE (MEDLARS Online) that does not require users to have special training in search commands or in MeSH (Medical Subject Headings), the con- trolled vocabulary used to provide subject access to NLM's indexed citations and cata- loging records.6 At the same time, NLM is exploring ways to modify the automated in- dexes to MEDLINE and enhance the synonym structure in MeSH to help health profes- sionals search the data bases more effectively. In order to expand the capability of U.S. medical libraries to provide service to health professionals regardless of their geographic location, NLM established the RML Network in the late 1960's. NLM currently contracts with seven major medical libraries in dif- ferent areas of the United States to develop and coordinate programs for their regions that give health professionals easier access to biomedical information.7 8 One of the important services of the RML Network is document location and delivery. If locally available collections cannot meet a reader's needs, the Network is organized to quickly refer and fill requests for documents from regional resource libraries. Over 2 mil- lion requests for documents are filled annu- ally by health sciences libraries throughout the United States. NLM serves as the ulti- mate backup resource for the network, filling over 100,000 interlibrary requests a year for documents that no other collection holds, mostly photocopies of journal articles. Net- work participants, including NLM, adhere to standards for turnaround time for processing requests. National maximum charges assure that reasonably priced service is available to health professionals whatever their geo- graphic location. The development of MEDLARS III, the next generation of NLM's automated system, will provide significant improvements in NLM's online services by providing single-search access to all categories of records via a sophisticated user-friendly query language and greater variety and flexibility in output formats. Both DOCLINE (an automated document request and referral system)8 and NLM's online public catalog will display data about the location of specific volumes and issues in the NLM collection. On-site requests for books or periodicals will be routed automatically to the appropriate stack locations within NLM. Improved statistical capabilities will permit more detailed ana- lyses concerning the usage of the medical literature. NLM Public Services FY 1975-1986 (In Thousands) 75 80 1986 Reader Requests Filled 20 75 80 Reference Inquiries 1986 MEDLARS Usage Purpose of Search Percentage of search purpose for each profession 6.9 Management 2.0 Regulation 2.6 Unknown 9 A Vision of the Future The following scenarios represent the Panel's views on what biomedical research, health professional education, health-care delivery and administration, and libraries will be like by the year 2006: Research Biomedical research will be limited to some degree by the same economic constraints that will be shared by all sectors of society, but new discoveries in basic science and clin- ical medicine will continue nevertheless. The pace and volume of research will be even greater, in part spurred by routine use of data bases and frequent communication among scientists via electronic systems. The trend toward a smaller percentage of M.D.s entering research careers will continue, but close links between clinical settings and the research community must and will be main- tained. Through advanced technology, more tests and procedures will be possible, but eco- nomic constraints will permit only some of the fruits of research to be applied. The necessity for increasingly sophisticated and cost-effective choice making will require the use of equally sophisticated diagnostic and decision-making models. Investigators in the health sciences will require greater access to more information than ever before. Data from multiple disciplines will be more accessible because of the development of computerized links among various scientific vocabularies. Thus, the investigations in agriculture, physics, chemistry, psychology, and veterinary medicine will be of potential interest and possible utility in, and more accessible to, human health-care research. The key challenges will be how to (1) index such data, (2) provide for its rapid acquisition in a for- mat tailored to the user, and (3) provide qual- ity filtering of data. Some clinical studies will involve researchers in geographically sepa- rated sites. They will share their results as those results develop, and electronic commu- nication will speed the exchange of clinical questions and answers. Education Based on current research about learning styles, new methods will be available to give learners rapid access to the massive volume of known "facts" and data. Conclusions gathered from research on differences in cog- nitive learning styles will allow users to access data in a format most appropriate to their learning ability—visual, auditory, and so forth. Instruction will be offered in multi- ple formats. The educator's role will be to teach students how to apply such facts and data to identify and solve clinical problems. The effective clinician, or other health-care worker, will understand and use such prob- lem solving principles. They will not attempt to rely on human memory for information, but will use electronically stored data bases. The purpose of students' meetings with teachers will not be to dispense and absorb information, but to achieve higher-level cog- nitive objectives and to meet goals in the affective and psychomotor domains. Facts and memory will be less important. Values clarification, appropriate affective behavior, the ability to understand complex conceptual relations, and the development of clinical judgment and clinical skills through experience will be the focus of faculty-student interactions. As knowledge expands exponentially, and technologies change with increasing rapidity, lifelong learning must be a commitment of every health professional. Clinical skills will be regularly updated. Humanistic qualities and ethical principles must be steadily rein- forced. Education will be a continuous pro- cess and a task never completed. Students at every level will be able to tailor the education "package" to their own needs, to access this material at a time and place of their own choosing, and to review, at will, any segment to improve comprehension and understand- ing. All types of media support will be used, including animation and extensive interac- tion, so as to portray and then reinforce con- cepts effectively. Professional meetings will continue to be important arenas in which the most up-to- date knowledge in a field is shared, critiqued, and disseminated to practitioners, but the form will be different. Health-care profes- sionals will continue to meet, but increasingly by electronic means; the need to travel to educational events will decrease. While the number of persons prepared in the traditional fields of health-care delivery (such as physicians and nurses) will continue to decline as admissions are further restricted, the actual number of people prepared in a much broader range of health fields (many as yet unknown to us) will be much larger. Strict standards of knowledge and competence will be continued for the traditional fields and established in the emerging fields as well. Since those health professionals who work in clinical settings will function more as part of a health-care team than is the case now, management and leadership skills that foster interdisciplinary efforts will be important. Teaching will occur in acute, ambulatory, and community settings; teachers will have received special instruction in the teach- ing/learning process, and they will be selected and evaluated in terms of their teaching abil- ities. The professional status of teaching will have been elevated. 11 Medical Practice Medical practice will have a three-pronged configuration: acute in-patient care; ambula- tory/community care; and health promotion and illness prevention. A relatively small population of patients—those who are severely or acutely ill—will receive the special care of a hospital. Hospitals will be smaller as the current trend continues toward concen- trating expensive equipment and highly trained staff. Each will focus its interests on relatively few areas, but collectively, hospitals will offer a broad spectrum of illness care. Thus, a patient will be sent to the hospital that has a special focus on his or her condi- tion. The hospital personnel will be fuUtime, expert in the requisite technology, and trained as specialists according to their par- ticular tasks. As knowledge and technology advance, a strong program of continuing edu- cation will be essential for personnel. They will receive their continuing education through a number of different educational formats based on their cognitive learning styles. Much clinical investigation will con- tinue in such settings, but few students will receive their basic clinical education there. The bulk of practitioners will work in am- bulatory and community settings that will serve the health-care needs of the larger population of patients. There will be elec- tronic linkage and transmission of all data about the patient. In both the acute care and ambulatory/community settings, as patients present clinical problems not readily solved, the health-care team members will be able to access data to address such problems, and to do so during the patient visit. When needed, consultative services will be available to the practitioner on a person-to-person basis, whether on-site or distant. Because varying levels and types of care are offered in a varie- ty of settings, the ability to provide timely and accurate information about the patient's health to experts anywhere will insure good health care. Health promotion and illness prevention will become increasingly important, and educa- tion of the public will assume a larger role in the activities of health-care professionals. The public will have more health-care choices to make, and will therefore need more health in- formation. Many health-care providers will be specialists or subspecialists, and most of their profes- sional time will be spent in delivery of these skills and services. There will be intense com- petition among health-care providers due to both personnel surpluses (at least for the next 10 to 15 years) and the reallocation of health- care activities among the various professional groups. While both of these trends may seem threatening to the health professionals of to- day, the outcomes may well be positive for the patient. Every health-care professional's actions and decisions will be recorded and compared to established standards of prac- tice. Deviation from standards of practice will result in a continuing education program specifically targeted to the health practi- tioner's deficiencies. Because of the over- supply of physicians, those persons requiring additional education will be glad to engage in the educational effort, as they will realize its advantage for their own careers. 12 Administration A more sophisticated management system will control the operation of health care—its facilities, personnel, education, accreditation, and certification. Delineation of privileges will continue to be critical and difficult to manage. The health-care system will be in- creasingly diverse—both in terms of geography and nature of services. Therefore, "mega-administration" structures and pro- cesses will be necessary to ensure the fiscal survival and functional integrity of the system elements. Library and Information Science In 2006, the "library" will be very different from the present day library as a physical entity, and will provide a home base for many services, products, and functions that may or may not be located in a central physi- cal location. Despite the highly technical nature of information transfer, there will also be a need for a "high touch" setting for information access. The physical space of the future library will contain a variety of mul- timedia formats, from printed materials to video disks, from three-dimensional models to computer terminals, and from holography to leisure reading. The resources will be linked by a coordinated system of human and electronic networks. The library will be both the hub of some networks and the gateway to others. The people who compose this informa- tion/communication/library system will be so- called "knowledge professionals," either specialists in a particular field (medicine, pharmacy, nursing, etc.) or process specialists (educators, instructional designers, librarians, information and communications people, etc.). These knowledge professionals will pro- vide "hard copy" material (on location) or the "soft" material (that which is located elsewhere) to users, assist with research, and offer technical consultation on and educate for lifelong learning skills. The knowledge professional may also help reformat and interpret the information for the user. The content or process specialists will offer value- added services within their institutions and beyond. Health professionals will be well aware of the important role that information plays in all aspects of the health-care system. Books, journals, films, video disks, slides, models, data bases, archives, and expert sys- tems will continue to form the hub of infor- mation available in the library information system. There will always remain a physical collection that records the current status of medical research and its historical roots. 13 Major Issues and Future Directions The scenarios created for the year 2006 sug- gest a number of opportunities for NLM. These opportunities fall within two major objectives—maintaining and improving the infrastructure for information transfer and providing support for users of health infor- mation. This report provides a descriptive model for each major objective. These models represent a conceptualization of resource cost/benefit and market analysis, which can serve as a structure for decisions. Thus, NLM's strategic position enables it to have direct influence on the infrastructure, standards, access, user population, and so forth. However, with a finite amount of resources, the models should help support resource allocation decisions for NLM. Addi- tionally, this report presents some issues and recommendations about access to health information internationally. Maintaining and Improving Access Through the Infrastructure of Organizational Units For the purpose of this report, infrastructure means all the components of a system by which information flows from source to user. Infrastructure for our purposes could include technology, people, the current RML Net- work, telecommunications systems, etc. The proposed infrastructure model for infor- mation access has a number of distinctive features. Principally, it will be a network link- ing services and sources through a number of "hubs" or "gateways" at various points, which allow access to ever-widening sources of information. The network will function at local, regional, national, or international levels, and ad hoc linkages could be estab- lished for specific purposes and discarded or changed as needed. The infrastructure will have an established set of standards that will enable the user to tap into various points within the system. It will also require a vari- ety of monitoring devices that control access, i.e. fee for service, membership, or policies. In 2006, access to biomedical information should meet these requirements: ■ Information should be acquired, organized and made available so that the user can select the most appropriate information needed in the most con- venient form and location. ■ Information access should appear as a one- step process from the user's perspective. ■ Information should be supplied directly to the user. ■ Access should not require complicated technical training; however, conceptual training in the principles of information management will be required. ■ Access should be cost-effective for the user. ■ Information should be accessible to all who need it regardless of their ability to pay. ■ Access to information should be accom- plished via pathways employing the most up-to-date and effective computer and telecommunications technology available. NLM is in a strategic position to provide infrastructure components for the medical information network in the United States and possibly the world. Several of the significant areas in which NLM can contribute are qual- ity, standards, and access—both physical and intellectual. NLM is also in an excellent posi- tion to establish national standards and influ- ence international standards for information organization, storage, and distribution. For example, in the access area, the current RML Network provides one potential path for phys- ical access and information services through- out the country. Private and not-for-profit sys- tems will provide other paths for access to information. The intellectual access is provided by the indexing and abstracting tools as well as the various data bases and refer- ence services, which in turn allow access to the information and/or the physical collections. 15 Linkages and Networks Regional Medical Library (RML) Program The support system for disseminating bio- medical information will consist of a variety of multitiered networks. Twenty years ago, the mechanisms for obtaining access to biblio- graphic and textual information and data were limited. Thus, the hierarchical RML Network developed by NLM starting in the late 1960's established repositories of resources, created bibliographic access sys- tems for learning about those resources, and defined protocols and policies for obtaining resources as needed. The mission of the RML Program is to provide health science practitioners, investigators, educators, and administrators in the United States with timely, convenient access to health-care and biomedical information resources. In 1986, the RML Program is one of the main components of the NLM infrastructure. Completed in 1970, the Network has been a major instrument through which NLM has carried out its mission. Through a national ILL (Interlibrary Loan) system managed by the RML Network, NLM provides health pro- fessionals with access to the actual documents of medical literature. The RMLs have brought NLM Acquire 25,000 Journals Reader and Reference Services Index and Catalog, Publish Document Delivery Build ONLINE Data Bases Central Computer Facility Grants to Libraries Audiovisual Programs Research and Development Training 16 Resource Libraries Acquire, 2,000-3,000 Journals Reader and Reference Services Document Delivery ONLINE Services Engage in Consortia the technologies of computerized searching to basic health science libraries and given health professionals access to medical litera- ture. NLM currently has contracts with seven major medical libraries throughout the United States to develop and coordinate regional programs that improve access to bio- medical information by health professionals. Panel Two recognizes that the pattern of referrals to larger collections may not be the most effective model for locating and obtain- ing biomedical information in the next 20 years. An updated regional network is needed that supports individual access to informa- tion through a variety of sources at multiple levels. For example, at one level, an individual can maintain personal data bases including raw data, synthesized data, bibliographic and textual data, and anything else deemed use- ful by that individual. At another level, the same individual would have access to exter- nal data bases that could contain biblio- graphic or full-text information, synthesized information in knowledge data bases, expert systems to support decision-making processes, and many other data bases of varying degrees of complexity. While the current DOCLINE system for ILL provides health professionals with access to RMLs Acquire, 3,000-5,000 Journals Reader and Reference Services Document Delivery ONLINE Coordination Training Encourage Consortia Local Libraries Acquire, 100-200 Journals Reader and Reference Services Document Delivery ONLINE Services Engage in Consortia biomedical literature, further links should be developed to provide relevant documents in any field required. These documents may be in subject areas related to biomedicine, e.g. biology, chemistry, sociology, or in support areas such as economics, facilities manage- ment, or statistics. Currently, this literature may only be available through other networks not easily accessed by the health professional. Integrated Academic Information Management Systems (IAIMS) In 1983, NLM began funding planning sites for the development of IAIMS prototypes. IAIMS refers to the basic principles of link- ing academic information (library-based infor- mation, information from specific data bases, and so forth) to specific purpose information (i.e. clinical information on specific patients). Each experimental site was required to develop a planning process for implementing this idea into the academic health center. After the planning process, centers were encouraged to apply for a model test grant in order to evaluate the results of the planning process. Although during the next two decades, infor- mation resources will increasingly be availa- ble in electronic form, much valuable infor- mation will still exist only in print. Thus, the physical repositories will be a vital part of the access model. Although libraries will be the primary site for these ever-growing repositories, libraries will continue to expand their service roles as well. Despite the increased availability of computer-based sup- port for individual health professionals, not all such individuals will feel equally comfort- able using these systems. These individuals, others who lack computer access, and still others who seek assistance in developing more sophisticated information-handling skills, will depend on libraries in the biomed- ical communications network to offer relevant support services. Reference Referral In addition to offering these two established programs, NLM is in a unique position to provide national reference referral services. Building on knowledge bases consisting of local, regional, national, and international resources that exist in a range of formats from traditional paper to electronic form, and combining this with expertise in infor- mation transfer and dissemination, NLM could provide linkages to bring requesting individuals and institutions together with an appropriate source of response in an effec- tive, timely, and cost-efficient manner. NLM Centers of Excellence Program Support University X Information Science Academic Unit Unified Medical Language System IAIMS Communication Programs ■ IAIMS Electronic Network IAIMS Medical Prototype Systems Integrated Academic Information Management System 17 Summary An updated model for information access should reflect the variety of information sup- pliers. These will include individuals, aca- demic and other institutions, commercial and nonprofit organizations, vendors, libraries of all types represented both as distinct organi- zations and as members of various types of networks and utilities (some including other types of libraries), and perhaps other sup- pliers, yet to be identified. The model should be flexible, allowing an individual to directly perform his or her own information search or relegate it to an intermediary. The updated model will identify pathways using the most current computer and telecommunications technology available to link an information requestor with the infor- mation source. The model will also employ other electronic networks as necessary to speed the information transfer process. NLM should ensure that an infrastructure is built that accommodates these several levels of information delivery and their related technologies. The continuing transition from the simple hierarchical infrastructure to a multidirectional one will be gradual. The IAIMS effort, the RML Network, and reference referral are logical mechanisms for assuring that the transition occurs as easily as possible, and also for assuring that health profes- sionals are brought into the infrastructure as both seekers and providers of information. Standards The networking infrastructure envisioned can be successful only if appropriate standards are established and broadly accepted in the biomedical information world. Only then can internodal transfer of data and knowledge be efficient, reliable, speedy, and effortless. Such standards must apply to health infor- mation products, systems and services, health facilities, and personnel. NLM will be in a position to set standards in some areas, and to support the efforts of national and interna- tional organizations in others. Standards serve the purposes of establishing qualitative or quantitative measures of value, as well as providing desirable uniformity and compara- bility in performance. Bibliographic Standards Health investigators, educators, and practi- tioners will have an increasing concern for the quality of the information they use, as more information from a wider variety of traditional and nontraditional sources becomes available. The issue of how to estab- lish valid quality filters at an intermediary level is likely to be addressed at many levels, including library/information specialists, professional health and information services organizations, journal publishers, and other data base creators. At NLM, some quality filtering occurs now through the process of selecting publications for indexing. This process creates a de facto standard. Since NLM's standards serve as a model for data base creators worldwide, NLM should strive to maintain the standards it has created for bibliographic information. Apart from the quality control issue, the transferability of data base contents depends upon the adoption of standard nomenclature that will link terms commonly used by vari- ous health-related disciplines and sectors. NLM should assume a leadership role in the development and application of a Unified Medical Language System. The Unified Med- ical Language System is one way to develop standards for data base content and format, whether the data bases be bibliographic, fac- tual, interpretive, raw research data, or patient records. Over the past two decades, NLM has sup- ported appropriate standards of quality for medical libraries, ranging from small hospital libraries to major academic medical center facilities, collections, and services. Most recently, NLM has supported the develop- ment of new guidelines for excellence in aca- demic health sciences libraries, and has encouraged such libraries to become leaders in designing IAIMS on their campuses. NLM should continue to support the development and application of standards for libraries as a major point of access in health information systems. Content and Systems Standards At the same time, the new means of access call for new standards of education for infor- mation management on the part of health professionals and information specialists. NLM should support the development of training programs that will raise the level of "literacy" in medical information systems to a higher standard. NLM should be aware of the potential for expert systems and clinical decision- assistance systems to enhance high standards of health-care practice. The ability to track a physician's performance, or to determine lia- bility based partially on use of an available information system, may be commonplace by 2006. NLM's role is to assure the develop- ment of high-quality information support sys- tems that can help achieve the highest stan- dard of health care for the Nation. Panel Two believes that at the present time, preservation of and access to the raw data upon which biomedical research conclusions depend is a neglected area of science that should be of concern to the research commu- nity. A great deal of data is effectively unavailable to the scientific community beyond the investigator who collects and ana- lyzes the data. No institution has responsibil- ity for maintaining files of primary data and making them available to other scientists. Panel Two believes that in 2006 researchers in the health sciences will require access to bodies of information that they themselves did not or cannot collect and will also need to access data from other disciplines. Investi- gations in agriculture, physics, chemistry, psy- chology, and veterinary medicine, as well as the biological sciences are of potential interest and possible utility in human health- care research. As investigators become more interdependent, it will be increasingly impor- tant for some centralized agency to assure the availability of primary research data and appendiceal files. By the year 2006, it will be possible to make this data more easily and readily available, leading to an increased activity in secondary analysis. En route to this end, there are many fundamental ques- tions of quality, standardization, and documentation to be addressed. Accordingly, Panel Two recommends that the leadership of NLM consider the issues of preserving and disseminating raw data and bring them to the attention of the concerned community. These issues might include guidelines appropriate for acquiring and storing research data, along with defined routes for users to access these data. 19 Nomenclature Standards For decades, a fundamental obstacle to wide- spread adoption of computer-based medical information systems has been the absence of standard vocabulary, terminology, definition, and criteria for recording the results of bio- medical research, the events of clinical patient care, and the managerial and busi- ness transactions of hospitals. However, there is the worldwide adoption of the NLM MED- LARS system for access to bibliographic cita- tions of the scientific literature, based on the MeSH thesaurus and indexing system. The breadth and depth of NLM's MED- LARS experience qualifies NLM for a leader- ship role among interested American health- care constituencies in designing and execut- ing a computer-based system for linking terms in various health-related vocabularies. This vocabulary system would be the founda- tion on which can be built the effective integration of information systems in the library, the clinic and hospital, the classroom, and the administrative center. With adequate additional funding for this program, some system parts would be availa- ble for testing within the first five years, and by 2006 the Unified Medical Language Sys- tem will have been developed and will be in widespread use throughout the health-care field. Technology It was the computer that made possible the speedier and more flexible search of biomedicine's ever-expanding data base; and it was the earth-orbiting satellites that ena- bled the world-wide transmission of the com- puter's signals. The future will undoubtedly involve NLM in the development of new dis- tribution methods. These may include high- speed satellite links; distribution of biblio- graphic data bases on compact disk; new cataloging methods such as workstations that semi-automated indexing; pre-indexing by authors or publishers at the time of composi- tion or issuance; and more sophisticated data base tools like full-text document retrieval and expert system data base management. Computer-Based Technology One new technology, parallel processing, can simplify information searching and will be of even more use with storage and retrieval of full-text documents. At the frontier of com- puter technology is the parallel processing used by special text-searching hardware to identify complex patterns in freely formatted text files. Three commercially available devices currently search free text at rates of 1, 2 and 10 million characters per second for complex patterns up to 10,000 characters in length, while allowing a considerable degree of arbitrary mismatch to the pattern (to accommodate misspellings, foreign spellings, etc.). These devices should be evaluated for use in direct serial searches that are used during cataloging, indexing, and during data base retrieval requests. The amount of medical information available in full-text machine-readable form will be growing exponentially during the next 20 years. This massive increase in publications that can be processed immediately by com- puter will present both a great problem and, at the same time, a great opportunity. On the one hand, the total amount of new informa- tion can be overwhelming; on the other hand, one can envision an automated proces- sor that indexes and catalogs journal articles. By establishing standards through relation- ships with publishers, it should be possible to obtain computer-readable versions of titles, authors, and abstracts in NLM-standardized formats even before obtaining the manuscripts themselves. Incentives for pub- lisher participation could include earlier inclusion of a journal's contents in NLM's data bases. Indexing and cataloging of computer- readable material could be greatly facilitated through rapid text-searching technology that relies on parallel serial search machines. Such rapid text-searching machines could be coupled with expert systems whose rules could identify familiar concepts and discover novel ones discussed in the paper. As heuris- tics for recognizing novel concepts in manuscripts are developed, the "technologi- cal" cataloging advisor could become a "cataloging expert." Development of expert- system driven text-analysis systems would take an immense burden off human catalogers and might be the only feasible way to keep up with the growing overload of information. The development of such expert systems should be undertaken in a carefully organized, highly focused manner, attempt- ing to capture the expertise of some of the better human catalogers in a relatively nar- row field. Once developed, the methods developed for generalizing the application to other fields will be much simpler. In particu- lar, the development of a Unified Medical Language System would be greatly aided by the use of both the rapid text-searching tools and expert systems. 21 Computer-Assisted Technology Many technological devices today are not computers, but receive assistance for opera- tions from a computer. The information infrastructure, for example, has several computer-assisted components, such as up-to- date telecommunications systems that allow connection with multiple data bases and elec- tronics and microwave systems that rely on satellites to transfer and retrieve information. Summary In order for NLM to be a leader in applying each of these technologies, Panel Two pro- poses that NLM evaluate their use in at least one project. The project would be aimed at solving at least one of NLM's own problems. One good example is the effort of NLM to record photographic documents on compact disk. Another example is current collabora- tion with an author and a publisher to pre- pare a completely cross-indexed and full-text searchable version of an exemplary reference work on a video disk. Texts heavily depen- dent on graphic materials in fields such as dermatology or radiology are ideal test mate- rial for trials. Such experiments could set de facto standards for access to a wide variety of full-text documents with a minimal invest- ment from NLM. Supporting Access to the Infrastructure by Users of Health Information Users: A Model Using the projection of the health-care and library/information science environments in the year 2006, Panel Two developed a model to depict the primary dimensions and ele- ments that shape NLM's role in providing access to health information. This model is presented in the figure on the opposite page. It has three primary compo- nents: (1) users of health information, (2) health information sources ordered according to the detail of the content, and (3) NLM's level of responsibility as it relates to the delivery of these types of information to iden- tified users. Users Potential users of NLM's health information resources fall into four categories: (1) health science and health-related libraries, (2) health science researchers, practitioners, educators, and students, (3) agencies and organizations that interpret and report health information and/or set health-related policy, and (4) the general public, including teachers and stu- dents. In general, the detail of health infor- mation content required by each of the user categories varies in a relatively predictable fashion. That is, researchers require informa- tion at its most detailed level, while users in the "public" category require information that has been condensed from a variety of sources. Timeliness is another significant variable. For example, among researchers and health practitioners, timeliness of information is critical; while timeliness is certainly desira- ble, it may not be as critical for other user groups. Users Libraries Health Sciences Libraries Other Libraries Health Professionals H.S. Researchers H.S. Practitioners H.S. Educators H.S. Students Information Intermediaries Professional Organizations Science Writers Media & Health Info. Del. Public & Priv. Decision Makers Public Teachers and Students Lay Public Raw Data / Data \ / Full Text \ / Synthesized literature \ / (e.g. review articles, texts, \ / knowledge bases) \ *V / Abstracts \ / Bibliographic Access (e.g. Medline, Catline) // Consumer Health Information Products (e.g. film, TV, journals, pamphlets, etc.) V Popular Magazines and Media A Model for Delineating the 2006 Directive Processed Information Health Information Sources In the model, health information is depicted as a triangle. At the apex lies the most detailed level of content—the raw data that result from research and clinical reports. The remaining information sources are catego- rized and arranged in descending level of detail and, correspondingly, increased amount of synthesis or interpretation, ending with the popular health-related literature. The information needs of the user groups at the top of that listing will, in general, correspond to the level of content detail that appears at the top of the triangle. However, it should be noted that the divisions between content detail levels are light lines, not heavy ones, which is intended to imply that any user may move up and down through the content detail levels as his/her varying health- information needs dictate. The choice of an upright triangle to "con- tain" the levels of health information sources has a number of graphic purposes. The width of the triangle is an indicator of the actual numbers of users—both individual and aggregate—that are likely to make use of the information at each level of complexity. Thus, while relatively few individuals— primarily researchers and practitioners—need raw data and full text, the general public has a growing awareness of and need for health information available in the media and popu- lar press. Second, the width of the triangle also is meant to convey broadly the actual volume of information that is available to users at the varying levels of content detail. Finally, the width of the triangle is an indica- tor of the number and variety of sources from which the information can be obtained. 23 Level of NLM Responsibility The level of responsibility that NLM should assume for activities that support the infor- mation needs of users' constituencies (both current and recommended) is the third pri- mary component of the Panel Two users' model. This element is depicted as an inverted triangle superimposed over the health information triangle. Where the trian- gle is widest (at the top of the figure), it is expected that primary support of the related information-management activities would be provided by NLM, and that the expenditure of resources would reflect the commitment accordingly. Near the lower tip of the inverted level of responsibility triangle it is expected that, while NLM fully recognizes that there are numerous users in need of less complex information, fewer NLM resources would be directed toward meeting those needs. For example, the lower end of the level of responsibility triangle extends into the information level containing sources in the popular press. Opportunities to Improve Access for Users Decision-Support Systems Panel Two's vision of the future includes a variety of decision-support mechanisms, such as expert systems that would improve or ease access to health and biomedical information at NLM. An area of great importance to NLM is the application of knowledge-based systems to the problems of simplifying access to biblio- graphic data. Knowledge-based systems sup- port decision making by linking technology and content. For example, the current MED- LARS index of the literature serves as a con- ceptual framework for the entire body of published medical literature. This framework contains both syntactic and lexical knowledge that is best represented in a knowledge-based system. Hence, putting MeSH terms into a knowledge base would be a relatively straightforward task. The knowledge-based system would maintain consistency and would perform all the cross checks necessary to maintain a proper hierarchical relationship between all the terms in a search and their attributes. In addition to advocating the use of expert systems for cataloging and indexing, NLM should consider giving strong support to groups that are using these methods to codify and represent knowledge in the bio- medical sciences. Expert systems can repre- sent, in an immediately useful form, very large amounts of information in a highly compact form that will be readily dissemi- nated. Along with the development of knowledge-based expert systems, methods for reviewing and judging the quality of the information they contain will have to evolve. That task should be left to the usual peer review methods, such as editorial boards or ad hoc reviews. Expert-system technology would also help index and formulate queries of the biblio- graphic data bases. These two processes, indexing and querying, are complementary, and both require large amounts of expertise and trained personnel for efficient use of the current data base. The complementary nature of these two processes suggests that it would be most profitable to prepare an expert sys- tem that could automatically index routine publications and that could serve as an expert adviser for other, less routine publica- tions containing novel concepts. Such expert systems for indexing publications will not be widely used until most publications are avail- able in computer-readable form. Lifelong Learning By 2006, it will be essential for all health sciences professionals to have the information-seeking skills necessary for life- long learning in a technological environment. NLM achieved the online revolution in health sciences libraries with its vision of intellec- tual access to medical literature for all health sciences professionals and its successful application of online, interactive computer technology to the searching of machine- readable data bases. Recognizing the impor- tance of training for successful implementa- tion of MEDLINE search services, NLM took the leadership in providing training programs for librarians as online searchers. Current advances in information technology provide health sciences professionals with the capability of meeting their information needs through direct access to a constantly expand- ing number of online information resources. There are hundreds of online data bases, increasing numbers of publications available online in full text, and advances in auto- mated systems to aid the practitioner in mak- ing clinical decisions. Given major advances in information technology, and concomitant rapid changes in the practice of medicine, it is most important that health sciences profes- sionals be prepared for their roles as lifelong learners. To use online information resources effectively, they must become expert in evalu- ating the limitations of information resources, retrieving information from a variety of online files, and organizing information for use. NLM shares responsibility with academic health centers and health sciences profes- sional societies for training health sciences professionals and students. This training can give new prominence to the role of the NLM- RML Network in disseminating new technol- ogies for health sciences information access and in providing educational programs on their applications. Health sciences faculty and librarians must also be trained in educa- tional methodologies and computer applica- tions, in order to take the lead in developing innovative teaching programs. Health Promotion The 2006 scenario indicates a more health- conscious consumer population, and thus a system of information to support the con- sumer needs must be in place. The American public desires information that can be useful in positively influencing each individual's own personal health outcomes and those of the family. The scope of health-related infor- mation needed includes information about previously diagnosed clinical conditions, as well as common health problems that can be prevented or controlled, such as diabetes, muscular dystrophy, use of alcohol and tobacco, and nutrition and diet. Some of this desired information now exists in formats and content intended for lay use. However, the range of subjects that has been addressed is small compared to the antici- pated demand. Further, access to this infor- mation is often difficult because the sources are largely unknown to the public. One or more agencies or institutions are needed as advocates for the public to identify and develop authoritative information sources designed for general public access, to encour- age expansion of such information to cover particular current public-information interests and needs, and to promote mechanisms for assessing the quality of such offerings. A variety of health information products that focus on prominent health problems already exists. General sources include health infor- mation offerings on radio and television. The quality of those sources varies greatly. All too frequently, the reporter or presenter lacks the appropriate background or adequate resources to present consistent, reliable health information. Other, more authoritative sources that focus specifically on health information include voluntary health organi- zations, local medical societies, medical insti- tutions, medical schools, hospitals, govern- ment agencies, and private entrepreneurs. Examples include videotapes produced jointly by the ACP (American College of Phy- sicians) and the Upjohn Corporation covering common medical problems such as pain, arthritis, and diabetes. Other examples are printed subscription health information sources that explain new advances on com- mon health conditions and problems (Mayo Health Letter, Harvard Health Letter). Some medical societies have offered videotapes explaining specific health conditions through local public libraries as a public service. At present, there is no one general source for identifying and locating available consumer health information products or services or for documenting the circumstances under which they have been produced. Unfortunately, newly produced information resources may duplicate or be of lower quality than existing material. A data base that identifies authoritative consumer health publications could reduce the cost of reproducing such information locally while enhancing the qual- ity nationally. Such a data base should iden- tify high-quality information appropriate for a variety of users. Some of these users could then provide accurate biomedical information directly to the public. Panel Two's view of 2006 predicts that health promotion and illness prevention will become increasingly important, and education of the public will assume a large role in the activi- ties of health-care professionals. The public will have more health-care options and partic- ipate more fully in their health-care deci- sions. Thus, for both health professionals and consumers, authoritative sources of informa- tion regarding health protection and health maintenance will be a significant body of literature. Assisting health-care deliverers to identify appropriate health information for consumers will be a highly valued service to the health-care industry and to American society in general. Panel Two believes that NLM should undertake a study to outline the scope of this effort. International Issues Information Transfer to Developing Countries There is a vast international need for health information. The range of information needs varies as much internationally as it does among U.S. health information users. Less- developed countries may have different needs for information or require different kinds of information than the developed world. The health information needs of developing coun- tries are enormous. Developing countries throughout the world lack adequate health information libraries, have sorely limited facilities for the training of health service personnel, and must depend on a large num- ber of field workers who may have inade- quate education. NLM is in a key position to enhance the health education and information status of these countries. NLM could provide access to information for the training of health providers, health educators, and health serv- ice administrators. Libraries in most develop- ing countries are usually very limited or nonexistent. Access to NLM or some regional equivalent could help compensate for those libraries. With ever-improving technology, information support for rural health workers could be provided. This would constitute a dramatic improvement over the kind of sup- port currently available to rural health deliv- ery. Links with NLM could provide health personnel throughout the world with consul- tation support that would otherwise not be accessible. NLM could help assess information needs and the type of communication channels necessary to deliver the information, design information systems of maximum utility, and train personnel of developing countries to establish information systems. Consideration should be given to identifying possible regional information centers linked to NLM, that would, in turn, support a group of coun- tries. This model could benefit from the suc- cessful experience NLM has had with its cur- rent national/regional library system. The following steps are proposed to imple- ment the necessary NLM linkage program with developing countries. ■ A series of international meetings to iden- tify the health information needs of devel- oping countries and to identify resources that might be available through NLM. ■ The establishment of criteria for creating NLM linkages, including telecommunica- tions infrastructure, host government sup- port requirements, personnel require- ments, etc. ■ The establishment of training programs in the design and support of health infor- mation systems. ■ Identification of sources of financial sup- port and technical assistance for the establishment of an electronic network with the Third World. While it is clear that considerable resources will be required, it should be pointed out that a number of communications satellite facili- ties and ground communication services already exist; usually these are owned by the government. Moreover, international donor agencies and foreign assistance programs often are interested in support- ing efforts to improve health services in developing countries. NLM is in an excel- lent position to contribute substantially to that improvement. Information Exchange Although a considerable portion of the research done abroad is published and avail- able to U.S. professionals through electronic data bases and published literature, there is still a great deal that is only available in more local data bases accessible through for- eign national systems and networks. Some of these files complement MEDLINE; others provide unique intellectual access to special- ized information sources, as well as biblio- graphic and factual data. NLM, with its long-established relationships with foreign national centers, is in a position to provide U.S. health-care professionals access to information and data generated outside the United States. Although the need for this information is broad-based, most health professionals would probably not have a need to access such files on a regular basis. NLM should be a source of information on the contents and features of international data bases and could provide a search service for both individuals and search intermedi- aries. NLM should also serve as a link to for- eign expertise. 27 Observations and Recommendations Getting from 1986 to the electronic world of the future will be a long evolutionary process. Progress will be achieved by building on the strengths of current programs and services as well as by providing aggressive support for the new initiatives recommended. Panel Two believes that NLM's role is to assure access to all forms of literature in order to facilitate the transition from the printed literature to electronic literature and to prevent any discontinuity in access to information during this transition. Maintain- ing continuity of cataloging information is also important so that we have a national resource and not only resources at the local level. This maintenance of the traditional role of NLM emphasizes the importance of its archival activity as the world's memory bank of the scholarly biomedical literature. NLM should protect the budget support for its traditional mission of acquiring, organizing, disseminating, and preserving the world's biomedical literature as visions of the future place greater and greater demands on scarce resources. NLM should provide health-care professionals and organizations with the services necessary to assure access to needed biomedical infor- mation. The Library may provide this access itself using the most up-to-date technologies and/or through other libraries or public and private organizations, as appropriate. The Library has the responsibility to assure that quality of access is the highest reasonably possible. The education and training of librarians/knowledge workers must be sup- ported to produce personnel capable of assuming leadership responsibility for the challenge at hand, promoting information management principles and practices, and supporting lifelong learning. NLM must continue its visionary support of basic and applied research. Its research pro- grams must lead to the establishment of the information networks, systems, and services that support and enhance the infrastructure and user services and programs. It must be recognized that the IAIMS initia- tive provides opportunities for the redefini- tion of the roles and responsibilities of libraries. Continued support for IAIMS development and research on that develop- ment ensures that the infrastructure will be grounded by sound planning, tested models, and empirical research. Projects such as IAIMS will provide information valuable in projecting the requirements and features of the future infrastructure and its users. The RML Network will provide the continuity necessary during the transitional period to make certain that NLM's programs and serv- ices maintain their present quality. During the transitional period, the RMLs will bear primary responsibility for technology and knowledge transfer. They should function to amplify NLM's products, services, and tech- nologies by disseminating them productively to resource and basic libraries. In turn, the Network participants will pass the resources of NLM on to the user. An effort must be made, however, to closely examine the RML model. NLM should continue to work with the Nation's libraries and other organizations as appropriate to establish overall principles under which information is made accessible. One such principle is the need to underscore the intrinsic worth of information apart from its economic value, which is subject to change. NLM must set de facto standards that will be followed by other information producers so that these groups and their products will be effectively linked into the infrastructure described in this report. Panel Two recommends some specific steps for NLM to take in moving toward its long- range goals. Those steps emerged from dis- cussions on the Panel's view of the year 2006 and the windows of opportunity it has identi- fied that lead to the 2006 goal. The recom- mendations section is parallel to the presen- tation in the report and is divided into two major groupings: those that support the infrastructure and those that support the user access grouping. Support for the Infrastructure (1) The RML Program, which is the major building block of NLM's current infrastruc- ture, is an invaluable tool for enhancing both direct and intellectual access in the future. NLM should use the RML Network as a vehicle for urging libraries and individuals into the new networking structure, as well as for maintaining appropriate existing systems of physical and intellectual access. NLM should evaluate the RML Network and develop a strategic plan for its future that will include the extension of the new technol- ogies for information and data base access to health professionals regardless of their geo- graphic location. (2) The current document delivery network should be linked to existing or new networks to provide comprehensive document delivery service to the individual. (3) NLM should provide national reference referral services. (4) NLM should expand its support of IAIMS planning, model development, and research efforts. A research program should be initiated by NLM to rigorously examine the various planning processes and models, frameworks, and outcomes produced by the IAIMS sites (for example, from 1983 to 1987). The goal of this research would be to deter- mine those aspects that can be generalizable to other institutions and those aspects that are functionally unique to the particular sites. (5) NLM should work cooperatively with data base producers and information services sup- pliers in the private sector to advance uses of technology that facilitate access to health sciences information. The cooperation with private and other public organizations will build a synergy for outstanding creative developments. (6) In support of building the infrastructure, NLM should support individualized access (both physical and intellectual) to needed information. NLM should encourage research projects in such areas as access to textual information via electronic means and other technology, such as telefacsimile, disks, etc., health information data base development, automated health information delivery sys- tems, medical library automation, and health information expert systems and consultation networks. Specifically, NLM should explore technologies that will facilitate the full integration of information. (7) NLM should develop knowledge-based expert systems to aid the novice user in retrieval from MEDLARS. (8) NLM should assume a leadership role in the development of a Unified Medical Lan- guage System. All elements of the health-care system in the United States should be encouraged to contribute to its development to facilitate more accurate and effective com- munication. Specifically, NLM should: ■ Obtain NIH interinstitutional support for a program initiative to create a Unified Medical Language System. ■ Seek additional funding for this research area. ■ Coordinate program efforts with the American Medical Association and other health professional associations; univer- sity investigators in academic depart- ments, including information science, computer science, and biomedical dis- ciplines; health sciences libraries; insur- ance and information industries; and other Government agencies. ■ Draw upon the resources of the Lister Hill National Center for Biomedical Communications as well as grants and contracts to stimulate a comprehensive research program with internal projects, extramural research at universities, and development work by private industries. (9) NLM should continue to exercise leader- ship in setting standards of quality through the programs it develops to accomplish its mission of access. NLM should also work with a broad representation of national and international standards organizations, profes- sional societies, public and private service organizations, and others to develop and maintain national and international standards for health information products, services, facilities, and personnel. It is envisioned that such standards will provide a beneficial pro- tective effect to the public by constituting a value system that encourages participants to strive for excellence. NLM should also recom- mend an approach to the standards required for the preservation, storage, and access to research data and appendiceal files. Support for User Access (10) NLM should undertake a study to con- sider the public's need for health informa- tion. This should be in conjunction with appropriate agencies, and the report should recommend an approach that will assure that appropriate health information is available and accessible to the public. Such an assess- ment might include, among other considera- tions, the following: ■ The scope and magnitude of current sources of consumer health information. ■ The range of subjects covered by authoritative, factual information. ■ Feasibility of bibliographic control of these information products and services. ■ Appropriate agencies or institutions that would disseminate such information to the general public. ■ What role the biomedical communica- tions network might have in disseminat- ing such information about the sources of health information. (11) NLM should encourage the establish- ment and maintenance of programs to edu- cate and train health information profes- sionals. The education programs should emphasize the focus of "integrated informa- tion" concepts. Attention should also be given to the management of information centers and the application of advances in computer technology to information dissemination. 30 References (12) NLM should provide all reasonable assistance to health-care professionals and organizations in other countries in identify- ing and gaining access to available health information sources worldwide, and in deter- mining the relative level of sophistication and general quality of such sources. NLM should also provide U.S. health-care professionals access to information and data generated outside the United States. (13) NLM should assess health professionals' need for information (how they get it, use it, evaluate it, and what impact their access to information has on the quality of care they provide). NLM should also encourage further study of the impact of electronic technology on health professional's "information processing" behavior. Summary In conclusion, Panel Two acknowledges the strong leadership role that NLM has played both directly and indirectly in influencing the high quality of information available in the health area. Panel Two encourages NLM to continue on its quest for excellence in the integration of information by continuing its commitment to developing an easily accessi- ble system for all who need information. Panel Two further encourages NLM to con- tinue the best possible services and products today and to research the services and products for the 21st century. 1 Matheson NW, Cooper JAD. Academic information in the Academic Health Sciences Center: Roles for the library in information management. / Med Educ 1982;57(Pt2):l-93. 2 Planning for Integrated Academic Information Management Systems. Proceedings of a symposium sponsored by the National Library of Medicine, October 17, 1984. Bethesda, MD: National Library of Medicine, 1985. 3 Beckelhimer MA, Cox JW, Hutchins JW, et al: The MEDLINE hardware and software. Med InffLond) 1978 Sep;3(3): 197-209. 4 McCarn DB. MEDLINE users, usage and eco- nomics. Med Inf(Lond) 1978 Sep;3(3): 177-83. 5 McCarn GH. The on-line user network: organization and working procedures. Med InffLond) 1978 Sep;3(3):211-23. 6 Thompson L. National medical library offers giant data bank to well-connected media. Sci Writers 1986 March;34(l):3-5. 7 Kasner L. The Regional Medical Library Program: a national medical information network. Sci Technol Libr 1980 Winter;l(2):43-51. 8 DOCLINE NLM News 1985 April;40(4):l-4. 31 Appendix A Medical Library Assistance Act: A Twenty-Year Review Alison Bunting, M.L.S. Biomedical Librarian Center for the Health Sciences University of California at Los Angeles Michael E. DeBakey, M.D. Chancellor and Chairman Department of Surgery Baylor College of Medicine Nicholas E. Davies, M.D. Chairman Department of Medicine Piedmont Hospital, Atlanta Judith Messerle, M.S.L.S. Director Medical Center Library St. Louis University School of Medicine Raymond Palmer, M.S.L.S. Executive Director Medical Library Association Introduction In 1965, the U.S. Congress enacted the MLA Act (Medical Library Assistance Act)—Public Law 89-291—to help health sciences libraries cope with the unprecedented expansion of biomedical knowl- edge and to insure delivery of this knowledge to health professionals, regardless of their geographic location or proximity to a developed library. The MLA Act was the direct result of the report of the Presi- dent's Commission on Heart Disease, Cancer, and Stroke, which recommended that the government recognize public information as a primary responsibil- ity and major instrument for preventing and control- ling disease. In light of the National Library of Medi- cine's 1986 planning efforts, it is appropriate to review the original intent of the MLA Act, record the progress made since 1965, and project the future needs of the nation's health sciences libraries. Part I of this report (1) reviews the information needs of health professionals before the MLA Act was enacted; (2) traces the evolution of those needs from 1965 to 1985; and, (3) projects future information needs. Part II of the report describes both developed and proposed programs that are necessary to meet those information needs now and in the future. Among the programs described are the original seven MLA Act initiatives, MEDLARS (Medical Literature Analysis and Retrieval System), and IAIMS (Integrated Academic Information Management Sys- tems). Summary of Recommendations For ready reference, the report's specific recommen- dations for future action in ten areas are summarized below. The recommendations center around the con- tinued need for a strong and effective BCN (biomedi- cal communications network). The continued growth in scientific information and the increasingly inter- disciplinary nature of scientific research require that health professionals have access to information from an overwhelming variety of sources. Thus, a BCN which provides access to this information through one gateway (to be defined as a set of simple computer commands that permits an individual to access all relevant information data bases, each of which usually must be searched by a unique search protocol) is essential to meet the information needs of tomorrow's health professionals. Health Information Needs Health professionals should be aware of and trained to use modern health information resources including bibliographic and factual data bases, hospital infor- mation systems, and expert systems. These resources should be accessible electronically, easy to use, not restricted to disciplinary lines, and accessible from multiple locations and by a variety of equipment. Medical information sciences (medical informatics) will be an increasingly important discipline. Training should be provided in health professional schools, and this specialty should be given appropriate academic recognition. 32 Given the current emphasis on individuals assuming a stronger role in their own health care, the shift in emphasis from disease treatment to prevention, and the patient's right to know, the Federal Government should develop an initiative which would support the collection, organization, evaluation, and cheap dis- semination of the widely varied lay health literature—a literature that presently is in disarray. Construction, Renovation, and Retool- ing of Health Sciences Libraries Health sciences libraries will shift their focus from being repositories of information to being switching stations for needed information. Planning must begin immediately to convert existing health sciences libraries into health information management centers and to build new ones where necessary. Technology will help health sciences libraries improve access to and delivery of information. Funding is needed to allow libraries to take advantage of techno- logical developments to meet the information needs of the health sciences community. Biomedical Information Management and Medical Informatics Training Information will be managed by individuals from vari- ous backgrounds who will need very specialized knowledge, skills, and expertise. Doctoral-level training programs should be developed or strengthened to pro- duce such information professionals. Incentives are needed to draw mid-career professionals to these pro- grams. Master's-level library and information science pro- grams should be upgraded to provide the knowledge needed to use automated libraries and data bases in biomedicine. Practicing information professionals need continuing education opportunities to upgrade their knowledge base. Distribution and Utilization of Bio- medical Knowledge Support should be continued for the compilation of all information related to a specific biomedical sub- ject in one publication. Support also should continue for original contributions to biomedical literature as it relates to the scientific, social, and cultural advancement in the health sciences. Research and Development in Bio- medical Information Management A research agenda in various aspects of biomedical information management is needed to better under- stand and fulfill the information needs of health professionals. Improvement and Expansion of the Basic Resources of Health Sciences Libraries As the biomedical knowledge base expands and infor- mation is available from disparate sources, it will be necessary for health sciences libraries to use technol- ogy to access remote information to engage in cooperative acquisitions, and to take part in preserva- tion programs to insure the availability of information on a local or regional level. In an era where some of the information needed can- not be purchased for use in a local library but must instead be accessed on a cost/use basis, it is essential for institutions to develop and implement policies regarding information access, costs, and fees. RML (Regional Medical Library) Network The RML Network will continue to use technological advances to improve and speed the delivery of infor- mation to health professionals. The RML Network must continue to link health sciences libraries and also must increase its efforts to directly serve health professionals without access to health sciences libraries. Special efforts to support information services in underserved areas should continue. Publications to Facilitate Access to the Biomedical Literature Support should continue for the publication of syn- theses and analyses of current developments in bio- medical research and practice and health sciences publications that are not commercially profitable. Support should be given to publications and develop- ment of data bases that describe and evaluate health sciences library collections, services, users, and needs. MEDLARS Expert systems (to be defined as the use of artificial intelligence through computers to make decisions that have heretofore only been possible for the human mind) to aid in evaluating, indexing, and synthesizing information for inclusion in MEDLARS should be developed. These expert systems, in combination with increased technological sophistication, will promote expanded MEDLARS data bases. Since health profes- sionals will be using the MEDLARS system directly, the search software will need to be "user friendly" to preclude the need for extensive training. The MEDLARS system should provide gateway access to other information data bases and libraries to facili- tate a retrieval of information from all relevant sources for health professionals. IAIMS Funding and testing for IAIMS models that reflect differing institutional patterns of organization should continue to promote the critical integration of infor- mation in health sciences institutions. Parti: Health Information Needs Historical Analysis John Shaw Billings foresaw the great need for access to the medical literature with far more prescience than most physicians of the late 19th century, indeed more than many people today. Billings' greatness lay in his ability to solve the problems that he encoun- tered. His solution to the medical literature access problem, when combined with several other sources of indexing and cataloging in the early 20th century (notably those maintained by the American Medical Association), served the health sciences community adequately until World War II. The modern age of medical science began during the late 1930's and 1940's with the sulfas and penicillin, new knowledge about treating the wounded (shock, renal failure, transfusions, and blood volume expanders), new vaccines and other public health advances, and the dawning of vast new technologies for studying and treating disease. Biomedical research burgeoned. Under the leadership of men such as Lister Hill, the son of an Alabama physician, whose prescience approached that of Billings, scientists were encouraged to find "the answers" to polio, heart dis- ease, stroke, cancer, and the other diseases affecting Americans. The results, in many instances, have been remarkably successful. In the early 1960's, it became clear that the vast amount of new information being developed by the biomedical community, most of it federally funded, was not reaching those people for whom it was intended: the academic community to an extent, but especially the practicing physician. Indeed, while many medical centers had developed into impressive generators of basic and clinical research, the health sciences libraries in many of these centers had been woefully neglected. Thus, paradoxically, as a greater need for health sciences information developed, espe- cially to satisfy researchers and educators, there was less and less funding for medical libraries whose specific task it was to meet these needs. A survey made in the early 1960's by the Public Health Associ- ation and the Association of American Medical Col- leges found that only 15 of 87 medical libraries had sufficient space, that more than one half of these libraries had been built before 1933, and that more than one half of the libraries were either filled to or exceeded their capacity. Clearly, something needed to be done. Things were being done at the NLM (National Library of Medicine). The new library building was started in 1958 and dedicated in 1961. Congress was relatively generous to the Library, attempting to sup- port its growth commensurate with the growth of the biomedical literature. Yet, because the final act of bio- medical research is the publication of results of the study, biomedical literature was growing exponentially. For example, between 1933 and 1963, NLM received some four million titles, while between 1836 and 1933 it received only three million. In addition, during the three years between 1958 and 1961, interlibrary loans at NLM increased 82 percent. This enormous growth in the medical literature and the inability of the existing system to handle it was studied by a number of committees in the 1950's and early 1960's. The results of these many studies are recorded in thousands of printed words. Perhaps the best one-line summary is found in the Weinberg Report. It said, "Transfer of information is an inseparable part of research and development [and all those involved] must accept responsibility for the transfer-in the same degree and spirit that they accept responsibility for research and development itself." The MLA Act was enacted to accept this responsi- bility. Health Information Needs in 1985 Unlike John Shaw Billings's very specific medical information needs in the early 1870's, the information needs of today's health scientists, who include researchers, educators, practitioners and paraprofes- sionals, are far more amorphous. Many of these peo- ple do not, until asked, realize that they do indeed have needs. Like the rower who has never used sail, or the sailor who has never used power, one often does not know one's needs until one finds that there are better ways to reach one's goal. So it is with most practitioners and paraprofessionals, and so it is likely to be with educators and researchers. Many of the problems facing the following four cate- gories of health-care personnel are common to all, yet each has its differences. Though the groupings are arbitrary, each will be treated separately. The needs of each group are discussed in the rank order of impor- tance. It must be noted that these opinions are not based on research data. Practicing Physicians Establishing and Using Modern Information Retrieval Ask perceptive practitioners what is their greatest medical information need and they will tell you it is quick, accurate, cheap answers to very specific ques- tions. Can you give amino-phyllin with cimetedine? What antibiotic is best for a penicillin-allergic elderly man with S. pneumoniae pneumonia? The quickest and cheapest method to find the answer, although not always the most accurate, is to ask a colleague. Most practitioners check their often outdated journals or texts, then call a consultant informally, or call their librarian. A good hospital librarian will find the answers to simple questions in five minutes. More complex questions may take hours or days. While these options are available to many physicians, unfortunately, there are more who do not have access to them or who fail to use those resources that are available. Clearly, a major need in 1986 is for better facilities for information retrieval, especially in the smaller rural hospitals. An even greater need, and perhaps one more difficult to resolve, is training the current generation of practitioners to use modern information-gathering sources. It is hoped that some of these people, especially the relatively large group who entered practice during the past decade, might respond well to such training. These litigious times have spawned concern about data bases that give specific, unequivocal answers to questions dealing with the human condition. All peo- ple who treat patients know that there is rarely one unequivocally correct answer to any question. Some answers consider only scientific fact, others consider economics, and still others interject regional or national bias. These problems must be addressed as data bases and expert systems for treating patients are developed. Computer Literacy Among Physicians In 1986 practicing physicians are becoming familiar with computers through two methods. First, many practitioners either now use or will use computing for their offices, for billing, patient records, and word processing. Second, many physicians are purchasing PCs (personal computers) for family use. There are enough computer users among certain groups of phy- sicians to establish a market for software packages for medical purposes—some good software and some obviously inadequate. Medical societies now are developing electronic mail systems and simple drug data bases and are putting self-assessment programs online. Computer clubs are forming. Physicians are beginning to access data bases directly, thus eliminat- ing the delays inherent in going through an inter- mediary. Systems Interface in Hospital Computers Today's practicing physicians are frustrated by their inability to collect all patient data at one terminal in their hospital or office. Hospitals with early involve- ment in computing often have one system for data processing, another for pharmacy, and yet another for their clinical laboratory. The hospital library can access NLM's MEDLINE or BRS/Saunders Col- league, but only at the library terminal and not on the wards. Many services, such as radiology, pulmo- nary function labs, and neurophysiology labs, have no computer access. Physicians must telephone or visit the department or await the typed report the follow- ing day. Frustrations from nonintegration of the hospital com- puter systems are causing community hospitals to contract with vendors for second- or third-generation computers for first-generation integrated hospital information systems. Hospitals are encountering, among other difficulties, the enormous problem of computers and computing systems being unable to communicate with each other. Teaching Information Retrieval The standard setters for the quality of the Nation's practicing physicians are the various boards that examine candidates to see if they are well trained in the science—and the art—of medicine. Today's boards realize that the ability to locate information quickly is an essential skill for all physicians. They currently are struggling with the problems of how to teach this skill in medical school and in postgraduate training programs and how to measure this ability. With the amount of information now available to practicing physicians and with the decreasing half-life of valid information about health care, clearly, all phy- sicians must become well versed in information retrieval from data bases as well as in information gathering from their patients, and in information evaluation—judging that which is valid and that which is spurious. Changing Emphasis in Medicine Competitive prepayment health-care programs are rapidly changing the method of health-care delivery in America. Most practicing physicians are no longer allowed the luxury of leisurely investigating each abnormality found in their patients. Rather, the emphasis is now on prevention, health maintenance, and health promotion. How this will change the health sciences information community remains to be seen, but it is clearly a factor to be observed carefully. Educators Concerns Shared with Practicing Physicians Many academic physicians also have some practice responsibilities, if only to make ward rounds with stu- dents and house officers. Therefore, many of the prob- lems facing the practicing physician in 1986 face the educator also: the need to obtain quick, cheap, and accurate clinical information; the familiarity with medical computing; the lack of integrated clinical information systems in their hospitals; the lack of standards by the various American medical boards for information retrieval; and the problems of pre- payment plans vis-a-vis health-care information. Establishing Programs in Medical Information Academic physicians in 1986 must face not only their own problems with information retrieval, they also must face the even larger problem of what to teach their students about health-care information. As the world enters the Information Age, it is increasingly clear that students cannot store in their brains all the information they will need to practice medicine. They must be taught how to do a few basic things well (his- tories, physicals, and certain procedures); how to rea- son clearly; then how to find the information they need. This last function is least well defined at this time. Among the approximately 130 American medical schools, fewer than 10 have advanced programs in the medical information sciences (to be called medical informatics). These programs are training the rela- tively few specialists in the field, some as M.D.s and others as Ph.D.s. Some graduates remain at their home institutions; others move quickly into industry; a few go to other academic centers to start their own programs. It is estimated that fewer than 40 medical schools have any formal training in medical infor- matics for students and house officers. The remaining schools either have ignored the problem, are making plans for the future, or are simply trying to stay alive. Improving Physical Facilities Today's library facilities in academic health sciences centers have improved considerably since 1965, yet there are still many centers with woefully inadequate facilities—little space, inaccessible books and jour- nals, and inadequate equipment. Centers that are modernizing their facilities are often calling them information centers, only one component of which is the traditional library containing books and journals. Added to this are learning centers with computer- based instruction, computer-based exams, and a remarkable array of audiovisual equipment. The Need for Medical Informatics Leaders As described so well by Matheson and Cooper,1 today there is the beginning of serious planning for an integrated information system in academic health sciences centers. These systems will include patient data, research data, educational information, and data bases of all sorts, with easy access at multiple sites. Initially, the system will be expensive, complex, and difficult to manage. There are as yet very few people capable of managing such an integrated information system, either because they lack training in the broad field of medical informatics, do not have academic qualifications, have no research background, or are not equipped by temperament to guide forcefully such a politically unstable project. Researchers Health sciences researchers sometimes face all of the problems encountered by practitioners and educators. Therefore, to varying degrees, the preceding com- ments should be considered valid for researchers also. The following comments discuss those problems that are unique to investigators, either basic or clinical, because it is these people who are concerned with answering the questions that come from practitioners and educators. Leisurely Search vs. Immediate Retrieval of Information Somewhat exclusive to researchers is the need for two distinct types of information: one that may be obtained leisurely, the other rapidly. The first is an extensive, thorough, somewhat reflective search of the literature on certain subjects. Unlike practitioners and most educators, researchers have a growing concern with studies worldwide. These searches may take weeks or months and produce voluminous, often irrelevant, material. The second type of information that has become increasingly important to researchers, especially those on the forefront of new diseases (e.g. AIDS) and new fields (e.g. recombinant DNA), is highly specific, notable for changing rapidly, often associated with a for-profit enterprise that is highly competitive. This information may be outdated within weeks; it must, therefore, be accessed very rapidly and updated daily. Much of this information is not available to the scientific community until it is patented, again a unique development of the past decade or two for the scientific community. Data Bases Beyond Medicine Biomedical research has extended into fields formerly unrelated to medicine. Whereas during the 1950s and 1960's researchers crossed boundaries within the medical specialties, now medical research has extended into such areas as engineering, physics, invertebrate zoology, sociology and other disparate fields and disciplines. A major problem in 1986 for health sciences investigators is how to access informa- tion in these fields by using the facilities and data bases that are already in place. Clearly, as the system now operates, this is not possible, to the detriment of researchers and their studies. Nurses and Paraprofessionals Nursing literature has developed over the years paral- lel to medical and other health sciences literature. Material unique to nursing is not of a quantity that constitutes a major problem in the immediate future. However, there is a vast array of technologies develop- ing in fields unknown 20 years ago (e.g., respiratory therapy, exercise physiology, and neurophysiology) as well as in nutrition, physical and occupational ther- apy, and other fields. Each of these has developed its own literature that must be available to others who need it. In addition, there are a number of parascien- tific fields that have developed literature of their own. These include traditional Chinese medicine and acupuncture, herbal medicine, and chiropractic. While the quantity is relatively small, there is the question of whether or not this literature should be indexed and cataloged. Lay Health Education Literature Many health sciences libraries are developing special collections for patients. Some hospitals use this as a marketing tool to attract both patients and physicians. These collections also can be helpful to personnel who produce local television programs for hospital educational channels. These collections usually con- tain movies and video tapes for in-service education. Unfortunately, it is often difficult to locate the mate- rial needed on a specific subject. There is also a great demand in public libraries for health information. Some public libraries do an excel- lent job in directing patrons to sound information. Others do not. It is difficult for a librarian to deter- mine what is scientifically sound, and in many instances, patrons alone decide what they wish to read. There is no national system whereby available literature is identified, reviewed for its scientific merit, cataloged, and published. History of Medicine Literature Currently, there are excellent resources for the history of medicine literature at NLM and at some of the older, larger health sciences libraries. These must be nurtured at all costs for they are the foundation on which a learned profession is built. Recommendations for the Future (1) Health professionals will continue to need quick, inexpensive, accurate answers to specific questions and will increase their direct access to this informa- tion. Literature references from online data bases will suffice until access to the information itself is readily available electronically. These electronic information systems should be "user friendly" to facilitate their use. (2) The problem of underutilization of modern infor- mation sources by a large number of physicians, espe- cially practitioners, must be addressed by those who set standards for excellence in health care in America. (3) In the development of computerized health-care data bases, the concern with litigation prompted by not using or not following the information available must be addressed. (4) Medical schools, medical societies, medical boards, and others who train physicians, nurses, and paraprofessionals must realize that the world has entered the Information Age, and train health profes- sionals to use and access information. (5) Major effort must be put into developing integrated hospital information systems that have mul- tiple access points to obtain all clinical information, search data bases, write patient care orders and nurs- ing plans, etc. Plans should be developed at the national level to link the various health-related infor- mation systems available in America. These systems must keep up with the rapid development of new medical technologies and the subsequent literature. (6) Each of our medical schools must look at its cur- rent situation in relation to medical informatics and decide its future course. Some will wish to become leaders in this new field, others simply proficient enough to teach their students well, assist their teachers and researchers, and little more. It is an expensive undertaking, yet it is essential and must not be delayed. Integrated academic medical information systems are the future. While there is no one standard that seems best for all, clearly, all must develop some system to survive. (7) There are currently few role models in medical informatics with whom students and house officers can identify. There must be more support from the universities for students wishing to specialize in medi- cal informatics. Medical school deans and vice presi- dents for health affairs must recognize the need for granting academic status to these specialists commen- surate with their training and scope of activities. (8) There is a great need in the research community for information retrieval systems that cross discipli- nary lines, so that a physician can search the engineering literature, for example, without having to go to an engineering library. (9) The health information system must continue to evaluate its position concerning certain nonstandard health-care practices such as traditional Chinese medicine and acupuncture, herbal medicine, and even chiropractic. (10) Lay health literature is in disarray because no person or group has ever had the responsibility for its organization. As self-help, health promotion, and dis- ease prevention become more prominent each year, the health information system in this country must face this problem and arrive at some answers: Who should be responsible for bringing order to this disar- ray? How much order is good and how far should it extend? How should this be financed? The Federal Government must play a major role in solving these problems. (11) The history of medicine collections and programs at NLM and other great health sciences libraries must be nurtured, for they are the foundation on which the learned profession of medicine is built. Medical history is being taught in more medical schools, and there is increasing interest in medical history and in all of the humanities in the medical curriculum. Part 2: Analysis of MLA Act Programs Assistance in New Construction and Renovation, Expansion, or Rehabilita- tion of Existing Medical Library Facilities. Status in 1965 Academic Health Sciences Libraries In 1953, the Deitrick-Berson Report, .Medical Schoob in the United States at Mid-Century, indicated a seri- ous problem in the Nation's medical school libraries. In particular, the demands resulting from the expan- sion of biomedical research activities were outstrip- ping support for current housing. Ten years later, Bloomquist2 confirmed the problem: The crowded and makeshift space in which these libraries are housed presents great ineffi- ciencies for library users and ultimately creates unwarranted expenses for the institution, just as an activity costs more to perform in inefficient quarters than it does in efficient quarters. The problem of damage to library collections should not be ignored either. The crowding, dirt, and improper heating, ventilating, and lighting in ill-housed libraries significantly diminish the life of library materials. Newer features of medi- cal libraries, such as space for storage and use of audio-visual materials, microfilm and photo- copy service, and electronic data-processing equipment, must be planned for in building renovation and construction. Again in 1965, the President's Commission on Heart Disease, Cancer, and Stroke pointed out the continu- ing strain on inadequate medical school library resources. Huang's study of medical library facilities3 found that only 19 medical school libraries were built or expanded from 1958 to 1964, all of them with space of less than 50,000 net area square feet. It was necessary to develop legislation to authorize the con- struction and renovation of medical library facilities. Congress responded with the passage of the MLA Act of 1965. Hospital Libraries In 1962, there were almost 5,500 short-term hospitals in the United States. Approximately 58 percent (3,192) were estimated to have libraries. It was more likely that a larger hospital would have a library than a smaller hospital.4 A 1957 report by the Committee on Hospital Library Architecture of the United Hospital Fund of New York noted the need for adequate library facilities. "Hospital libraries are frequently located in areas which are inaccessible, cramped and generally unsuited to the full development of library service. This is sometimes the case even with recently con- structed libraries."5 Progress in 1965-1985 Academic Health Sciences Libraries Between 1967 and 1970, eleven institutions received construction grants under the MLA Act: nine medical schools, one school of optometry, and and one school of veterinary medicine. Data suggested that the con- struction and renovation program needed to expand to meet the needs of other deserving institutions (approximately 25 to 40).6 The five-year extension of the MLA Act, which occurred in 1970, provided only modest support for medical library construction. New construction was phased out at the end of the extension, although some renovation funds were available as part of the NLM extramural grants program.7 Between 1966 and 1975, 86 medical school libraries were built, expanded, or were under construction. This period represents the greatest expansion in the history of medical school libraries. Some of this build- ing and renovation activity was supported by the MLA Act.3 Much of it sprang from institutional and other support of medical school libraries brought about by the MLA Act and the growing awareness that the handling of technical information is an integral part of science. 39 Hospital Libraries Hospital libraries received relatively few Federal con- struction and renovation dollars during this period, with support coming principally from Resource Pro- ject Grants. A 1969 analysis showed that the average hospital library occupied approximately 1,536 square feet in community hospitals. The same study noted a posi- tive relationship between the total size of the hospital and the space allotted to library activity.8 Sparked by new attention to health-care libraries via RML activi- ties and consortial development, many hospital libraries did grow during the 1970's, expanding both space and facilities. Unmet/Future Needs (1) During the last decade, the role of the health sciences library began shifting from repository of information to switching station, connecting each health professional to a large array of information systems—from book, to media, to remote data bases. To continue moving in this new technological direc- tion, planning must begin immediately to convert existing health sciences libraries into health informa- tion management centers, and to build new ones where necessary. This planning must consider that the traditional structural function of library facilities will be altered dramatically by sophisticated commu- nications technology. Storage space for books and journals in paper and other formats will be reduced. In turn, space will have to be expanded or converted for communications technology (e.g. computers, satel- lite and television transmission capabilities, etc.), classrooms, and auditoriums. (2) Procedures must be developed and implemented that allow existing health sciences libraries to be retooled in order to install appropriate systems to meet the information needs of the health sciences community. This will require conversion of material to electronic form, to optical video disks and display devices, and installation of professional workstations and data base management systems. (3) Despite the growth and recognition of the hospital library, many are still in serious need of physical space renovation, particularly as they move further into an information management environment. Hospi- tal library facilities require alterations to house new computer hardware and wiring components with link- ages to other parts of the hospital's data systems. Space requirements for end user (to be defined as an individual who does his or her own data base search- ing without the assistance of an information special- ist) training and service will increase the demand for square footage in the hospital library as a primary access point for information. Assistance in the Training of Medical Librarians and Other Information Specialists in the Health Sciences. Status in 1965 The Lister Hill report to Congress in 1965 identified a "critical shortage of professional personnel trained to meet the special needs of health science libraries and the medical community they serve." This short- age of medical librarians was, in part, underscored by the lack of special programs for medical librarianship. With only two medical librarianship programs availa- ble prior to 1966, too few professionals were entering the field, and library services in the medical field were understaffed and losing staff. The MLA Act Training Grants were designed to address this problem. Progress in 1965-1985 From 1966 to 1971, a total of 115 training grants were awarded to medical librarians. Several studies and an NLM external evaluation of the program indicated that by 1971, the shortage of medical librarians had been largely eliminated. In 1972, eight masters pro- grams in medical librarianship were operational and the eight internship programs developed through grant funding had been reduced to four. Not all graduates in the field were able to find employment and many were hired into nonlibrary settings. Recom- mendations were made to five new focuses to the training programs. Since 1972, the thrust in training has been in apply- ing computer technology in the health sciences. The trainees are health scientists, educators, and others with biomedical experience. Through 1984, 109 awards had trained 345 individuals in this area. Unmet/Future Needs (1) Availability of qualified library graduates and positions in the field. The original impetus for the Training Grant program was the "critical shortage of medical librarians." While that need was filled in terms of numbers by 1971, several changes in the broader environment have made this need apparent again. Of the 16 programs (8 masters, 8 internships) in medi- cal librarianship designed between 1966 and 1971, a few masters programs still offer one or two medical library courses. Internships were altered in 1972 to focus on computers in medicine with trainers prin- cipally health scientists, educators, and others with biomedical experience. As a result, newly employed library graduates are fully dependent on an intensive internal orientation and often must participate in external training programs in order to be able to function as professionals in the field. Finally, the number of available positions appears to be increasing as health science libraries expand and reshape their efforts to provide access to information. This increase in options for the graduate librarian makes recruitment of "qualified" librarians more difficult. (2) Upgraded Knowledge Base Required— Master's Level The current library school graduate has been edu- cated in a program that does not provide an adequate knowledge base for practice in today's changing health sciences library. These programs again need examination and upgrading to provide baseline knowl- edge for entry-level graduates who are equipped to practice in automated libraries and search data bases in biomedicine. (3) New Knowledge Bases—Post Masters, Doc- toral, Predoctoral The future demands health science librarians whose knowledge, skills, and expertise transcend what can be taught in a masters-level program. Attention should focus on major research agendas in the organization, accessibility, and transfer of knowledge. There is a need to develop bright individuals, credentialing them through doctoral programs, to lead in designing and implementing integrated information programs. New doctoral programs addressing these needs will draw candidates from a broad spectrum of back- grounds. Initially, however, they should be structured to attract current library administrative personnel and middle management so that the development agenda will be quick and effective. Incentives should draw midcareer individuals to the problems. (4) Continuing Education Opportunities A large number of today's medical librarians have been forced to learn about new technology on the job. Such learning often has neglected an understanding of the basic theoretical underpinnings of the technology. Additionally, while many middle managers and library administrators may have had a course in library administration and management, theory and tech- niques have seen major changes in the last 20 years. Most are hard pressed to design new and improved systems since they lack background in technological advancements, systems theory, and management tech- niques. Individual efforts to stay current through reading, collegial exchange, and periodic training pro- grams have been inadequate. Formal continuing education for practicing librarians should advance the dissemination of information to health-care researchers, practitioners, and students as the newly trained individuals become more effective at solving information management problems. (5) Health Professional Training Since 1972, NLM has offered training grants in the application of computer technology to the health sciences. Trainees have been health scientists, educa- tors, and others with biomedical experience. The need for such training continues. The publication of the GPEP report9 underscored the importance of increas- ing the understanding of information management among practitioners, researchers, and students. As access to computer technology grows and information mushrooms, management of information becomes more and more critical to daily functioning for the health practitioner. Training must be targeted to the health science stu- dent so that skills are learned early and become part of everyday life. Attention also must be paid to train- ing trainers—educators who can teach theory, princi- ples, and practice of information organization and access. Beyond that, some continuing training assistance is necessary for potential leaders in health care who can transform knowledge into useful products or who can seek new knowledge in biomedi- cal information management. Assistance in Compiling Existing and Creating Additional Material That Dis- seminates Information on Scientific, Social, and Cultural Advancements in the Health Sciences, Through Fellow- ships to Physicians, Health-Care Professionals, and Scientists. Status in 1965 One of the mandates of NLM is the organization of biomedical information. This mandate is fulfilled by cataloging, indexing, and abstracting portions of the literature. Organization of the literature also is achieved by producing reviews, synopses, handbooks, and other forms of secondary publications. Their chief purpose is to provide additional means of access to the literature. With the rapid expansion of scien- tific knowledge following World War II, NLM recog- nized the need for the scholarly compilation of all information related to a specific biomedical subject that had not been brought together in one publica- tion, and for original contributions related to the scientific, social, and cultural advancement in the health sciences. Progress in 1965-1985 The purpose of this program has been to assist in the preparation of major scholarly treatises and book- length analyses of the literature on important health topics by highly qualified scientists, practitioners, and scholars. The program began in 1966 as fellowship awards under the training grant authority and changed to project grants under research authority with the first extension of the MLA Act in 1970. In 1973, the emphasis again changed to that of individual authorship, one-year projects, and the requirement for detailed plans and publication arrangements. Since 1976, 10 books have been published, all of them favorably reviewed. Distinguished scholar- scientists have completed major works on toxicity of the liver, epidemiology of diabetes, and most recently, the evolution of the American municipal hospital sys- tem, as well as a major introduction to legal princi- ples, issues, and controversies affecting health profes- sionals in the performance of their work. Unmet/Future Needs (1) Continue to provide treatises that integrate all of the literature on a single subject where none has been written. (2) Continue to emphasize single-author works of authoritative scholars to insure readability and the perspective of a special point of view. (3) Continue to support the development of works that emphasize health and social issues (e.g. informed consent). Assistance in Conducting Research and Investigations in the Field of Medical Library Science and Related Activities and in Developing New Techniques, Systems, and Equipment for Processing, Storing, Retrieving, and Distributing Information in the Sciences Related to Health. Status in 1965 In 1965, nearly all tasks in medical libraries were per- formed manually—checking out books, recording the receipt of journal issues, preparing bibliographic searches by checking indexes and abstracts. The latest in technology was the photocopy machine. This lack of technology contrasted with the volume and charac- ter of a biomedical literature that posed peculiar problems of storage, retrieval, and transmission. Nearly all aspects of current medical library science needed to be reassessed to develop more efficient skills and technology. In addition, research was needed in related areas such as the uses of informa- tion by scientists, teachers, and practitioners and their searching techniques and attitudes; medical terminol- ogy and classification; machine indexing; and graphic image storage and retrieval. Progress in 1965-1985 The MLA Act research grant program supports basic research into fundamental issues of health informa- tion generation, reorganization, and utilization. It also helps develop new methods of information processing and testing those methods in operating conditions. Research grants have been of two kinds: health sciences librarianship and computers in medicine. The first five-year assessment of this MLA Act pro- gram suggested that the funds invested were unrewarding for libraries, since results were, at best, only peripherally relevant. It was also determined that there were numerous straightforward studies that could improve the quality and quantity of service offered by health sciences libraries, but that individuals best able to identify instances of need and opportunity are frequently ill-equipped because of their service training, orientation, and responsibility to design, earn out, and evaluate first-rate research projects. The program took on new vigor in 1979 with the adoption of the computers-in-medicine program, which involved research in the application of com- puter technology in clinical problem solving and deci- sion making, in assisting health personnel in better utilizing research results and health knowledge, and in innovative health data management. During the last decade, research in the field of medi- cal information science has become synonymous with medical informatics. This emerging field of inquiry is concerned with improving the organization, accessi- bility, and transfer of biomedical knowledge. Unmet/Future Needs Develop a research agenda including but not limited to the following: development of new techniques, sys- tems, and equipment for processing, storing, and retrieving biomedical information; study of the uses of information by scientists, teachers, and practi- tioners; comparing information systems developed for the same purpose but where little or no objective comparison of their strengths or weaknesses is attempted; studying information-handling skills of stu- dents, practitioners, and researchers in order to develop scientific methods to promote lifelong learn- ing; designing techniques and systems to effectively assimilate new information and provide access to knowledge in order to improve receptivity and utiliza- tion; developing ways to deliver the knowledge desired rather than a document; developing techniques for the interpretation of data and the evaluation of the biomedical literature; developing the methodology to promote relevant research; and establishing the means to expand the knowledge base on which further advances must be made. Assistance in Improving and Expand- ing the Basic Resources of Medical Libraries and Related Facilities. Status in 1965 Academic Health Sciences Libraries In 1965, the plight of the Nation's health science libraries was critical. Only 14 of the 87 existing medi- cal school libraries had the recommended level of 100,000 volumes, and few had the recommended 1,500 journal subscriptions. Organization of the mate- rial in these libraries suffered from a lack of resources and trained staff, resulting in large backlogs of tasks such as cataloging and indexing. The libraries were able to support only rudimentary reference services that were insufficient to meet the research and educa- tional needs of the institutions they served. They also lacked the equipment (photoduplication) necessary to speed the processing of materials.1 Hospital Libraries A review by the American Hospital Association in 1962 revealed that, of the 5,500 short-term hospitals surveyed, only 58.6 percent had professional libraries. The collections in these libraries were inadequate; the average number of books held was 561, and current journal subscriptions numbered an average of 33. Library staff was primarily part-time and provided a minimal level of service and organization of library materials. In 1962, only 70 percent of the libraries provided reference services, 26 percent borrowed material on interlibrary loan, 21 percent provided bib- liographies, and 12 percent provided photocopies of materials.4 Progress in 1965-1985 MLA Act Programs Instituted To help public or private nonprofit health sciences libraries establish, expand, or improve their resource and information services, NLM began a program of Resource Grants. In 1966, formula grants provided assistance for acquiring library resources. In 1970, emphasis was changed to improving and extending information services with support from Project Grants and to establishing or improving libraries in commu- nity health facilities, primarily hospitals, through Improvement Grants. Approximately 16 percent of the Nation's hospitals have received support and devel- oped basic information services. Academic Health Sciences Libraries The 1983/84 Annual Statistics for Medical School Libraries in the United States and Canada11 reveals that the collections, organization, and services of aca- demic health sciences libraries have improved sub- stantially since 1965. Of the 132 libraries reporting statistics, the mean number of volumes in the collec- tions is 152,006 and the average number of journal subscriptions is 2,159. A full range of services is provided by all libraries, including interlibrary lend- ing and borrowing (over 293,000 items borrowed and 969,716 items lent); over 2 million information con- tacts, and over 130 million photocopies made. Most impressive is the fact that all academic health sciences libraries provide MEDLINE searches (over 242,000 accesses), and all but a few search other online data bases (over 40,000 accesses). Management of academic health sciences libraries collections has improved considerably; most notable is the fact that 89 percent of the reporting libraries use an auto- mated cataloging system. Equipment such as pho- tocopiers, so badly needed in 1965, have become com- monplace in academic health sciences libraries, as have online search equipment and equipment to access bibliographic utilities. A number of academic health sciences libraries have begun to use computers for internal functions such as circulation and journal receipt. Hospital Libraries Hospital libraries have increased in both number and sophistication since 1965. The number of libraries grew by 13 percent between 1969 and 1979,12 and most impressive is the number and educational back- ground of staff managing these libraries. A study of libraries in Region Seven of the RML Network reveals that in 1969, only 40 percent of the hospital libraries were staffed by librarians, as compared to 69 percent in 1984. The number of hospital librarians with MLS degrees increased from 35 in 1969 to 249 in 1984. Collection size also improved significantly in Region Seven. In 1969, 70 percent of the hospital libraries subscribed to 50 or fewer journal titles; in 1984 only 40 percent of the libraries subscribed to this small number of titles. As in the academic health sciences libraries, the range and amount of services provided also have increased. In 1969, 30 percent of Region Seven libraries provided manual or computerized bib- liographies, as compared to 100 percent in 1984; 41 percent provided photocopy services in 1969, and 84 percent provided this service in 1984. Interlibrary lending and borrowing activities increased substan- tially in this same period, with 56 percent of the libraries borrowing over 200 items in 1984 as opposed to 14 percent in 1969, and 47 percent of the libraries lending over 200 items in 1984 as compared to 4 per- cent in 1969. Organization and currency of the collec- tions also improved with more libraries regularly weeding and cataloging their collections.13 Access to needed equipment also has improved considerably, with photocopy machines, computer search equip- ment, and microcomputers becoming more prevalent in hospital libraries. An evaluation of the Improvement Grant Program by Matheson and West14 shows that grant recipients out- performed libraries that applied for but did not receive awards and libraries that did not participate in the grant program. For example, libraries receiving grants showed a 109-percent gain in the number of current journal subscriptions, as compared to a 20- percent gain in the unsuccessful applicant group, and an 86-percent gain in the nonparticipant group. 44 Unmet/Future Needs (1) Through the use of technology, health sciences li- braries will be able to improve access to and delivery of information. Funding is needed to allow libraries to take advantage of the latest technological developments. (2) The diversity and sources of information needed by today's health professional are expanding. Health sciences libraries must be prepared to serve as switch- ing stations to access remote information, in addition to continuing in their roles of organizing and storing information. (3) Expert systems should be developed to aid in the organization and control of information collected by libraries. (4) As biomedical knowledge and sources of knowl- edge expand, it will become increasingly difficult for libraries to collect all pertinent information needed by their users. Cooperative acquisitions and preservation programs that provide access to needed information on a regional level should be encouraged. (5) Continued service to health professionals requires continued analysis of their needs and the ability of health sciences libraries to meet them. It is recom- mended that national mechanisms for the effective collection and analysis of data concerning informa- tion services and their related research and educa- tional activities be developed. (6) In an era where much of the information that health professionals need cannot be purchased for use in a local library but must be accessed on a cost/use basis, it is essential for institutions to develop and implement policies regarding information access, costs, and fees. Assistance in Developing a National System of RMLs Each With Facilities of Sufficient Depth and Scope To Sup- plement the Services of Other Medical Libraries Within the Region It Serves. Status in 1965 Expanding medical literature and the need for increased speed in accessing it calls for measures to improve availability of books, journals, and other informational material. A network of regional medical libraries would help fill this need and also would relieve NLM of its responsibility as the "libraries' library." Progress in 1965-1985 The RML Program was developed to provide health professionals with timely, convenient access to health- care and biomedical resources. Initially, the Nation was divided into 11 regions. In 1983, the 11 regions were reconfigured into 7 to increase the proportion of funds going into direct information services. Over the years, the program developed a network of health sciences libraries that shares collection and staff resources, and uses modern technology to improve access to information. Major accomplishments of the RML Program include: (1) Developing an effective network of over 3,300 libraries to provide information in books, journal arti- cles, and audiovisuals to health professionals. (2) Using modern technology to locate serial titles not held locally so requests can be filled quickly and cost- effectively. (3) Expanding access to online systems to more than six million health professionals and assisting in providing online training to librarians and health professionals. (4) Training hospital library personnel in basic library management and delivery of information services. Each year this training improves information services to more than 1,000 health professionals. (5) Fostering the development of more than 270 hospi- tal library consortia for resource sharing. (6) Developing and implementing a regional auto- mated document request and routing system that han- dles an estimated 50,000 requests a year. (7) Identifying geographic areas in which health professionals are underserved and implementing pro- grams to provide services in 10 to 20 of these each year. Unmet/Future Needs Although the RML Network has improved dramati- cally the access to information by health profes- sionals, technological advances will continue this improvement. Specific programs that should be undertaken by NLM and the RML Network include: (1) Development and implementation of a nationwide automated document request and routing system to save money and provide more rapid delivery of requests. (2) Development of improved mechanisms for the delivery of documents, including telefacsimile and optical disk technology. (3) Development of improved mechanisms to locate books and audiovisuals, including access to this infor- mation directly by end users. (4) Development of improved information on the con- tents of books so that requestors can select the trans- mission of needed portions only, rather than the loan of the physical volume. (5) Development of improved mechanisms for access to non-health-related materials needed by health professionals by facilitating linkages with other infor- mation networks and data bases. (6) Facilitation of improved regional collection development to insure that needed materials are acquired and retained in the regions. (7) Development of a National Biomedical Reference Referral Network. (8) Development of a National Biomedical Preserva- tion Plan integrated into the national preservation plan for the entire scholarly record. Access to the RML Network is readily available to health sciences libraries with trained personnel. Use of the network by individuals without access to health sciences libraries and by institutions without devel- oped information services and sources needs improve- ment. In addition, developed libraries need to employ technology to improve their provision of information services. This can be accomplished by: (1) Developing ways to provide basic levels of informa- tion services to isolated geographic areas in which health professionals are still underserved. (2) Encouraging enhanced participation in the RML Network by network libraries and health professionals. (3) Examining appropriate levels of access to the net- work in light of technological developments. (4) Encouraging the use of current technological advances in order to disseminate information. (5) Implementing at test sites and evaluating the use of modern technologies that can enhance the delivery and use of health-related information. Provide Financial Support to Biomedical Scientific Publications. Status in 1965 The fundamental rationale for NLM's mission recog- nizes that health is a national priority and that health research is a major national investment. Not only is NLM responsible for collecting and organizing, but it also must disseminate health research information. To realize the full benefit from the investment in research, every possible means must be taken to stimulate the effective dissemination of that informa- tion.15 For NLM to accomplish that objective, it had always been necessary to catalog its books and index its journals. With the increasing volume and complex- ity of biomedical literature, it became necessary to develop other secondary publications that provide efficient and targeted access to selective parts of the literature: reviews, indexes, abstracts, and handbooks. Translation of foreign biomedical literature was also important so that U.S. health-care professionals could keep abreast of medical research and progress made elsewhere. Supporting works that would improve the coverage of the biomedical literature, but were not commercially viable, was also an objective. Progress in 1965-1985 The grants and contracts under this MLA Act pro- gram have assisted in preparing, producing, and dis- seminating biomedical scientific publications that were not commercially viable and that would provide new points of entry to the medical literature. This effort, begun in 1966, supported a wide range of pub- lication activities, including critical, analytic review of the status of medical research and practice; secondary literature tools such as handbooks and biomedical bibliographies; studies in the history of medicine; and translations. In 1974, the program emphasis was changed to------------------ providing selective, short-term support for critical MEDLARS reviews and monographs in special areas of health research and practice Status in 1965 Studies supported include critical reviews and mono- graphs on current and past developments in medical research and services, publications in biomedical communications and health information science, translation of significant foreign-language mono- graphs, and proceedings of symposia important to U.S. health interests. Over 439 publications have been supported by this program. Reviews indicate these publications have been of consistently high quality and that a signifi- cant number of the publications have direct relevance to improvement of health-care delivery. Unmet/Future Needs (1) Continue to make important contributions to bio- medical knowledge by providing syntheses and ana- lyses of current development in biomedical research and practice for the use of U.S. health practitioners, health researchers, and medical educators. (2) Continue to provide low cost, time-limited support for important publications in the health sciences that are not commercially profitable. (3) Continue to identify areas of the literature where such publications are needed. (4) Support publications and the development of data bases that would describe health sciences library col- lections, services, users, and needs. Since its founding in 1836, NLM (formerly the Library of the Surgeon General's Office) has had the major responsibility for publishing the national bib- liographies in medicine, the Index Catalog of the Library of the Surgeon General's Office (the first vol- ume of which was produced in 1880), and the Index Medicus (which began publication in 1879). The pur- pose of the Index Catalog and the Index Medicus was to provide bibliographic access to the book and jour- nal literature owned by the Library. Similarly, the need for bibliographic control of a burgeoning medi- cal literature resulted in the development of NLM's MEDLARS. By the late 1950's, the number of jour- nals and articles to be indexed by NLM for inclusion in the Current List of Medical Literature (at that time the monthly index serving as the key to medicine's journal literature) strained NLM's ability to keep the publication current. The Library developed a plan to produce the Current List using punch cards contain- ing bibliographic information describing each journal article. The plan called for the cards to be sorted by machine and the output photographed by a high- speed camera to produce a photo-offset negative for printing. While the system produced the Current List, the machines' card-sorting capabilities were too slow to make the selective retrieval of bibliographic infor- mation practical.6 16 Not long after that, advances in data processing tech- niques enabled NLM to produce Index Medicus (the union of NLM's Current List of Medical Literature and the American Medical Association's Quarterly Cumulative Index Medicus) and paved the way for the development of a computerized bibliographic service for the health professional. The indexing information for each journal article was fed into the system and stored on magnetic tape, which was then manipulated by a digital computer. The processed tape activated a high-speed composing device that produced photo- graphic masters for printing Index Medicus. The same bibliographic data base used for publishing Index Medicus was now accessible for selective infor- mation retrieval, and the first MEDLARS searches were formulated in 1964. Progress in 1965-1985 Trained search analysts in selected academic health center libraries sites around the country formulated MEDLARS searches and sent them to NLM for com- puter processing. Results were returned to the initiat- ing institution for distribution to the library's clien- tele. While it was possible to process only 20,000 MEDLARS searches per year, the ability to selectively search the Index Medicus data base was a milestone in the application of technology. It improved immeas- urably the speed and quality of literature searches that health sciences libraries could provide their users. MEDLARS foreshadowed the development of myriad data bases whose selective access for the pur- pose of developing tailor-made bibliographies trans- formed not only health sciences library service, but also service in public, special, and academic libraries in the United States and abroad. NLM's current system, MEDLINE (MEDLARS online), permits the user to hold a dialog—online— with the computer by typing in responses to prompts and queries at the computer terminal keyboard. MEDLARS data bases are accessed mainly through MeSH (Medical Subject Headings), the controlled vocabulary used to index journals for Index Medicus. When the searcher finds the necessary references, citations may be printed at the terminal. NLM's data bases are used by private vendors who lease MED- LARS tapes and individual health professionals, as well as over 2,000 health sciences institutions (mainly libraries). MEDLINE, which contains about 800,000 citations from 3,000 biomedical journals, is the largest and most frequently used of NLM's data bases. Unmet/Future Needs (1) Expert systems to help evaluate, index, and synthe- size information for inclusion in MEDLARS should be developed. Examples of such expert systems include: (a) Quality filters (i.e., means to evaluate articles, protocols, and other information in the literature of biomedicine. Such an evaluation might be based on citation analysis and randomized control trials.) (b) Automated indexing (2) Increased technological sophistication and expert systems make it possible to store larger quantities of information economically and facilitate the addition of this information to the MEDLARS data base. Examples of types of information that could be included are: (a) Contents of books (b) Synthesized biomedical information (3) As health professionals become directly involved in information management and access, it will be essen- tial to design increasingly "user friendly" software that does not require extensive training. (4) The increasingly interdisciplinary nature of scien- tific investigation will require access to information in other data bases and libraries. The MEDLARS sys- tem should be designed to provide access to these information sources through gateways. Integrated Academic Information Management 1983-1985 In 1983, NLM began a special new initiative: IAIMS. This program was developed in response to recom- mendations in an NLM-supported study by the Association of American Medical Colleges, Academic Information in the Academic Health Sciences Center: Roles for the Library in Information Management. The study addressed the need for networks that would facilitate the flow of recorded biomedical knowledge throughout academic health science centers and hospitals. It specifically recommended support for prototypes that would encourage such net- working. The IAIMS initiative originally awarded contracts to four academic health science centers to plan proto- type design and implementation. Since the original four awards, other institutions have received funding for IAIMS planning activities. 48 Unmet/Future Needs There is widespread recognition that health sciences libraries across the country will need to develop net- works for biomedical institutional environments. All such networks, however, may not be identical or even parallel since institutional and individual needs differ from location to location. Funding for IAIMS models and for alternatives to IAIMS should be continued if biomedical information is to be accessible and usable within the academic health center and within the hospital. References 1. Matheson NW, Cooper JAD. Academic information in the Aca- demic Health Sciences Center. Roles for the library in information management. / Med Educ 1982;57(Pt2):l-93. 2. Bloomquist H. The status and needs of medical school libraries in the United States. / Med Educ 1963;38:145-63. 3. Huang CK. Physical facilities of medical school libraries in the United States, 1966-1975: a statistical review. Bull Med Libr Assoc 1976;173-8. 4. Giesler R, Yast H. A survey of current hospital library resources. Hospitals 1964;38:55-8. 5. Committee on Hospital Library Architecture. Planning the hospi- tal library. New York: United Hospital Fund of N.Y., 1957. 6. Cummings MM, Corning ME. The medical library assistance act: an analysis of NLM extramural programs, 1965-1970. Bull Med Libr Assoc 1971;59:375-91. 7. Miles WD. A history of the National Library of Medicine. Bethesda, MD: National Library of Medicine, 1982; NIH publication no. 85-1904. 8. Miller JD. Health science libraries in hospitals. Bull Med Libr Assoc 1972;60(Suppl). 9. Physicians for the twenty-first century. Report of the Panel on the General Education of the Physician and College Preparation for Medicine. / Med Educ 1984;59(Pt2):l-208. 10. Report of Task Force on Research Grants Program. Bethesda, MD: National Library of Medicine, 1977. (Unpublished) 11. Annual statistics of medical school libraries in the United States and Canada, 1983-84. 7th ed. Houston: Association of Aca- demic Health Sciences Library Directors and the Houston Academy of Medicine-Texas Medical Center Library, 1984. 12. Crawford S. Health science libraries in the United States: HI. Hospital health science libraries, 1969-1979. Bull Med Libr Assoc 1983;71:30-6. 13. Van Vuren DD, Graham E, Flack V. Hospital library develop- ment and the impact of PSRMLS services: report of an evaluation project. Los Angeles: UCLA Biomedical Library, Pacific Southwest Regional Medical Library Service, 1985. 14. Matheson NW, West RT. NLM medical library resource improve- ment grant program: an evaluation. Bull Med Libr Assoc 1976;64:309-19. 15. Werner G. The cost recovery issue and the government's role in provision of information services. Bull Med Libr Assoc 1982;70:244-5. 16. Mehnert RB, Leiter J. The National Library of Medicine: history and organization. In: Darling L, ed. Handbook of medical library practice. 4th ed. Chicago: Medical Library Association. 1982. 49 Appendix B: NLM Planning Process In January, 1985 the Board of Regents of the National Library of Medicine resolved to develop a long-range plan to guide the Library in wisely using its human, physical, and finan- cial resources to fulfill its mission. The Board recognized the need for a well-formulated plan because of rapidly evolving information technol- ogy, continued growth in the literature of biomedicine, and the need to make informed choices of intermediate objectives that would lead NLM toward its strategic, long-range goals. Not only would a good plan generate goals and checkpoints for management, actually a map of program directions, but it would also inform the various constituencies among the Library's users about the future it sought and could help to enlist their support in achieving that future. At the Board's direction, a broadly based proc- ess was begun involving the participation of librarians, physicians, nurses, and other health professionals; biomedical scientists; computer scientists; and others whose interests are inter- twined with the Library's. A total of 77 experts in various fields accepted invitations to serve on one of the five planning panels. Each panel addressed the future in one of the five domains that encompass NLM's current programs and activities. The domains, which provided the panels a framework for thinking about the future are: 1. Building and organizing the Library's collection 2. Locating and gaining access to medical and scientific literature 3. Obtaining factual information from data bases 4. Medical informatics 5. Assisting health professions education through information technology The Library chose a planning model with three components. First, it incorporates a general, somewhat indistinct vision of the future 20 years from now in medicine, library and information science, and computer-communications technol- ogy. That environment cannot be forecast pre- cisely, but we can speak of a "distant" goal. That goal is seen as a societal objective whose attainment involves many organizations and agencies. NLM has a major role to play in achieving the goal and must plan its part. Sec- ond, while the 20-year goals are indistinct, there are opportunities for and impediments against achieving them. The opportunities and impedi- ments can be more clearly envisioned because they appear to lie roughly 10 years away. Third, the specific steps that should be taken to remove the impediments and take advantage of the opportunities should be programmed for 3 to 5 years. The planning process also involved participation within the Library. The Director provided his ver- sion of the future in the form of a "Scenario: 2005," which was distributed to panel members and Library staff. NLM staff prepared back- ground documents that reported NLM achieve- ments in the five domains, identified issues, and reviewed current planning. Senior NLM staff members also acted as resource persons to the planning panels. At the end of the planning process, each panel formulated recommendations and priorities for future NLM programs and activities in the domain under its purview. The five panel reports were reviewed by the Board of Regents in June 1986. The Board then asked the NLM staff to analyze and reconcile their findings, eliminating any duplications and consolidating the recom- mendations. This synthesized plan is presented in this volume. Together with the planning panel reports, it represents the official Long-Range Plan of the Board of Regents of the National Library of Medicine. Photographs were obtained from the several Bureaus, Institutes, and Divi- sions of the National Institutes of Health (including the Office of the Director, NIH, the Warren G. Magnuson Clinical Center, and the National In- stitute on Aging), the Uniformed Serv- ices University of the Health Sciences, the World Health Organization, and William A. Yasnoff, M.D., Ph. D.. ... - mI,( *V December 1986 —. '•• - Awtii.- v *L