The Road to IFIP-TC4 and IMIA Through Cybernetic Medicine and the Elsinore Meetings C

Historical Roots of International Biomedical and Health Informatics: The Road to IFIP-TC4 and IMIA through Cybernetic Medicine and the Elsinore Meetings C. A. Kulikowski Department of Computer Science, Rutgers University, USA

Summary 1 Introduction ematics and physics, who worked with biologists and physicians. It started its own Background: It is 50 years since the International Federation of This paper gives a brief overview of how series of Cybernetic Medicine Congresses Information Processing (IFIP) Societies approved the formation international activities in biomedical and in 1960 in Naples [17]. In 1962, Gustav of a new Technical Committee (TC) 4 on Medical Information health informatics started and evolved Wagner in Germany founded the jour- Processing under the leadership of Professor Francois Grémy, after World War II, together with a brief ac- nal Methods of Information in Medicine which was the direct precursor of the International Medical count of literature focusing on the history (MIM), which was the first journal oriented Informatics Association (IMIA).

of the field in its precursor period before towards documentation and medical statis- Objectives: The goals of this paper are to give a very brief the formation of the International Federa- tics [20], as well as information processing overview of early international developments leading to tion of Information Processing Technical in the broadest sense of medical work, and informatics in medicine, with the origins of the applications of Committee 4 (IFIP-TC4) which was the which later became the official journal of computers to medicine in the USA and Europe, and two meetings predecessor of the International Medical IMIA. Figure 1 illustrates the cover of the - of the International Society of Cybernetic Medicine, and the Informatics Association (IMIA). IFIP-TC4 first issue of MIM, with a part of its index. Elsinore Meetings on Hospital Information Systems - that took came about primarily as the result of the The first fully medical-oriented interna- place in 1966. These set the stage for the formation of IFIP-TC4 initiative of Franҫois Grémy [1, 2]. It was tional meeting of pioneering investigators the following year, with later sponsorship of the first MEDINFO in driven by antecedent work from computers working on computer data processing 1974, setting the path for the evolution to IMIA. in medicine, especially related to diagnostic methods for hospitals and medicine more Methods: This paper reviews and analyzes some of the earliest logic and probabilistic reasoning and health generally was held in Elsinore in 1966 [21]. research and publications, together with two critical contrasting care [2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14], These first steps toward an international meetings in 1966 involving international activities in what and more general approaches to computa- organization led, over the next decade, to evolved into biomedical and health informatics in terms of their tion in biology and medicine [15, 16], as the first World Conference on Medical probable influence on the formation of IFIP-TC4. well as the attempts to model the processes Informatics (MEDINFO) in Stockholm in Conclusion: The formation of IFIP-TC 4 in 1967 by Francois of communication and control under the 1974 [22], and the change in emphasis and Grémy arose out of his concerns for merging, at an international rubric of cybernetic medicine [17], follow- designation from cybernetics and medical level, the diverse strands from the more abstract work on ing the work of Wiener [18], and related to computer or electronic data processing, cybernetic medicine and its basis in biophysical and neural the information theory of Shannon [19]. to the naming of medical informatics as modeling, with the more concrete and health-oriented medical This involved mathematical approaches a major professional endeavor with its information processing that was developing at the time for taken by researchers from biophysics, broader emphasis on the full spectrum of hospitals and clinical decision-making. bioengineering, epidemiology, biometry, biomedical and healthcare information. and clinical documentation, together with This led to a distinct discipline in the Keywords clinicians who experimented with formal 1970’s, foreshadowing the worldwide International Biomedical and Health Informatics, IFIP-TC4, History, models of medical decision-making with acceptance of the term informatics as de- Cybernetic Medicine, Elsinore Meetings, IMIA, François Grémy various logical and statistical approaches, fining the core of the new discipline in the starting in the 1950’s. 1980’s, and the universal adoption of this Yearb Med Inform 2017:257-62 In 1958, an International Society of Cy- designation for all biomedical and health http://dx.doi.org/10.15265/IY-2017-001 bernetic Medicine was founded, drawing information-related work by the start of Published online May 8, 2017 on a wide range of researchers in math- the new millennium.

IMIA Yearbook of Medical Informatics 2017 258 Kulikowski

Fig. 1 Cover and part of the index of the first volume of the journal Methods of Information in Medicine, founded by Gustav Wagner in 1962.

connected later to the work on syntactical lowing the popularity of the book of the same 2 Early International information theory by Shannon at Bell Labs name [18], and when applied to medicine, this Activities of Computers [19], dealing with the efficiency of signal led to the First Congress on Cybernetic Med- transmission or electronic communication icine, organized by Aldo Masturzo in Naples in Biology and Medicine: across noisy channels. This in turn attracted in 1960, followed by another in Amsterdam Cybernetics and Medical the attention of scientists and technologists in 1962, and a re-visit to Naples in 1964 [17]. who saw links to general human reasoning However, it is the 4th Congress held in Nice Information Processing mechanisms related to the McCulloch-Pitts in 1966 [26] that likely had the greatest influ- In many countries, primarily in the USA and models of neuronal transmission [23]. They ence on the eventual emergence of medical Europe, after World War II, several distinct developed computer-based implementations informatics, since it was attended by Francois scientific, engineering, and philosophical of mathematical models for information Grémy and his colleague JC Pagès, who could lines of mathematical research and its appli- coding, communication and control, as well not have but helped observe the contrast be- cation to practical computer-related work, as statistical models for signal processing - de- tween the very general and theoretical concerns were developed at the interfaces of biology, tection, classification, and prediction - which with learning that was the focus of much of the medicine, and computer and communication in turn gave rise to pattern classification and work presented on cybernetics, as opposed to technology. These were largely inspired by recognition methods in the 1960’s [24, 25]. the more concrete applications of statistics and the cybernetics approaches pioneered by In Europe, Weiner’s and Shannon’s ideas logic to medical diagnosis, which François Norbert Weiner at MIT during the war [18], were generally designated as cybernetics fol- Grémy summarized at the Nice meeting.

IMIA Yearbook of Medical Informatics 2017 259 Historical Roots of International Biomedical and Health Informatics: The Road to IFIP-TC4 and IMIA through Cybernetic Medicine and the Elsinore Meetings

In Germany, there was a strong tradition of leadership of Bruce Waxman in 1961 [27], the idea that statistical decision-making medical documentation and analysis, which leading to the emergence of ad-hoc study was the major “way to go” in formalizing viewed information as central to medical sections to evaluate grants for funding the clinical as well as epidemiological reason- recording, statistics, and organization field in the decade of the 1960’s. In Japan, ing. Around the same time, the Collen and processes [20], and led to the founding of one of the other early computer adopting Garfield group at Kaiser Permanente [8], Methods of Information in Medicine. Wag- countries, work began also in the 1960’s on building on the earlier work on preventive ner’s decision to start the journal came from medical electronic documentation, as well and occupational medicine, had developed his participation in the post-war German as on computational modeling for scientific a large database of cases from which they research on medical documentation (both inquiry and technological instrumentation. derived the frequencies of occurrence of paper and early computer-based) with In the late 1940’s, strands of technolo- various health problems under different its own professional society, the German gy-oriented research had gradually arisen conditions, which eventually led to one of Society for Medical Documentation and that was aimed at what is now usually the most systematic statistical likelihood ra- Statistics (GMDS.) Around this time, the known as decision-support for medicine. tio approaches to estimating risk of disease documentation focus was reinforced by Robert Ledley, a dentist who joined the and other health-adverse conditions [28]. the consolidation of medical biometry National Institute of Standards in the US Collen’s emphasis on preventive medicine, and epidemiology research into academic stands out during this period for his remark- the efficiency of documentation and the departments. GMDS incorporated the able insights, proposals, and actual system analysis of intervention outcomes, had led work on medical documentation on which implementations showing how emerging to many statistical screening methods, such biometrical and epidemiological analyses computational technologies could assist as the Cornell Index [29]. were based, and was later re-named the not only medicine and dentistry, but all the Deutsche Gesellschaft für Mediz- healthcare disciplines and their underlying inische Informatik, Biometrie und biomedical sciences [16]. He pioneered Epidemiologie while maintaining its the formalization of medical diagnostic 3 The 1966 Elsinore GMDS abbreviation. reasoning as statistical decision-making us-

In the USA, which led the post-war ing Bayes/Laplace’s Rule, and, joining with Meetings – The First world in the development of computer Dr. Lee Lusted, wrote the break-through International Symposia technology, science, and their applications, paper in Science in 1959 [6] which brought a most insightful physician and engineer, formal statistical approaches for modeling for Medical Information Dr. Morris Collen, who had worked on medical decision-making to world-wide maintaining the health of workers at the attention and encouraged a strong thread of Processing Kaiser Industries that produced the Liberty research on the topic that persists in various With the benefit of hindsight, it can be seen Ships in Oakland, California, developed the formulations to this very day. Earlier in the how, by 1966, the time was ripe for a truly systematic documentation of health condi- 1950’s, Nash [3] had experimented with a medically-oriented international confer- tions on cards, demonstrating the essential slide-rule model for capturing the logic ence on the use of computers in the practice role of careful records in managing medical of diagnostic combinations of signs and of medicine. The international meetings problems, as well as helping prevent future symptoms that could be aligned to “read held in Elsinore, Denmark, in the shadow health problems through the multiphasic out” a diagnostic category by an ingenious of the castle of Hamlet fame, were held in screening process he pioneered [8]. spatial placement of the slides to align with two parts from April 22 to May 3, 1966. It was from all the above roots, that we a plausible diagnostic category. A more They brought together many researchers can trace the emphasis on the centrality wide-ranging set of approaches to the use and investigators from Europe, North of information in supporting medical and of computers in support of clinical practice America, and Japan, to discuss a range of healthcare practice tasks, which contrasted had been carried out, and encouraged, by diverse topics such as hospital and clinical with the different and more abstract empha- Dr. Ralph Engle of Columbia University computer systems, mathematical and sta- ses that arose from the mathematical mod- [11] who had worked with researchers at tistical modeling and analysis of biomed- eling and simulation research by biophys- IBM Yorktown Height’s TJ Watson Re- ical problems, cardiology, and laboratory icists and bioengineers, and the statistical search Center. The series of IBM Medical data. Many of the participants contributed epidemiological investigations, and meth- Symposia lasted over a decade and exam- papers to the proceedings [21]. While odological researches in medical statistics ined various computational methods and the Proceedings title is “Automated Data which dated back to the 19th and earlier systems. That “the time had come” to focus Processing in Hospitals”, the meetings part of the 20th century before the advent on medical diagnosis as a major scientific comprised two separate programs, with the of computers. All these strands of research and practical problem was emphasized by first one from April 20 to April 23, focusing caught the attention of the National Insti- John Jacquez’s organization of The Diag- on the hospital data processing theme and tutes of Health (NIH) in the USA, which nostic Process conference at Ann Arbor subtitled “An International Conference on set up an advisory committee under the in 1961 [13], which further strengthened the Interface Problems”, while the second

IMIA Yearbook of Medical Informatics 2017 260 Kulikowski

one running from April 25 to May 3 was talk on The Information Problem in Medi- The first subject of the Cybernetic more general, and entitled “International cine, though his written contribution was on Medicine meeting in Nice was “Devices of Advanced Symposium on Data Processing Quality Control in Medicine. The last day Command and Control in Learning”, which in Medicine”. of the first set of meetings was devoted to had three parts –“Learning Processes in The meetings were organized and chaired hospital information issues. Machines”, “Brain Mechanisms of Learn- by Dr. A. Tybjaerg Hansen of Denmark, who The second part of the Elsinore meetings ing Communications”, and “Man-Machine gave a first address on “The role of comput- covered seven themes: computer-assisted Interrelationship from the Point of View ers in the hospital – the interface problems in processing of biomedical information, of Learning”. The first two parts reflected medical and administrative patient manage- cybernetics in medicine, computers in pa- the early and immature stage of knowledge ment”, thus highlighting a challenge which tient management and the use of operations about learning by machines, such as using has persisted and endures to this day on the research methods, analog information, simple perceptrons or feed-forward linear often conflicting electronic data processing man-machine communication in the hospi- artificial neural networks that had not yet (EDP) goals and the requirements for clin- tal, education in computers, and instruction been sufficiently understood in terms of ical or health care management purposes in automated data processing (ADP) for their limitations [30], as effective math- vs. administrative, financial, insurance, and hospital personnel. The first day was chaired ematical models for learning concepts or hospital logistics purposes. Already in 1966, by Dr. Remond and featured a questioning hypotheses from evidence. This was mixed it was clear that different data and interfaces of the aims of data processing in medicine with much conjectural discussion about pu- were required for the very different clinical by E. Dessau of Copenhagen, with talks by tative models of biological learning in the vs. administrative processes, as discussed Dr. Donald A.B. Lindberg of the University brain, the role of esthetics, and “molecular by many participants following Dr. Han- of Missouri on “Processing and Evaluation memory”, all treated at a most superficial sen’s presentation, such as Charles Flagle of Hospital Laboratory Data”, by Dr. Homer level reflecting the poor understanding of Johns Hopkins who gave an operations Warner on “The Digital Computer as Tool of neural mechanisms of the period. The research perspective on the requirements for the Analysis of Physiological Systems”, third part on Man-Machine Inter-rela- for information systems in hospitals, Homer and concluded by Dr. Hansen speaking tionships for learning was very concrete,

Warner of the University of Utah who pre- about “Clinician’s Wishes and Expecta- featuring specific xamplese of computer sented computer time-sharing for processing tions about Data Processing”. The next teaching machines, discussing the use of patient data from the bedside (reflecting his day covered hybrid systems and intensive the then-ubiquitous microfilm, electronic pioneering work on cardiology in the ICU), observation and care units with many of the calculators, and very basic computers and Henry Yellowlees from the United King- speakers of the first set of meetings, while for instruction, as well as a discussion of dom who talked about the computer in the it was followed by one on Analog Data misunderstandings based on Weizenbaum’s National Health Service. Acquisition and Processing which ended deliberately ironic simulation of a Roge- Analog-Digital Data Processing, which with a demonstration of the then-novel on- rian psychiatric interaction through the was still a major concern for interfacing line time-sharing service between Elsinore chatbot ELIZA [31]. However, the papers instrumentation and interpreting their ana- and the Systems Development Corporation were almost all about generic computer log signals, was emphasized by Dr. Antoine (SDC) in Santa Monica, California by Dr. processing, and non-specific to medicine Remond from Pitié - Salpétrière Hospital Anne Summerfield. All these presentations despite the title of the conference. in Paris, and Josiah Macy, Jr, from Yeshiva were very concretely based on working clin- The second subject of the conference, University in New York, who talked about ical systems and the experience of clinicians however, did cover a wide range of topics “Hybrid Systems for a Hybrid World” during with their operation. about collecting and processing medical the second day of the meeting. Specific in- data, starting with the basics of computer stances were then given by the Drs. Caceres, representation of numbers and text, with Pipberger, and Adey who described their medical examples from general statistical work on the analysis of ECGs, and by Dr. diagnosis, electrocardiogram interpreta- Weil for the management of the critically 4 The 1966 International tion, automating documentation in med- ill. The third day covered computers in Congress of Cybernetic icine, the automation of medical records clinical laboratory, radiation planning, and using mark-sense documents, and several the diagnosis of congenital heart disease and Medicine in Nice: A Study in cardiovascular modeling and simulation brain tumors with speakers from the USA, methods as well as psychiatric models. Sweden, the Netherlands, and Japan. There Contrasts This part of the conference was medically was also a discussion on Man-Machine The Fourth Cybernetic Medicine Congress oriented, but it is almost impossible to see Communication in the Hospital by Octo held in Nice from the 19th to the 22nd of connections with the first part on cybernet- Barnett, the pioneer from Massachussetts September of 1966 was, like the Elsinore ics and learning. In this sense, the confer- General Hospital and Harvard, concluded meetings, divided into two very different ence would have split the audience in two: by Gustav Wagner, from Heidelberg, with a segments, denoted as subjects or themes. those interested in the abstract issues of

IMIA Yearbook of Medical Informatics 2017 261 Historical Roots of International Biomedical and Health Informatics: The Road to IFIP-TC4 and IMIA through Cybernetic Medicine and the Elsinore Meetings

learning, and those interested in practical applications of computers to move the medical applications of computing. While field forward, rather than the extremely this may have been a reasonable strategy conjectural and abstract topics of learning at the time, given the primitive state-of-the in humans and machines that characterized art in cybernetics, one cannot but feel that the direction of cybernetics at the time. it must have left the medical audience, The success of the Elsinore meetings in and their presenters, like Francois Grémy, dealing with practical issues of informa- with little doubt that any real impact on tion processing for hospitals and clinical medicine by cybernetics would remain far tasks and the more specific advances of in the future. statistical, especially Bayesian, methods The contrast in the split at the Nice for modeling medical decision-making meeting between the conjectural cybernet- likewise drew Grémy’s attention, as his ics learning discussions and the practical talk in Nice summarized developments clinically-oriented discussions that Grémy of statistical methods for diagnosis [2]. and Pagès experienced during their at- Such observations most likely motivated tendance at the Elsinore meetings might him to consider the formation of a pro- have impressed them greatly - it was at fessional organization to advance a more Elsinore that there had been so many spe- synthetic discipline bridging the practical Fig. 2 Francois Grémy addressing the IFIP-TC 4 Conference held cific, concrete, examples presented that work on medical information processing in Lyon in 1970. bridged hospital, laboratory, and clinical combined with the mathematical modeling applications of computers with the practice for measurable biological systems and the of healthcare. The highlighting of so many statistical modeling for decision-making, many meetings [34, 35] like the one on of these technologically-grounded projects while putting aside for the longer term the information processing in medical records would have likely reinforced what Grémy automated learning aspirations of cyber- [36] held in Lyon in 1970 illustrated in the had experienced during a visit to MIT two netics. From the perspective of 50 years picture displayed as Figure 2. years earlier [32]. later, where we are still struggling to under- Other IFIP-TC 4 meetings were held in stand learning, this turned out to be a wise the following years on the topics of signal decision. Patrice Degoulet, who met and processing, mathematical models in biology talked with Grémy for the first time around and medicine [37], education informatics 1970, recalls that the factors that plausibly for healthcare personnel with J Anderson 5 Conclusion: The influenced Grémy’s decision to form TC4 and JC Pagès [38], and decision support formation of IFIP-TC 4 and included: a) the truly international nature of with FT de Dombal [39]. IFIP, 2) the federated nature of IFIP’s orga- The first MEDINFO, organized by IF- the First World Conference nization with specific technical committees IP-TC4, was held in 1974 in Stockholm on Medical Informatics – subsequently inspiring IMIA’s mode of concurrently with the IFIP meeting from organization, 3) the strong link of IFIP with August 5th to 10th [22]. François Grémy – MEDINFO 1974 in UNESCO - the “biological father “of IFIP, was chair of the Program Committee, while and 4) the fact that IFIP agreed to create J. Anderson and J.M. Forsythe were editorial Stockholm, Sweden IFIP TC4 under the leadership of François committee co-chairs. Grémy had negotiated In 1967, Professor Francois Grémy formed Grémy with Forsythe as Secretary [33]. with the IFIP leadership a separate confer- the Technical Committee 4 (TC4) in Med- The first meeting of the TC4 was ence dedicated to healthcare data processing, ical Information Processing within the held in Paris in April 1968, with a dozen and this was accepted since there was great International Federation for Information nations represented and Grémy as presi- hope at that time that computers in medicine Processing (IFIP). He had already, during dent. During the following years, he was would help not only improve healthcare the Fourth Congress on Cybernetic Med- instrumental in the development and for- but also lead to dramatic new discoveries icine held in Nice in September 1966, mation of several working groups covering in biomedicine – like a cure for cancer. emphasized the historical importance of various subfields of medical information As a result, the seeds for the evolution of “informatique” and the “reasons” for a processing. In 1969, Grémy founded the IFIP-TC4 into IMIA were sown by this tran- marriage between medicine and informat- INSERM Unit U88 entitled Informatics sition at MEDINFO 1974 towards a focus ics [2]. As mentioned above, Grémy and and Statistics Methodology in Medicine on practical computer models and methods Pagès had attended the highly contrast- as an environment in which close collab- for information processing in biomedicine ing presentations at the Nice Congress, orators developed informatics, statistics, and health care more generally, which later and may have realized how important it and decision support systems in medicine. became the central concern of IMIA – which would be to emphasize concrete medical Grémy used TC4 as the vehicle to organize now gets ready to celebrate its 50th Birthday.

IMIA Yearbook of Medical Informatics 2017 262 Kulikowski

Acknowledgments 10. Feinstein AR. Boolean algebra and clinical 27. Waxman BD. Public Health Service support of bio- I would like to thank my many friends and taxonomy. I. Analytical synthesis of the general medical computing. Proceedings 3rd IBM Medical spectrum of a human disease. New Eng J Med Symposium, Endicott, NY IBM; 1961:199-202. colleagues, and especially the members of 1963;269:929-38. 28. Collen M. Automated multiphasic screening as the IMIA History Working Group, and of the 11. Engle RL, Davis BJ. Medical Diagnosis: Past, a diagnostic method for preventive medicine. American College of Medical Informatics Present and Future: III: Diagnosis in the future, Methods Inf Med 1965;4:71-4. (ACMI) Committee of Historians for their including a critique on the use of electronic com- 29. Brodman K, van Woerkom AJ. Computer-aided puters as diagnostic aids to the clinician. Arch Int diagnostic screening for 100 common diseases. suggestions, support, and help in starting and Med 1963;112:530-43. JAMA 1966;197:901-5. pursuing the projects related to the history of 12. Jacquez JA. The Diagnostic Process: Proceedings of 30. Minsky M. Perceptrons. MIT Press; 1968. biomedical and health informatics. I would a Conference at the University of Michigan; 1963. 31. Weizenbaum J. Computer Power and Human 13. Lusted LB. Introduction to Medical Decision Reason: From Judgment to Calculation. New York: also like to thank the reviewers and editors Making. Springfield, IL: Charles C. Thomas; 1968. W.H. Freeman and Company; 1976. of the IMIA Yearbook for their valuable and 14. Lindberg DAB. The Computer and Medical Care. 32. Grémy I. Personal Communication from JC Pagès untiring help and suggestions. Springfield, IL: Charles C. Thomas; 1968 transmitted by B Séroussi. 15. Enslein K, editor. Data Acquisition and Pro- 33. Degoulet P. Personal Communication via B. cessing in Biology and Medicine. New York: Séroussi. Pergamon; 1964. 34. Degoulet P, Haux R, Kulikowski C, Lun KC. References 16. Ledley RS. Use of Computers in Biology and François Grémy and the Births of IMIA. 1st IMIA/ Medicine. New York: McGraw Hill; 1965. UMIT Medical Informatics Award of Excellence 1. Grémy F, Leterrier F. Eléments de biophysique. 17. Masturzo A. Cybernetic Medicine. Springfield IL: given to Professor Grémy. Methods Inf Med Tome 1. Paris: Flammarion Médecine-Sciences; Charles C Thomas; 1965. 2005;44:349-51. 1966. 18. Weiner N. Cybernetics or Control and Communi- 35. Degoulet P, Fieschi M, Goldberg M, Salomon R. 2. Grémy F, Joly H, Pages J-C. Application des cation in the Animal and the Machine. New York: Francois Grémy, a Humanist and Information Sci- Machines à Traiter l’Information au Diagnostic MIT Press and J Wiley; 1948. ences Pioneer. Yearb Med Inform 2014;9(1):3–5. Médical. In Proceedings IVth Congres Interna- 19. Shannon CE. A Mathematical Theory of Com- 36. Anderson J, Forsythe JM, editors. Information tional de Médecine Cybernetique (Masturzo A, munication. Bell System Technical Journal Processing of Medical Records. Proc of the Nayrac P, eds), Nice, France, September 1966. 1948;27:379–442 and 27:623–56, and Shannon IFIP-TC4 Working Conference on Information p. 289-97. CE, Weaver W. The Mathematical Theory of Processing of Medical Records, Lyon, April 6-10, 3. Nash FA. Differential Diagnosis, an apparatus to Communication. Urbana IL: Univ of Illinois 1970. Amsterdam: North-Holland; 1970. assist the logical faculties. Lancet 1954;266:874-5. Press; 1949. 37. Zywietz C, Schneider B, editors. Computer Ap-

4. Paycha F. Critique scientifique des démarches 20. Oberhoffer G. Report on an international seminar plication on ECG and VCG Analysis. Proc 2nd intellectuelles du diagnostic. Presse Med 1955; for medical documentation and statistics. Methods IFIP-TC4 Working Conference, Hanover 1971. 63:1753-6. Inf Med 1962;1:27-31. Amsterdam: North-Holland; 1973. 5. Lipkin M, Hardy JD. Mechanical correlation of 21. Hansen T. Proceedings of the First International 38. Anderson J, Grémy F, Pagès JC, editors. Education data in differential diagnosis of hematological Hospital Data Processing Conference, Elsinore; in Informatics of Health Personnel. Amsterdam: diseases. JAMA 1958; 166: 113-25. 1966. North-Holland; 1974. 6. Ledley RS, Lusted LB. Reasoning Foundations of 22. Proceedings of the First International Conference 39. De Dombal FT, Grémy F, editors. Decision Making on Medical Informatics, Stockholm. North-Hol- and Medical Care. Amsterdam: North-Holland; Medical Diagnosis Science. 1959;130:9-21. land; 1974. 1976. 7. Overall JE, Williams CM. Models for medical 23. McCulloch WS, Pitts WA. Logical Calculus of diagnosis. Behav Sci 1961;6:134-41. Ideas Immanent in Nervous Activity. Bulletin of 8. Collen MF, Rubin L, Neyman J, Dantzig GB, Mathematical Biophysics; 1943;5:115–33. Baer RM, Siegelaub AB. Automated multiphasic 24. Uhr L. Pattern Recognition Psychol Bulletin screening and diagnosis. Amer J Public Health 1963;60:40-73. Correspondence to: 1964;54:741-50. 25. Watanabe S. Knowing and Guessing, New York: Casimir A. Kulikowski 9. Warner HR, Toronto AF, Veasy LG, Stephenson Wiley; 1968. Department of Computer Science R. A mathematical approach to medical diagnosis. 26. Masturzo A, Nayrac P, editors. Proceedings IVth Rutgers - The State University of New Jersey Application to congenital heart disease. JAMA Congres International de Medecine Cybernetique. Piscataway, New Jersey, USA 1961;177:177-83. Nice, France: Editions C.R.E.N.A.F.; September 1966. E-mail: [email protected]

IMIA Yearbook of Medical Informatics 2017