Computers in Medicine, Research, and Training: Government Orphan?, 1983
Page  1

computers in medicine research and training government orphan what perceptive observers have been predicting has gradually occurred computers and communications technology have transformed us from an industrial into an information society machines as extensions of the human brain have in many respects become more important to us than machines as extensions of human muscle this transformation is inevitable irreversible and like the gift of fire from prometheus will profoundly affect the future of mankind although those in the biomedical community acknowledge that such a transformation is occurring they have not yet been willing to admit that it is truly relevant to their endeavors the greatest problem we face with these brain extending machines is not how to build them but how to apply them unly when we have designed an application that the community recognizes universally as contributing to the improvement of medicine will they concede its relevance to their endeavors hy concern is not the need to train more programmers systems analysts computer operators and technicians although the growth rate in these occupations by 1990 will be three times that for all occupations in the economy 1 such training is now being vigorously pursued in industry universities community colleges even in high schools what we require in the health sciences are talented men and women who will devote themselves at the highest level to fundamental issues of knowledge management that computers make it possible for the health professions to address keynote address for the american association for medical systems and informatics aafisi 1983 congress may 2 1983 by martin h cummings m d director national library of medicine

Page  2 2 the problem as highlighted in the consensus statement by a panel of experts assembled by the national science foundation there is a major crisis in training computer scientists there is a lack of support to graduate students which leads to a shortage of faculty 2 john r opel president of ibm warns that the united states is slipping in the race to strengthen not its capacity in buildings and machines vital as they are but the capabilities of its people talented educated and trained human beings the ultimate resource in any nation 3 in the united states last year we graduated more than 16000 medical doctors doctorates in computer science numbered 250 the imbalance is not righting itself but is getting worse universities find it increasingly difficult to offer advanced computer science training because of a lack of faculty at the doctoral level talented students are being siphoned off into well-paying jobs rather than completing higher studies one federal expert has estimated that a person with a two year degree can get a programming joe for 20003 to 22030 a year in industry a person with an eight year ph d might make 20000 to 521003 a year as an assistant professor 4 the picture is somewhat brighter coming at the problem from the other direction that is beginning with the much larger pool of biomedically trained researchers who with proper training and in congenial settings would be available to make important contributions in applying computer technology to the health sciences this is the approach the national library of medicine took some 10 years ago when it began its computer technology training grant program

Page  3 is was augmented in 19/9 with expanded support programs for research applying computers to medical problems as valuable as these programs are they have not cd n'de the same prioriy federal support as categorical studies of disease in laboratories clinics and hospitals nevertheless a review of accomplishments to date will show just how grave their loss would be the computers in medicine programs the medical library assistance act of 1965 is the legislative authority for nlm grants training grants were an important part of the original legislation and for the first five years a wide variety of training programs received support health science librarianship some computer technology application's information science science writing and the history of medicine in 1972 the library conducted a thorough review of its training program and concluded that it had been successful in meeting the originally identified need for more trained library professionals the staff recommended and the regents agreed that the training program be redirected to concentrate on applying computer technology to information problems within the health sciences the program was designed for health professionals and scientists who could apply their creative talents to various aspects of information handling for example knowledge representation database construction indexing and retrieval modeling of biologic systems and clinical decision making the ultimate goal of the training was to promote the complete and effective integration of computer technology into all phases of clinical medicine teaching research and practice

Page  4 since nrl has supported training programs at thirteen academic institutions none are active today they vary considerably their character being influenced by the administrative setting of the program the special emphases of their curricula and the style of leadership the site of successful programs has ranged from university graduate schools to medical school departments and even to divisions within a single medical school department the breadth of the training offered and the variety of persons selected for fellowships have been governed by these different settings some programs have enrolled exclusively postgraduates others have included predoctoral fellows as well the number of individuals enrolled in these various programs past and present is about 350 those selected for training have been physicians about 30 percent ph d's 10 percent and holders of predoctoral degrees in the computer or biological sciences 60 percent we are encouraged by the results of a recent survey which show that only 20 percent of those trained have gone to industry the remaining 80 percent have faculty positions are employed in the delivery of health services or are pursuing additional training research grants began in the late 1960's as a rather broad program encompassing research related to library and information services biomedical information and communications science health sciences education and knowledge transfer and studies in the history of medicine in 1977 an evaluation of these research activities found that they were on a plane with other rln supported research grant programs in terms of quality and contributions to the field however rln's program was judged to be highly heterogeneous with projects

Page  5 scattered throughout the large domain encompassed in the term biomedical communications the data showed there was no cohesive or readily identifiable research constituency and that few individuals applied for or were awarded more than one research grant among the recommendations of the reviewers was that nlm build on its earlier research accomplishments by concentrating ok the computer sciences related to health this would be done in several ways first traditional investigator initiated research grants would be encouraged with one area of strong emphasis being computers in medicine second a new category of program project grants would be instituted that would fund a small number of major studies into new methods for representing medical knowledge although this research would draw from the computer and cognitive sciences it would be health related and require close collaboration with highly qualified health scientists and clinical investigators the program would support not only basic research but also the demonstration of practical applications of research findings in operating conditions it was recommended that supported projects should emphasize bibliographically validated material rather than data from patient records or diagnostic tests although such research must necessarily involve categorical diseases and clinical experiences the results of the proposed projects should contribute to advances in the health information sciences projects involving computer technology where the primary advance would be better understanding of biomedical phenomena should be referred to other components of the national institutes of health among the kinds of research eligible for funding

Page  6 6 knowledge representation principles of knowledge base design and database management attributes of language processing in medical contexts computer reasoning including artificial intelligence as applied to knowledge bases and so called expert systems human factors relating to the use of computerized knowledge and consultant systems the role of knowledge in medical decision making and problem solving and the impact of advanced computer systems on accepted patterns of acquiring transmitting and using medical knowledge as a result of this recommendation five program project grants were added each funded for five years through 1984 the supported projects deal with research in artificial intelligence and clinical problem solving medical databases and clinical investigation biomedical knowledge representation and investigations in clinical decision making the researchers involved in these projects included physicians 20 computer scientists and 7 research associates who are paid candidates assessing research productivity is a difficult and sensitive issue but if we rely on the traditional yardstick of published output the results are most encouraging since january 1981 the grantees have published 28 refereed articles have written 14 invited book chapters and two monographs and in addition have made approximately 50 presentations at national and international scientific meetings

Page  7 the total invesment by the national library of medicine in all aspects of grant support for computers in medicine training and research has been approximately $25 million a modest sum in the context of a $4 billion annual researo budget for nlm as a whole technology aside there have been some dramatic changes in the environment of the computers in medicine field in the ten years since nlm began to emphasize its support today there is an acute awareness of the potential of computer technology as reflected in the biomedical literature at the meetings of such organizations as the aamsi and other professional societies the issue is being addressed through workshops symposia and papers on a wide variety of topics related to decision analysis computer applications and computer literacy in medical education unfortunately academic health science institutions have not kept up with this trend by establishing departments that would foster research in medical commuting ten years ago there was one department of medical computing in this country today there is none the place of medical computer scientists in academic medicine and the availability of research support are vital concerns in short a career ladder is not available as it is for other established health disciplines even so the present environment is clearly conducive to a more widespread recognition of the importance of this field by the entire health science community however except for the programs of the library which i have just outlined there is little federal support of research and career development in this field

Page  8 related ferderal programs there are several federal agencies that support other aspects of the computers in medicine field the division of research resources of the national institutes of health has provided extensive support over the years for computer resources used in research the biotechnology resources program as it is called originally concentrated on supporting large general purpose computer centers in medical schools the highly regarded sumex aim network is a noteworthy accomplishment more recently the program has shifted its emphasis to supporting a broad and innovative array of other biomedically relevant technologies because this support is for computer and other technology based resources it is thus fundamentally different from nlms support for research projects in computer science the biotechnology resources program operates at a level of about $20 million annually but its influence is far out of proportion to the dollars spent in 1980 for example the various agencies of the public health service supported more than 1000 grants almost 120 million for investigators whose research drew on the biotechnology resources supported under this program 5 among these were seven investigators supported by nlms modest computers in medicine program another source of grant suppore are seve are several federal agencies that support other aspects of the computers in medicine field the division of research resources of the national institutes of health has provided extensive support over the years for computer resources used in research the biotechnology resources program as it is called originally concentrated on supporting large general purpose computer centers in medical schools the highly regarded sumex aim network is a noteworthy accomplishment more recently the program has shifted its emphasis to supporting a broad and innovative array of other biomedically relevant technologies because this support is for computer and other technology based resources it is thus fundamentally different from nlms support for research projects in computer science the biotechnology resources program operates at a level of about $20 million annually but its influence is far out of proportion to the dollars spent in 1980 for example the various agencies of the public health service supported more than 1000 grants almost 120 million for investigators whose research drew on the biotechnology resources supported under this program 5 among these were seven investigators supported by nlms modest computers in medicine program another source of grant suppore are several federal agencies that support other aspects of the computers in medicine field the division of research resources of the national institutes of health has provided extensive support over the years for computer resources used in research the biotechnology resources program as it is called originally concentrated on supporting large general purpose computer centers in medical schools the highly regarded sumex aim network is a noteworthy accomplishment more recently the program has shifted its emphasis to supporting a broad and innovative array of other biomedically relevant technologies because this support is for computer and other technology based resources it is thus fundamentally different from nlms support for research projects in computer science the biotechnology resources program operates at a level of about $20 million annually but its influence is far out of proportion to the dollars spent in 1980 for example the various agencies of the public health service supported more than 1000 grants almost 120 million for investigators whose research drew on the biotechnology resources supported under this program 5 among these were seven investigators supported by nlms modest computers in medicine program another source of grant support for computers in medicine research is the public health services national center for health services research the research supported by this center differs from that of nlms in that it is aimed prim arily at practical payoffs for patient care since its inception in 1958 the centers support has resulted in a number of noteworthy accomplishments

Page  9 including a comprehensive hospital information system the design of a commercially marketable automated ambulatory medical record system the establishment of clinical databanks for coronary artery disease and rheumatic diseases and the development of a computer assisted electrocardiogram analysis program 6 this program valuable as it is is in no sense a substitute for nlm's own support for basic computers in medicine research and training the center supports a variety of research related to health services in addition to computer applications unfortunately the center's overall budget has undergone drastic cuts in recent years from a high of 55 million in 1974 to 14 million in 1980 to less than 6 million in 1982 its support for computer related projects has likewise dwindled to less than 1 million in 1982 the national science foundation through its division of information science and technology also funds basic and applied research in information and communications although some of the supported projects may ultimately have a health related application most are broader more generalizable studies into the principles governing the generation transmission and use of information nsf divides its support into three areas information science which is concerned with increasing the fundamental knowledge necessary for understanding information processes information technology which is concerned with research on applying information science to the design of advanced information systems and information impact which seeks to improve our understanding of the pervasive impact of information and information technology on the economic and social fabric of society 7 the level of support for these grants has remained fairly constant in recent years at about 5 million annually

Page  10 conclusion most observers agree that we have barely scratched the surface of the computer's potential in medical applications at the national library of medicine we have seen our 20 year investment in computerized bibliography repaid many times over in vastly improved services to the nation we can now foresee the possibility of similar advances in the broad arena of computers in medicine it is thus particularly unfortunate that the one federal program that has sought to support research and training in this important area cannot acquire sufficient funds to maintain its level of support the library's board of regents at its meeting in january 1983 assessed the prospects for the program and reluctantly agreed that in the absence of help from other federal sources the nlm computers in medicine efforts be reduced the regents reviewed a plan that would phase out the computers in medicine research project grants subsume the separate computers in medicine category under other investigator initiated research grants and narrow the scope of the computers in medicine training program they recognized that protecting young investigators is critical consequently our training program will concentrate on research career training and we will continue to offer special awards for young scientists the board's chairman william d mayer m d summed it up for his colleagues when he said the entire concept of the computers in medicine program is going down the drain at the very time when we have developed a core of experts who

Page  11 could lead us in the health community to the information age we are facing someone must get the message across that it is essential to the future of the biomedical research enterprise and the u s leadership role to invest a small but critical amount of our resources into managing and disseminating the information we generate after a successful birth and robust infancy the library's computers in medicine program finds itself an orphan it deserves adoption by caring foster parents

Page  12 references 1 employment trends in computer occupations bulletin 2101 bureau of labor satistics u s department of labor washington d c 1982 2 a consensus statement the panel on computing and higher education irving shain chairman national science foundation washington d c march 30 1981 3 opel j r education science and national economic competitiveness science 217 17 september 1982 1116 4 andrew holnar national science foundation quoted in the wall street journal january 14 1983 5 division of research resources annual report fiscal year 1982 national institutes of health bethesda m d 6 computer applications in health care nchsr research report series dhhs publication no phs 80 3251 public health service hyattsville md 7 research in information science and technology program announcement national science foundation washington d c